The European Journal of Public Health Advance Access originally published online on August 26, 2005
The European Journal of Public Health 2005 15(5):470-474; doi:10.1093/eurpub/cki029
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Health Services Research |
Influence of local structural factors on physicians' sick-listing practice: a population-based study
Britt E. Arrelöv1, Lars Borgquist2 and Kurt F. Svärdsudd1
1 Uppsala University, Department of Public Health and Caring Sciences, Family Medicine and Clinical Epidemiology Section, Uppsala, Sweden
2 Linköping University, Faculty of Health Sciences, Department of Medicine and Care, Primary Care, Linköping, Sweden
Correspondence: Dr Britt Arrelöv, Department of Public Health and Caring Sciences, Family Medicine Section, Uppsala Science Park, 751 85 Uppsala, Sweden, tel: +46 18 611 34 26, fax: +46 18 51 16 57, e-mail: britt.arrelov{at}pubcare.uu.se
Received April 9, 2003, accepted June 21, 2004
 |
Abstract |
|---|
 Top Abstract MethodsResultsDiscussionReferences |
|---|
Background: Physicians have a central role as gatekeepers to the social security system, including sick-listing. Variation in physicians' sick-listing practices has been demonstrated in several studies. The objective of this study was to determine whether local structural factors affect sick-listing practice. Methods: A total of 57 563 consecutive sick-listing certificates, issued during 4 months in 1995 and 2 months in 1996, were collected from the local branches of the National Social Insurance Office in eight Swedish counties. County code, local community population size and presence of a hospital in the area were used as indicators of local structural factors. Length of the sick-listing certificates and of the sick-listing episodes were used as outcome variables. Results: After ajustment for the influence of category of issuing physician, patients' age, sex and diagnosis (‘case mix’), and type of certificate there was a large variation of the length of the sick-listing certificates and of the sick-listing episodes between counties, between communities of various size and between communities with or without a hospital in the area. All these factors were independently and significantly correlated to the length of the certificate and of the sick-listing episode. Conclusions: The results support the hypothesis that physicians' sick-listing practice is influenced by local structural factors.
Keywords: epidemiology, sick-leave, sick-listing, sickness certification practice, structural factors, Sweden
Sickness benefits are meant to reduce the economic burden for a person with reduced work capacity due to illness. A person's absence from work depends on a combination of the relative strength of the tendency to appear at work and the barriers to attendance.1 Sickness absence may be understood as a practice in which subjects take into account both their health and the rules of the community in which they are living.2
Physicians play a central role as gatekeepers to social security systems, when illness or disability is used as an eligibility criterion. The role of physicians is to mediate between the claims of individuals and the formal rules of organizations.3 In their role as rational agents, physicians are expected to act in the best interest of both patients and society,4 and satisfy a number of goals.5
Sick-listing practice variations between individual physicians, physician categories and physicians in various geographical areas have been demonstrated in several studies.6–11 Socio-economic conditions, the labour market, culture and attitudes have been found to explain some of the local variation of sickness absence.2,12,13
It has also been proposed that physicians adjust their certification practices according to local sickness absence practice.2,13 This may be even more pronounced in countries like Sweden, where health care is owned and run by local authorities (counties). We therefore performed this study to determine whether physicians adjust their sick-listing practice according to local structural factors, such as county, municipality size and presence of a hosptal in the area, which in turn may be proxies for structure and organization of health care, socio-economic conditions, labour market and views on the legitimacy of absence from work.
|
Methods |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
The Swedish National Insurance Act covers all residents 16–64 years of age and regulates sickness benefit. The first 2 weeks of sickness absence were, during the study period, for employed persons, compensated by their employer. Thereafter, and for unemployed persons during the whole period, the sickness benefits were paid by the national social insurance system. Except for the first week, a sickness certificate must be issued by a physician to obtain sickness benefit. The first certificate in a sickness episode that was received by the local social insurance office was called the initial certificate, and any following certificates were called prolongation certificates.
