Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The development of instruments to measure the work disability assessment behaviour of insurance physicians

The development of instruments to measure the work disability assessment behaviour of insurance... Background: Variation in assessments is a universal given, and work disability assessments by insurance physicians are no exception. Little is known about the considerations and views of insurance physicians that may partly explain such variation. On the basis of the Attitude - Social norm - self Efficacy (ASE) model, we have developed measurement instruments for assessment behaviour and its determinants. Methods: Based on theory and interviews with insurance physicians the questionnaire included blocks of items concerning background variables, intentions, attitudes, social norms, self-efficacy, knowledge, barriers and behaviour of the insurance physicians in relation to work disability assessment issues. The responses of 231 insurance physicians were suitable for further analysis. Factor analysis and reliability analysis were used to form scale variables and homogeneity analysis was used to form dimension variables. Thus, we included 169 of the 177 original items. Results: Factor analysis and reliability analysis yielded 29 scales with sufficient reliability. Homogeneity analysis yielded 19 dimensions. Scales and dimensions fitted with the concepts of the ASE model. We slightly modified the ASE model by dividing behaviour into two blocks: behaviour that reflects the assessment process and behaviour that reflects assessment behaviour. The picture that emerged from the descriptive results was of a group of physicians who were motivated in their job and positive about the Dutch social security system in general. However, only half of them had a positive opinion about the Dutch Work and Income (Capacity for Work) Act (WIA). They also reported serious barriers, the most common of which was work pressure. Finally, 73% of the insurance physicians described the majority of their cases as ‘difficult’. Conclusions: The scales and dimensions developed appear to be valid and offer a promising basis for future research. The results suggest that the underlying ASE model, in modified form, is suitable for describing the assessment behaviour of insurance physicians and the determinants of this behaviour. The next step in this line of research should be to validate the model using structural equation modelling. Finally, the predictive value should be tested in relation to outcome measurements of work disability assessments. Background occur in the diagnosis and treatment of patients. A Variation in assessments by professionals is a well- degree of uncertainty is inherent in the profession and it known phenomenon which occurs in cases where is very easy to reach different conclusions in comparable assessments are carried out by several raters and in var- cases [1,2]. Inter-doctor variation in diagnosis and/or ious disciplines. In the case of physicians, variations treatment is found in different medical disciplines [3-7]. Specific research into variation among GPs shows varia- tion in diagnosis [8,9], request for interventions * Correspondence: romy.steenbeek@tno.nl [8,10-12], treatment [13,14] and rate of referral to spe- TNO Work and Employment, PO Box 718, 2130 AS Hoofddorp, the cialists [15]. Literature on insurance physicians is less Netherlands Full list of author information is available at the end of the article © 2011 Steenbeek et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Steenbeek et al. BMC Public Health 2011, 11:1 Page 2 of 18 http://www.biomedcentral.com/1471-2458/11/1 extensive but variation in the assessment of functional 1970-1980 by Fishbein and Ajzen in their ‘Theory of capacity for work exists [16-18]. In this study we con- Reasoned Action (TRA)’ [22,23]. In the 1980 s this the- centrate on how insurance physicians assess workers’ ory was taken a step further by Azjen [24] in his ‘Theory claims to compensation for loss of work capacity. Insur- of Planned Behaviour (TPB)’. He added ‘perceived beha- ance physicians have to judge a claim and in doing so vioural control’ (self-efficacy) as a factor that moderates base themselves on the available information (file) and behaviour. In the second half of the 1980 s Azjen’s information provided by the “client”, i.e. the patient who model was supplemented by elements from the ‘Social claims a work disability benefit, and others. The out- CognitiveTheory(SCT)’ of Bandura[25] in theso- called ASE model [26]. ASE is a model that has general come of the assessment, i.e. the functional capacity for work, is variable. An important aspect of this assessment scientific acceptance and explains behaviour by linking is the client’s medical situation in the context of the attitude, social influence and self-efficacy with behaviour current social security legislation. In making the assess- and behavioural intention [27]. In addition to the three ment the insurance physician must therefore deal with determinants of intention and behaviour, intermediary the characteristics of both the legal and the medical factors such as ‘knowledge’ and ‘barriers’ can play a role. decision-making process. Assessing work disability is TPB and the ASE model are used in the Netherlands to therefore a complex and specialised process that also explain, among other things, the behaviour of physicians gives rise to variation in outcomes. Ydreborg and Ekberg [28] and patients in an occupational health context [16] found variations in the extent to which applicants [29-32] and the health behaviour of individuals who for disability pension were rejected in practice. Spanjer belong to a particular target group [33,34]. On the basis et al. [17] evaluated inter-rater reliability between insur- of TPB research Croon & Langius [29] have studied atti- ance physicians in respect of physical disability and tudes and working styles (behavioural intentions) among mental disability assessment as reasonable to good. insurance physicians who assessed employees’ sick leave However, inter-rater reliability in respect of the assess- not exceeding one year. The present survey takes the ment of the number of hours clients could function ASE model as the basis for possible explanations of the daily was low. Spanjer et al. [18] found a significant dif- behaviour of insurance physicians in assessing work dis- ference in various scores on assessed work limitation ability after sick leave lasting one year (Disability Insur- items by insurance physicians. ance Act -WAO), two years (Work and Income The outcome of work disability assessments by insur- (Capacity for Work) Act - WIA) or more years ance physicians can be seen as the result of behaviour (Adapted Reassessment Act - HERBO, see next influenced by various factors, including behavioural paragraph). determinants of the physicians in relation to the In the Netherlands, if you are partially or fully incap- intended object of their assessment. Little is known of able of working after two years of illness, you may be what considerations and views of insurance physicians eligible to receive a benefit under the Work and Income may partly account for variation in the outcome of (Capacity for Work) Act (WIA). The WIA succeeded assessments. the Disability Insurance Act (WAO) in January 2006. The WAO was not repealed by the WIA, but now Conceptualisation applies only to those who were already receiving a It is evident from the above that variation generally WAO benefit on 1 January 2006. The Adapted occurs in assessments by physicians. Research shows Re-assessment Act (HERBO) was introduced in August that this variation is connected with, among other 2004 for the reassessment of WAO benefits clients, i.e. things, certain personal characteristics [13,19,20] and the claimants (< 50 years), on the basis of new, stricter behavioural characteristics, such as personal style or criteria that put the emphasis on the client’sresidual attitude [11,13,21]. Systematic variation between insur- functional capacities. These stricter assessment rules ance physicians in the outcomes of assessments (i.e. under HERBO also apply to the WIA. Young disabled grant or reject the claim) can be regarded as the result people may be eligible to receive a benefit under the of assessment behaviour which - in addition to other Invalidity Insurance (Young Disabled Persons) Act factors - is determined in part by the attitude which the (Wajong). The WAO and WIA differ in the time of insurance physician has towards the intended purpose assessment. The WAO provides for assessments after of his assessment. Thus, assessment behaviour is defined one year of illness, whereas the WIA provides for assess- as all behaviour that may influence the outcome of the ments after two years of illness. In this study, we will assessment, including the collection and evaluation of only use disability assessment outcomes under WAO, information about the client. HERBO and WIA. For parts of the ASE model (see figure 1) research has A theory of the relationship between attitudes and shown that there is a correlation with variation in behaviour was elaborated in the literature of the period Steenbeek et al. BMC Public Health 2011, 11:1 Page 3 of 18 http://www.biomedcentral.com/1471-2458/11/1 Barriers Attitude Assessment Social Norm Intention behaviour Self-efficacy Knowledge Background Figure 1 The ASE model. assessments by physicians. Research among GPs shows that few evidence-based and peer-reviewed articles exist. varying conclusions as regards attitude. Taylor con- This is why we also used ‘grey literature’. Literature on cluded as long ago as 1977 in a review [13] that varia- insurance physician’s behaviour is scarce. That is why tion in prescribing behaviour was associated with the we added research on GPs, because in many countries personal attitudes of GPs. In addition, the results of the GP is the physician issuing sickness certificates. Grytten and Sorensen [11] indicate that practice style reflects a deeply rooted behaviour with respect to how Background variables to practise medicine. However, Tellnes et al. [35] and The gender of the physician appears to influence assess- Thies-Zajong et al. [36] found no association between ment behaviour. Female insurance physicians more fre- doctors’ attitudes and some measures of variation. Social quently restrict the number of hours a client can work norms also appear to play a role. The degree to which as part of their assessment [40]. Court [41] showed this insurance physicians are required to achieve a certain with specific regard to Dutch insurance physicians: male level of production and the pressure exerted on them to insurance physicians conclude less frequently than achieve the targets differ from office to office [37-39]. female insurance physicians that clients are not able to No literature was found on the relationship between work. self-efficacy and assessments by physicians. Research Age is closely connected with experience: the older among GPs shows that a higher workload (operationa- the doctor the more experience he has. Both factors lised as list size) functions as a barrier; it leads to more influence assessment behaviour. Physicians with more referrals to specialists [15]. A lack of sufficient knowl- experience in family medicine issued more sickness cer- edge and information can also serve as a barrier. Davis tificates [42] and the duration of episodes of sickness et al. [40] found that diagnostic uncertainty in GPs was certification was longer for patients of older doctors associated with higher rates of investigation and follow- [35]. In the case of insurance physicians, greater experi- up. Finally, client or patient characteristics such as gen- ence is associated with greater optimism about the der [20], unemployment and age [16] or the type of patient’s return to work [43], better quality assessment patient visit [4] were found to be associated with varia- in the case of mental complaints [44], more frequent tions in outcome. allowance of a reduction in working hours [45], a ten- dency to assess suitability for work as higher in the Applying the conceptualisation event of reassessments [43] and a shift in thinking from Below we describe the factors in respect of which an the ‘seriousness of the complaint’ to the ‘extent to association has been found with variations in assess- which the complaints influence daily functioning’ [46]. ments by physicians for each concept from the ASE Training [35,42] and specialisation [19] are also con- model and which have been included in our model. Our nected with the assessment behaviour of physicians. search for literature on the specific relationship between Physicians who were trained in social insurance medi- certain factors and variation in assessments revealed cine as undergraduates issued more sickness certificates Steenbeek et al. BMC Public Health 2011, 11:1 Page 4 of 18 http://www.biomedcentral.com/1471-2458/11/1 [42]. The duration of episodes of sickness certification Mooreetal. [53] concludedthat fearofclaimscan was shorter in patients of doctors with postgraduate influence the relationship with the client and hence the training [35]. decision as well. If the physician is afraid that the client The number of working hours is also a factor that will appeal, he will be less inclined to rule against the influences sickness certification: physicians working part client’s wishes [54]. Insurance physicians who are con- time issued more sickness certificates than physicians fronted by stricter medical professional norms are more working full time [42]. inclined to decide that the client is no longer able to Assessments also vary according to locations or perform any work whatever [38]. region: the variation between these units has been shown to be greater than within them [12,16,47,48]. Self-efficacy Self-efficacy (a person’s belief about his ability and capa- Attitude city to accomplish a task) moderates behaviour [24]. Taylor concluded as long ago as 1977 [13] that more desirable prescribing patterns by GPs were associated Barriers with a more psychosocial orientation towards medical Work pressure, autonomy, emotional workload and care. In case of insurance physicians, desirable behaviour emotional exhaustion are known to be related to health could be associated with perceived justness of the social complaints and influence job participation [55-57]. security system. Office culture can strongly influence job satisfaction and The attitude towards the perceived quality of the hence either hinder or stimulate the physicians in their assessment may also influence it. Insurance physicians work. are dissatisfied about the scope for development and Factors connected with quality are also important. In refresher training and complain about a lack of clear the case of insurance physicians, managerial emphasis instructions in the legislation, regulations and instru- on quantity can act as a barrier [58]. Some physicians ments [49,50]. Views on quality are also influenced by consider that the quality of the assessments suffers from increased production pressure and the feeling of having work pressure and changes in the organisation. Lack of too little time for the assessments [41]. time means that duties are carried out with less care Finally, the physician’s attitude towards the personal [49,50]. Choij [59] concludes that the attitude of profes- needs and circumstances of the client (including his sionals to the organisation is negative. They feel that recovery) can influence the assessment. Physicians tend they are not heard, recognised or appreciated. Manage- to have fairly differing views on the time a client needs ment is seen as the major culprit because it exerts for recovery and also differ in the extent to which they undue production pressure. Stricter requirements and take account of personal circumstances. Some take no an increase in the number of clients with more serious account of this and others apply as a criterion that the sicknesses reduces the physician’s feelings of autonomy client should not be exhausted after finishing work or and increases stress through time pressure [60]. Guide- that theclientshouldbeproperlyrestedwhenhe lines and protocols can be an aid in enhancing quality resumes work the next day [49,51]. [61]. However, physicians consider that they receive too little guidance in applying guidelines for disorders that Social norm are difficult to assess objectively [62] or in applying leg- In the case of insurance physicians the influence of islation, regulations and instruments [50]. 40% of the social norms may emanate from the office and the files are not in keeping (or completely in keeping) with environment. Within the office the norm for the ‘strict- the statutory requirements [63]. ness’ of the assessments may differ [37]. Other factors More difficult assessments may result in more varia- are how much pressure is brought to bear in terms of tion between insurance physicians. The assessment is production and promptness targets and how strictly perceived as more difficult in relation to certain groups these are checked [38,39]. of clients. This applies to older clients [37,64], clients Changing social norms can influence assessments [52]. with mental problems [37,39], clients with impairments The stricter the norm that states that society’s interest that are difficult to determine objectively [65,66] and cli- should be guarded, the higher the probability that insur- ents with psychosomatic disorders [64]. Interviews with ance physicians will find a sufficient job for the client insurance physicians during the preparation of this and the higher the residual earning capacity. However, if questionnaire also showed that they regard assessment the norm to which insurance physicians are subject of the following categories as extra difficult: clients with states that society is responsible for its citizens if they a poor command of Dutch, clients who act aggressively become incapacitated for work, the probability that suf- or manipulatively, and cases where poor preliminary work has been done by the occupational health service. ficient jobs are available will be lower [38]. In addition, Steenbeek et al. BMC Public Health 2011, 11:1 Page 5 of 18 http://www.biomedcentral.com/1471-2458/11/1 Knowledge insurance physician must determine the client’s capacity Deciding whether or not to request and use information for work. This is done by reference to an instrument of third parties may influence the assessment, and this known as FAL (Functional Ability List). On this list the is done to a very varied extent [62,67,68]. Davis et al. physician enters the client’s scores for limitations and [69] found that diagnostic uncertainty among GPs was abilities. These findings serve as the input for the labour associated with higher rates of investigation and follow- expert in determining the extent to which the client is up. Having more information increased the physician’s able to earn income and able to work. As an instrument self-confidence [43]. However, Kerstholt et al. [70] sug- the FAL comes within the statutory framework of dis- gest that assessments of disability are largely based on ability assessments in the Netherlands. If the client is the initial view formed after reading the file. The main permanently and fully disabled, he is classified as such pitfall is that the final view is based on general beliefs (in the Netherlands known as GBM). This is therefore a rather than on actual client information. The insurance dichotomy measurement. The literature shows that stat- physician’s job experience, competencies and interests utory rules are implemented in different ways [74]. determine in part how difficult it is to ‘translate’ mental Insurance physicians interpret the guideline on perma- complaints into limitations and residual capacities [44]. nent disability in a wide variety of ways [49,75]. The Physicians need additional instruments, knowledge and definition given in the guideline - namely an incapacity experience in relation to disorders that are difficult to to function socially and personally - is considered inade- assess objectively [66]. quate. The criterion of permanent full disability is also used as a safety net in order to compensate for loss of Intention income in difficult situations. The possibility of indicat- Studies by De Boer et al. [65] and Spanjer [71,72] show ing that the client can work a limited number of hours that the object, the physician’s interpretation of his (limitation of hours) is also applied in a wide variety of duties and the basic premises have an important bearing ways [45,49]. Likewise, the time required for recovery is on the proposed assessment. assessed in a variety of ways [49]. The above concepts are described in the model under ‘Behaviour: Behaviour assessment’. We will distinguish two types of behaviour: 1) behaviour that reflects the process of assessment, such as the col- Aim lection of information about the client and 2) behaviour The first aim of this study was to develop measurement directly connected with the assessment itself, such as instruments that can potentially affect disability assess- the use of assessment instruments in order to evaluate ment behaviour by insurance physicians. The descrip- the information. tion of a conceptual model includes the relevant Spanjer [18] found that process variables hardly (behavioural) variables. We expect that this model will affected assessment outcomes. Information on participa- contribute towards understanding and explaining varia- tion and activity limitations provided by the patient had tion in assessments of functional capacity by insurance only limited influence on inter-rater reliability by insur- physicians. A second aim was to discuss the descriptive ance physicians. However, there was a significant differ- results of the insurance physicians’ scores on the differ- ence in scores on assessed work limitation items ent concepts. In accordance with these aims, the compared with medical history-taking alone. It follows research questions were: 1) can we construct measure- that in disability assessment interviews physicians should ment instruments to measure assessment behaviour and ask for medical information as well as detailed informa- its potential determinants? and 2) What are the charac- tion on participation and activity. Dedication (an aspect teristics of the assessment behaviour of Dutch insurance of ‘work engagement’ [73]) is a reflection of the inten- physicians and its potential determinants according to tion to carry out tasks and is therefore placed under these instruments? behaviour. This also applies to dealing with conflicts. A client may have a different view on the outcome of Methods the assessment. This may result in a conflict during a Study procedure disability evaluation. The above concepts are described