Explanations for gender differences in sickness absence: evidence from middle-aged municipal employees from Finland

Transcription

1 Explanations for gender differenes in sikness absene: evidene from middle-aged muniipal employees from Finland M Laaksonen, 1 P Martikainen, 2,3 O Rahkonen, 1 E Lahelma 1 Original artile 1 Department of Publi Health, University of Helsinki, Helsinki, Finland; 2 Helsinki Collegium for Advaned Studies, University of Helsinki, Helsinki, Finland; 3 Population Researh Unit, Department of Soiology, University of Helsinki, Helsinki, Finland Correspondene to: Dr M Laaksonen, Department of Publi Health, PO Box 41, FIN University of Helsinki, Finland; mikko.t.laaksonen@ helsinki.fi Aepted 14 September 2007 ABSTRACT Objetives: To examine gender differenes in sikness absene spells of various lengths and to explain these differenes by health status, working onditions and family-related fators. Methods: The study inluded 5470 female and 1464 male employees of the City of Helsinki surveyed at baseline in These survey data were linked to the employer s sikness absene reords until the end of 2005, providing a mean follow-up time of 3.9 years. Explanations for gender differenes in self-ertified (1 3 days) and medially onfirmed absene spells of various lengths (4 days or more, more than 2 weeks, and more than 60 days) were examined using Poisson regression. Results: Women had 46% higher risk for self-ertified sikness absene than men. In medially onfirmed spells there was 34% female exess whih gradually weakened with lengthening absene, and no differenes were observed in spells longer than 60 days. Adjusting for physial funtioning and self-reported diagnosed diseases learly attenuated gender differenes in sikness absene spells shorter than two weeks and fully explained them in longer absene spells. Physial work demands explained female exess in medially onfirmed absene spells of all lengths, as did work fatigue in spells longer than two weeks. Psyhosoial working onditions and familyrelated fators did not affet the gender differenes. Physial health problems, physial work demands and work fatigue were somewhat more prevalent in women than in men, but their impat on sikness absene was similar in both genders. Conlusions: The overall gender differenes in sikness absene are due to relatively short absene spells being more ommon among women. In longer sikness absene spells the female exess is mainly explained by heavier burden of ill-health and to a lesser extent by higher physial work demands among women. The authors found no support for greater vulnerability to health- and work-related problems among women as reasons for sikness absene. Most previous studies have shown that women have more sikness absene than men. 1 2 Female exess in sikness absene has been observed in ountries with differenes in the overall levels of sikness absene, the soial insurane system and the level of female labour fore partiipation, although there are also exeptions to this general finding. 3 While the studies typially show a larger number of spells and longer umulative duration of sikness absene in women than in men, with respet to the length of absene the findings tend to vary. A Frenh study showed about twofold risk for women in medially onfirmed sikness absene spells of all lengths. 4 A Norwegian study showed the largest female exess in sikness absene spells of medium duration. 5 British studies based on the Whitehall II data have shown larger gender differenes in self-ertified than in medially onfirmed sikness absene. 6 7 It is therefore not lear how spells of different lengths ontribute to the overall gender differenes in sikness absene. The female exess in sikness absene might be explained by a number of fators. In the first plae, sikness absene an be onsidered as an indiator of health problems. Sikness absene is, by definition, absene from work as a result of illness, and medially onfirmed sikness absene has been shown to predit future disability pension 8 and mortality. 