The sampling unit was sickness certificates. A total of 57 563 certificates were received by the local social insurance offices in 27 municipalities in eight counties during 4 months in 1995 and 2 months in 1996. Four of the counties used capitation for reimbursement of primary care and four used overall budget. At least one city or municipality with a hospital and at least one with no hospital in each county was included. The area covered included urban and rural districts.
Variables used for this report were insurance office code, county code, the patient's age and sex, main diagnosis, category of certifying physician, type of certificate (initial or prolongation), length of the total sickness episode until the present certificate and length of the current certified period after adjustment for degree of sickness absence, expressed as number of net days. Physician category was classified as general practitioner (GP), hospital physician, occupational medicine physician or other physician (mainly private practitioners).
Diagnoses were grouped into musculoskeletal disorders, psychiatric disorders, injuries, gastrointestinal disorders, cardiovascular disorders, respiratory disorders and miscellaneous conditions. In addition, information on presence of a hospital in the municipality (none, small or large) and municipality population size in the age segment 16–64 years was obtained.
Statistical considerations
The analyses were performed with the SAS14 and JMP15 programme packages. The partial non-response rate (missing data within sampled certificates) was <1%. Summary statistics were computed with standard parametric methods. Univariate analyses of group differences were performed with Student's t-test or analysis of variance. Multivariate analyses were performed with ordinal logistic regression or standard least-squares regression technique.
All tests were two-tailed. Probability values <0.05 were regarded as statistically significant. Very small P values were denoted as <0.0001, even when they were much smaller.
|
Results |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
The study population
The county share of certificates was 7–30% (table 1). The distribution of certificates was similar for the three municipality population size groups. Forty-six percent of the certificates came from municipalities with a large hospital. GPs and hospital physicians issued 80% of the certificates. Mean patient age was 45 years, and 60% of the certificates were issued to women. The largest diagnostic group was musculoskeletal disorders, followed by psychiatric diseases and injuries. On average, 31 net days were issued in the current certificate and 193 days for the total sick-listing episode.
|
Effects on number of net days and length of episode
The range of crude net days issued in the various counties was 29–35, and for the sick-listing episodes was 151–217. However, there were significant differences between the counties in the distribution of type of certificate, patient age and sex, diagnostic groups, category of issuing physician and type of certificate. When the influence of these variables was taken into account, the rank order between the counties was moderately changed and the range of net days and episode of sickness increased somewhat.
Certificates from small municipalities tended to have fewer crude net days and shorter episodes than certificates from large municipalities. However, GPs issued most certificates in the smallest and largest municipalities, and hospital physicians in middle-sized municipalities, and the largest municipalities had a higher proportion of psychiatric and a lower proportion of musculoskeletal diagnoses. After adjustment for these differences, the medium-sized communities had the smallest number of adjusted net days and shortest episodes.
Certificates in municipalities with no hospital had the smallest number of crude net days and the shortest episodes. There were also other differences, for instance in the proportion of certificates issued by the various physician categories, and in degree of sick-listing. After adjustment, the rank order according to net days remained, whereas the municipalities with a small hospital had the longest episodes.
GPs issued the shortest certificates and shortest episodes, also after adjustment for all other factors. The number of certified days and episode length increased with patient age, and men had more certified days than women. Respiratory diseases had the shortest certificates and episodes and cardiovascular diseases the longest certificates, and musculoskeletal diseases the longest episodes.
The combined effect of physician category and presence of a hospital in the municipality adjusted for the influence of patient age and sex, diagnostic group, county and population size is shown in figure 1. GPs in municipalities with no hospital had the lowest number of net days and hospital physicians in municipalities with a large hospital the highest.
|
|
Discussion |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
We thus found practice variations, as reflected in length of sick-listing certificates, in relation to local structural factors. The counties were chosen from a strategic point of view to represent a variety of health-care organization, geography, and urban and rural districts with and without hospitals. There was some unavoidable clustering, for instance large hospitals are mostly located in cities, whereas small municipalities usually have no hospital. However, these effects were largely taken into account in the multivariate analyses.
