Objectives To investigate sick leave and disability pension in working-age subjects with knee osteoarthritis (OA) compared with the general population.
Methods Population-based cohort study: individual-level inpatient and outpatient Skåne Health Care Register data were linked with data from the Swedish Social Insurance Agency. In 2009 all working-age (16–64 years) Skåne County residents who in 1998–2009 had been diagnosed with knee OA (International Classification of Diseases-10 code M17) were identified and their sick leave and disability pension in 2009 related to those of the general working-age population (n=789 366) standardised for age.
Results 15 345 working-age residents (49.6% women) with knee OA were identified. Compared with the general population, the RR (95% CI) of having had one or more episodes of sick leave during the year was 1.82 (1.73 to 1.91) for women and 2.03 (1.92 to 2.14) for men with knee OA. The corresponding risk for disability pension was 1.54 (1.48 to 1.60) for women and 1.36 (1.28 to 1.43) for men with knee OA. The annual mean number of sick days was 87 for each patient with knee OA and 57 for the general population (age- and sex-standardised). Of all sick leave and disability pension in the entire population, 2.1% of days were attributable to knee OA or associated comorbidity in the patients with knee OA (3.1% for sick leave and 1.8% for disability pension).
Conclusions Subjects with doctor-diagnosed knee OA have an almost twofold increased risk of sick leave and about 40–50% increased risk of disability pension compared with the general population. About 2% of all sick days in society are attributable to knee OA.
Statistics from Altmetric.com
Osteoarthritis (OA) is one of the most common causes of pain and disability in the musculoskeletal system,1,–,4 and the knee joint is one of the principal joints affected.5 ,6 Pain, stiffness, impaired function, reduced activity and participation are the main recognised symptoms, consequences and health indicators.7 ,8 Even though the prevalence of knee OA, and OA of other sites, increases rapidly with age5 ,6 the disease is still not uncommon in subjects of working age.6 ,9 Despite the high prevalence of knee OA2 ,10 ,11 there is limited knowledge of the burden on society with respect to sick leave and disability pension.3 ,12 Studies have suggested that the economic burden of OA is substantial.3 ,13,–,19 However, most reports mainly include older subjects14,–,16 ,20 ,21 of whom many are above retirement age.17 ,22 Estimates of sick leave and work loss are therefore often based on limited numbers of subjects and results are inconclusive.23 Furthermore, most studies are based on survey data with the potential for selection and recall bias, and few studies include comparative figures of sick leave and disability in the general population which put the estimates in a broader perspective.19 ,24
The main purpose of this study was to investigate the extent of sick leave and disability pension in subjects with doctor-diagnosed knee OA, and to put these figures in perspective by comparison with the general population. We used comprehensive prospectively ascertained cohort data covering a geographically well-defined population of 1.2 million inhabitants.
Material and methods
The Skåne Health Care Register (SHCR)—All inpatient and outpatient healthcare provided in the southernmost county of Sweden, Skåne County, is registered in the SHCR (covering a population of 1 231 062 at 31 December 2009). Reporting data to the register constitutes the basis for reimbursement. Diagnoses are classified by the doctors themselves according to the Swedish translation of International Classification of Diseases and Related Health Problems 10 (ICD-10) system. Validation of the diagnostic accuracy in the register has so far been shown to be high for rheumatoid arthritis and spondyloarthritis.25 ,26 About 30% of all visits to a doctor in outpatient healthcare are made to private providers and for those visits diagnoses are not automatically reported to the SHCR.
The Social Insurance Register—Swedish social insurance is administered by the Swedish Social Insurance Agency (SSIA) and covers everyone who lives or works in Sweden. All sickness benefit exceeding 14 days and all disability pension payments are administered and registered by the SSIA. The register also includes one main diagnosis for sickness benefits and two main diagnoses for disability pension.
The Swedish Population Register—The population register is a national register containing statistics of vital events such as births, deaths, change of residential address. The register is administrated by the Swedish Tax Agency.