The research group of the organisations participating in in the model under ‘Behaviour: process’. this study - TNO Quality of Life, the EMGO Institute of Behaviour directly connected with the assessment the VU Medical Centre and the Employee Benefits itself was defined as the use of instruments to assess Insurance Agency (UWV) - drafted the questionnaire capacity for work and the importance the insurance for insurance physicians. At the start of 2008 UWV physician attached to the client’s opinion about his/her drew up a list of addresses of all insurance physicians own functional capacities. In the Netherlands the working for the agency. In March 2008 UWV sent the Steenbeek et al. BMC Public Health 2011, 11:1 Page 6 of 18 http://www.biomedcentral.com/1471-2458/11/1 questionnaire, together with a covering letter containing We included 11 items about the perceived justness of an invitation to participate in the research, to the home the social security system, the agency that administers addresses of 750 insurance physicians. A reminder was the scheme (UWV) and the Permanent Full Disability sent two weeks later. Not all the physicians belonged to Standard, FAL and the implementation of the Work and our target group, but it was not possible to make a Income (Capacity for Work) Act (WIA). The items have selection in the mailing. In total we wrote to 750 insur- five response categories, ranging from (1) I totally dis- ance physicians. Our estimate is that the target group agree to (5) I totally agree. consisted of 450 insurance physicians. The criteria for We included nine items on the attitude to quality in inclusion were mentioned in the accompanying letter. relation to the importance which insurance physicians We only included insurance physicians who were attach to the development of skills, to refresher training, actively employed by UWV in May 2008 and who had to guidance by management, to the development and also performed work disability assessments in 2007 or in use of protocols and guidelines and to updating the case preceding years. The participants sent the completed file. The items have five response categories ranging questionnaire to TNO (Netherlands Organisation for from (1) I totally disagree to (5) I totally agree. Applied Scientific Research). The response consisted of We included six items on the physician’s attitude 231 questionnaires (estimated response approximately towards recovery time, the personal circumstances of 51%). As this study was based on a survey under (insur- the client and the physician’s efforts to build a good ance) physicians only, approval by a Medical Ethical relationship with the client. The items have five Commission was not necessary under Dutch law. response categories ranging from (1) I totally disagree to (5) I totally agree or (1) never to (5) always. Questionnaire Social norm In drawing up the questionnaire we used existing and In the case of insurance physicians the influence of newly developed concepts. These concepts were chosen social norms may emanate from the office and the on the basis of literature studies and four interviews environment. We included four items on management with insurance physicians. In a pilot study two insurance attitudes to quantity as opposed to quality, to the use of physicians completed the questionnaire while thinking protocols and guidelines and to production and out- aloud in order to enable us to test whether the items comes. The items have five response categories ranging were correctly understood. Finally, two other insurance from (1) I totally disagree to (5) I totally agree. physicians were timed while they completed the ques- In addition we included 13 items on the importance tionnaire. An English translation of the original Dutch which insurance physicians attach to the exercise of questionnaire is accessible (additional file 1). their profession, to the opinion of the Employee Benefits Insurance Agency (UWV), the government authorities Concept measurements questionnaire and professional organisations such as the Dutch Asso- Background variables ciation for InsuranceMedicine(NVVG)and theDutch We measured gender, age, number of years’ experience, Association for Insurance Physicians at the UWV training and specialisation. In the case of training and (UWVA), friends and family, colleagues in the office specialisation we included two items. First, we asked and elsewhere, public opinion, professional publications, whether the insurance physician is registered as such quality assessment and the trade unions. The items have (and is not still in training). Second, whether he prac- four response categories, ranging from (1) not important tises or has practised in another area of medicine. In to (4) very important. The higher the score, the greater order to register differences between offices or regions the importance attached to this opinion or view. we recorded the location of the insurance physician’s Self-efficacy office. We also asked how many hours they work each Self-efficacy was measured by the ten items formulated week, how many assessments they make each week, by Scholz et al. [77], adjusted to measure the insurance from which industry the majority of their clients come physician’s belief about his ability and capacity to carry and the statutory background of the assessments of the out work disability assessments. The items relate specifi- majority of their clients, namely the Work and Income cally to self-efficacy during the disability assessment (Capacity for Work) Act (WIA), the Disability Insurance interview. The items have four response categories, ran- Act (WAO) or another statutory regime. ging from (1) completely incorrect to (4) completely Attitude correct. Job satisfaction was measured by three items with five Barriers response categories ranging from (1) never to (5) always Work pressure was measured by means of a four-item (I am satisfied with my work; my work suits me; I like scale drawn up by Smulders, Andries and Otten [78]. A my work) from Van Dijk et al. [76]. sample item is ‘Do you have to get through a lot of Steenbeek et al. BMC Public Health 2011, 11:1 Page 7 of 18 http://www.biomedcentral.com/1471-2458/11/1 work?’ A four-point answering scale was used ranging doctor) regarding the attempt to return to work, the from (1) never to (4) always. diagnosis and information from the parties. We also Emotional workload was measured using a three-item asked whether information from the reintegration report scale from the Copenhagen Psychological Questionnaire (drawn up by the occupational physician and sent with [79]. A sample item is: ‘Does your work put you through the WIA benefits application) was decisive and whether emotionally difficult situations?’.Answers were scored the physician received sufficient feedback from the on a four-point scale ranging from (1) never to (4) claims manager about the outcome of his assessments. always. The items had four response categories, ranging from Decision-making authority (4 items) was measured (1) never to (4) always. using a Dutch version of the Job Content Questionnaire, Intention aimed at assessments [80,81]. A sample item is: ‘Do you For practical reasons we chose to measure only ‘the determine the order in which you carry out your tasks?’ object, the physician’s interpretation of his duties and Answers were scored on a four-point scale ranging from the basic premises’ in relation to intention. A question- (1) never to (4) always. naire in which intention is measured in respect of all Emotional exhaustion was measured using the five- behavioural items would be much too long. The 15 item emotional exhaustion scale of the Dutch version of items had five response categories ranging from (1) not the Maslach Emotional Exhaustion Inventory [82]. at all important to (5) very important. In the case of Answer categories varied from (1) never to (5) almost object/interpretation of duties we asked how important daily. The higher the score, the greater the exhaustion. the following objects are in relation to the assessment: We included 12 items on cooperation, office atmo- determination of physical capacities and cause of sick- sphere, consultation, being taken seriously by the man- ness, promotion of behaviour conducive to recovery, agement, and influence on workload. [76]. The higher return to work, client’s self-insight and reintegration. the score the greater is the extent of the cooperation or We also asked how important the following factors are co-determination. The items have five response cate- in the assessment of claims: health complaints, impair- gories ranging from (1) I totally disagree to (5) I totally ments, limitations or handicaps of the client, an intern- agree. ally consistent and plausible account provided by the We included 11 items on factors that could hinder or client, thorough questioning of the account given by the promote the quality of the assessment: legislation, reor- client of his daily activities, work capacity, chances in ganisations, support and guidance by staff physicians the labour market and information about the client’s and management, reporting requirements, protocols/ home situation. guidelines and standards, production requirements, Behaviour process refresher training and other measures to promote exper- Engagement is a concept that refers to being fully tise, and mutual consultation. Each of the items has immersed in an activity (absorption), being highly acti- three response categories: adverse influence, no influ- vated (vigour) and identifying with the work (dedica- ence or beneficial influence. tion). We used the four items of the subscale of We included 16 items concerning ‘difficult clients’.We dedication on the engagement scale developed by asked whether the following eight categories constitute Schaufeli et al. [73]. The items have five response cate- an important proportion of the physician’s clients and gories, ranging from (1) never to (5) always. The higher whether the physician considers the assessments of the score, the greater is the work dedication. these categories to be extra difficult: clients with disor- Research by De Boer et al. [83] shows that although ders that are difficult to determine objectively, clients there are various interview models, they are not used as with mental disorders, clients with a poor command of such. We did not therefore ask about the models, but Dutch, clients who are aggressive, clients who are included nine items on different core elements from the manipulative, clients who have problems at home or models, such as who determines what is discussed in work, older clients and cases in which poor preparatory the interview (physician or client), whether the physician work has been done by the occupational health service. asks questions in a fixed order, whether the physician Knowledge asks questions about subjects raised by the client, and In order to form a picture of the need for knowledge/ whether the physician asks for concrete examples of information, the actual information received and the use barriers and examines whether barriers result in limita- of this information for the purposes of the assessment tions. The items had four response categories, ranging we choose to include 11 general items (i.e. not specifi- from (1) never to (4) always. cally relating to diagnosis) as to whether physicians had To measure conflict handling we used the Dutch Test sufficient medical knowledge, medical information, for Conflict Handling [84], after modifying the items to information from the occupational physician (company confine them to the disability evaluation and conflicts Steenbeek et al. BMC Public Health 2011, 11:1 Page 8 of 18 http://www.biomedcentral.com/1471-2458/11/1 with a client. This test measures to what extent five female. Insurance physicians of the total population strategies are applied for handling conflicts, namely worked on average 32 hours per week. Although distri- yielding, problem-solving, compromising, avoiding and bution measures of these population means could not forcing. We used three items for each strategy (total 15 be calculated, even if the (unknown) population confi- items). dence intervals were smaller than those of the respon- Behaviour assessment dent group, the respondent group in our study would As far as the ‘permanent full disability’ criterion is con- not significantly differ from the population of insurance cerned we included five items on the extent to which physicians in terms of age, gender, and working hours per week. the rules are followed and how physicians assessed a cli- ent who is completely unable to work but can still func- Imputation of missing values tion in at least one social role. The items had four With listwise deletion, only 122 cases of the 231 cases response categories, ranging from (1) never to (4) would be left. We therefore decided to impute for miss- always. ing values. Because year of birth was not answered in 40 As regards the FAL we included nine items in order of the 231 cases, we imputed 38 of these missing cases to estimate to what extent physicians focus on a) limita- by the predictions of an OLS regression equation for tions, impairments or complaints; b) what the client can age (i.e. year of birth minus 2008). In the regression do; c) difficult home circumstances; d) internal and equation (listwise, enter procedure, n = 185) we used external consistency; e) worsening health; and f) consul- the other background variables as independent variables: tation with the labour expert. The items had five sex (dummy), registered as insurance physician response categories, ranging from (1) I totally disagree (dummy), (formerly) registered as curative specialist to (5) I totally agree. (dummy), working hours per week, number of assess- As regards behaviour in relation to the client, it is evi- ments per week, type of statutory scheme applicable to dent from the study by Nagtegaal [85] that the client’s most of the assessments (three dummies for WIA, account of daily activities is a useful instrument in asses- WAO and the Invalidity Insurance (Young Disabled Per- sing the extent of his physical capacities. We included sons Act (Wajong)) and sector (ten dummies for eleven 10 items with four response categories, ranging from (1) sectors). The multiple correlation of the predicted age never to (4) always. We asked how often the interview with the observed age was 0.696; the standardised resi- lasted as long as necessary, whether the client was trea- duals had a completely normal distribution. The SPSS ted with respect, whether the physician felt involved 15.0 program [86] was used for this regression analysis. with the client, whether the physician took an indepen- The remaining missing values for the background vari- dent position and did not allow himself to be affected ables, the scale variables and the object scores of the by the client’s interests and whether the physician took HOMALS dimensions (see the next paragraph) were the time to question the client thoroughly about his imputed using the ‘expected maximisation’ algorithm account of his daily activities, to provide good reasons [87]. There were three variables with eleven to seven- for his conclusion and to write a good report. teen imputed cases, six variables with six to ten imputed cases and thirteen variables with two to five imputed Analyses cases. The remaining variables had no or only one Response imputed case. The interactive Lisrel program with Prelis In total we wrote to 750 insurance physicians. Our esti- 2.72 [88] was used for this imputation procedure. mate is that the target group consisted of 450 insurance Construction of scales and dimensions for the ASE concepts physicians. The response consisted of 231 questionnaires The answers of the 231 insurance physicians were used (estimated response approximately 51%). As we lacked to determine which concepts from the questionnaire the necessary data of the target population to do a full were suitable for further analysis. The responses given non-response analysis, we checked whether the group of by the insurance physicians were inspected. For some participants (N = 231) was representative of the total items it was necessary to recode the original items in population of insurance physicians working for UWV fewer categories as some categories were empty or (N= approximately 900, including staff-members and almost empty. Negatively formulated items were recoded physicians not performing disability assessments) in positively. terms of age, gender, and working hours per week. The Scales were formed for the following already validated mean age and 95% confidence interval (CI) of the scales: job satisfaction, self-efficacy, work pressure, emo- respondents was 50.8 years (95% CI [49.1;51.7]) and tional workload, decision-making authority, emotional 41.1% were female. The respondents worked on average exhaustion and engagement. Cronbach’s alpha was com- 32.5 hours per week (95% CI [31.5-33.4]). The total puted for each of these scales. For the remaining items, population’s mean age was 49 years and 41.7% were factor analyses with principal components analysis and Steenbeek et al. BMC Public Health 2011, 11:1 Page 9 of 18 http://www.biomedcentral.com/1471-2458/11/1 varimax or oblique rotation per block of items, were transformations [90], they are not scales, and reliability performed to extract factors for each theoretical con- analysis cannot be performed. Therefore, we call these cept. Oblique rotation was chosen only if there was a variables ‘dimensions’, contrary to the variables which we significant correlation between the extracted factors. constructed as additive scales, which we call ‘scales’.We Where this was not the case, we decided to use varimax used the SPSS 15.0 program [86] for the factor analyses, rotation. Bartlett’s test of sphericity was used to test reliability analyses and the HOMALS analyses. whether the correlation matrix was an identity matrix. The sampling adequacy was inspected by means of the Results Kaiser-Meyer-Oikin measure (KMO) and found to be Descriptives of all measured scales are presented in table greater than 0.6. The number of extracted factors was 1 and those of the measured object scores resulting decided on the basis of the scree test, the Eigenvalue from the HOMALS analyses in table 2, including the and, most of all, the interpretability of the extracted fac- finalnumberofitems.Whennot alloriginalitems are tors. For each extracted factor, reliability analysis, included in the scales and dimensions, we report it in including item analysis, was performed to construct this section. A summary of scales and dimensions for additive scales from the items of the factors. An additive each ASE concept is presented in figure 2. scale is constructed of numerical categories of items that can be meaningfully added. In the item analysis, the New scales and dimensions contribution of each item to the reliability of an additive Attitude scale can be estimated. If an item did not contribute to We developed two scales on the theme of justness (10 an additive scale, this item was deleted from this scale. of the 11 original items): The higher the score on the When Cronbach’s alpha was equal to or larger than 0.6, scale ‘Positive attitude towards the WIA’, the more posi- additive scales of the selected items were calculated. We tive the opinion about the justness of the WIA. The nonetheless also decided to use additive scales in three higher the score on the scale ‘Social security system is cases where Cronbach’s alpha was less than 0.6 (0.560, just’, the more positive the opinion about the agency 0.566 and 0.594, respectively). These three scales were that administers the scheme (UWV), the Permanent Full considered to be theoretically important. For each addi- Disability Standard and FAL. tive scale we also calculated the percentage of respon- We developed two attitude scales for quality (eight of dents who, on average, scored above the theoretical the nine original items). The higher the score on the mean of the additive scale. This means that in case of scale ‘Quality: development of skills important’, an additive scale consisting of four Likert scale items the greater the importance attached by the physician to ranging from 1 to 5, we report the percentage of the promotion of expertise, consultation with colleagues, respondents with a scale average above 3*4 = 12.0. In working in accordance with protocols and properly the remaining text when we refer to scales, we mean updating the case file. The higher the score on the scale ‘additive scales’. ‘Quality: support by management important’, the greater For some blocks of items and for some individual items the importance attached by the physician to support it was not possible to construct a scale for several rea- and management by the immediate superior and sup- sons: the correlation matrix was not an identity matrix, port by the staff physician. and/or the sampling adequacy was not good, or Cron- We developed the two attitude dimensions on recov- bach’s alpha was too small. We grouped these ‘lost’ items ery time (six items). The higher the score on the ‘Recov- on a theoretical basis, recoded them, if necessary, into ery time: client still has some energy left after work’ two or three categories and used HOMALS (homogene- dimension, the more the insurance physician agrees the ity analysis by means of alternating least squares) to ana- client should not be completely exhausted after work. lyse the dimensions behind these grouped items [89]. The higher the score on the ‘Recovery time: good rela- The number of dimensions was decided on basis of the tionship with client’ dimension, the more the insurance sum of the Eigenvalues of the dimensions. We estimated physician tries to establish good relations with the client for each dimension the discrimination measures of the and takes account of personal circumstances. items, the category quantifications of categories of items, Social norm and the object scores of the cases. We used the discrimi- We developed three scales (12 of the 13 original items) nation measures and the category quantifications to and two dimensions (four items). The following three interpret both poles (negative and positive) of the dimen- scales are about the influence which the opinions and sions. The object scores of the dimensions that were views of certain persons/authorities have on the perfor- meaningful and gave additional information were mance of the profession. The higher the score on the selected as variables. Because object scores of multiple scale, the greater the importance attached to this opi- Homals dimensions are constructed with non-linear nion or view. The ‘Opinion of UWV and employee Steenbeek et al. BMC Public Health 2011, 11:1 Page 10 of 18 http://www.biomedcentral.com/1471-2458/11/1 Table 1 Description of scales (n = 231) ASE Scale # % yes/ Theor. Median Mean sd Cronbach’s 1 2 items