Since the number of certified days is the focus of this report, sick-listing certificates were the sampling unit. The identity of individual patients or physicians was not available and the number of sick-listing physicians, sick-listed persons or the number of certificates per patient was unknown. The certificates were collected during 6 months, scattered over 2 years, to account for seasonal variation and variation from one year to another. The insurance benefit came from one source, the national social insurance office, with a standardized basis for forming judgements. The large sample size provides a considerable statistical power, with an option to make subgroup analyses with enough power.
Local influencing factors
Sex segregation, socio-cultural factors, epidemics, socio-economic conditions, demographic variables, type of labour market, local practice in medical care, type of social insurance and organization of the local community and health care have been found to influence sickness absence in various directions on the regional and municipality level.2,6,16 The demography, labour market, socio-economic levels and organization of health care differs between small communities and cities. The same is valid for communities that do or do not have a hospital in the area. Furthermore, the counties included in this study differ in their geographical, demographic and labour market structure, and the regional health-care structure is governed by the county councils. Differences in sick-listing practice between individual physicians and groups of physicians might arise because of all the above mentioned factors, as well as local sickness absence culture, values and traditions.
In a study of practice variation in Swedish primary care, it was found that longer sick-listings were prescribed by GPs working at a distance from a hospital.11 In the present study, we found, if anything, that all physician categories in municipalities without a hospital issued shorter certificates. In the county with the highest mean for net days of sickness absence, hospital physicians and occupational health-care physicians issued much longer certificates than in other counties. A possible explanation for the higher mean number of net days might be that a group of physicians with special interests, knowledge or views about sickness absence were working in that county, i.e. the enthusiasm hypothesis.17
Sick-listing practice variation
Sick-listing practice variation exists between physician categories and physicians in different regions.2,10,11,18 Physicians working in general practice issued shorter certificates than other physician categories. This may be due to their more profound training in sickness certification and their training as gate-keepers to medical and social benefits.
Physicians have been found to adjust their certification practices according to local sickness absence practice.2 It seems that they adjust their certification practice in different ways, due to differences in their values regarding sick-listing and their way of meeting patients' needs.19 The findings in this study support the view that it is essential to take into account the social context in which the decision-making activities take place.20
Influence of attitudes and absence culture
The various attitudes towards sickness absence in workplaces, society and professional groups influence physicians' attitudes, and the predominant attitude will be incorporated into the sick-listing decisions. However, the context in which the physicians work will have an effect on the influence, demonstrated by differences between physician categories and the findings regarding the influence of the presence of a large or small hospital on the different kinds of physicians. Physicians living and working near their patients, such as GPs and physicians working in small hospitals or local surgeries, will probably be more strongly affected by norms and tradition in the local society than other physicians, and will perhaps be part of the local social life and concerned about the impact on their status in the community.21 Despite the fact that most physicians working in occupational health organizations are in the same situation, the attitudes and culture of their company probably has an influence on them. Physicians working in large hospitals are not so close to the local society, and the professional systems' views and opinions will have a larger impact on their practice.22
Sick-listing decisions include a social control function, patients' expectations, the preservation of the patient–physician relationship and the moral values held by the profession.23 Physicians will adjust their sick-listing practice in order to reduce dissonance between their own beliefs and others' expectations.24 The preservation of the doctor–patient relationship is important, as well as the views of peers.7,25
In conclusion, these results support the hypothesis that physicians' sick-listing practice is influenced by local structural factors. The relative importance of these and other influencing factors needs further investigation.
Key points
|
|
Acknowledgments |
|---|
We are indebted to Mr Dan Ljungberg, statistician at the National Social Insurance Board at the time of the study, for valuable contributions. This study was supported by grants from the Swedish National Board of Health and Welfare, the National Social Insurance Board and 28 of its local offices.
|
References |
|---|
|
TopAbstractMethodsResultsDiscussionReferences |
|---|
1 Ås D. Absenteeism—a social fact in need of a theory. Acta Sociol 1962;6:278–86.