Sick leave —We defined sick leave as days with sickness benefit registered by the SSIA. In Sweden you are entitled to sickness benefit when you cannot work owing to illness or injury. Sickness benefit covers everyone who legally works in Sweden or who is registered as unemployed. It is generally limited to 1 year but can be extended in certain cases. For the first sick day you do not receive any compensation. If you are employed, your employer will pay sick pay for the next 13 days and from day 15 you receive sickness benefit and from then on are registered in the SSIA registers.
Disability pension—In Sweden if your work ability is permanently reduced by at least one-quarter you can receive a disability pension.
Both sick leave and disability pension can be granted for a full, three-quarters, half, or one-quarter of a day depending on the extent to which your work ability is reduced. In this study we calculated sick leave and disability pension as the proportion of individuals being granted such compensation and as net sick days. Net sick days (full-time sick days) are the total number of days for which sickness benefit or disability pension payment is received from the SSIA, multiplied by the extent of the sick leave or disability pension for each day (eg, 20 sick days with one-quarter of a day extent are equal to 5 net sick days).
Defining cases of knee OA
Using the SHCR for the 12-year period, from 1998 to 2009, we identified all men and women who had been diagnosed at least once in primary care, secondary care, or in hospital care with OA of the knee (ICD-10 code M17). We required subjects to be aged 16–64 in 2009 as the general age of retirement in Sweden is 65 years and the rules for sick leave and disability pension change thereafter. We further required them to be resident in the Skåne County on 31 December 2008 or 31 December 2009, via cross-linkage with the population register.
Defining the general population
For comparison we used the general population including all men and women aged 16–64 years (in 2009) and registered as a resident in the Skåne County by 31 December 2008 or 31 December 2009.
In a sensitivity analysis, we evaluated the effect on the risk estimates by excluding subjects from the general population who had not consulted a doctor in 1998 to 2009.
Study measures and statistical methods
Using subjects' unique 10-digit personal identification number, we linked (on an individual level) year 2009 social insurance data with data from the SHCR.
First, we calculated the proportion of male and female patients with knee OA who during 2009 had received sickness benefit and/or a disability pension payment.
Second, we estimated the increase in risk of one or more episodes of sick leave or disability pension during 2009 for those with knee OA compared with the general population, standardised for age. We also evaluated the proportion of men and women who had received sickness benefit or disability pension payment in different age categories.
Third, for 2009 we calculated the mean number of net sick days with sick leave and disability pension compensated by SSIA for each subject with knee OA and compared those with the numbers for the general population standardised for age. (For the reason for using the mean as a summary measure see Neovius et al27 ,28) We calculated the mean number of sick days per year and per person for different age categories. To estimate the mean length of the periods, we also calculated the mean number of net sick days per compensated individual—that is, dividing the net sick days only by those who had received sick leave or disability pension.
Fourth, we were interested in the proportion of the total number of days of sick leave and disability pension due to knee OA. To determine this we first calculated the ‘excess’ number of sick days in patients with knee OA. Hence, we calculated the total amount of days observed in patients with knee OA and subtracted the total amount of days expected considering the same rate of sick leave/disability pension as in the general population without knee OA. This ‘excess’ can conceptually be considered as attributable to knee OA or to associated comorbidity, life style or socioeconomic background factors in these patients. We then divided the number of ‘excess days’ by the total number of sick days seen in the entire population.
So far, all of our analyses have been based on data on sick leave and disability pension, irrespective of the cause. However, fifth, we also calculated the proportion of days with sick leave and days with disability pension for which knee OA was specifically registered as the main cause by the listing doctor.
For age-standardised estimates, the proportions and the number of net sick days for the general population are standardised according to the age distribution of the subjects with knee OA—that is, to the pattern ‘expected’ for the knee OA cohort provided that they had the same sick-listing pattern as the general population.
We tested differences in proportions using χ2 test and considered a two-tailed p value <0.05 to be statistically significant. We calculated standardised RRs and their 95% CIs by dividing the observed by the expected proportion of subjects with knee OA with sick leave and (or) disability pension.