high max alpha Attitude Job satisfaction 3 78 15 12 11.41 2.51 0.875 Positive attitude towards WIA 5 53 25 16 15.98 3.63 0.797 Social security system just 5 70 25 17 17.43 3.25 0.636 Quality: development of skills important 5 99 25 22 22.11 2.27 0.648 Quality: support by management important 3 68 15 11 10.46 2.32 0.643 Social Norm Opinion of UWV and employee representative bodies 6 43 24 15 15.01 2.75 0.697 important Colleagues’ opinion important 5 66 20 14 13.40 2.36 0.679 Society’s opinion important 3 10 12 6 5.77 1.53 0.560 Self-efficacy Self-efficacy 10 40 32 32.81 4.21 0.908 Barriers Work pressure 4 44 16 10 10.18 2.06 0.771 Emotional workload 3 20 12 6 6.38 1.42 0.702 Decision making authority 4 61 16 11 10.96 2.65 0.724 Emotional exhaustion 5 12 25 10 10.79 3.97 0.892 Office culture: good cooperation 8 83 40 30 29.62 5.65 0.900 Office culture: sufficient co-determination 4 20 20 10 9.78 3.25 0.814 Quality: influence of refresher training and 2 97 6 6 5.90 0.38 0.665 consultation beneficial Quality: influence of staff physician beneficial 2 80 6 6 5.38 0.89 0.675 Quality : influence of manager beneficial 2 27 6 4 4.03 1.05 0.647 Many difficult clients/cases 16 73 16 15.62 2.82 0.675 Knowledge Sufficient information from the occupational physician 3 44 12 7 7.20 1.33 0.769 Intention Stimulate recovery and return to work 4 94 20 17 16.71 2.77 0.852 Basic premises: residual capacity 6 99 30 28 27.03 2.75 0.809 Basic premises: client’s account and home 4 97 20 18 17.37 2.17 0.727 circumstances Behaviour Process Dedication 4 73 20 14 13.94 2.74 0.874 Technical interview: describe object and procedure 2 84 8 7 6.71 1.15 0.594 Conflict handling: seek compromise 7 5 35 15 15.57 3.46 0.733 Behaviour Assessment Comply with permanent full disability rules 2 72 8 6 6.09 1.23 0.734 FAL: take account of client 5 31 25 14 14.16 2.47 0.566 FAL: consult with labour expert when not necessary 2 63 10 8 7.09 2.00 0.761 yes/high = % of respondents whose score is above the theoretical scale average, e..g. the % of respondents who on average ‘(totally) agree’, think of something as ‘(very) important’, or score ‘often/always’. The theoretical maximum value of the scale. The percentage of insurance physicians who classify the majority of their clients/cases as difficult. representative bodies important’ concerns the influence important’ scale concerns the influence of fellow insur- of the UWV (supervisor, staff physician, objection and ance physicians, both in the office and elsewhere. The appeal), and employee representative bodies such as the ‘Society’s opinion important’ scale concerns the influ- Dutch Association for Insurance Medicine (NVVG), the ence of family, friends, TV and government. Dutch Association for Insurance Physicians at the UWV Thehigherthe scoreonthe dimension ‘Managing by (UWVA) and the trade union. The ‘Colleagues opinions reference to quality rather than quantity’,the more Steenbeek et al. BMC Public Health 2011, 11:1 Page 11 of 18 http://www.biomedcentral.com/1471-2458/11/1 Table 2 Description of Homals object scores for dimensions (n = 231) # Min Max Median Mean* Sd* Eigen items value Attitude Recovery time: client still has some energy left after work 6 -2.50 1.79 -0.0784 0.00 1.00 0.254 Recovery time: good relationship with client 6 -3.36 2.07 0.0931 0.00 1.01 0.188 Social norm Managing by reference to quality rather than quantity 4 -1.44 1.70 -0.0953 0.00 1.00 0.489 Managing less by reference to production targets and outcomes 4 -1.16 4.03 -0.1753 0.01 1.00 0.371 Barriers Quality: influence of legislation and reorganisations not adverse 4 -1.75 2.13 -0.0833 0.00 1.02 0.352 Quality: influence of guidelines not adverse and production target not beneficial 4 -1.25 2.90 -0.6381 0.01 1.02 0.269 Knowledge Possessing, requesting and using insufficient information 8 -2.65 2.20 0.0174 -0.01 1.02 0.231 Insufficient medical information and knowledge 8 -2.77 1.80 0.0751 0.00 1.01 0.173 Sufficient knowledge, reintegration report less often supplements medical 8 -2.67 2.77 -0.0985 0.00 1.01 0.154 information Behaviour Process Interview management: client decisive 6 -1.66 2.46 -0.2199 0.00 1.00 0.263 Interview: limitations not checked 6 -1.32 3.09 -0.4262 0.00 1.00 0.214 Interview: respond to client 6 -2.32 2.61 -0.0496 0.00 1.00 0.170 Conflict handling: engage in confrontation 8 -1.91 2.63 0.0086 0.00 1.00 0.235 Conflict handling: play down differences 8 -1.92 2.87 -0.1829 0.00 1.00 0.209 Behaviour Assessment FAL and recovery time: strict/formalistic approach 6 -9.13 1.82 0.0851 -0.07 1.27 0.292 FAL and recovery time: focus on impairments 6 -8.94 2.27 0.0507 -0.28 1.65 0.288 Client approach: involved with and time for 8 -2.24 2.25 -0.0536 0.00 1.00 0.263 Client approach: time for account of daily activities and reporting 8 -2.68 1.58 0.0614 0.00 1.00 0.178 Client approach: too little time, but involved with 8 -7.66 2.74 -0.2466 -0.07 1.23 0.168 * Because of imputation, object scores can deviate from mean = 0 and sd = 1. importance is attached in relative terms to quality-based We constructed one scale named ‘Many difficult cli- management and the less importance to quantity-based ents’, based on 16 items. The higher the score, the more management. The higher the score on the dimension the insurance physician is confronted with clients whom ‘Managing less by reference to production targets and he experiences as difficult. outcomes’, the less importance is attached to manage- The higher the score on the ‘Quality: influence of leg- ment based on production targets and outcomes. islation and reorganisations not adverse’ dimension, the Barriers lower the adverse impact of legislation and UWV reor- We developed four scales and two dimensions for bar- ganisations on quality. The higher the score on the riers. Of the 11 original items about quality of the ‘Quality: influence of guidelines not adverse and produc- assessment, six were used in three scales of two items tion requirement not beneficial’ dimension, the lower each and four were used in two dimensions. The higher the adverse impact of the guidelines and the lower the the score on the ‘Quality: influence of staff physician is beneficial impact of the production requirement on beneficial’ scale, the more the assessment benefits in quality. terms of quality from support and professional guidance Knowledge provided by the staff physician. The higher the score on We developed one scale (three items) and three dimen- the ‘Quality: influence of refresher training and consul- sions (eight items). The higher the score on the ‘Suffi- tation beneficial’ scale, the greater the beneficial effect cient information from the occupational physician’ scale, of refresher training, mutual consultation and measures the more sufficient the available information from the to promote expertise support. Finally, the higher the OP. The higher the score on the ‘ Possessing, requesting score on the ‘Quality: influence of manager beneficial’ and using insufficient information’ dimension, the more scale, the greater the beneficial effect of the support and the physician considers that he does not always have professional guidance provided by the manager. sufficient medical information available, does not always Steenbeek et al. BMC Public Health 2011, 11:1 Page 12 of 18 http://www.biomedcentral.com/1471-2458/11/1 Barriers •Work pressure (1S) •Emotional workload (1S) •Decision authority (1S) Attitude •Emotional exhaustion (1S) •Job satisfaction (1S) •Office culture (2S) •Justness of system (2S) •Quality (3S, 2D) •Importance of skills and support (2S) •Difficult clients (1S) •Attitude towards recovery time (2D) Intention Assessment behaviour: process Social Norm •Recovery and return to •Dedication (1S) •Influence of representative bodies, work (1S) •Collecting information (1S, 3D) colleagues, society (3S) •Basic premises (2S) •Conflict handling (1S, 2D) •Production quality versus quantity (2 (2D) D) Knowledge Assessment behaviour: assessment Self-efficacy •Information from OP (1S) •Use of assessment instruments (3S, 2D) •Self-efficacy (1S) •Possessing, requesting •Client approach (3D) and using sufficient information (3D) Background Gender, age, experience, training, specialisation, location, working hours, production, client industry, assessment type Figure 2 The ASE model with a summary of scales and dimensions. S = Scale; D = Dimension, the number refers to the number of constructed scales and dimensions. request information from third parties and does not the importance of residual capacity, sickness, impair- always take this into account in making the assessment. ments, limitations and handicaps in the assessment. The higher the score on the ‘Insufficient medical infor- Finally, the ‘Basic premises: client’s account and home mation and knowledge’ dimension, the more the physi- circumstances’ scale concerns the importance of a con- cian considers that he does not always have sufficient sistent account of daily activities, thorough questioning medical information and medical knowledge to make about this account and information about the client’s the assessments. The higher the score on the ‘Sufficient home circumstances in the assessment. knowledge, reintegration report less frequently supple- Behaviour process ments medical information’ dimension, the more We developed one scale and three dimensions concern- sufficient is medical information and the less frequently ing the collection of information about functional capa- the reintegration report supplements the medical cities (eight of the nine original items). The higher the information. score on the ‘Technical interview: describe object and Intention procedure’ scale, the greater the emphasis which the We developed three intentionscales(14of theoriginal physician puts at the beginning of the interview on the 15 items). The higher the score on these scales, purpose of the interview and the procedure to be fol- the more importance is attached to the subjects. The lowed during it. The higher the score on the ‘Interview ‘Stimulate recovery and return to work’ scale concerns management: client decisive’ dimension, the more the the importance to promote recovery behaviour, return client determines the order of events rather than the to work, self-insight and reintegration in the assessment. insurance physician. The higher the score on the ‘Inter- The ‘Basic premises: residual capacity’ scale concerns view: limitations not checked’ dimension, the less Steenbeek et al. BMC Public Health 2011, 11:1 Page 13 of 18 http://www.biomedcentral.com/1471-2458/11/1 frequently the insurance physician checks what limita- client. The higher the score on the ‘Client approach: tions the client faces. The higher the score on the ‘Inter- time for account of daily activities and reporting’ dimen- view: respond to client’ dimension, the more frequently sion, the more often the insurance physician takes time the insurance physician responds to subjects raised by to thoroughly question the client about his daily activ- the client. ities and to report on this. The higher the score on the The Dutch Test for Conflict Handling [84] measures ‘Client approach: too little time but involved with’ to what extent five strategies are applied for handling dimension, the more likely it is that the insurance physi- cian has too little time to draw up a proper report and conflicts, namely yielding, problem-solving, compromis- to question the client about his daily activities, but feels ing, avoiding and forcing. It is noteworthy that insur- ance physicians evidently did not use all five of these involved with the client. strategies during the disability evaluation. We were able to measure one scale (based on seven items) and two Descriptive results dimensions (calculated from the remaining eight items). Descriptive results for background variables are sum- The higher the score on the ‘Conflict handling: seek marised in table 3. 58.9% of the insurance physicians in compromise’ scale, the more often the physician this survey were men. On average the respondents were searches for a compromise with a client in the event of aged 50.8 years and had 16.2 years of experience. 85.7% a difference of opinion. The higher the score on the were registered and almost two third worked 33 or ‘Conflict handling: engage in confrontation’ dimension, more hours per week and carried out on average 9.1 the more likely the physician is to engage in confronta- disability assessments per week. The patients came from tionwith theclient intheevent of adifferenceofopi- all industries. It is noteworthy that 53.7% of the insur- nion. The higher the score on the ‘Conflict handling: ance physicians reported that a substantial proportion of play down differences’ dimension, the more often the their clients were temporary workers. physician will try to circumvent differences of opinion As regards attitude, table 3 shows that 78% of the and play down their importance. insurance physicians were motivated by the job, 70% Behaviour assessment considered that the social security system was just and We developed three scales and five dimensions (23 of the 24 original items). The higher the score on the Table 3 Background variables ‘Comply with permanent full disability rules’ scale, themorestrictly theinsurance physician complies with % mean sd the permanent full disability rules. The higher the score Gender (%man) 58.9 on the ‘FAL: take account of the client’ scale, the more Age 50.8 7.0 often the physician focuses on the complaints raised by Registered as insurance physician 85.7 the client, what the client can really do, the client’s diffi- Extra medical speciality 15.2 cult home circumstances and limitations experienced by Working hours (week) % up to 24 hrs 16.0 the client. The higher the score on the ‘FAL: consult % 25-32 hors 23.8 with labour expert when not necessary’ scale, the more % 33 hrs or more 60.2 often the insurance physician will consult with the N assessments (week) 9.1 4.0 labour expert in circumstances where the client is Years of experience 16.2 7.7 unable to work or does not belong in the benefits Assessments mainly under WIA 37.7 category. Assessments mainly under WAO 26.4 The higher the score on the ‘FAL and recovery time: Assessments mainly under Wajong 13.0 strict/formalistic approach’ scale, the more the insurance Clients mainly from the agriculture, fishing and food 13.0 industries physician takes a formalistic and strict approach to Clients mainly from the construction and timber 19.5 drawing up the FAL and takes no account of the client industries and his recovery time. The higher the score on the ‘FAL Clients mainly from manufacturing industry 39.4 and recovery time: focus on impairments’ scale, the Clients mainly from the retail and wholesale sectors 41.6 greater the attention which the insurance physician pays Clients mainly from the transport sector 24.2 when drawing up the FAL to limitations caused by Clients mainly from the financial services sector 26.8 impairments, particularly in the light of consistency, and Clients mainly from the temporary work sector 53.7 takes no account of a possible deterioration in the cli- Clients mainly from the health sector 35.1 ent’s health. Clients mainly from the education sector 22.1 The higher the score on the ‘Client approach: involved Clients mainly from the rest of the public sector 13.0 with and time for’ dimension, the more often the insur- Clients mainly from the professions and other sectors 33.8 ance physician takes time for and is involved with the Steenbeek et al. BMC Public Health 2011, 11:1 Page 14 of 18 http://www.biomedcentral.com/1471-2458/11/1 only 53% had a positive opinion about the WIA. The model is an extension of the model designed by Croon results for ‘social norm’ show that insurance physicians and Langius [29] in their study of the process of sickness are mostly influenced by colleagues (66%) and by their certification assessment by social insurance physicians. employer (UWV, 43%) and much less by society/the They took the theory of planned behaviour as a starting public (10%). The results on barriers show that 44% of point. The concept of barriers and stimuli experienced by the insurance physicians experienced substantial work physicians, their own effectiveness and the availability of pressure and 20% substantial emotional workload, 12% sufficient knowledge (concepts which are recognised in were emotionally exhausted and 73% reported that they the ASE model) are also included in their model. Our analysis model divides Croon and Langius’ concept of the viewed the majority of their clients/cases as ‘difficult’. The influence of refresher training and the staff physi- ‘influence of the environment’ into the concept of the cian is viewed as conducive to quality, whereas only 27% social norm (which influences the intention) and barriers of the insurance physicians consider that the manager (which have an intermediary effect between intention promotes quality. As far as knowledge is concerned, less and behaviour). It could be argued that the conceptual than half of the physicians consider that they receive model of the theory of planned behaviour is problematic sufficient information from the occupational physician in that its concepts are not specific enough [92]. We have (company doctor). The scores for intentions show that countered this argument in our proposed model by most insurance physicians intend to carry out the pro- focusing the concepts specifically on the subject of work fession in the manner expected of them as professionals. disablement assessment. As regards behavioural process, we see that three quar- One particular strength of this study is the extent of ters of the physicians are dedicated and inform the cli- the good response to the survey by the insurance physi- ent of the object of the interview and the procedure and cians, which was considerably higher than we had that only 5% indicate that they seek a compromise in expected. A weakness of the study is its cross-sectional theevent of adifferenceofopinionwith theclient. As design, which does not allow for analysis of causal rela- regards behavioural assessment, we see that 72% of the tionships between attitude, social influence, intention insurance physicians follow the rules, 31% consider that and behaviour. Another weakness may be the fact our they have taken account of the client and 63% fre- explanation of measured scales and dimensions in rela- quently consult with the labour expert in circumstances tion to the ASE concepts is only based on theoretical where this is not mandatory, namely in situations of grounds. It is therefore possible that certain scales and ‘medical incapacity for work’ or ‘capacity for own work’. dimensions may not fit in with the ASE concept. Furthermore, the study does not investigate the struc- tural relationships between the measured constructs. Discussion Discussion of the methods Further study is therefore needed in order to demon- In this article we have presented the development of strate whether the ASE model is the best model to instruments for measuring and explaining variations in explain insurance physician’s behaviour. the behaviour of insurance physicians in relation to assessments of functional capacities. Data from 231 Discussion of content questionnaires were analysed and used as a basis for fill- The descriptive results may give rise to some concern. ing the ASE model with 29 scales and 19 dimensions. We see a professional group that is highly motivated We identified scales and dimensions that represent Atti- about the job and positive about the Dutch social secur- tude, Social norm, Self-efficacy, Barriers, Knowledge and ity system. However, only half of them have a positive Intention. We slightly modified the underlying ASE opinion about the Work and Income (Capacity for model by dividing Behaviour into two blocks, the first Work) Act (WIA). The views of the insurance physi- reflecting the process and the second reflecting assess- cians about the social security system and legislation ment-related behaviour. The value of the instruments are, in principle, separate from the manner in which proposed in this article lies in their specificity for insur- they carry out their professional duties and endeavour ance physicians and their sound psychometric character- to achieve a high quality. Furthermore, insurance physi- istics. The extensive literature study, in combination cians experience serious barriers, the most frequent of with the interviews safeguarded the internal validity. which is work pressure. Work pressure, emotional work- While our instruments and the underlying concepts load and emotional exhaustion are positively correlated. show considerable similarities to the study of the com- Finally, 73% of the insurance physicians describe a munication of insurance physicians with their clients majority of their clients/cases as ‘difficult’.Inorder to conducted by Van Rijssen et al. [91], the operationalisa- determine if these scores were relatively high, we com- tion of the underlying concepts was specifically designed pared our outcomes with the same scales of a large to meet the objective of the present study. Our analysis survey among employees (NEA 2008). This comparison Steenbeek et al. BMC Public Health 2011, 11:1 Page 15 of 18 http://www.biomedcentral.com/1471-2458/11/1 reveals that the insurance physicians in this study do not modified form, is suitable for describing the assessment differ significantly from the NEA group ‘physicians, den- behaviour of insurance physicians and the determinants tists and veterinary surgeons (N = 240)’ in terms of of this behaviour. The next step in this line of research work pressure, emotional workload and emotional should be to validate the model using structural equa- exhaustion. Insurance physicians were found to have tion modelling. Finally, the predictive value should be higher levels of autonomy than other ‘physicians, den- tested in relation to work disability assessment tists and veterinary surgeons’.The negative aspectsdo outcomes, i.e. grant or reject the claim. not therefore differ from those of a comparable group of professionals. Additional material The answers to the questionnaire also indicate that insurance physicians are primarily bound, as regards their Additional file 1: Questionnaire insurance physicians. English translation of the original Dutch questionnaire for insurance physicians. professional conduct, by the norms and views of insurance physicians as a professional group. In this way, frameworks are set for the discretionary power of the insurance physi- cians which is necessary in order to do justice in special Acknowledgements This research project has been funded by the Dutch ‘Stichting Instituut Gak’, cases. A fellow insurance physician must be able to come a foundation that initiates and supports innovative projects in the Dutch to the same assessment (reproducibility). welfare sector. Additional funding