2 Virtanen P, Nakari R, Ahonen H, et al. Locality and habitus: the origin of sickness absence practices. Soc Sci Med 2000;50:27–39.[CrossRef][ISI][Medline]
3 Coe J. The physician's role in sickness absence certification: a reconsideration. J Occup Med 1975;17:722–4.[ISI][Medline]
4 Wennberg JE. Practice variations and the challenge to leadership. Spine 1996;21:1472–8.[CrossRef][ISI][Medline]
5 Eisenberg JM. Doctors decision and the cost of medical care. The reason for doctors' practice patterns and ways to change them. Ann Arbor, MI: Health Administration Press Perspective, 1986.
6 Rutle O, Forsen L. Inter-doctor variation in general practice. Report no. 8. Oslo: National Institute of Public Health, Department of Health Services Research, 1984.
7 Mabeck CE, Kragstrup J. Is variation a quality in general practice? Scand J Prim Health Care 1993;11:32–5.
8 Condren L, Cox J, McCormick JS, Sullivan A. Certification of unfitness for work. Ir Med J 1984;77:159–60.[ISI][Medline]
9 Alexandersson K, Leijon M, Akerlind I, et al. Epidemiology of sickness absence in a Swedish county in 1985, 1986 and 1987. A three year longitudinal study with focus on gender, age and occupation. Scand J Soc Med 1994;22:27–34.[ISI][Medline]
10 Arrelöv B, Borgquist L, Ljungberg D, Svärdsudd K. Do GPs sick-list to a lesser extent than other physician categories. Fam Pract 2001;18:393–8.
11 Peterson S, Eriksson M, Tibblin G. Practice variation in Swedish primary care. Scand J Prim Health Care 1997;15:68–75.[ISI][Medline]
12 Moncrieff J, Pomerleau J. Trends in sickness benefits in Great Britain and the contribution of mental disorders. J Public Health Med 2000;22:59–67.
13 Isacsson A, Hansson BS, Janzon L, Kugelberg G. The epidemiology of sick leave in an urban population in Malmo, Sweden. Scand J Soc Med 1992;20:234–9.[ISI][Medline]
14 SAS language and procedures. Version 6.12. Cary, NC: SAS Institute Inc., 2000.
15 JMP version 4.0.2. Statistics and graphics guide. Cary, NC: SAS Institute Inc., 2000.
16 Alexandersson K. Sickness absence: a review of performed studies with focused on levels of exposures and theories utilized. Scan J Soc Med 1998;26:241–9.
17 Chassin MR. Explaining geographic variations. The enthusiasm hypothesis. Med Care 1993;31:37–44.
18 Englund L, Tibblin G, Svärdsudd K. Variations in sick-listing practice among male and female physicians of different specialities based on case vignettes. Scan J Prim Health Care 2000;1:48–52.
19 Wennberg JE, Barnes BA, Zubkoff M. Professional uncertainty and the problem of supplier-induced demand. Soc Sci Med 1982;16:811–24.[CrossRef][ISI][Medline]
20 Humphrey P, Berkeley D. Problem structuring calculi and levels of knowledge representation in decision making. In: Scholtz RW, editor. Decision making under uncertainty. Amsterdam: Elsevier, 1983:121–57.
21 Eisenberg JM. Physician utilization. The state of research about physicians' practice patterns. Med Care 1985;23:461–83.[CrossRef][ISI][Medline]
22 Lopez Fernandez LA, Jimenez Martin JM, Luna del Castillo J de D, et al. Sources of influence on medical practice. J Epidemiol Commun Health 2000;54:623–30.
23 Murray TH. Divided loyalties for physicians: social context and moral problems. Soc Sci Med 1986;23:827–32.[CrossRef][ISI][Medline]
24 Festinger L. A theory of cognitive dissonance. Stanford, CA: Stanford University Press, 1957.
25 Bradley CP. Factors which influence the decision whether or not to prescribe: the dilemma facing general practitioners. Br J Gen Pract 1992;42:454–8.[ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||