7 9 Several indiators, suh as selfreported symptoms and visits to physiian, suggest that women may have more health problems than men, although the findings on women s greater morbidity are not entirely onsistent Women may also be better than men at reognising signs of illness and be more ative about seeking medial help As subjetive assessment of one s health plays a role in sikness absene behaviour, different levels of health problems as well as responsiveness to suh problems may ontribute to gender differenes in sikness absene. Differenes in working onditions may provide alternative explanations for the gender differenes in sikness absene. Various kinds of working onditions have shown assoiations with sikness absene and the assoiations may be different in women and men If women are more exposed to physial and psyhosoial job strain or poorer workplae limate this ould ontribute to their exess absene. Furthermore, various fators relating to home and private life have been suggested to explain female exess in sikness absene. Women often bear the main responsibility of household tasks whih inreases their total work load and may ause diffiulties in ombining work and family life, and even ontribute to mental and physial health problems. However, previous results on the assoiations of domesti obligations, the double burden, and work-home onflit, with sikness absene have been inonsistent. Stressful episodes in private life, suh as eonomi diffiulties, divore or serious illness in the family, have been more strongly assoiated with sikness absene in women than in men. This study examined gender differenes in sikness absene and potential explanations for these differenes. We identify health status, Oup Environ Med 2007;0:1 6. doi: /oem om33910 Module 1 Oupational and Environmental Mediine 14/11/07 14:48:34

2 Original artile Table 1 Desriptive information of the sample and the number of sikness absene spells of various lengths per 100 person-years by gender and age n Personyears Total number of absene days/ person-year Number of sikness absene spells/100 person-years Self-ertified Medially onfirmed 4 days or more Over two weeks Women by age (years) All ages* Men by age (years) All ages* *Number of sikness absenes spells adjusted for age. Over 60 days working onditions and family-related fators as the main andidates in explaining gender differenes in sikness absene, although previous evidene on gender differenes in these fators and their assoiations with sikness absene has been inonsistent. Beause the determinants of shorter and longer sikness absene spells may differ, we separately analysed absene spells of various lengths. Attention was paid to the prevalene of the possible explanatory fators as well as their potentially differential impat in women and men. A large prospetive dataset was used, with register-based sikness absene reords on middle-aged women and men employed by the City of Helsinki. The speifi aims of the study were to examine: (1) gender differenes in sikness absene of various lengths; (2) whether differenes in health status, working onditions and familyrelated fators explain gender differenes in sikness absene; and (3) whether gender differenes in sikness absene are due to differential exposure or responsiveness to these fators. METHODS Data The baseline data were olleted by independent ross-setional surveys in 2000, 2001 and 2002 among middle-aged employees of the City of Helsinki. 27 The City of Helsinki is the largest employer in Finland, with nearly employees. The personnel register was used to identify employees who in these onseutive years reahed the age of 40, 45, 50, 55 or 60, and a self-administered questionnaire was mailed to them (n = ). The overall response rate was 67%. The data are generally representative of the target population although men, younger people and manual workers were slightly underrepresented among the respondents The questionnaire data were prospetively ombined with sikness absene reords derived from the City of Helsinki registers using the unique personal identifiation number assigned to eah Finnish itizen. However, the linkage was only possible for those respondents who had given written permission for it when returning the questionnaire (78% of the respondents). Bakground harateristis of the study sample (age, oupational lass, inome, type of employment ontrat, oupational setor) were similarly assoiated with sikness absene among those who gave and did not give permission for the linkage. Sikness absene data ould not be sought for 16 respondents beause they had removed the ode number from the questionnaire. Thus, the study inludes 5470 women and 1464 men, refleting the gender distribution among the employees of the City of Helsinki. Measurement of sikness absene The number of sikness absene