came from TNO Work and Employment, The management of UWV (Dutch Benefits Insurance Hoofddorp, and the Research Center for Insurance Medicine AMC-UWV- VUmc, the Netherlands. Agency), which focuses above all on the work processes and production, is often seen as setting norms, but is Author details not regarded as supporting the quality of the work. This TNO Work and Employment, PO Box 718, 2130 AS Hoofddorp, the Netherlands. VU University Medical Center, Department of Public and is not a unique finding, but an illustration of the pro- Occupational Health, EMGO Institute for Health and Care Research, blem of managing professionals in general [93]. In his Amsterdam, the Netherlands. UWV, Employee Benefits Insurance Authority, international comparative study into work disability Amsterdam, the Netherlands. Research Center for Insurance Medicine, AMC- UWV-VUmc, Amsterdam, the Netherlands. assessments De Boer [94] also concludes that the profes- sional definition of quality of evaluation of work disabil- Authors’ contributions ity is ‘performance according to professional standards’. RS participated in the design of the study and its coordination, developed the questionnaire, performed the statistical analysis and drafted the He emphasises that in the Netherlands the requirement manuscript. AS participated in the design of the study, participated in of a fair trial is also a central part of the quality of claim developing the questionnaire, performed the statistical analysis and assessment. We see this reflected in our results. The participated in drafting the manuscript. HM participated in the design of the study and its coordination, developed the questionnaire and participated in results for social norms show that the insurance physi- drafting the manuscript. JRA participated in the design of the study and cians attach most importance to the views of their fel- participated in developing the questionnaire. HK participated in the design low professionals and thereafter to those of their of the study and its coordination and provided the logistics for the questionnaire. JB participated in the design of the study and its employer (UWV). They attach the least importance to coordination, participated in developing the questionnaire and drafting the the views of society. Many insurance physicians believe manuscript. All authors read and approved the final manuscript. that the quality of their assessments is positively influ- Competing interests enced by good cooperation with colleagues, refresher All authors declare that there are no financial or other relationships that training and consultation, as well as guidance by staff might lead to a conflict of interest. physicians. Many insurance physicians score highly in Received: 20 April 2010 Accepted: 3 January 2011 terms of following rules during the assessment, so that a Published: 3 January 2011 fair process is possible, while only few insurance physi- cians indicate that their work style is to look for com- References promises in the event of a difference of opinion with the 1. Eddy DM: Variation in physician practice. The role of uncertainty. Health Affairs 1984, 3:74-89. client. This would detract from their independent pro- 2. Eisenberg JM: Physician Utilization. The state of research about fessional status and the requirements of a fair trial. physicians’ practice patterns. Med Care 1985, 23:461-83. Nonetheless, this does not prevent one third of the 3. Wilson JRM, Clarke MG, Ewings P, Graham JD, MacDonagh R: The assessment of patient life-expectancy: how accurate are urologists and insurance physicians from indicating that they take oncologists? BJU Int 2005, 95:794-798. account of the client’s specific circumstances when 4. Zandbelt LC, Smets EMA, Oorta FJ, Godfried MH, De Haes HCJM: drawing up the functional capacity assessment. Determinants of physicians’ patient-centred behaviour in the medical specialist encounter. Soc Sci Med 2006, 63:899-910. 5. Shahinian VB, Kuo Y, Freeman JL, Goodwin JS: Determinants of Androgen Conclusions Deprivation Therapy Use for Prostate Cancer: Role of the Urologist. J The scales and dimensions developed appear to be valid Natl Cancer Inst 2006, 98:839-845. 6. Vogels AGC, Jacobusse GW, Hoekstra F, Brugman E, Crone M, Rijneveld SA: and offer a promising basis for future research. The Identification of children with psychosocial problems differed between results suggest that the underlying ASE model, in Steenbeek et al. BMC Public Health 2011, 11:1 Page 16 of 18 http://www.biomedcentral.com/1471-2458/11/1 preventive child health care professionals. J Clin Epidemiol 2008, 31. Van Oostrom, Anema JR, Terluin B, Vet HCW de, Knol DL, Van Mechelen W: 61:1144-1151. Cost-effectiveness of a workplace intervention for sick-listed employees 7. Chang LW, Fung TY, Leung TY, Sahota DS, Lau TK: Volumetric (3D) with common mental disorders: design of a randomized controlled trial. imaging reduces inter- and intraobserver variation of fetal biometry BMC Public Health 2008, 8:12. measurements. Ultrasound Obstet Gynecol 2009, 33:447-452. 32. Vermeulen SJ, Anema JR, Schellart AJ, Van Mechelen W, Van der Beek AJ: 8. Marinus AMF: Interdoktervariatie in de huisartsenpraktijk (Inter-doctor Intervention mapping for development of a participatory return-to-work variation in GPs’ practices). PhD thesis University of Amsterdam; 1993. intervention for temporary agency workers and unemployed workers 9. Van der Weijden T, Hutten JBF, Brandenburg BJ, Grol RPTM, Van der sick-listed due to musculoskeletal disorders. BMC Public Health 2009, Velden K: Cholesterol management in Dutch general practice. A 9:216. comparison with national guidelines. Scand J Prim Health Care 1994, 33. De Vries H: Determinanten van gedrag. (Determinants of behaviour). In 12:281-288. Gezondheidsvoorlichting en gedragsverandering. Edited by: Damoiseaux V, 10. Janssen HAM, Borghouts JAJ, Muris JWM, Metsemakers JFM, Koes BW, Van der Molen H, Kok GJ. Assen: Van Gorcum; 1993:109-132. Knottnerus JA: Health status and management of chronic non-specific 34. De Vries H, Mudde AN: Predicting stage transitions for smoking cessation abdominal complaints in general practice. Br J Gen Pract 2000, applying the attitude-social influence-efficacy model. Psychol Health 1998, 50:375-379. 13:369-385. 11. Grytten J, Sørensen R: Practice variation and physician-specific effects. J 35. Tellnes G, Sandvik L, Moum T: Inter-doctor variation in sickness Health Econ 2003, 22:403-418. certification. Scan J Prim Health Care 1990, 8:45-52. 12. Verstappen WHJM, Ter Riet G, Dubois WI, Winkens R, Grola RTPM, Van der 36. Thies-Zajong S, Szecsenyi J, Kochen MM: Attitudes and empirical referral Weijden T: Variation in test ordering behaviour of GPs: professional or data of West German family physicians. Gesundheitswesen 1993, context-related factors? Fam Pract 2004, 21:387-395. 55:635-640. 13. Taylor RJ: General-practitioner prescribing. J R Coll Gen Prac 1977, 37. Besseling JJM, Bockting AJV, Franquinet JMWAF, Sprenger WJ: Evaluatie 27:79-82. stelselherziening (System revision evaluation) Amsterdam: GMD (Joint Medical 14. Martens JD, Van der Weijden T, Severens JL, De Clercq PA: The effect of Examination Service); 1990. computer reminders on GPs’ prescribing behaviour: A cluster- 38. Van de Goor AG: Effects of regulation on disability duration. PhD thesis randomised trial. Int J Med Iinfor 2007, 76(Suppl 3):403-416. Amsterdam, Thesis Publishers; 1997. 15. Delnoij DMJ, Spreeuwenberg PMM: Variation in GPs’ referral rates to 39. Aarts L, De Jong P, Van der Veen R: Met de beste bedoelingen wao 1975- specialists in internal medicine. Eur J Public Health 1997, 7:427-435. 1999 trends, onderzoek en beleid. (With the nest intentions WAO 1975-1999: 16. Ydreborg BAM, Ekberg K: Disqualified for disability pension -a case/ trends, research and policies) Doetinchem: Elsevier bedrijfsinformatie; 2002. referent study. Disabil Rehabil 2004, 26:1079-1086. 40. Wevers CWJ, Vinke H: Bedrijfsartsen over verzuimbegeleiding, een 17. Spanjer J, Krol B, Groothoff JW: Inter-rater reliability in disability casusonderzoek (Company doctors on sick leave management; a case assessment based on a semi-structured interview report. Disabil Rehabil study). TBV 1999, 7:78-83. 2008, 30:1885-1890. 41. De la Court E: WAO volume-effect van de eerstejaarsherbeoordeling en zijn 18. Spanjer J, Krol B, Popping R, Groothoff JW, Brouwer S: Disability disability determinanten: reïntegratie in het eerste WAO jaar (WAO volume effect of the evaluation: the role of detailed information on functioning in addition first-year reassessment and its determinants: reintegraton in the first WAO to medical history taking. J Rehabil Med 2009, 41:267-272. year) Hoofddorp: TNO Arbeid; 2000. 19. Arrelov B, Borgquist L, Ljungberg D, Svardsudd K: Do GPs sick-list patients 42. Norrmen G, Svardsudd K, Andersson D: Impact of physician-related factors to a lesser extent than other physician categories? A population based on sickness certification in primary health care. Scand J Prim Health Care study. Fam Pract 2001, 18:393-398. 2006, 24:104-109. 20. Shiels C, Gabbay M: The influence of GP and patient gender interaction 43. Kerstholt JH, De Boer WEL, Jansen EJM: Psychologische aspecten van on the duration of certified sickness absence. Fam Pract 2006, 23:246-253. claimbeoordeling (Psychological aspects of claims assessment) Hoofddorp: 21. Hofstee WKBH, Kroneman H, De Boer WEL: Representatieve beoordeling TNO Arbeid; 2002. van arbeidsvermogen (Representative assessment of work capacity). TBV 44. Laitinen-Krispijn SM, Nicolaï LC: Verzekeringsgeneeskundige beoordeling 2009, 17:406-409. bij psychische problematiek (Insurance medicine assessment in the case 22. Fishbein M, Ajzen I: Belief, attitude, intention and behaviour; An introduction of mental problems). TBV 2006, 14:167-170. to theory and reserch Reading (MA): Addison-Wesley; 1975. 45. Spanjer J: De inter- en intra-beoordelaarsbetrouwbaarheid van WAO- 23. Ajzen I, Fishbein M: Understanding attitudes and predicting social behavior beoordelingen (The inter-rater and intra-rater reliability of WAO Englewood Cliffs NJ: Prentice Hall; 1980. assessments). TBV 2001, 9:234-241. 24. Ajzen I: The Theory of Planned Behavior. Organ Behav Hum Decis Process 46. Razenberg PPA: Verzekeringsgeneeskundige oordeelsvorming. Inzicht in 1991, 50:179-211. de praktijk (Insurance medicine assessments. Insight into practice). PhD 25. Bandura A: Social foundations of thought and action Englewood Cliffs NJ: thesis University of Amsterdam; 1992. Prentice Hall; 1986. 47. De Jong JD, Westert GP, Lagoe R, Groenewegen PP: Variation in Hospital 26. De Vries H, Dijkstra M, Kuhlman P: Self-efficacy: the third factor besides Length of Stay: Do Physicians Adapt Their Length of Stay. Decisions to attitude and subjective norm as a predictor of behavorial intention. What Is Usual in the Hospital Where They Work? Health Serv Res 2006, Health Educ Res 1988, 3:273-282. 41:374-394. 27. De Vries H, Backbier E, Kok GJ, Dijkstra M: Measuring the impact of social 48. Solomon DH, Brookhart MA, Gandhi TK, Karson A, Gharib S, Orav J, influences on smoking onset in a longitudinal study: an integration of Shaykevich S, Licari A, Cabral D, Bates DW: Adherence with Osteoporosis social psychological approaches. J Appl Soc Psychol 1995, 25:237-257. Practice Guidelines: A Multilevel Analysis of Patient, Physician, and 28. Rammeloo KC, Tamminga SJ, Anema JR, Schellart AJM: De implementatie Practice Setting Characteristics. Am J Med 2004, 117:919-924. van een verzekeringsgeneeskundig protocol Gebruik van Intervention 49. Van Deursen CGL: Claimbeoordeling WAO: het handelen van Mapping bij het ontwikkelen van een implementatiestrategie voor het verzekeringsartsen in de praktijk (WAO Claims assessment: the actions of verzekeringsgeneeskundig protocol Depressie (The implementation of insurance physicians in practice) Amsterdam: National Institute of Social an insurance medicine protocol for the use of intervention mapping in Insurance Agencies (Lisv); 2000. the development of a strategy for the depression protocol). TBV 2009, 50. Van Deursen, Veerman TJ: Randvoorwaarden van de sociaal-medische 17:95-102. begeleiding : onderzoek onder verzekeringsartsen en arbeidsdeskundigen die lid 29. Croon NHTh, Langius SWTh: (Work and health. A study about the zijn van het Lisv uitvoerderspanel (Parameters for sociomedical counselling: practices in social insurance medicine at the work insurance boards). survey among insurance physicians and labour experts who are members of PhD Thesis University of Amsterdam, Institute for Social Medicine; 1993. the Lisv implementation panel) Amsterdam: National Institute of Social 30. Rebergen D, Hoenen J, Heinemans A, Bruinvels D, Bakker A, Van Insurance Agencies (Lisv); 1997. Mechelen W: Adherence to mental health guidelines by Dutch 51. Spanjer J: Tien stellingen rond herstelgedrag (Ten propositions occupational physicians. Occup Med 2006, 56:461-468. concerning recovery behavior). TBV 1999, 7:233-235. Steenbeek et al. BMC Public Health 2011, 11:1 Page 17 of 18 http://www.biomedcentral.com/1471-2458/11/1 52. Besseling JJM: Burgerlijke staat en arbeidsongeschikheid 1977-1983 (Marital 70. Kerstholt JH, De Boer WEL, Jansen EJM: Disability assessments: Effects of status and work disability) Amsterdam: GMD (Joint Medical Examination response mode and experience. Disabil Rehabil 2006, 28:111-115. Service); 1986. 71. Spanjer J: In De reproduceerbaarheid van WAO-beoordelingen, een 53. Moore PJ, Adler NE, Robertson PA: Medical malpractice: the effect of literatuuronderzoek (The reproducability of WAO assessments, a literature doctor-patient relations on medical patient perceptions and malpractice study). Volume 10. TBV; 2002:195-198. intentions. West J Med 2000, 173:244-250. 72. Spanjer J: In Voorbeelden: Een hulpmiddel om de ernst van de beperkingen te 54. Mudde EC: Ziek in de zin der wet. De interpretatie van ziekteverzuim beoordelen (Examples: an aid for assessing the seriousness of limitations). door verzekeringsgeneeskundigen en rapporteurs (Sickness within the Volume 10. TBV; 2002:109-111. meaning of the law. The interpretation of sick leave by insurance 73. Schaufeli WB, Salanova M, González-Romá V, Bakker AB: The measurement physicians and rapporteurs). PhD thesis University of Amsterdam; 1995. of engagement and emotional exhaustion: A two sample confirmatory 55. Weijman I, Ros WJG, Rutten GEHM, Schaufeli WB, Schabracq MJ, factor analytic approach. J Happiness Stud 2002, 3:71-92. Winnubst JAM: Fatigue in employees with diabetes: its relation with 74. De Bont AA, Berendsen L, Boonk MPA, Van de Brink JC: In de spreekkamer work characteristics and diabetes related burden. Occup Environ Med van de verzekeringsarts : een onderzoek naar het verzekeringsgeneeskundige 2003, 60(suppl 1):93-98. deel van de WAO claimbeoordeling.(In the surgery of the insurer’s medical 56. De Croon EM, Sluiter JK, Nijssen TF, Kammerijer M, Dijkmans BA, adviser: a study of the insurance medicine part of the WAO claims assessment) Lankhorst GJ, Frings-Dresen MH: Work ability of Dutch employees with Zoetermeer: CTSV; 2000. rheumatoid arthritis. Scand J Rheumatol 2005, 34:277-283. 75. Maas-Oostenbrink AJ, Simon-de Zwart AJ: Standaard ‘Geen Duurzaam 57. Ybema JF, Evers M: Profiel Arbeid en Gezondheid 2005 (Labour and Health Benutbare Mogelijkheden’ altijd benutbaar? een onderzoek naar het gebruik Profile 2005) Hoofddorp: TNO Work and Employment; 2005. van de standaard GDBM bij maligne aandoeningen met een infauste 58. Berendsen L, Mullenders P: De manager de baas? Een onderzoek naar prognose (Permanent Full Disability Standard always useable? A study of the WAO-managers en hun integrale verantwoordelijkheid voor het werk use of the standard in the case of malignant disorders with an unfavourable van verzekeringsartsen (Is the manager the boss? A study of WAO prognosis) Hoofddorp: TNO Arbeid; 1999. managers and their integral responsibility for the work of insurance 76. Van Dijk JK, Messchendorp HJ, Koopman MI, Steenbeek R, Van Til CT: physicians). Den Haag: Inspectie Werk en Inkomen; 2004. Personeel in beeld. De arbeidsbeleving van medewerkers in ziekenhuizen, 59. Choy J: Rapportage van de raadplegingen door bureau SMZ met de revalidatiecentra en andere instellingen die vallen onder de CAO-Z (Personnel verzekeringsartsen en arbeidsdeskundigen (Reports of assessments by bureau in the picture. The work perception of employees in hospitals, rehabilitation SMZ with insurance physicians and labour specialists) Amsterdam: Bureau centres and other institutions covered by the ‘Z’ collective agreement) Utrecht: SMZ; 2005. Prismant; 2002. 60. Edlund C, Dahlgren L: The physician’s role in the vocational rehabilitation 77. Scholtz U, Gutierrez Dona B, Sud S, Schwarzer R: Is general self-efficacy az process. Disabil Rehabil 2002, 24:727-733. Universal construct? Psychometric findings from 25 countries. Eur J 61. Willems JHBM, Kroneman H: In Arts of advocaat? Nieuwe wao verandert de Psychol Assess 2002, 18:242-251. rol van verzekeringsarts (Physician or lawyer? New WAO alters the role of the 78. Smulders PGW, Andries F, Otten F: Hoe denken Nederlanders over hun werk? insurance physician). Volume 9. MDC; 2003:337-340. Opzet, kwaliteit en eerste resultaten van de TNO Arbeidssituatie Survey. (What 62. De Bont A, Van den Brink JC, Berendsen L, Boonk M: In De beperkte controle do Dutch people think about their work? Design, quality and first results of the van de informatie voor de arbeidsongeschiktheidsbeoordeling (The limited TNO Working Conditions Survey) Hoofddorp: TNO Arbeid; 2001. checks on information for work disability assessment). Volume 146. Ned 79. Kristensen T, Borg V: Copenhagen Psychosocial Questionnaire (COP-SOQ) Tijdschr Geneeskd; 2002:27-30. Kopen-hagen: National Institute of Occupational Health; 2000. 63. Van Glabbeek RHJ: Houden verzekeringsartsen zich aan de wet? Het 80. Goudswaard A, Dhondt S, Kraan K: Flexibilisering en Arbeid in de Informatie- Schattingsbesluit en de Standaard Geen Duurzaam Benutbare Mogelijkheden maatschappij; werknemersvragenlijst, bestemd voor werknemers van bedrijven in de praktijk (Do insurance physicians comply with the law? The Estimation die deelnemen aan het SZW-Werkgeverspanel 1998 (Flexibilisation and Work in Decree and the Permanent Full Disability Standard) Dissertation for the the Information Society; questionnaire for employees of businesses that take occupational health and safety training course for doctors) Netherlands part in the 1998 Employers Panel of the Ministry of Social Affairs and School of Public and Occupational Health (NSPOH): Utrecht; 2003. Employment) Hoofddorp: TNO Arbeid; 1998. 64. Sprenger WJ: Arbeidsdeskundigen en verzekeringsgeneeskundigen van de 81. Houtman I, Goudswaard A, Dhondt S, Van der Grinten M, Hildebrandt V, gemeenschappelijke medische dienst over de gevolgen van de stelselherziening Van der Poel EGT: Evaluation of the monitoring study of stress and physical voor hun praktijk (Labour experts and insurance physicians of the Joint load The Hague: VUGA; 1995. Medical Examination Service (GMD) on how a revision of the system would 82. Schaufeli WB, Van Dierendonck D: The construct validity of two emotional affect their practice) Amsterdam: GMD (Joint Medical Examination Service); exhaustion measures. J Organ Behav 1993, 14:631-47. 1989. 83. De Boer WEL, Wijers J, Spanjer J, Van der Beijl I, Zuidam W: 65. De Boer WEL, Croon NHTh, Van der Toorn M, Csánky-Achilles HW: In Gespreksmodellen in de verzekeringsgeneeskunde (Discussion models Enquête Medisch Arbeidsongeschiktheidscriterium (Medical Work Disability in insurance medicine). TBV 2006, 14:17-23. Criterion Survey). Volume 4. TBV; 1996:98-104. 84. De Dreu CKW, Evers A, Beersman B, Kluwer ES, Nauta A: A theory based 66. Brenninkmeijer V, Lagerveld SE, Blonk RWB: In Moeilijk objectiveerbare measure of conflict management strategies in the workplace. J Organ klachten in de praktijk van de bedrijfs- en verzekeringsarts. Een empirisch Behav 2001, 22:645-668. onderzoek (Complaints in the practice of company doctors and insurer’s 85. Nagtegaal GH: De rol van het dagverhaal in de oordeelsvorming van de medical advisers that are hard to assess objectively: an empirical study). verzekeringsarts ten aanzien van AAW/WAO beoordelingen (The role of the Volume 14. TBV; 2006:354-359. client’s account of daily activities in assessment by insurance physicians for the 67. Van Berkel NA, Njoo T: Overleg met de behandelende sector: het effect van purposes of the AAW/WAO) Utrecht: Netherlands School of Public Health informatie van de behandelaar op de WAO-beoordeling, vergelijkend (NSPH); 1997. 86. SPSS 15.01 Chicago (Il): SPSS Inc; 2006. onderzoek tussen twee UVI’s. (Consultation with the treating sector: the effect of information from the treating physician on the WAO assessment: 87. Du Toit M, Du Toit S: Interactive LISREL: User’s guide Lincolnwood (Il): comparative surveys between two social security administration agencies) Scientific Software International Inc; 2001. Amsterdam: Netherlands School of Occupational Health (NSOH); 2001. 88. Jöreskog KG, Sörbom D: Interactive Lisrel (version 8.72) Lincolnwood (Il): 68. Mortelmans AK, Donceel P, Lahaye D, Bulterys S: Does enhanced Scientific Software International Inc; 2004. information exchange between social insurance physicians and 89. SPSS Categories Chicago (Il): SPSS Inc; 1990. occupational physicians improve patient work resumption? A controlled 90. Gifi A: Homals users Guide Department of Datatheory, University of Leiden; intervention study. Occup Environ Med 2006, 63:495-502. 1981. 69. Davis P, Gribben B, Scott A, Yee RL: The ``supply hypothesis’’ and medical 91. Van Rijssen HJ, Schellart AJM, Anema JR, Van der Beek AJ: A theoretical practice variation in primary care: testing economic and clinical models framework to describe communication processes during medical of inter-practitioner variation. Soc Sci Med 2000, 50:407-418. disability disability evaluations. BMC Public Health 2009, 9:375. Steenbeek et al. BMC Public Health 2011, 11:1 Page 18 of 18 http://www.biomedcentral.com/1471-2458/11/1 92. Ogden J: Some problems with social cognition models: a pragmatic and conceptual analysis. Health Psychol 2003, 22:424-428. 93. Weggemans M: Leiding geven aan professionals? Niet doen! (Providing leadership to professionals. Don’t do it!) Schiedam: Scriptum; 2008. 94. De Boer WEL: Quality of evaluation of work disability. PhD Thesis University of Amsterdam; 2010. Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1471-2458/11/1/prepub doi:10.1186/1471-2458-11-1 Cite this article as: Steenbeek et al.: The development of instruments to measure the work disability assessment behaviour of insurance physicians. BMC Public Health 2011 11:1. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Public Health Springer Journals

The development of instruments to measure the work disability assessment behaviour of insurance physicians

Loading next page...