We identified 15 345 people (49.6% women) who had been diagnosed with knee OA during 1998 to 2009 and were of working age and resident in Skåne County during 2009. They had a mean (SD) age of 55 (8.2) years for women and 53 (9.2) years for men (table 1). Among women with knee OA, a total share of 48% had received either sickness benefit or disability pension payment, or both, during the year 2009, and among men with knee OA the corresponding proportion was 31% (p<0.001).
Twenty-one per cent of women and 17% of men with knee OA had received sickness benefit at some point in time during the year. The risk of sick leave was increased in all age categories for both genders (figure 1). The overall age-standardised and sex-specific RR for sick leave compared with the general population was 1.82 (95% CI 1.73 to 1.91) for women and 2.03 (95% CI 1.92 to 2.14) for men (table 2).
During 2009 more women with knee OA than men with knee OA had received disability pension payment to some extent (32% vs 16%, p<0.001, table 1). The share of patients with knee OA with disability pension payment increased with increasing age (figure 2). The overall age-standardised RR for having received disability pension was 1.54 (95% CI 1.48 to 1.60) for women and 1.36 (95% CI 1.28 to 1.43) for men (table 2).
Mean numbers of sick days
The mean (SD) number of days with sick leave and (or) disability pension (irrespective of the cause) in subjects with diagnosed knee osteoarthritis was 87 (140) days (number of days divided by those both with and without sick leave/disability pension during the year). In the general population the corresponding age- and sex-standardised number of days was 57 (78).
Women with knee OA had a mean number of 18 (57) days with sick leave and 94 (149) days with disability pension per subject and year. Men had a mean number of 15 (52) days of sick leave and 47 (115) days of disability pension. The corresponding age-standardised number of net sick days for sick leave and disability pension in the general population was 9 (41) and 62 (128) days for women and 7 (37) and 37 (104) days for men (table 3).
The mean extent of the sick leave and disability pension during the year (irrespective of cause) calculated as mean (SD) number of net days per compensated individual was 86 (98) days of sick leave and 291 (103) days of disability pension among women and 88 (96) days of sick leave and 292 (99) days of disability pension among men with knee OA. For sick leave these numbers were slightly higher than for the general population and for disability pension slightly lower (data not shown).
Days attributable to knee OA
Of all sick leave and disability pension in the entire population, 2.1% of days were estimated to be attributable to knee OA or associated comorbidity, life style or socioeconomic background in the patients with knee OA (sick leave=3.1%, 2.7% for women and 3.6% for men, and disability pension=1.8%, 2.3% for women and 1.1% for men).
According to the main diagnosis on doctor's sick listing certificate, 1.4% (1.2% for women and 1.6% for men) of all net days with sick leave and 1.2% (1.2% for women and 1.3% for men) of all net days with disability pension were due to knee OA (table 3).
In the general working-age population 747 815 of the 789 366 subjects (94.7%) had consulted a doctor at least once in 1998–2009. Using this alternative reference population yielded, as expected, marginally lower risk ratios. However, the estimates were still similar—for example, risk of sick leave for those with knee OA was 1.79 (95% CI 1.70 to 1.88) for women and 1.96 (95% CI 1.86 to 2.07) for men (the previous point estimates were 1.82 and 2.03, respectively).
To our knowledge this is the first study presenting prospectively ascertained comprehensive population-based data on the increased risk of both sick leave and disability pension for those with knee OA. We also estimate the proportion of sick leave and disability pension in the Swedish society that was attributable to knee OA. Men and women with a diagnosis of knee OA have an almost twofold increased risk of sick leave and about 40–50% increased risk of disability pension compared with the general population. In the entire population about 2% of sick days were attributable to knee OA, comorbidities or other associated factors in the patients with knee OA.