spells during the follow-up period was used as the outome variable. This outome effetively uses the information when one individual has several sikness absene spells and is not dominated by only a few prolonged absene spells. We examined separately self-ertified (1 3 days) and medially onfirmed absene spells of various lengths (4 days or more, more than 2 weeks, and more than 60 days). The absene spells of various lengths do not seriously overlap, for example, a vast majority of over two-week spells are shorter than 60 days (table 1). The follow-up time was started from the day the respondent returned the questionnaire and ontinued until the end of 2005 or until the work ontrat had terminated. Absene due to hildren s sikness was not inluded and all interruptions in working due to reasons other than own illness were subtrated from the follow-up time. The overall number of person-years in the analyses was with a mean follow-up time of 3.9 years. Potential explanatory fators Health status was assessed using five indiators. Physial and mental funtioning were measured by the omponent summaries of the Short-Form (SF-36) health questionnaire. 30 For the purposes of this study the ontinuous sores were divided into deiles. Limiting longstanding illness was eliited by asking whether the respondent had any longstanding illness, disability or infirmity, and a follow-up question asking whether this illness/disability restrited work or limited daily ativities. Furthermore, the partiipants were asked to rate their health on a five-point sale with the response alternatives ranging from exellent to poor. A sum of 28 self-reported diseases diagnosed by a dotor were alulated and the sum was trunated at four points. 2 Oup Environ Med 2007;0:1 6. doi: /oem om33910 Module 1 Oupational and Environmental Mediine 14/11/07 14:48:41

3 Original artile Table 2 Rate ratios (95% CI) for the exess risk of sikness absene among women as ompared to men (RR 1.00) (n = 6275): the effet of adjusting for health status, working onditions and family-related fators Self-ertified Medially onfirmed 4 days or more Over two weeks Over 60 days Base model* 1.46 (1.34 to 1.59) 1.34 (1.22 to 1.47) 1.25 (1.08 to 1.45) 1.06 (0.79 to 1.44) Health status Physial funtioning 1.39 (1.28 to 1.51) 1.22 (1.12 to 1.33) 1.11 (0.97 to 1.27) 0.91 (0.69 to 1.20) Mental funtioning 1.46 (1.35 to 1.59) 1.35 (1.23 to 1.48) 1.25 (1.09 to 1.45) 1.04 (0.78 to 1.39) Limiting longstanding illness 1.46 (1.34 to 1.58) 1.34 (1.23 to 1.46) 1.24 (1.08 to 1.42) 1.05 (0.81 to 1.37) Self-rated health 1.47 (1.36 to 1.60) 1.35 (1.24 to 1.48) 1.27 (1.11 to 1.45) 1.07 (0.82 to 1.38) Diagnosed diseases 1.34 (1.23 to 1.45) 1.17 (1.07 to 1.28) 1.08 (0.94 to 1.24) 0.86 (0.65 to 1.14) All health status measures 1.35 (1.25 to 1.46) 1.16 (1.07 to 1.27) 1.09 (0.96 to 1.24) 0.87 (0.67 to 1.13) Working onditions Physial demands 1.43 (1.31 to 1.56) 1.24 (1.13 to 1.36) 1.13 (0.97 to 1.31) 0.95 (0.70 to 1.29) Mental demands 1.46 (1.34 to 1.59) 1.34 (1.22 to 1.48) 1.25 (1.08 to 1.45) 1.06 (0.79 to 1.43) Job strain (Karasek) 1.46 (1.34 to 1.58) 1.32 (1.21 to 1.45) 1.24 (1.07 to 1.44) 1.06 (0.79 to 1.43) Work fatigue 1.44 (1.32 to 1.56) 1.31 (1.19 to 1.43) 1.19 (1.04 to 1.38) 1.00 (0.76 to 1.31) Job satisfation 1.48 (1.37 to 1.61) 1.37 (1.25 to 1.51) 1.30 (1.12 to 1.50) 1.10 (0.83 to 1.47) Bullying at workplae 1.45 (1.33 to 1.57) 1.33 (1.21 to 1.46) 1.24 (1.07 to 1.43) 1.04 (0.78 to 1.40) Being bullied 1.46 (1.34 to 1.58) 1.34 (1.22 to 1.47) 1.25 (1.08 to 1.44) 1.05 (0.79 to 1.41) All working onditions 1.44 (1.32 to 1.57) 1.23 (1.12 to 1.35) 1.13 (0.98 to 1.31) 0.94 (0.71 to 1.26) Family-related fators Marital status 1.43 (1.32 to 1.55) 1.32 (1.20 to 1.45) 1.23 (1.06 to 1.42) 1.05 (0.77 to 1.42) Having hildren in the family 1.46 (1.35 to 1.59) 1.34 (1.22 to 1.47) 1.25 (1.08 to 1.45) 1.06 (0.79 to 1.43) Work/home onflit 1.48 (1.36 to 1.61) 1.35 (1.23 to 1.49) 1.27 (1.10 to 1.48) 1.09 (0.81 to 1.47) Soial networks 1.46 (1.35 to 1.59) 1.34 (1.22 to 1.47) 1.25 (1.08 to 1.45) 1.07 (0.79 to 1.44) Soial support 1.47 (1.36 to 1.60) 1.36 (1.24 to 1.49) 1.27 (1.10 to 1.48) 1.08 (0.80 to 1.46) Negative events in the family 1.45 (1.33 to 1.58) 1.33 (1.21 to 1.46) 1.24 (1.07 to 1.44) 1.04 (0.77 to 1.41) All family-related fators 1.44 (1.33 to 1.56) 1.35 (1.23 to 1.48) 1.27 (1.09 to 1.46) 1.04 (0.77 to 1.41) All adjustments 1.34 (1.23 to 1.45) 1.12 (1.02 to 1.20) 1.02 (0.89 to 1.16) 0.82 (0.62 to 1.07) *All models adjusted for age and oupational lass. Working onditions inluded a question asking physial demands of work on a four-point sale. A similar question was used for mental work demands. Job demands and job ontrol were measured by the Framingham version of the Karasek questionnaire. 