 
/lp/springer-journals/the-development-of-instruments-to-measure-the-work-disability-J5O7wR1maF

References (233)

Publisher
Springer Journals
Copyright
Copyright © 2011 by Steenbeek et al; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Public Health; Medicine/Public Health, general; Epidemiology; Environmental Health; Biostatistics; Vaccine
eISSN
1471-2458
DOI
10.1186/1471-2458-11-1
pmid
21199570
Publisher site
See Article on Publisher Site

Abstract

Background: Variation in assessments is a universal given, and work disability assessments by insurance physicians are no exception. Little is known about the considerations and views of insurance physicians that may partly explain such variation. On the basis of the Attitude - Social norm - self Efficacy (ASE) model, we have developed measurement instruments for assessment behaviour and its determinants. Methods: Based on theory and interviews with insurance physicians the questionnaire included blocks of items concerning background variables, intentions, attitudes, social norms, self-efficacy, knowledge, barriers and behaviour of the insurance physicians in relation to work disability assessment issues. The responses of 231 insurance physicians were suitable for further analysis. Factor analysis and reliability analysis were used to form scale variables and homogeneity analysis was used to form dimension variables. Thus, we included 169 of the 177 original items. Results: Factor analysis and reliability analysis yielded 29 scales with sufficient reliability. Homogeneity analysis yielded 19 dimensions. Scales and dimensions fitted with the concepts of the ASE model. We slightly modified the ASE model by dividing behaviour into two blocks: behaviour that reflects the assessment process and behaviour that reflects assessment behaviour. The picture that emerged from the descriptive results was of a group of physicians who were motivated in their job and positive about the Dutch social security system in general. However, only half of them had a positive opinion about the Dutch Work and Income (Capacity for Work) Act (WIA). They also reported serious barriers, the most common of which was work pressure. Finally, 73% of the insurance physicians described the majority of their cases as ‘difficult’. Conclusions: The scales and dimensions developed appear to be valid and offer a promising basis for future research. The results suggest that the underlying ASE model, in modified form, is suitable for describing the assessment behaviour of insurance physicians and the determinants of this behaviour. The next step in this line of research should be to validate the model using structural equation modelling. Finally, the predictive value should be tested in relation to outcome measurements of work disability assessments. Background occur in the diagnosis and treatment of patients. A Variation in assessments by professionals is a well- degree of uncertainty is inherent in the profession and it known phenomenon which occurs in cases where is very easy to reach different conclusions in comparable assessments are carried out by several raters and in var- cases [1,2]. Inter-doctor variation in diagnosis and/or ious disciplines. In the case of physicians, variations treatment is found in different medical disciplines [3-7]. Specific research into variation among GPs shows varia- tion in diagnosis [8,9], request for interventions * Correspondence: romy.steenbeek@tno.nl [8,10-12], treatment [13,14] and rate of referral to spe- TNO Work and Employment, PO Box 718, 2130 AS Hoofddorp, the cialists [15]. Literature on insurance physicians is less Netherlands Full list of author information is available at the end of the article © 2011 Steenbeek et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Steenbeek et al. BMC Public Health 2011, 11:1 Page 2 of 18 http://www.biomedcentral.com/1471-2458/11/1 extensive but variation in the assessment of functional 1970-1980 by Fishbein and Ajzen in their ‘Theory of capacity for work exists [16-18]. In this study we con- Reasoned Action (TRA)’ [22,23]. In the 1980 s this the- centrate on how insurance physicians assess workers’ ory was taken a step further by Azjen [24] in his ‘Theory claims to compensation for loss of work capacity. Insur- of Planned Behaviour (TPB)’. He added ‘perceived beha- ance physicians have to judge a claim and in doing so vioural control’ (self-efficacy) as a factor that moderates base themselves on the available information (file) and behaviour. In the second half of the 1980 s Azjen’s information provided by the “client”, i.e. the patient who model was supplemented by elements from the ‘Social claims a work disability benefit, and others. The out- CognitiveTheory(SCT)’ of Bandura[25] in theso- called ASE model [26]. ASE is a model that has general come of the assessment, i.e. the functional capacity for work, is variable. An important aspect of this assessment scientific acceptance and explains behaviour by linking is the client’s medical situation in the context of the attitude, social influence and self-efficacy with behaviour current social security legislation. In making the assess- and behavioural intention [27]. In addition to the three ment the insurance physician must therefore deal with determinants of intention and behaviour, intermediary the characteristics of both the legal and the medical factors such as ‘knowledge’ and ‘barriers’ can play a role. decision-making process. Assessing work disability is TPB and the ASE model are used in the Netherlands to therefore a complex and specialised process that also explain, among other things, the behaviour of physicians gives rise to variation in outcomes. Ydreborg and Ekberg [28] and patients in an occupational health context [16] found variations in the extent to which applicants [29-32] and the health behaviour of individuals who for disability pension were rejected in practice. Spanjer belong to a particular target group [33,34]. On the basis et al. [17] evaluated inter-rater reliability between insur- of TPB research Croon & Langius [29] have studied atti- ance physicians in respect of physical disability and tudes and working styles (behavioural intentions) among mental disability assessment as reasonable to good. insurance physicians who assessed employees’ sick leave However, inter-rater reliability in respect of the assess- not exceeding one year. The present survey takes the ment of the number of hours clients could function ASE model as the basis for possible explanations of the daily was low. Spanjer et al. [18] found a significant dif- behaviour of insurance physicians in assessing work dis- ference in various scores on assessed work limitation ability after sick leave lasting one year (Disability Insur- items by insurance physicians. ance Act -WAO), two years (Work and Income The outcome of work disability assessments by insur- (Capacity for Work) Act - WIA) or more years ance physicians can be seen as the result of behaviour (Adapted Reassessment Act - HERBO, see next influenced by various factors, including behavioural paragraph). determinants of the physicians in relation to the In the Netherlands, if you are partially or fully incap- intended object of their assessment. Little is known of able of working after two years of illness, you may be what considerations and views of insurance physicians eligible to receive a benefit under the Work and Income may partly account for variation in the outcome of (Capacity for Work) Act (WIA). The WIA succeeded assessments. the Disability Insurance Act (WAO) in January 2006. The WAO was not repealed by the WIA, but now Conceptualisation applies only to those who were already receiving a It is evident from the above that variation generally WAO benefit on 1 January 2006. The Adapted occurs in assessments by physicians. Research shows Re-assessment Act (HERBO) was introduced in August that this variation is connected with, among other 2004 for the reassessment of WAO benefits clients, i.e. things, certain personal characteristics [13,19,20] and the claimants (< 50 years), on the basis of new, stricter behavioural characteristics, such as personal style or criteria that put the emphasis on the client’sresidual attitude [11,13,21]. Systematic variation between insur- functional capacities. These stricter assessment rules ance physicians in the outcomes of assessments (i.e. under HERBO also apply to the WIA. Young disabled grant or reject the claim) can be regarded as the result people may be eligible to receive a benefit under the of assessment behaviour which - in addition to other Invalidity Insurance (Young Disabled Persons) Act factors - is determined in part by the attitude which the (Wajong). The WAO and WIA differ in the time of insurance physician has towards the intended purpose assessment. The WAO provides for assessments after of his assessment. Thus, assessment behaviour is defined one year of illness, whereas the WIA provides for assess- as all behaviour that may influence the outcome of the ments after two years of illness. In this study, we will assessment, including the collection and evaluation of only use disability assessment outcomes under WAO, information about the client. HERBO and WIA. For parts of the ASE model (see figure 1) research has A theory of the relationship between attitudes and shown that there is a correlation with variation in behaviour was elaborated in the literature of the period Steenbeek et al. BMC Public Health 2011, 11:1 Page 3 of 18 http://www.biomedcentral.com/1471-2458/11/1 Barriers Attitude Assessment Social Norm Intention behaviour Self-efficacy Knowledge Background Figure 1 The ASE model. assessments by physicians. Research among GPs shows that few evidence-based and peer-reviewed articles exist. varying conclusions as regards attitude. Taylor con- This is why we also used ‘grey literature’. Literature on cluded as long ago as 1977 in a review [13] that varia- insurance physician’s behaviour is scarce. That is why tion in prescribing behaviour was associated with the we added research on GPs, because in many countries personal attitudes of GPs. In addition, the results of the GP is the physician issuing sickness certificates. Grytten and Sorensen [11] indicate that practice style reflects a deeply rooted behaviour with respect to how Background variables to practise medicine. However, Tellnes et al. [35] and The gender of the physician appears to influence assess- Thies-Zajong et al. [36] found no association between ment behaviour. Female insurance physicians more fre- doctors’ attitudes and some measures of variation. Social quently restrict the number of hours a client can work norms also appear to play a role. The degree to which as part of their assessment [40]. Court [41] showed this insurance physicians are required to achieve a certain with specific regard to Dutch insurance physicians: male level of production and the pressure exerted on them to insurance physicians conclude less frequently than achieve the targets differ from office to office [37-39]. female insurance physicians that clients are not able to No literature was found on the relationship between work. self-efficacy and assessments by physicians. Research Age is closely connected with experience: the older among GPs shows that a higher workload (operationa- the doctor the more experience he has. Both factors lised as list size) functions as a barrier; it leads to more influence assessment behaviour. Physicians with more referrals to specialists [15]. A lack of sufficient knowl- experience in family medicine issued more sickness cer- edge and information can also serve as a barrier. Davis tificates [42] and the duration of episodes of sickness et al. [40] found that diagnostic uncertainty in GPs was certification was longer for patients of older doctors associated with higher rates of investigation and follow- [35]. In the case of insurance physicians, greater experi- up. Finally, client or patient characteristics such as gen- ence is associated with greater optimism about the der [20], unemployment and age [16] or the type of patient’s return to work [43], better quality assessment patient visit [4] were found to be associated with varia- in the case of mental complaints [44], more frequent tions in outcome. allowance of a reduction in working hours [45], a ten- dency to assess suitability for work as higher in the Applying the conceptualisation event of reassessments [43] and a shift in thinking from Below we describe the factors in respect of which an the ‘seriousness of the complaint’ to the ‘extent to association has been found with variations in assess- which the complaints influence daily functioning’ [46]. ments by physicians for each concept from the ASE Training [35,42] and specialisation [19] are also con- model and which have been included in our model. Our nected with the assessment behaviour of physicians. search for literature on the specific relationship between Physicians who were trained in social insurance medi- certain factors and variation in assessments revealed cine as undergraduates issued more sickness certificates Steenbeek et al. BMC Public Health 2011, 11:1 Page 4 of 18 http://www.biomedcentral.com/1471-2458/11/1 [42]. The duration of episodes of sickness certification Mooreetal. [53] concludedthat fearofclaimscan was shorter in patients of doctors with postgraduate influence the relationship with the client and hence the training [35]. decision as well. If the physician is afraid that the client The number of working hours is also a factor that will appeal, he will be less inclined to rule against the influences sickness certification: physicians working part client’s wishes [54]. Insurance physicians who are con- time issued more sickness certificates than physicians fronted by stricter medical professional norms are more working full time [42]. inclined to decide that the client is no longer able to Assessments also vary according to locations or perform any work whatever [38]. region: the variation between these units has been shown to be greater than within them [12,16,47,48]. Self-efficacy Self-efficacy (a person’s belief about his ability and capa- Attitude city to accomplish a task) moderates behaviour [24]. Taylor concluded as long ago as 1977 [13] that more desirable prescribing patterns by GPs were associated Barriers with a more psychosocial orientation towards medical Work pressure, autonomy, emotional workload and care. In case of insurance physicians, desirable behaviour emotional exhaustion are known to be related to health could be associated with perceived justness of the social complaints and influence job participation [55-57]. security system. Office culture can strongly influence job satisfaction and The attitude towards the perceived quality of the hence either hinder or stimulate the physicians in their assessment may also influence it. Insurance physicians work. are dissatisfied about the scope for development and Factors connected with quality are also important. In refresher training and complain about a lack of clear the case of insurance physicians, managerial emphasis instructions in the legislation, regulations and instru- on quantity can act as a barrier [58]. Some physicians ments [49,50]. Views on quality are also influenced by consider that the quality of the assessments suffers from increased production pressure and the feeling of having work pressure and changes in the organisation. Lack of too little time for the assessments [41]. time means that duties are carried out with less care Finally, the physician’s attitude towards the personal [49,50]. Choij [59] concludes that the attitude of profes- needs and circumstances of the client (including his sionals to the organisation is negative. They feel that recovery) can influence the assessment. Physicians tend they are not heard, recognised or appreciated. Manage- to have fairly differing views on the time a client needs ment is seen as the major culprit because it exerts for recovery and also differ in the extent to which they undue production pressure. Stricter requirements and take account of personal circumstances. Some take no an increase in the number of clients with more serious account of this and others apply as a criterion that the sicknesses reduces the physician’s feelings of autonomy client should not be exhausted after finishing work or and increases stress through time pressure [60]. Guide- that theclientshouldbeproperlyrestedwhenhe lines and protocols can be an aid in enhancing quality resumes work the next day [49,51]. [61]. However, physicians consider that they receive too little guidance in applying guidelines for disorders that Social norm are difficult to assess objectively [62] or in applying leg- In the case of insurance physicians the influence of islation, regulations and instruments [50]. 40% of the social norms may emanate from the office and the files are not in keeping (or completely in keeping) with environment. Within the office the norm for the ‘strict- the statutory requirements [63]. ness’ of the assessments may differ [37]. Other factors More difficult assessments may result in more varia- are how much pressure is brought to bear in terms of tion between insurance physicians. The assessment is production and promptness targets and how strictly perceived as more difficult in relation to certain groups these are checked [38,39]. of clients. This applies to older clients [37,64], clients Changing social norms can influence assessments [52]. with mental problems [37,39], clients with impairments The stricter the norm that states that society’s interest that are difficult to determine objectively [65,66] and cli- should be guarded, the higher the probability that insur- ents with psychosomatic disorders [64]. Interviews with ance physicians will find a sufficient job for the client insurance physicians during the preparation of this and the higher the residual earning capacity. However, if questionnaire also showed that they regard assessment the norm to which insurance physicians are subject of the following categories as extra difficult: clients with states that society is responsible for its citizens if they a poor command of Dutch, clients who act aggressively become incapacitated for work, the probability that suf- or manipulatively, and cases where poor preliminary work has been done by the occupational health service. ficient jobs are available will be lower [38]. In addition, Steenbeek et al. BMC Public Health 2011, 11:1 Page 5 of 18 http://www.biomedcentral.com/1471-2458/11/1 Knowledge insurance physician must determine the client’s capacity Deciding whether or not to request and use information for work. This is done by reference to an instrument of third parties may influence the assessment, and this known as FAL (Functional Ability List). On this list the is done to a very varied extent [62,67,68]. Davis et al. physician enters the client’s scores for limitations and [69] found that diagnostic uncertainty among GPs was abilities. These findings serve as the input for the labour associated with higher rates of investigation and follow- expert in determining the extent to which the client is up. Having more information increased the physician’s able to earn income and able to work. As an instrument self-confidence [43]. However, Kerstholt et al. [70] sug- the FAL comes within the statutory framework of dis- gest that assessments of disability are largely based on ability assessments in the Netherlands. If the client is the initial view formed after reading the file. The main permanently and fully disabled, he is classified as such pitfall is that the final view is based on general beliefs (in the Netherlands known as GBM). This is therefore a rather than on actual client information. The insurance dichotomy measurement. The literature shows that stat- physician’s job experience, competencies and interests utory rules are implemented in different ways [74]. determine in part how difficult it is to ‘translate’ mental Insurance physicians interpret the guideline on perma- complaints into limitations and residual capacities [44]. nent disability in a wide variety of ways [49,75]. The Physicians need additional instruments, knowledge and definition given in the guideline - namely an incapacity experience in relation to disorders that are difficult to to function socially and personally - is considered inade- assess objectively [66]. quate. The criterion of permanent full disability is also used as a safety net in order to compensate for loss of Intention income in difficult situations. The possibility of indicat- Studies by De Boer et al. [65] and Spanjer [71,72] show ing that the client can work a limited number of hours that the object, the physician’s interpretation of his (limitation of hours) is also applied in a wide variety of duties and the basic premises have an important bearing ways [45,49]. Likewise, the time required for recovery is on the proposed assessment. assessed in a variety of ways [49]. The above concepts are described in the model under ‘Behaviour: Behaviour assessment’. We will distinguish two types of behaviour: 1) behaviour that reflects the process of assessment, such as the col- Aim lection of information about the client and 2) behaviour The first aim of this study was to develop measurement directly connected with the assessment itself, such as instruments that can potentially affect disability assess- the use of assessment instruments in order to evaluate ment behaviour by insurance physicians. The descrip- the information. tion of a conceptual model includes the relevant Spanjer [18] found that process variables hardly (behavioural) variables. We expect that this model will affected assessment outcomes. Information on participa- contribute towards understanding and explaining varia- tion and activity limitations provided by the patient had tion in assessments of functional capacity by insurance only limited influence on inter-rater reliability by insur- physicians. A second aim was to discuss the descriptive ance physicians. However, there was a significant differ- results of the insurance physicians’ scores on the differ- ence in scores on assessed work limitation items ent concepts. In accordance with these aims, the compared with medical history-taking alone. It follows research questions were: 1) can we construct measure- that in disability assessment interviews physicians should ment instruments to measure assessment behaviour and ask for medical information as well as detailed informa- its potential determinants? and 2) What are the charac- tion on participation and activity. Dedication (an aspect teristics of the assessment behaviour of Dutch insurance of ‘work engagement’ [73]) is a reflection of the inten- physicians and its potential determinants according to tion to carry out tasks and is therefore placed under these instruments? behaviour. This also applies to dealing with conflicts. A client may have a