Satisfaction of staff of Swiss insurance companies with medical appraisals: a cross sectional study

Included insurances and medical appraisals

We prospectively included all consecutive medical appraisals with arrival at Swiss insurances from February 1st to April 30th, 2008. The date of commission of the appraisals was not relevant. Included Swiss insurance companies comprised the Swiss disability insurance (a public and compulsory insurance responsible for occupational reintegration measures and invalidity pensions), loss of income insurances (responsible for payments due to intermediate inability to work), accident insurances (responsible for benefits due to damages) and liability insurances (responsible for third party liabilities).

A medical appraisal was defined as any written medical assessment (with or without direct patient contact) used for appraisal of patients with insurance claims due to acute or chronic medical conditions. Thus, second opinion appraisals of previous medical certificates and appraisals for court cases were also included. We excluded reports about bio-mechanical issues in road accidents, interim reports of treatment progress for sickness insurances or simple physician attestations.

Questionnaire development and data collection

The questionnaire was developed in a stepwise approach. Initially, we implemented an expert board of specialists for medical appraisals with practice knowledge from different fields (medical expertise; design of teaching programs for medical assessors, disability and sickness insurance knowledge; forensic expertise; experience as patient ombudsman). A literature search revealed no suitable validated questionnaire that was applicable in our context. To assure relevance of the items of the questionnaire, the expert group initially defined domains, where problems in the medical appraisal field typically occur (such as coordination between medical expert and insurance staff; formal requirements; response to key questions; usefulness of the appraisal for the commissioner). For each domain, several draft items were formulated. Draft items were amended in the expert group in an iterative process and additional useful items from the literature were incorporated. To deem face and content validity of the items and comprehensibility of formulations by persons outside the expert group, staff of six insurance companies from different fields (such as officers, insurance physicians) checked the items and amendments were made accordingly. Reliability of the questionnaire was not formally tested, but items were grouped along to the usual work-flow of appraisal generation to allow standardised response for insurance staff.

Finally, a pilot test was performed by several participating insurance companies to test the electronic version of the questionnaire using authentic appraisals. The scope was on data entry problems and possible technical difficulties, as well as on comprehensibility of items. No relevant new input for item content or formulations was generated.

The final questionnaire comprised 41 questions, including 10 questions that explored the satisfaction of the commissioner with the final appraisal. Responding was possible on a 3 to 4 point likert scale. The remaining items of the questionnaire covered insurance characteristics, process information and patient data. The questionnaire was translated into the two other main Swiss languages, French and Italian, by translators with bi-lingual mother tongue and medical background. A second bi-lingual translator checked for content consistence.

Insurance staff answered the questionnaires anonymously in a prospective manner when the companies received the appraisals. As this was a country wide study, several measures were implemented to increase data quality and participation rate (such as teaching sessions and continuous data controlling). Ethical approval was provided by the local ethics committee.

Statistical analysis

For our descriptive analysis, we used means (SD) and medians (IQR with 25%-75% percentiles) for continuous variables and proportions for categorical data. For inferential analysis we applied parametric and non-parametric tests. The unit of analysis was the medical appraisal. For comparison of predefined subgroups we grouped appraisals of the Swiss disability insurance as "disability insurance" and appraisals of private accident, liability and loss of income insurances as "private insurances". These groups often represent different patient problems (chronic morbidity vs. damage events) as well as different organisational and legal features (public, compulsory insurance vs. private insurance companies). The two groups together comprised 98.6% of all included appraisals. Appraisals from the psychiatric specialty comprised another predefined subgroup, as our expert group judged psychiatric appraisals as specifically complex. For this subgroup, we included only mono-disciplinary psychiatric appraisals to avoid contamination from other specialties. To evaluate the contribution of single factors to overall satisfaction, we performed a multivariable logistic regression analysis (dependent variable: "very satisfied with appraisal" yes/no; examples of independent variables: "good formal structure and content" yes/no, "conclusions sound and comprehensible" yes/no; with controlling for potential confounders such as patient age, sex and involved number of medical specialties per appraisal). Data analysis was done with SPSS for Windows, version 18.0 (SPSS Inc., Chicago, Illinois); for regression analysis we used Stata 9.0 (StataCorp 2004, Stata Statistical Software, College Station, TX).

Medical appraisals and patients

The mean patient age was 45.9 years and 68% of the patients (2115/3165) were in the age group between 41 and 60 years. 50.7% of the patients were female. In the past, most of the patients had been employed (n = 2617; 83%) or self-employed (n = 206; 6.5%).

Satisfaction with medical appraisals

Satisfaction of the staff of the Swiss disability insurance with appraisals of medical assessors for their patients was generally lower compared to private insurances. The most striking differences were found for the domains "timeliness of appraisals" (not satisfied: disability insurance 30.2% vs. private insurances 11.4%; p < 0.001), "completeness of response to key questions" (incomplete response: 14.6% vs. 3.4%; p < 0.001) and "soundness and comprehensibility of conclusions" (not sufficient: 9.3% vs. 4.9%; p < 0.001). Overall satisfaction with 1159 mono-disciplinary psychiatric appraisals was basically the same as for the total population (86% vs.87% "very satisfied" or "satisfied"). Also for items that covered specific sub-domains, no relevant differences were found in comparison to the total population.

Implications for practice and research

The results of our study should be incorporated into the continuous education programme of physicians, who perform medical appraisals. This can increase their awareness that sound and comprehensible conclusions in the medical appraisal are of outstanding importance. Specifically in court cases this is of relevance, as non-medical legal specialist have to base their decisions on comprehensible medical conclusions [7]. Furthermore, standards of time limits for generation of appraisals should be established. Any time loss due to late appraisals will reduce the probability of early reintegration measures for the patient, if indicated.

A specific sub-project of the MGS-study has assessed the correlation between overall satisfaction of the commissioners and the professional quality of appraisals as judged by trained medical experts [10]. Commissioners were able to identify good appraisals sufficiently well, but had problems to identify low quality appraisals. Thus, the high overall satisfaction of insurance staff with appraisals, as seen in our data, may give an optimistic picture concerning the quality of medical appraisals. Satisfaction research in hospitals has shown, that overall satisfaction rates can be partly misleading [11]. Given these findings, continuous educational measures are also necessary for the staff of insurances. This will enable them to select assessors who provide state of the art appraisals in due time. Nevertheless, assessment of satisfaction from the viewpoint of insurance staff remains important. Some relevant issues can only be judged by the "clients" (i.e. the staff of insurances), such as coordination issues [9] and timeliness.

A follow-up study would contribute valuable information about improvements over time. For example, in the context of current reforms of the Swiss disability law, that promotes reintegration measures to avoid invalidity pensions [12]. Similar studies in the neighbour countries of Switzerland may create useful information to learn from each other. We hypothesize, that conditions in countries such as Germany or Austria may not be long apart.