31 Job strain was defined as the ratio between job demands and job ontrol and divided into deiles. Work fatigue was measured using a sale of six questions onstruted at the Finnish Institute of Oupational Health on the basis of the emotional exhaustion subsale of the Maslah Burnout Inventory. 32 Job satisfation was measured using a single-item question with seven response alternatives derived from an inventory asking satisfation with different areas of life. Bullying at work was measured using two questions. The first one onerned whether bullying existed in the workplae; the seond whether the respondent was urrently being subjeted to bullying. Six indiators relating to family and private life were inluded. Marital status was divided into three ategories: married or ohabiting, single, and previously married (divored, separated or widow/widower). The number of hildren (aged less than 18 years) was ategorised as 0, 1 and 2 or more. Work/ home onflit was measured with an item asking how satisfied the respondent is with ombining paid work and family life. Soial networks were measured by asking how often the respondents met friends or relatives outside the nulear family. The ategories were almost daily, about one a week, about one a month, less often, and never. Soial support was measured by the Sarason inventory 33 and divided into deiles. Negative events in the familyinludeddeathofspouseorhild,deathofanotherlose person, serious illness of a family member, divore or separation, rupture of long-lasting relationship, and inreased onflits with spouse. A sum of events during the last 12 months was alulated and trunated at two. Confounders Age and oupational lass were inluded in the analyses as potential onfounders. Age was analysed using the five age groups of the questionnaire. Oupational lass was ategorised into managers and professionals, semi-professionals, routine non-manuals, and manual workers. 27 Statistial methods Sikness absene rates for spells of different lengths were first alulated in five-year age-groups among men and women. The rates are reported per 100 person-years. The effets of health status, working onditions and familyrelated fators on gender differenes in sikness absene were examined using Poisson regression taking differenes in the individual follow-up times into aount. We first fitted a base model showing the rate ratio with 95% onfidene intervals for women ompared with men, adjusted for age and oupational lass. We then examined how adjusting for health status, working onditions and family-related fators affeted this assoiation. In the Poisson models there was moderate overdispersion that was orreted by saling. This does not affet the point estimates but inreases standard errors and thus inflates the onfidene intervals. 34 For variables that were found to explain gender differenes in sikness absene, interations were fitted to examine whether their effets were different in women and men. In these analyses the explanatory fators were used as ontinuous. The rate ratios for the interations therefore desribe gender differenes in the effets per one unit hange in Oup Environ Med 2007;0:1 6. doi: /oem om33910 Module 1 Oupational and Environmental Mediine 14/11/07 14:48:42

4 Original artile Table 3 Differenes in the level of fators explaining gender differenes in sikness absene and rate ratios (95% CI) for interations desribing different effets of the explanatory fators among women and men on absenes of different lengths (n = 6275) Mean (SD) RR (95% CI) for gender interations* p Value for Explanatory fator (range) Men Women differene Self-ertified 4 days or more Over two weeks Over 60 days Physial funtioning (1 4) 1.40 (1.08) 1.58 (1.12), (0.93 to 1.07) 0.97 (0.90 to 1.05) 0.91 (0.81 to 1.03) 0.82 (0.63 to 1.07) Diagnosed diseases (0 4) 1.66 (1.33) 2.12 (1.41), (0.95 to 1.07) 1.03 (0.96 to 1.09) 1.04 (0.94 to 1.14) 1.03 (0.84 to 1.26) Physial demands (1 4) 1.92 (0.71) 2.28 (0.80), (0.93 to 1.18) 0.96 (0.84 to 1.09) 0.92 (0.76 to 1.12) 1.00 (0.66 to 1.51) Work fatigue (0 6) 1.10 (1.65) 1.36 (1.78) (0.94 to 1.03) 