different view on the outcome of Methods the assessment. This may result in a conflict during a Study procedure disability evaluation. The above concepts are described The research group of the organisations participating in in the model under ‘Behaviour: process’. this study - TNO Quality of Life, the EMGO Institute of Behaviour directly connected with the assessment the VU Medical Centre and the Employee Benefits itself was defined as the use of instruments to assess Insurance Agency (UWV) - drafted the questionnaire capacity for work and the importance the insurance for insurance physicians. At the start of 2008 UWV physician attached to the client’s opinion about his/her drew up a list of addresses of all insurance physicians own functional capacities. In the Netherlands the working for the agency. In March 2008 UWV sent the Steenbeek et al. BMC Public Health 2011, 11:1 Page 6 of 18 http://www.biomedcentral.com/1471-2458/11/1 questionnaire, together with a covering letter containing We included 11 items about the perceived justness of an invitation to participate in the research, to the home the social security system, the agency that administers addresses of 750 insurance physicians. A reminder was the scheme (UWV) and the Permanent Full Disability sent two weeks later. Not all the physicians belonged to Standard, FAL and the implementation of the Work and our target group, but it was not possible to make a Income (Capacity for Work) Act (WIA). The items have selection in the mailing. In total we wrote to 750 insur- five response categories, ranging from (1) I totally dis- ance physicians. Our estimate is that the target group agree to (5) I totally agree. consisted of 450 insurance physicians. The criteria for We included nine items on the attitude to quality in inclusion were mentioned in the accompanying letter. relation to the importance which insurance physicians We only included insurance physicians who were attach to the development of skills, to refresher training, actively employed by UWV in May 2008 and who had to guidance by management, to the development and also performed work disability assessments in 2007 or in use of protocols and guidelines and to updating the case preceding years. The participants sent the completed file. The items have five response categories ranging questionnaire to TNO (Netherlands Organisation for from (1) I totally disagree to (5) I totally agree. Applied Scientific Research). The response consisted of We included six items on the physician’s attitude 231 questionnaires (estimated response approximately towards recovery time, the personal circumstances of 51%). As this study was based on a survey under (insur- the client and the physician’s efforts to build a good ance) physicians only, approval by a Medical Ethical relationship with the client. The items have five Commission was not necessary under Dutch law. response categories ranging from (1) I totally disagree to (5) I totally agree or (1) never to (5) always. Questionnaire Social norm In drawing up the questionnaire we used existing and In the case of insurance physicians the influence of newly developed concepts. These concepts were chosen social norms may emanate from the office and the on the basis of literature studies and four interviews environment. We included four items on management with insurance physicians. In a pilot study two insurance attitudes to quantity as opposed to quality, to the use of physicians completed the questionnaire while thinking protocols and guidelines and to production and out- aloud in order to enable us to test whether the items comes. The items have five response categories ranging were correctly understood. Finally, two other insurance from (1) I totally disagree to (5) I totally agree. physicians were timed while they completed the ques- In addition we included 13 items on the importance tionnaire. An English translation of the original Dutch which insurance physicians attach to the exercise of questionnaire is accessible (additional file 1). their profession, to the opinion of the Employee Benefits Insurance Agency (UWV), the government authorities Concept measurements questionnaire and professional organisations such as the Dutch Asso- Background variables ciation for InsuranceMedicine(NVVG)and theDutch We measured gender, age, number of years’ experience, Association for Insurance Physicians at the UWV training and specialisation. In the case of training and (UWVA), friends and family, colleagues in the office specialisation we included two items. First, we asked and elsewhere, public opinion, professional publications, whether the insurance physician is registered as such quality assessment and the trade unions. The items have (and is not still in training). Second, whether he prac- four response categories, ranging from (1) not important tises or has practised in another area of medicine. In to (4) very important. The higher the score, the greater order to register differences between offices or regions the importance attached to this opinion or view. we recorded the location of the insurance physician’s Self-efficacy office. We also asked how many hours they work each Self-efficacy was measured by the ten items formulated week, how many assessments they make each week, by Scholz et al. [77], adjusted to measure the insurance from which industry the majority of their clients come physician’s belief about his ability and capacity to carry and the statutory background of the assessments of the out work disability assessments. The items relate specifi- majority of their clients, namely the Work and Income cally to self-efficacy during the disability assessment (Capacity for Work) Act (WIA), the Disability Insurance interview. The items have four response categories, ran- Act (WAO) or another statutory regime. ging from (1) completely incorrect to (4) completely Attitude correct. Job satisfaction was measured by three items with five Barriers response categories ranging from (1) never to (5) always Work pressure was measured by means of a four-item (I am satisfied with my work; my work suits me; I like scale drawn up by Smulders, Andries and Otten [78]. A my work) from Van Dijk et al. [76]. sample item is ‘Do you have to get through a lot of Steenbeek et al. BMC Public Health 2011, 11:1 Page 7 of 18 http://www.biomedcentral.com/1471-2458/11/1 work?’ A four-point answering scale was used ranging doctor) regarding the attempt to return to work, the from (1) never to (4) always. diagnosis and information from the parties. We also Emotional workload was measured using a three-item asked whether information from the reintegration report scale from the Copenhagen Psychological Questionnaire (drawn up by the occupational physician and sent with [79]. A sample item is: ‘Does your work put you through the WIA benefits application) was decisive and whether emotionally difficult situations?’.Answers were scored the physician received sufficient feedback from the on a four-point scale ranging from (1) never to (4) claims manager about the outcome of his assessments. always. The items had four response categories, ranging from Decision-making authority (4 items) was measured (1) never to (4) always. using a Dutch version of the Job Content Questionnaire, Intention aimed at assessments [80,81]. A sample item is: ‘Do you For practical reasons we chose to measure only ‘the determine the order in which you carry out your tasks?’ object, the physician’s interpretation of his duties and Answers were scored on a four-point scale ranging from the basic premises’ in relation to intention. A question- (1) never to (4) always. naire in which intention is measured in respect of all Emotional exhaustion was measured using the five- behavioural items would be much too long. The 15 item emotional exhaustion scale of the Dutch version of items had five response categories ranging from (1) not the Maslach Emotional Exhaustion Inventory [82]. at all important to (5) very important. In the case of Answer categories varied from (1) never to (5) almost object/interpretation of duties we asked how important daily. The higher the score, the greater the exhaustion. the following objects are in relation to the assessment: We included 12 items on cooperation, office atmo- determination of physical capacities and cause of sick- sphere, consultation, being taken seriously by the man- ness, promotion of behaviour conducive to recovery, agement, and influence on workload. [76]. The higher return to work, client’s self-insight and reintegration. the score the greater is the extent of the cooperation or We also asked how important the following factors are co-determination. The items have five response cate- in the assessment of claims: health complaints, impair- gories ranging from (1) I totally disagree to (5) I totally ments, limitations or handicaps of the client, an intern- agree. ally consistent and plausible account provided by the We included 11 items on factors that could hinder or client, thorough questioning of the account given by the promote the quality of the assessment: legislation, reor- client of his daily activities, work capacity, chances in ganisations, support and guidance by staff physicians the labour market and information about the client’s and management, reporting requirements, protocols/ home situation. guidelines and standards, production requirements, Behaviour process refresher training and other measures to promote exper- Engagement is a concept that refers to being fully tise, and mutual consultation. Each of the items has immersed in an activity (absorption), being highly acti- three response categories: adverse influence, no influ- vated (vigour) and identifying with the work (dedica- ence or beneficial influence. tion). We used the four items of the subscale of We included 16 items concerning ‘difficult clients’.We dedication on the engagement scale developed by asked whether the following eight categories constitute Schaufeli et al. [73]. The items have five response cate- an important proportion of the physician’s clients and gories, ranging from (1) never to (5) always. The higher whether the physician considers the assessments of the score, the greater is the work dedication. these categories to be extra difficult: clients with disor- Research by De Boer et al. [83] shows that although ders that are difficult to determine objectively, clients there are various interview models, they are not used as with mental disorders, clients with a poor command of such. We did not therefore ask about the models, but Dutch, clients who are aggressive, clients who are included nine items on different core elements from the manipulative, clients who have problems at home or models, such as who determines what is discussed in work, older clients and cases in which poor preparatory the interview (physician or client), whether the physician work has been done by the occupational health service. asks questions in a fixed order, whether the physician Knowledge asks questions about subjects raised by the client, and In order to form a picture of the need for knowledge/ whether the physician asks for concrete examples of information, the actual information received and the use barriers and examines whether barriers result in limita- of this information for the purposes of the assessment tions. The items had four response categories, ranging we choose to include 11 general items (i.e. not specifi- from (1) never to (4) always. cally relating to diagnosis) as to whether physicians had To measure conflict handling we used the Dutch Test sufficient medical knowledge, medical information, for Conflict Handling [84], after modifying the items to information from the occupational physician (company confine them to the disability evaluation and conflicts Steenbeek et al. BMC Public Health 2011, 11:1 Page 8 of 18 http://www.biomedcentral.com/1471-2458/11/1 with a client. This test measures to what extent five female. Insurance physicians of the total population strategies are applied for handling conflicts, namely worked on average 32 hours per week. Although distri- yielding, problem-solving, compromising, avoiding and bution measures of these population means could not forcing. We used three items for each strategy (total 15 be calculated, even if the (unknown) population confi- items). dence intervals were smaller than those of the respon- Behaviour assessment dent group, the respondent group in our study would As far as the ‘permanent full disability’ criterion is con- not significantly differ from the population of insurance cerned we included five items on the extent to which physicians in terms of age, gender, and working hours per week. the rules are followed and how physicians assessed a cli- ent who is completely unable to work but can still func- Imputation of missing values tion in at least one social role. The items had four With listwise deletion, only 122 cases of the 231 cases response categories, ranging from (1) never to (4) would be left. We therefore decided to impute for miss- always. ing values. Because year of birth was not answered in 40 As regards the FAL we included nine items in order of the 231 cases, we imputed 38 of these missing cases to estimate to what extent physicians focus on a) limita- by the predictions of an OLS regression equation for tions, impairments or complaints; b) what the client can age (i.e. year of birth minus 2008). In the regression do; c) difficult home circumstances; d) internal and equation (listwise, enter procedure, n = 185) we used external consistency; e) worsening health; and f) consul- the other background variables as independent variables: tation with the labour expert. The items had five sex (dummy), registered as insurance physician response categories, ranging from (1) I totally disagree (dummy), (formerly) registered as curative specialist to (5) I totally agree. (dummy), working hours per week, number of assess- As regards behaviour in relation to the client, it is evi- ments per week, type of statutory scheme applicable to dent from the study by Nagtegaal [85] that the client’s most of the assessments (three dummies for WIA, account of daily activities is a useful instrument in asses- WAO and the Invalidity Insurance (Young Disabled Per- sing the extent of his physical capacities. We included sons Act (Wajong)) and sector (ten dummies for eleven 10 items with four response categories, ranging from (1) sectors). The multiple correlation of the predicted age never to (4) always. We asked how often the interview with the observed age was 0.696; the standardised resi- lasted as long as necessary, whether the client was trea- duals had a completely normal distribution. The SPSS ted with respect, whether the physician felt involved 15.0 program [86] was used for this regression analysis. with the client, whether the physician took an indepen- The remaining missing values for the background vari- dent position and did not allow himself to be affected ables, the scale variables and the object scores of the by the client’s interests and whether the physician took HOMALS dimensions (see the next paragraph) were the time to question the client thoroughly about his imputed using the ‘expected maximisation’ algorithm account of his daily activities, to provide good reasons [87]. There were three variables with eleven to seven- for his conclusion and to write a good report. teen imputed cases, six variables with six to ten imputed cases and thirteen variables with two to five imputed Analyses cases. The remaining variables had no or only one Response imputed case. The interactive Lisrel program with Prelis In total we wrote to 750 insurance physicians. Our esti- 2.72 [88] was used for this imputation procedure. mate is that the target group consisted of 450 insurance Construction of scales and dimensions for the ASE concepts physicians. The response consisted of 231 questionnaires The answers of the 231 insurance physicians were used (estimated response approximately 51%). As we lacked to determine which concepts from the questionnaire the necessary data of the target population to do a full were suitable for further analysis. The responses given non-response analysis, we checked whether the group of by the insurance physicians were inspected. For some participants (N = 231) was representative of the total items it was necessary to recode the original items in population of insurance physicians working for UWV fewer categories as some categories were empty or (N= approximately 900, including staff-members and almost empty. Negatively formulated items were recoded physicians not performing disability assessments) in positively. terms of age, gender, and working hours per week. The Scales were formed for the following already validated mean age and 95% confidence interval (CI) of the scales: job satisfaction, self-efficacy, work pressure, emo- respondents was 50.8 years (95% CI [49.1;51.7]) and tional workload, decision-making authority, emotional 41.1% were female. The respondents worked on average exhaustion and engagement. Cronbach’s alpha was com- 32.5 hours per week (95% CI [31.5-33.4]). The total puted for each of these scales. For the remaining items, population’s mean age was 49 years and 41.7% were factor analyses with principal components analysis and Steenbeek et al. BMC Public Health 2011, 11:1 Page 9 of 18 http://www.biomedcentral.com/1471-2458/11/1 varimax or oblique rotation per block of items, were transformations [90], they are not scales, and reliability performed to extract factors for each theoretical con- analysis cannot be performed. Therefore, we call these cept. Oblique rotation was chosen only if there was a variables ‘dimensions’, contrary to the variables which we significant correlation between the extracted factors. constructed as additive scales, which we call ‘scales’.We Where this was not the case, we decided to use varimax used the SPSS 15.0 program [86] for the factor analyses, rotation. Bartlett’s test of sphericity was used to test reliability analyses and the HOMALS analyses. whether the correlation matrix was an identity matrix. The sampling adequacy was inspected by means of the Results Kaiser-Meyer-Oikin measure (KMO) and found to be Descriptives of all measured scales are presented in table greater than 0.6. The number of extracted factors was 1 and those of the measured object scores resulting decided on the basis of the scree test, the Eigenvalue from the HOMALS analyses in table 2, including the and, most of all, the interpretability of the extracted fac- finalnumberofitems.Whennot alloriginalitems are tors. For each extracted factor, reliability analysis, included in the scales and dimensions, we report it in including item analysis, was performed to construct this section. A summary of scales and dimensions for additive scales from the items of the factors. An additive each ASE concept is presented in figure 2. scale is constructed of numerical categories of items that can be meaningfully added. In the item analysis, the New scales and dimensions contribution of each item to the reliability of an additive Attitude scale can be estimated. If an item did not contribute to We developed two scales on the theme of justness (10 an additive scale, this item was deleted from this scale. of the 11 original items): The higher the score on the When Cronbach’s alpha was equal to or larger than 0.6, scale ‘Positive attitude towards the WIA’, the more posi- additive scales of the selected items were calculated. We tive the opinion about the justness of the WIA. The nonetheless also decided to use additive scales in three higher the score on the scale ‘Social security system is cases where Cronbach’s alpha was less than 0.6 (0.560, just’, the more positive the opinion about the agency 0.566 and 0.594, respectively). These three scales were that administers the scheme (UWV), the Permanent Full considered to be theoretically important. For each addi- Disability Standard and FAL. tive scale we also calculated the percentage of respon- We developed two attitude scales for quality (eight of dents who, on average, scored above the theoretical the nine original items). The higher the score on the mean of the additive scale. This means that in case of scale ‘Quality: development of skills important’, an additive scale consisting of four Likert scale items the greater the importance attached by the physician to ranging from 1 to 5, we report the percentage of the promotion of expertise, consultation with colleagues, respondents with a scale average above 3*4 = 12.0. In working in accordance with protocols and properly the remaining text when we refer to scales, we mean updating the case file. The higher the score on the scale ‘additive scales’. ‘Quality: support by management important’, the greater For some blocks of items and for some individual items the importance attached by the physician to support it was not possible to construct a scale for several rea- and management by the immediate superior and sup- sons: the correlation matrix was not an identity matrix, port by the staff physician. and/or the sampling adequacy was not good, or Cron- We developed the two attitude dimensions on recov- bach’s alpha was too small. We grouped these ‘lost’ items ery time (six items). The higher the score on the ‘Recov- on a theoretical basis, recoded them, if necessary, into ery time: client still has some energy left after work’ two or three categories and used HOMALS (homogene- dimension, the more the insurance physician agrees the ity analysis by means of alternating least squares) to ana- client should not be completely exhausted after work. lyse the dimensions behind these grouped items [89]. The higher the score on the ‘Recovery time: good rela- The number of dimensions was decided on basis of the tionship with client’ dimension, the more the insurance sum of the Eigenvalues of the dimensions. We estimated physician tries to establish good relations with the client for each dimension the discrimination measures of the and takes account of personal circumstances. items, the category quantifications of categories of items, Social norm and the object scores of the cases. We used the discrimi- We developed three scales (12 of the 13 original items) nation measures and the category quantifications to and two dimensions (four items). The following three interpret both poles (negative and positive) of the dimen- scales are about the influence which the opinions and sions. The object scores of the dimensions that were views of certain persons/authorities have on the perfor- meaningful and gave additional information were mance of the profession. The higher the score on the selected as variables. Because object scores of multiple scale, the greater the importance attached to this opi- Homals dimensions are constructed with non-linear nion or view. The ‘Opinion of UWV and employee Steenbeek et al. BMC Public Health 2011, 11:1 Page 10 of 18 