Several studies have estimated the burden associated with OA.3 ,13,–,19 ,24 Most of them agree that the economic burden of OA is substantial. The French COART study, using official healthcare and government health insurance data, suggested that OA of all sites contributed to about 2.6% of all sick leave days among employed individuals and compared it with that of coronary heart disease.18 As OA is often considered a disease of older people, several studies focus on people of whom only a minority is still working.14,–,17 ,20,–,23 Owing to differences in social security systems and labour markets, but also differences in study design, comparisons with previous studies are hard to make. Most studies on the burden of OA are based on survey data3 ,5 ,22 or on structured interviews,14,–,16 and data on sick leave are based on the patients' own retrospective recall. Despite this, the self-reported estimates of sick leave from some of these studies are fairly consistent with the findings of our study. One study estimated that 20.5% of patients with knee OA reported having missed a work day because of knee OA22 as compared with 19.1% with sick leave in our study (not including sick leave periods <15 days for employees). In another study patients reported 12–19 mean days of sick leave a year because of their OA (including all OA)17 as compared with 17 days a year in our study.19 ,24 In one previous study comparing patients with OA with subjects drawn from the general population, 10.5% of the former reported reduction in work hours and 13.7% reported having retired early owing to illness compared with 1.7% and 3.4% of their non-arthritic controls.
In general, estimates for disability pension in patients with OA are hard to find. One reason for this is probably the large differences in the way in which compensation is paid for permanent work loss.
This study has some important limitations. First, the identification of patients with knee OA is based solely on patient-driven clinic visits and the doctor's diagnostic coding. Although no validation of medical records has been performed, the validity of other chronic rheumatic disease in the SHCR has so far proved to be high.25 ,26 Importantly, the demographics of the knee OA cohort fulfil the expected age and sex distribution, and the proportion of subjects who had had total knee replacement is far higher than in the general population, which supports high accuracy of the diagnostic coding (data not shown). Any misclassification of disease is expected to be non-differential biasing the estimates of effect toward the null (less difference).
Second, patients with knee OA who have visited only private-care providers will be missed as diagnoses by those providers are not reported to the SHCR. This might lead to an underestimation in the calculation of days attributable to knee OA.
Third, for employees, periods of sick leave of ≤14 days are not registered by the SSIA, and hence not included. This limitation leads to a slight underestimation of the total proportions having had sick leave and the number of sick days, but to a slight overestimation of the mean length of the sick leave.
Finally, the increased risk for sick leave and disability pension in these patients, as well as the calculation of days attributable to knee OA might be affected by other factors, not included in this study, such as comorbidity, life style or socioeconomic factors. Unfortunately, we did not have the opportunity to adjust or standardise the analysis for these factors in this study. However, we see a need for future studies on their effect.
The strengths of the study include the use of comprehensive data covering an entire population and the opportunity to make direct comparisons with the general population. We can, for example, obtain both absolute and relative estimates on the burden on society with respect to sick leave. We can also study combined estimates on sick leave and disability pension which is important as sick leave often decreases as disability pension increases. Further, variations due to social security system changes or labour market factors probably affect patients with knee OA in a similar manner to the rest of the population and data are not affected by survey response bias or subjects' recall bias.
In conclusion this study provides robust data to show that patients with knee OA have an almost twofold increased risk of sick leave and about 40–50% increased risk of disability pension compared with the general population. Further, in the Swedish population, despite knee OA often being viewed as a ‘disease of the older people’, over 2% of the total amount of sick days in the society are attributable to knee OA or associated comorbidities, life style or socioeconomic background in the patients with knee OA.
The authors would like to acknowledge Charlotte Bergknut and Anna Jöud for their assistance in the data extraction. The study was funded by grants from the regional health service authorities in Skåne (Region Skåne), the Swedish Social Insurance Agency, The Swedish Research Council, Kock Foundations, King Gustaf V 80-Year Birthday Foundation, and the Faculty of Medicine at Lund University.
Funding The regional health service authorities Region Skåne, the Swedich Social Insurance Agency, The Swedish Research Council, Kock Foundations, King Gustaf V 80-Year Birthday Foundation, the Faculty of Medicine at Lund University.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.