http://www.biomedcentral.com/1471-2458/11/1 Table 1 Description of scales (n = 231) ASE Scale # % yes/ Theor. Median Mean sd Cronbach’s 1 2 items high max alpha Attitude Job satisfaction 3 78 15 12 11.41 2.51 0.875 Positive attitude towards WIA 5 53 25 16 15.98 3.63 0.797 Social security system just 5 70 25 17 17.43 3.25 0.636 Quality: development of skills important 5 99 25 22 22.11 2.27 0.648 Quality: support by management important 3 68 15 11 10.46 2.32 0.643 Social Norm Opinion of UWV and employee representative bodies 6 43 24 15 15.01 2.75 0.697 important Colleagues’ opinion important 5 66 20 14 13.40 2.36 0.679 Society’s opinion important 3 10 12 6 5.77 1.53 0.560 Self-efficacy Self-efficacy 10 40 32 32.81 4.21 0.908 Barriers Work pressure 4 44 16 10 10.18 2.06 0.771 Emotional workload 3 20 12 6 6.38 1.42 0.702 Decision making authority 4 61 16 11 10.96 2.65 0.724 Emotional exhaustion 5 12 25 10 10.79 3.97 0.892 Office culture: good cooperation 8 83 40 30 29.62 5.65 0.900 Office culture: sufficient co-determination 4 20 20 10 9.78 3.25 0.814 Quality: influence of refresher training and 2 97 6 6 5.90 0.38 0.665 consultation beneficial Quality: influence of staff physician beneficial 2 80 6 6 5.38 0.89 0.675 Quality : influence of manager beneficial 2 27 6 4 4.03 1.05 0.647 Many difficult clients/cases 16 73 16 15.62 2.82 0.675 Knowledge Sufficient information from the occupational physician 3 44 12 7 7.20 1.33 0.769 Intention Stimulate recovery and return to work 4 94 20 17 16.71 2.77 0.852 Basic premises: residual capacity 6 99 30 28 27.03 2.75 0.809 Basic premises: client’s account and home 4 97 20 18 17.37 2.17 0.727 circumstances Behaviour Process Dedication 4 73 20 14 13.94 2.74 0.874 Technical interview: describe object and procedure 2 84 8 7 6.71 1.15 0.594 Conflict handling: seek compromise 7 5 35 15 15.57 3.46 0.733 Behaviour Assessment Comply with permanent full disability rules 2 72 8 6 6.09 1.23 0.734 FAL: take account of client 5 31 25 14 14.16 2.47 0.566 FAL: consult with labour expert when not necessary 2 63 10 8 7.09 2.00 0.761 yes/high = % of respondents whose score is above the theoretical scale average, e..g. the % of respondents who on average ‘(totally) agree’, think of something as ‘(very) important’, or score ‘often/always’. The theoretical maximum value of the scale. The percentage of insurance physicians who classify the majority of their clients/cases as difficult. representative bodies important’ concerns the influence important’ scale concerns the influence of fellow insur- of the UWV (supervisor, staff physician, objection and ance physicians, both in the office and elsewhere. The appeal), and employee representative bodies such as the ‘Society’s opinion important’ scale concerns the influ- Dutch Association for Insurance Medicine (NVVG), the ence of family, friends, TV and government. Dutch Association for Insurance Physicians at the UWV Thehigherthe scoreonthe dimension ‘Managing by (UWVA) and the trade union. The ‘Colleagues opinions reference to quality rather than quantity’,the more Steenbeek et al. BMC Public Health 2011, 11:1 Page 11 of 18 http://www.biomedcentral.com/1471-2458/11/1 Table 2 Description of Homals object scores for dimensions (n = 231) # Min Max Median Mean* Sd* Eigen items value Attitude Recovery time: client still has some energy left after work 6 -2.50 1.79 -0.0784 0.00 1.00 0.254 Recovery time: good relationship with client 6 -3.36 2.07 0.0931 0.00 1.01 0.188 Social norm Managing by reference to quality rather than quantity 4 -1.44 1.70 -0.0953 0.00 1.00 0.489 Managing less by reference to production targets and outcomes 4 -1.16 4.03 -0.1753 0.01 1.00 0.371 Barriers Quality: influence of legislation and reorganisations not adverse 4 -1.75 2.13 -0.0833 0.00 1.02 0.352 Quality: influence of guidelines not adverse and production target not beneficial 4 -1.25 2.90 -0.6381 0.01 1.02 0.269 Knowledge Possessing, requesting and using insufficient information 8 -2.65 2.20 0.0174 -0.01 1.02 0.231 Insufficient medical information and knowledge 8 -2.77 1.80 0.0751 0.00 1.01 0.173 Sufficient knowledge, reintegration report less often supplements medical 8 -2.67 2.77 -0.0985 0.00 1.01 0.154 information Behaviour Process Interview management: client decisive 6 -1.66 2.46 -0.2199 0.00 1.00 0.263 Interview: limitations not checked 6 -1.32 3.09 -0.4262 0.00 1.00 0.214 Interview: respond to client 6 -2.32 2.61 -0.0496 0.00 1.00 0.170 Conflict handling: engage in confrontation 8 -1.91 2.63 0.0086 0.00 1.00 0.235 Conflict handling: play down differences 8 -1.92 2.87 -0.1829 0.00 1.00 0.209 Behaviour Assessment FAL and recovery time: strict/formalistic approach 6 -9.13 1.82 0.0851 -0.07 1.27 0.292 FAL and recovery time: focus on impairments 6 -8.94 2.27 0.0507 -0.28 1.65 0.288 Client approach: involved with and time for 8 -2.24 2.25 -0.0536 0.00 1.00 0.263 Client approach: time for account of daily activities and reporting 8 -2.68 1.58 0.0614 0.00 1.00 0.178 Client approach: too little time, but involved with 8 -7.66 2.74 -0.2466 -0.07 1.23 0.168 * Because of imputation, object scores can deviate from mean = 0 and sd = 1. importance is attached in relative terms to quality-based We constructed one scale named ‘Many difficult cli- management and the less importance to quantity-based ents’, based on 16 items. The higher the score, the more management. The higher the score on the dimension the insurance physician is confronted with clients whom ‘Managing less by reference to production targets and he experiences as difficult. outcomes’, the less importance is attached to manage- The higher the score on the ‘Quality: influence of leg- ment based on production targets and outcomes. islation and reorganisations not adverse’ dimension, the Barriers lower the adverse impact of legislation and UWV reor- We developed four scales and two dimensions for bar- ganisations on quality. The higher the score on the riers. Of the 11 original items about quality of the ‘Quality: influence of guidelines not adverse and produc- assessment, six were used in three scales of two items tion requirement not beneficial’ dimension, the lower each and four were used in two dimensions. The higher the adverse impact of the guidelines and the lower the the score on the ‘Quality: influence of staff physician is beneficial impact of the production requirement on beneficial’ scale, the more the assessment benefits in quality. terms of quality from support and professional guidance Knowledge provided by the staff physician. The higher the score on We developed one scale (three items) and three dimen- the ‘Quality: influence of refresher training and consul- sions (eight items). The higher the score on the ‘Suffi- tation beneficial’ scale, the greater the beneficial effect cient information from the occupational physician’ scale, of refresher training, mutual consultation and measures the more sufficient the available information from the to promote expertise support. Finally, the higher the OP. The higher the score on the ‘ Possessing, requesting score on the ‘Quality: influence of manager beneficial’ and using insufficient information’ dimension, the more scale, the greater the beneficial effect of the support and the physician considers that he does not always have professional guidance provided by the manager. sufficient medical information available, does not always Steenbeek et al. BMC Public Health 2011, 11:1 Page 12 of 18 http://www.biomedcentral.com/1471-2458/11/1 Barriers •Work pressure (1S) •Emotional workload (1S) •Decision authority (1S) Attitude •Emotional exhaustion (1S) •Job satisfaction (1S) •Office culture (2S) •Justness of system (2S) •Quality (3S, 2D) •Importance of skills and support (2S) •Difficult clients (1S) •Attitude towards recovery time (2D) Intention Assessment behaviour: process Social Norm •Recovery and return to •Dedication (1S) •Influence of representative bodies, work (1S) •Collecting information (1S, 3D) colleagues, society (3S) •Basic premises (2S) •Conflict handling (1S, 2D) •Production quality versus quantity (2 (2D) D) Knowledge Assessment behaviour: assessment Self-efficacy •Information from OP (1S) •Use of assessment instruments (3S, 2D) •Self-efficacy (1S) •Possessing, requesting •Client approach (3D) and using sufficient information (3D) Background Gender, age, experience, training, specialisation, location, working hours, production, client industry, assessment type Figure 2 The ASE model with a summary of scales and dimensions. S = Scale; D = Dimension, the number refers to the number of constructed scales and dimensions. request information from third parties and does not the importance of residual capacity, sickness, impair- always take this into account in making the assessment. ments, limitations and handicaps in the assessment. The higher the score on the ‘Insufficient medical infor- Finally, the ‘Basic premises: client’s account and home mation and knowledge’ dimension, the more the physi- circumstances’ scale concerns the importance of a con- cian considers that he does not always have sufficient sistent account of daily activities, thorough questioning medical information and medical knowledge to make about this account and information about the client’s the assessments. The higher the score on the ‘Sufficient home circumstances in the assessment. knowledge, reintegration report less frequently supple- Behaviour process ments medical information’ dimension, the more We developed one scale and three dimensions concern- sufficient is medical information and the less frequently ing the collection of information about functional capa- the reintegration report supplements the medical cities (eight of the nine original items). The higher the information. score on the ‘Technical interview: describe object and Intention procedure’ scale, the greater the emphasis which the We developed three intentionscales(14of theoriginal physician puts at the beginning of the interview on the 15 items). The higher the score on these scales, purpose of the interview and the procedure to be fol- the more importance is attached to the subjects. The lowed during it. The higher the score on the ‘Interview ‘Stimulate recovery and return to work’ scale concerns management: client decisive’ dimension, the more the the importance to promote recovery behaviour, return client determines the order of events rather than the to work, self-insight and reintegration in the assessment. insurance physician. The higher the score on the ‘Inter- The ‘Basic premises: residual capacity’ scale concerns view: limitations not checked’ dimension, the less Steenbeek et al. BMC Public Health 2011, 11:1 Page 13 of 18 http://www.biomedcentral.com/1471-2458/11/1 frequently the insurance physician checks what limita- client. The higher the score on the ‘Client approach: tions the client faces. The higher the score on the ‘Inter- time for account of daily activities and reporting’ dimen- view: respond to client’ dimension, the more frequently sion, the more often the insurance physician takes time the insurance physician responds to subjects raised by to thoroughly question the client about his daily activ- the client. ities and to report on this. The higher the score on the The Dutch Test for Conflict Handling [84] measures ‘Client approach: too little time but involved with’ to what extent five strategies are applied for handling dimension, the more likely it is that the insurance physi- cian has too little time to draw up a proper report and conflicts, namely yielding, problem-solving, compromis- to question the client about his daily activities, but feels ing, avoiding and forcing. It is noteworthy that insur- ance physicians evidently did not use all five of these involved with the client. strategies during the disability evaluation. We were able to measure one scale (based on seven items) and two Descriptive results dimensions (calculated from the remaining eight items). Descriptive results for background variables are sum- The higher the score on the ‘Conflict handling: seek marised in table 3. 58.9% of the insurance physicians in compromise’ scale, the more often the physician this survey were men. On average the respondents were searches for a compromise with a client in the event of aged 50.8 years and had 16.2 years of experience. 85.7% a difference of opinion. The higher the score on the were registered and almost two third worked 33 or ‘Conflict handling: engage in confrontation’ dimension, more hours per week and carried out on average 9.1 the more likely the physician is to engage in confronta- disability assessments per week. The patients came from tionwith theclient intheevent of adifferenceofopi- all industries. It is noteworthy that 53.7% of the insur- nion. The higher the score on the ‘Conflict handling: ance physicians reported that a substantial proportion of play down differences’ dimension, the more often the their clients were temporary workers. physician will try to circumvent differences of opinion As regards attitude, table 3 shows that 78% of the and play down their importance. insurance physicians were motivated by the job, 70% Behaviour assessment considered that the social security system was just and We developed three scales and five dimensions (23 of the 24 original items). The higher the score on the Table 3 Background variables ‘Comply with permanent full disability rules’ scale, themorestrictly theinsurance physician complies with % mean sd the permanent full disability rules. The higher the score Gender (%man) 58.9 on the ‘FAL: take account of the client’ scale, the more Age 50.8 7.0 often the physician focuses on the complaints raised by Registered as insurance physician 85.7 the client, what the client can really do, the client’s diffi- Extra medical speciality 15.2 cult home circumstances and limitations experienced by Working hours (week) % up to 24 hrs 16.0 the client. The higher the score on the ‘FAL: consult % 25-32 hors 23.8 with labour expert when not necessary’ scale, the more % 33 hrs or more 60.2 often the insurance physician will consult with the N assessments (week) 9.1 4.0 labour expert in circumstances where the client is Years of experience 16.2 7.7 unable to work or does not belong in the benefits Assessments mainly under WIA 37.7 category. Assessments mainly under WAO 26.4 The higher the score on the ‘FAL and recovery time: Assessments mainly under Wajong 13.0 strict/formalistic approach’ scale, the more the insurance Clients mainly from the agriculture, fishing and food 13.0 industries physician takes a formalistic and strict approach to Clients mainly from the construction and timber 19.5 drawing up the FAL and takes no account of the client industries and his recovery time. The higher the score on the ‘FAL Clients mainly from manufacturing industry 39.4 and recovery time: focus on impairments’ scale, the Clients mainly from the retail and wholesale sectors 41.6 greater the attention which the insurance physician pays Clients mainly from the transport sector 24.2 when drawing up the FAL to limitations caused by Clients mainly from the financial services sector 26.8 impairments, particularly in the light of consistency, and Clients mainly from the temporary work sector 53.7 takes no account of a possible deterioration in the cli- Clients mainly from the health sector 35.1 ent’s health. Clients mainly from the education sector 22.1 The higher the score on the ‘Client approach: involved Clients mainly from the rest of the public sector 13.0 with and time for’ dimension, the more often the insur- Clients mainly from the professions and other sectors 33.8 ance physician takes time for and is involved with the Steenbeek et al. BMC Public Health 2011, 11:1 Page 14 of 18 http://www.biomedcentral.com/1471-2458/11/1 only 53% had a positive opinion about the WIA. The model is an extension of the model designed by Croon results for ‘social norm’ show that insurance physicians and Langius [29] in their study of the process of sickness are mostly influenced by colleagues (66%) and by their certification assessment by social insurance physicians. employer (UWV, 43%) and much less by society/the They took the theory of planned behaviour as a starting public (10%). The results on barriers show that 44% of point. The concept of barriers and stimuli experienced by the insurance physicians experienced substantial work physicians, their own effectiveness and the availability of pressure and 20% substantial emotional workload, 12% sufficient knowledge (concepts which are recognised in were emotionally exhausted and 73% reported that they the ASE model) are also included in their model. Our analysis model divides Croon and Langius’ concept of the viewed the majority of their clients/cases as ‘difficult’. The influence of refresher training and the staff physi- ‘influence of the environment’ into the concept of the cian is viewed as conducive to quality, whereas only 27% social norm (which influences the intention) and barriers of the insurance physicians consider that the manager (which have an intermediary effect between intention promotes quality. As far as knowledge is concerned, less and behaviour). It could be argued that the conceptual than half of the physicians consider that they receive model of the theory of planned behaviour is problematic sufficient information from the occupational physician in that its concepts are not specific enough [92]. We have (company doctor). The scores for intentions show that countered this argument in our proposed model by most insurance physicians intend to carry out the pro- focusing the concepts specifically on the subject of work fession in the manner expected of them as professionals. disablement assessment. As regards behavioural process, we see that three quar- One particular strength of this study is the extent of ters of the physicians are dedicated and inform the cli- the good response to the survey by the insurance physi- ent of the object of the interview and the procedure and cians, which was considerably higher than we had that only 5% indicate that they seek a compromise in expected. A weakness of the study is its cross-sectional theevent of adifferenceofopinionwith theclient. As design, which does not allow for analysis of causal rela- regards behavioural assessment, we see that 72% of the tionships between attitude, social influence, intention insurance physicians follow the rules, 31% consider that and behaviour. Another weakness may be the fact our they have taken account of the client and 63% fre- explanation of measured scales and dimensions in rela- quently consult with the labour expert in circumstances tion to the ASE concepts is only based on theoretical where this is not mandatory, namely in situations of grounds. It is therefore possible that certain scales and ‘medical incapacity for work’ or ‘capacity for own work’. dimensions may not fit in with the ASE concept. Furthermore, the study does not investigate the struc- tural relationships between the measured constructs. Discussion Discussion of the methods Further study is therefore needed in order to demon- In this article we have presented the development of strate whether the ASE model is the best model to instruments for measuring and explaining variations in explain insurance physician’s behaviour. the behaviour of insurance physicians in relation to assessments of functional capacities. Data from 231 Discussion of content questionnaires were analysed and used as a basis for fill- The descriptive results may give rise to some concern. ing the ASE model with 29 scales and 19 dimensions. We see a professional group that is highly motivated We identified scales and dimensions that represent Atti- about the job and positive about the Dutch social secur- tude, Social norm, Self-efficacy, Barriers, Knowledge and ity system. However, only half of them have a positive Intention. We slightly modified the underlying ASE opinion about the Work and Income (Capacity for model by dividing Behaviour into two blocks, the first Work) Act (WIA). The views of the insurance physi- reflecting the process and the second reflecting assess- cians about the social security system and legislation ment-related behaviour. The value of the instruments are, in principle, separate from the manner in which proposed in this article lies in their specificity for insur- they carry out their professional duties and endeavour ance physicians and their sound psychometric character- to achieve a high quality. Furthermore, insurance physi- istics. The extensive literature study, in combination cians experience serious barriers, the most frequent of with the interviews safeguarded the internal validity. which is work pressure. Work pressure, emotional work- While our instruments and the underlying concepts load and emotional exhaustion are positively correlated. show considerable similarities to the study of the com- Finally, 73% of the insurance physicians describe a munication of insurance physicians with their clients majority of their clients/cases as ‘difficult’.Inorder to conducted by Van Rijssen et al. [91], the operationalisa- determine if these scores were relatively high, we com- tion of the underlying concepts was specifically designed pared our outcomes with the same scales of a large to meet the objective of the present study. Our analysis survey among employees (NEA 2008). This comparison Steenbeek et al. BMC Public Health 2011, 11:1 Page 15 of 18 http://www.biomedcentral.com/1471-2458/11/1 reveals that the insurance physicians in this study do not modified form, is suitable for describing the assessment differ significantly from the NEA group ‘physicians, den- behaviour of insurance physicians and the determinants tists and veterinary surgeons (N = 240)’ in terms of of this behaviour. The next step in this line of research work pressure, emotional workload and emotional should be to validate the model using structural equa- exhaustion. Insurance physicians were found to have tion modelling. Finally, the predictive value should be higher levels of autonomy than other ‘physicians, den- tested in relation to work disability assessment tists and veterinary surgeons’.The negative aspectsdo outcomes, i.e. grant or reject the claim. not therefore differ from those of a comparable group of professionals. Additional material The answers to the questionnaire also indicate that insurance physicians are primarily bound, as regards their Additional file 1: Questionnaire insurance physicians. English translation of the original Dutch questionnaire for insurance physicians. professional conduct, by the norms and views of insurance physicians as a professional group. In this way, frameworks are set for the discretionary power of the insurance physi- cians which is necessary in order to do justice in special Acknowledgements This research project has been funded by the Dutch ‘Stichting Instituut Gak’, cases. A fellow insurance physician must be able to come a foundation that initiates and supports innovative projects in the Dutch to the same assessment (reproducibility). welfare sector. Additional funding came from TNO Work and Employment, The management of UWV (Dutch Benefits Insurance Hoofddorp, and the Research Center for Insurance Medicine AMC-UWV- VUmc, the Netherlands. Agency), which focuses above all on the work processes and production, is often seen as setting norms, but is Author details not regarded as supporting the quality of the work. This TNO Work and Employment, PO Box 718, 2130 AS Hoofddorp, the Netherlands. VU University Medical Center, Department of Public and is not a unique finding, but an illustration of the pro- Occupational Health, EMGO Institute for Health and Care Research, blem of managing professionals in general [93]. In his Amsterdam, the Netherlands. UWV, Employee Benefits Insurance Authority, international comparative study into work disability Amsterdam, the Netherlands. Research Center for Insurance Medicine, AMC- UWV-VUmc, Amsterdam, the Netherlands. assessments De Boer [94] also concludes that the profes- sional definition of quality of evaluation of work disabil- Authors’ contributions ity is ‘performance according to professional standards’. RS participated in the design of the study and its coordination, developed the questionnaire, performed the statistical analysis and drafted the He emphasises that in the Netherlands the requirement manuscript. AS participated in the design of the study, participated in of a fair trial is also a central part of the quality of claim developing the questionnaire, performed the statistical analysis and assessment. We see this reflected in our results. The participated in drafting the manuscript. HM participated in the design of the study and its coordination, developed the questionnaire and participated in results for social norms show that the insurance physi- drafting the manuscript. JRA participated in the design of the study and cians attach most importance to the views of their fel- participated in developing the questionnaire. HK participated in the design low professionals and thereafter to those of their of the study and its coordination and provided the logistics for the questionnaire. JB participated in the design of the study and its employer (UWV). They attach the least importance to coordination, participated in developing the questionnaire and drafting the the views of society. Many insurance physicians believe manuscript. All authors read and approved the final manuscript. that the quality of their assessments is positively influ- Competing interests enced by good cooperation with colleagues, refresher All authors declare that there are no financial or other relationships that training and consultation, as well as guidance by staff might lead to a conflict of interest. physicians. Many insurance physicians score highly in Received: 20 April 2010 Accepted: 3 January 2011 terms of following rules during the assessment, so that a Published: 3 January 2011 fair process is possible, while only few insurance physi- cians indicate that their work style is to look for com- References promises in the event of a difference of opinion with the 1. Eddy DM: Variation in physician practice. The role of uncertainty. Health Affairs 1984, 3:74-89. client. This would detract from their independent pro- 2. Eisenberg JM: Physician Utilization. The state of research about fessional status and the requirements of a fair trial. physicians’ practice patterns. Med Care 1985, 23:461-83. Nonetheless, this does not prevent one third of the 3. Wilson JRM, Clarke MG, Ewings P, Graham JD, MacDonagh R: The assessment of patient life-expectancy: how accurate are urologists and insurance physicians from indicating that they take oncologists? BJU Int 2005, 95:794-798. account of the client’s specific circumstances when 4. Zandbelt LC, Smets EMA, Oorta FJ, Godfried MH, De Haes HCJM: drawing up the functional capacity assessment. Determinants of physicians’ patient-centred behaviour in the medical specialist encounter. Soc Sci Med 2006, 63:899-910. 5. Shahinian VB, Kuo Y, Freeman JL, Goodwin JS: Determinants of Androgen Conclusions Deprivation Therapy Use for Prostate Cancer: Role of the Urologist. J The scales and dimensions developed appear to be valid Natl Cancer Inst 2006, 98:839-845. 6. Vogels AGC, Jacobusse GW, Hoekstra F, Brugman E, Crone M, Rijneveld SA: and offer a promising basis for future research. The Identification of children with psychosocial problems differed between results suggest that the underlying ASE model, in Steenbeek et al. BMC Public Health 2011, 11:1 Page 16 of 18 http://www.biomedcentral.com/1471-2458/11/1 preventive child health care professionals. J Clin Epidemiol 2008, 31. Van Oostrom, Anema JR, Terluin B, Vet HCW de, Knol DL, Van Mechelen W: 61:1144-1151. Cost-effectiveness of a workplace intervention for sick-listed employees 7. Chang LW, Fung TY, Leung TY, Sahota DS, Lau TK: Volumetric (3D) with common mental disorders: design of a randomized controlled trial. imaging reduces inter- and intraobserver variation of fetal biometry BMC Public Health 2008, 8:12. measurements. Ultrasound Obstet Gynecol 2009, 33:447-452. 32. Vermeulen SJ, Anema JR, Schellart AJ, Van Mechelen W, Van der Beek AJ: 8. Marinus AMF: Interdoktervariatie in de huisartsenpraktijk (Inter-doctor Intervention mapping for development of a participatory return-to-work variation in GPs’ practices). PhD thesis University of Amsterdam; 1993. intervention for temporary agency workers and unemployed workers 9. Van der Weijden T, Hutten JBF, Brandenburg BJ, Grol RPTM, Van der sick-listed due to musculoskeletal disorders. BMC Public Health 2009, Velden K: Cholesterol management in Dutch general practice. A 9:216. comparison with national guidelines. Scand J Prim Health Care 1994, 33. De Vries H: Determinanten van gedrag. (Determinants of behaviour). In 12:281-288. Gezondheidsvoorlichting en gedragsverandering. Edited by: Damoiseaux V, 10. Janssen HAM, Borghouts JAJ, Muris JWM, Metsemakers JFM, Koes BW, Van der Molen H, Kok GJ. Assen: Van Gorcum; 1993:109-132. Knottnerus JA: Health status and management of chronic non-specific 34. De Vries H, Mudde AN: Predicting stage transitions for smoking cessation abdominal complaints in general practice. Br J Gen Pract 2000, applying the attitude-social influence-efficacy model. Psychol Health 1998, 50:375-379. 13:369-385. 11. Grytten J, Sørensen R: Practice variation and physician-specific effects. J 35. Tellnes G, Sandvik L, Moum T: Inter-doctor variation in sickness Health Econ 2003, 22:403-418. certification. Scan J Prim Health Care 1990, 8:45-52. 12. Verstappen WHJM, Ter Riet G, Dubois WI, Winkens R, Grola RTPM, Van der 36. Thies-Zajong S, Szecsenyi J, Kochen MM: Attitudes and empirical referral Weijden T: Variation in test ordering behaviour of GPs: professional or data of West German family physicians. Gesundheitswesen 1993, context-related factors? Fam Pract 2004, 21:387-395. 55:635-640. 13. Taylor RJ: General-practitioner prescribing. J R Coll Gen Prac 1977, 37. Besseling JJM, Bockting AJV, Franquinet JMWAF, Sprenger WJ: Evaluatie 27:79-82. stelselherziening (System revision evaluation) Amsterdam: GMD (Joint Medical 14. Martens JD, Van der Weijden T, Severens JL, De Clercq PA: The effect of Examination Service); 1990. computer reminders on GPs’ prescribing behaviour: A cluster- 38. Van de Goor AG: Effects of regulation on disability duration. PhD thesis randomised trial. Int J Med Iinfor 2007, 76(Suppl 3):403-416. Amsterdam, Thesis Publishers; 1997. 15. Delnoij DMJ, Spreeuwenberg PMM: Variation in GPs’ referral rates to 39. Aarts L, De Jong P, Van der Veen R: Met de beste bedoelingen wao 1975- specialists in internal medicine. Eur J Public Health 1997, 7:427-435. 1999 trends, onderzoek en beleid. (With the nest intentions WAO 1975-1999: 16. Ydreborg BAM, Ekberg K: Disqualified for disability pension -a case/ trends, research and policies) Doetinchem: Elsevier bedrijfsinformatie; 2002. referent study. Disabil Rehabil 2004, 26:1079-1086. 40. Wevers CWJ, Vinke H: Bedrijfsartsen over verzuimbegeleiding, een 17. Spanjer J, Krol B, Groothoff JW: Inter-rater reliability in disability casusonderzoek (Company doctors on sick leave management; a case assessment based on a semi-structured interview report. Disabil Rehabil study). TBV 1999, 7:78-83. 2008, 30:1885-1890. 41. De la Court E: WAO volume-effect van de eerstejaarsherbeoordeling en zijn 18. Spanjer J, Krol B, Popping R, Groothoff JW, Brouwer S: Disability disability determinanten: reïntegratie in het eerste WAO jaar (WAO volume effect of the evaluation: the role of detailed information on functioning in addition first-year reassessment and its determinants: reintegraton in the first WAO to medical history taking. J Rehabil Med 2009, 41:267-272. year) Hoofddorp: TNO Arbeid; 2000. 19. Arrelov B, Borgquist L, Ljungberg D, Svardsudd K: Do GPs sick-list patients 42. Norrmen G, Svardsudd K, Andersson D: Impact of physician-related factors to a lesser extent than other physician categories? A population based on sickness certification in primary health care. Scand J Prim Health Care study. Fam Pract 2001, 18:393-398. 2006, 24:104-109. 20. Shiels C, Gabbay M: The influence of GP and patient gender interaction 43. Kerstholt JH, De Boer WEL, Jansen EJM: Psychologische aspecten van on the duration of certified sickness absence. Fam Pract 2006, 23:246-253. claimbeoordeling (Psychological aspects of claims assessment) Hoofddorp: 21. Hofstee WKBH, Kroneman H, De Boer WEL: Representatieve beoordeling TNO Arbeid; 2002. van arbeidsvermogen (Representative assessment of work capacity). TBV 44. Laitinen-Krispijn SM, Nicolaï LC: Verzekeringsgeneeskundige beoordeling 2009, 17:406-409. bij psychische problematiek (Insurance medicine assessment in the case 22. Fishbein M, Ajzen I: Belief, attitude, intention and behaviour; An introduction of mental problems). TBV 2006, 14:167-170. to theory and reserch Reading (MA): Addison-Wesley; 1975. 45. Spanjer J: De inter- en intra-beoordelaarsbetrouwbaarheid van WAO- 23. Ajzen I, Fishbein M: Understanding attitudes and predicting social behavior beoordelingen (The inter-rater and intra-rater reliability of WAO Englewood Cliffs NJ: Prentice Hall; 1980. assessments). TBV 2001, 9:234-241. 24. Ajzen I: The Theory of Planned Behavior. Organ Behav Hum Decis Process 46. Razenberg PPA: Verzekeringsgeneeskundige oordeelsvorming. Inzicht in 1991, 50:179-211. de praktijk (Insurance medicine assessments. Insight into practice). PhD 25. Bandura A: Social foundations of thought and action Englewood Cliffs NJ: thesis University of Amsterdam; 1992. Prentice Hall; 1986. 47. De Jong JD, Westert GP, Lagoe R, Groenewegen PP: Variation in Hospital 26. De Vries H, Dijkstra M, Kuhlman P: Self-efficacy: the third factor besides Length of Stay: Do Physicians Adapt Their Length of Stay. Decisions to attitude and subjective norm as a predictor of behavorial intention. What Is Usual in the Hospital Where They Work? Health Serv Res 2006, Health Educ Res 1988, 3:273-282. 41:374-394. 27. De Vries H, Backbier E, Kok GJ, Dijkstra M: Measuring the impact of social 48. Solomon DH, Brookhart MA, Gandhi TK, Karson A, Gharib S, Orav J, influences on smoking onset in a longitudinal study: an integration of Shaykevich S, Licari A, Cabral D, Bates DW: Adherence with Osteoporosis social psychological approaches. J Appl Soc Psychol 1995, 25:237-257. Practice Guidelines: A Multilevel Analysis of Patient, Physician, and 28. Rammeloo KC, Tamminga SJ, Anema JR, Schellart AJM: De implementatie Practice Setting Characteristics. Am J Med 2004, 117:919-924. van een verzekeringsgeneeskundig protocol Gebruik van Intervention 49. Van Deursen CGL: Claimbeoordeling WAO: het handelen van Mapping bij het ontwikkelen van een implementatiestrategie voor het verzekeringsartsen in de praktijk (WAO Claims assessment: the actions of verzekeringsgeneeskundig protocol Depressie (The implementation of insurance physicians in practice) Amsterdam: National Institute of Social an insurance medicine protocol for the use of intervention mapping in Insurance Agencies (Lisv); 2000. the development of a strategy for the depression protocol). TBV 2009, 50. Van Deursen, Veerman TJ: Randvoorwaarden van de sociaal-medische 17:95-102. begeleiding : onderzoek onder verzekeringsartsen en arbeidsdeskundigen die lid 29. Croon NHTh, Langius SWTh: (Work and health. A study about the zijn van het Lisv uitvoerderspanel (Parameters for sociomedical counselling: practices in social insurance medicine at the work insurance boards). survey among insurance physicians and labour experts who are members of PhD Thesis University of Amsterdam, Institute for Social Medicine; 1993. the Lisv implementation panel) Amsterdam: National Institute of Social 30. Rebergen D, Hoenen J, Heinemans A, Bruinvels D, Bakker A, Van Insurance Agencies (Lisv); 1997. Mechelen W: Adherence to mental health guidelines by Dutch 51. Spanjer J: Tien stellingen rond herstelgedrag (Ten propositions occupational physicians. Occup Med 2006, 56:461-468. concerning recovery behavior). TBV 1999, 7:233-235. Steenbeek et al. BMC Public Health 2011, 11:1 Page 17 of 18 http://www.biomedcentral.com/1471-2458/11/1 52. Besseling JJM: Burgerlijke staat en arbeidsongeschikheid 1977-1983 (Marital 70. Kerstholt JH, De Boer WEL, Jansen EJM: Disability assessments: Effects of status and work disability) Amsterdam: GMD (Joint Medical Examination response mode and experience. Disabil Rehabil 2006, 28:111-115. Service); 1986. 71. Spanjer J: In De reproduceerbaarheid van WAO-beoordelingen, een 53. Moore PJ, Adler NE, Robertson PA: Medical malpractice: the effect of literatuuronderzoek (The reproducability of WAO assessments, a literature doctor-patient relations on medical patient perceptions and malpractice study). Volume 10. TBV; 2002:195-198. intentions. West J Med 2000, 173:244-250. 72. Spanjer J: In Voorbeelden: Een hulpmiddel om de ernst van de beperkingen te 54. Mudde EC: Ziek in de zin der wet. De interpretatie van ziekteverzuim beoordelen (Examples: an aid for assessing the seriousness of limitations). door verzekeringsgeneeskundigen en rapporteurs (Sickness within the Volume 10. TBV; 2002:109-111. meaning of the law. The interpretation of sick leave by insurance 73. Schaufeli WB, Salanova M, González-Romá V, Bakker AB: The measurement physicians and rapporteurs). PhD thesis University of Amsterdam; 1995. of engagement and emotional exhaustion: A two sample confirmatory 55. Weijman I, Ros WJG, Rutten GEHM, Schaufeli WB, Schabracq MJ, factor analytic approach. J Happiness Stud 2002, 3:71-92. Winnubst JAM: Fatigue in employees with diabetes: its relation with 74. De Bont AA, Berendsen L, Boonk MPA, Van de Brink JC: In de spreekkamer work characteristics and diabetes related burden. Occup Environ Med van de verzekeringsarts : een onderzoek naar het verzekeringsgeneeskundige 2003, 60(suppl 1):93-98. deel van de WAO claimbeoordeling.(In the surgery of the insurer’s medical 56. De Croon EM, Sluiter JK, Nijssen TF, Kammerijer M, Dijkmans BA, adviser: a study of the insurance medicine part of the WAO claims assessment) Lankhorst GJ, Frings-Dresen MH: Work ability of Dutch employees with Zoetermeer: CTSV; 2000. rheumatoid arthritis. Scand J Rheumatol 2005, 34:277-283. 75. Maas-Oostenbrink AJ, Simon-de Zwart AJ: Standaard ‘Geen Duurzaam 57. Ybema JF, Evers M: Profiel Arbeid en Gezondheid 2005 (Labour and Health Benutbare Mogelijkheden’ altijd benutbaar? een onderzoek naar het gebruik Profile 2005) Hoofddorp: TNO Work and Employment; 2005. van de standaard GDBM bij maligne aandoeningen met een infauste 58. Berendsen L, Mullenders P: De manager de baas? Een onderzoek naar prognose (Permanent Full Disability Standard always useable? A study of the WAO-managers en hun integrale verantwoordelijkheid voor het werk use of the standard in the case of malignant disorders with an unfavourable van verzekeringsartsen (Is the manager the boss? A study of WAO prognosis) Hoofddorp: TNO Arbeid; 1999. managers and their integral responsibility for the work of insurance 76. Van Dijk JK, Messchendorp HJ, Koopman MI, Steenbeek R, Van Til CT: physicians). Den Haag: Inspectie Werk en Inkomen; 2004. Personeel in beeld. De arbeidsbeleving van medewerkers in ziekenhuizen, 59. Choy J: Rapportage van de raadplegingen door bureau SMZ met de revalidatiecentra en andere instellingen die vallen onder de CAO-Z (Personnel verzekeringsartsen en arbeidsdeskundigen (Reports of assessments by bureau in the picture. The work perception of employees in hospitals, rehabilitation SMZ with insurance physicians and labour specialists) Amsterdam: Bureau centres and other institutions covered by the ‘Z’ collective agreement) Utrecht: SMZ; 2005. Prismant; 2002. 60. Edlund C, Dahlgren L: The physician’s role in the vocational rehabilitation 77. Scholtz U, Gutierrez Dona B, Sud S, Schwarzer R: Is general self-efficacy az process. Disabil Rehabil 2002, 24:727-733. Universal construct? Psychometric findings from 25 countries. Eur J 61. Willems JHBM, Kroneman H: In Arts of advocaat? Nieuwe wao verandert de Psychol Assess 2002, 18:242-251. rol van verzekeringsarts (Physician or lawyer? New WAO alters the role of the 78. Smulders PGW, Andries F, Otten F: Hoe denken Nederlanders over hun werk? insurance physician). Volume 9. MDC; 2003:337-340. Opzet, kwaliteit en eerste resultaten van de TNO Arbeidssituatie Survey. (What 62. De Bont A, Van den Brink JC, Berendsen L, Boonk M: In De beperkte controle do Dutch people think about their work? Design, quality and first results of the van de informatie voor de arbeidsongeschiktheidsbeoordeling (The limited TNO Working Conditions Survey) Hoofddorp: TNO Arbeid; 2001. checks on information for work disability assessment). Volume 146. Ned 79. Kristensen T, Borg V: Copenhagen Psychosocial Questionnaire (COP-SOQ) Tijdschr Geneeskd; 2002:27-30. Kopen-hagen: National Institute of Occupational Health; 2000. 63. Van Glabbeek RHJ: Houden verzekeringsartsen zich aan de wet? Het 80. Goudswaard A, Dhondt S, Kraan K: Flexibilisering en Arbeid in de Informatie- Schattingsbesluit en de Standaard Geen Duurzaam Benutbare Mogelijkheden maatschappij; werknemersvragenlijst, bestemd voor werknemers van bedrijven in de praktijk (Do insurance physicians comply with the law? The Estimation die deelnemen aan het SZW-Werkgeverspanel 1998 (Flexibilisation and Work in Decree and the Permanent Full Disability Standard) Dissertation for the the Information Society; questionnaire for employees of businesses that take occupational health and safety training course for doctors) Netherlands part in the 1998 Employers Panel of the Ministry of Social Affairs and School of Public and Occupational Health (NSPOH): Utrecht; 2003. Employment) Hoofddorp: TNO Arbeid; 1998. 64. Sprenger WJ: Arbeidsdeskundigen en verzekeringsgeneeskundigen van de 81. Houtman I, Goudswaard A, Dhondt S, Van der Grinten M, Hildebrandt V, gemeenschappelijke medische dienst over de gevolgen van de stelselherziening Van der Poel EGT: Evaluation of the monitoring study of stress and physical voor hun praktijk (Labour experts and insurance physicians of the Joint load The Hague: VUGA; 1995. Medical Examination Service (GMD) on how a revision of the system would 82. Schaufeli WB, Van Dierendonck D: The construct validity of two emotional affect their practice) Amsterdam: GMD (Joint Medical Examination Service); exhaustion measures. J Organ Behav 1993, 14:631-47. 1989. 83. De Boer WEL, Wijers J, Spanjer J, Van der Beijl I, Zuidam W: 65. De Boer WEL, Croon NHTh, Van der Toorn M, Csánky-Achilles HW: In Gespreksmodellen in de verzekeringsgeneeskunde (Discussion models Enquête Medisch Arbeidsongeschiktheidscriterium (Medical Work Disability in insurance medicine). TBV 2006, 14:17-23. Criterion Survey). Volume 4. TBV; 1996:98-104. 84. De Dreu CKW, Evers A, Beersman B, Kluwer ES, Nauta A: A theory based 66. Brenninkmeijer V, Lagerveld SE, Blonk RWB: In Moeilijk objectiveerbare measure of conflict management strategies in the workplace. J Organ klachten in de praktijk van de bedrijfs- en verzekeringsarts. Een empirisch Behav 2001, 22:645-668. onderzoek (Complaints in the practice of company doctors and insurer’s 85. Nagtegaal GH: De rol van het dagverhaal in de oordeelsvorming van de medical advisers that are hard to assess objectively: an empirical study). verzekeringsarts ten aanzien van AAW/WAO beoordelingen (The role of the Volume 14. TBV; 2006:354-359. client’s account of daily activities in assessment by insurance physicians for the 67. Van Berkel NA, Njoo T: Overleg met de behandelende sector: het effect van purposes of the AAW/WAO) Utrecht: Netherlands School of Public Health informatie van de behandelaar op de WAO-beoordeling, vergelijkend (NSPH); 1997. 86. SPSS 15.01 Chicago (Il): SPSS Inc; 2006. onderzoek tussen twee UVI’s. (Consultation with the treating sector: the effect of information from the treating physician on the WAO assessment: 87. Du Toit M, Du Toit S: Interactive LISREL: User’s guide Lincolnwood (Il): comparative surveys between two social security administration agencies) Scientific Software International Inc; 2001. Amsterdam: Netherlands School of Occupational Health (NSOH); 2001. 88. Jöreskog KG, Sörbom D: Interactive Lisrel (version 8.72) Lincolnwood (Il): 68. Mortelmans AK, Donceel P, Lahaye D, Bulterys S: Does enhanced Scientific Software International Inc; 2004. information exchange between social insurance physicians and 89. SPSS Categories Chicago (Il): SPSS Inc; 1990. occupational physicians improve patient work resumption? A controlled 90. Gifi A: Homals users Guide Department of Datatheory, University of Leiden; intervention study. Occup Environ Med 2006, 63:495-502. 1981. 69. Davis P, Gribben B, Scott A, Yee RL: The ``supply hypothesis’’ and medical 91. Van Rijssen HJ, Schellart AJM, Anema JR, Van der Beek AJ: A theoretical practice variation in primary care: testing economic and clinical models framework to describe communication processes during medical of inter-practitioner variation. Soc Sci Med 2000, 50:407-418. disability disability evaluations. BMC Public Health 2009, 9:375. Steenbeek et al. BMC Public Health 2011, 11:1 Page 18 of 18 http://www.biomedcentral.com/1471-2458/11/1 92. Ogden J: Some problems with social cognition models: a pragmatic and conceptual analysis. Health Psychol 2003, 22:424-428. 93. Weggemans M: Leiding geven aan professionals? Niet doen! (Providing leadership to professionals. Don’t do it!) Schiedam: Scriptum; 2008. 94. De Boer WEL: Quality of evaluation of work disability. PhD Thesis University of Amsterdam; 2010. Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1471-2458/11/1/prepub doi:10.1186/1471-2458-11-1 Cite this article as: Steenbeek et al.: The development of instruments to measure the work disability assessment behaviour of insurance physicians. BMC Public Health 2011 11:1. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit

Journal

BMC Public HealthSpringer Journals

Published: Jan 3, 2011

There are no references for this article.