1.00 (0.95 to 1.05) 1.01 (0.94 to 1.09) 1.05 (0.91 to 1.22) *RR above 1.00 desribes stronger effet of the explanatory fator among women, RR below 1.00 desribes stronger effet of the explanatory fator among men. The RRs are adjusted for age and oupational lass. the explanatory fator. All analyses were onduted using SAS version 8.02 for Windows (SAS Institute In, Cary, NC). RESULTS Table 1 presents basi harateristis of the study population and the number of sikness absene spells of different lengths stratified by gender and age-group. Women had higher average number of sikness absene days than men in all age-groups, but gender differenes strengthened by age. When all age groups were ombined, women learly had more self-ertified sikness absene spells and any medially onfirmed sikness absene spells than men, but in spells longer than two weeks the differenes were smaller and no differenes were found in spells longer than 60 days. In self-ertified absene spells and any medially onfirmed absene spells the female exess was evident in all age groups. In ontrast, in spells longer than two weeks and longer than 60 days gender differenes within age groups were small. Self-ertified absene spells and any medially onfirmed absene spells were more ommon in younger age groups, whereas spells of over two weeks and over 60 days were more ommon in the older age groups. After adjusting for age and oupational lass, women had 46% higher risk for self-ertified sikness absene than men (table 2). In medially onfirmed spells there was 34% female exess whih gradually weakened with lengthening absene, and in spells longer than 60 days no statistially signifiant gender differenes were found. Adjusting for physial funtioning slightly attenuated the female exess in self-ertified sikness absene spells and more strongly in medially onfirmed spells. Diagnosed diseases had even stronger effet, explaining about one third of female exess in self-ertified absene spells and about half in medially onfirmed spells. The effets of both of these health status measures were gradually stronger for lengthening sikness absene. Among working onditions, adjusting for physial work demands learly attenuated gender differenes in medially onfirmed spells of all lengths. Work fatigue attenuated the gender differenes in absene spells longer than two weeks. Other working onditions had no effets on the gender differenes in sikness absene. The effets of family-related fators on gender differenes in sikness absene were minimal. When all health status measures, working onditions and family-related fators were simultaneously adjusted for, the female exess in self-ertified sikness absene was attenuated by a third and that in any medially onfirmed sikness absene by two thirds. There were no longer statistially signifiant gender differenes in spells longer than two weeks and the differenes in spells longer than 60 days were rather turned into a male exess, although this did not reah statistial signifiane. Table 3 shows differenes in the prevalene of the four fators that most learly explained gender differenes in sikness absene and whether these fators were similarly assoiated with sikness absene in women and men. On average, women had poorer physial funtioning and more diagnosed diseases than men. Women also reported higher exposure to physial work demands and work fatigue than men. Instead, there was little evidene that the effets of these variables on sikness absene would be different in women and men. None of the rate ratios desribing gender differenes in the effets of these explanatory fators on sikness absene spells of different lengths was statistially signifiant at the 0.05 level. DISCUSSION The main findings of this study an be summarised as follows: among middle-aged muniipal employees, women had more self-ertified sikness absene spells than men, but for lengthening medially onfirmed absene spells the gender differenes gradually weakened and no differenes were observed in spells longer than 60 days. Physial funtioning, self-reported diagnosed diseases, physial work demands and work fatigue explained the differenes, espeially in longer spells. This was beause problems in these areas were more ommon among women, but there were no gender differenes in responsiveness to these problems. Our finding that women have more sikness absene than men agrees with several previous studies from Finland and elsewhere. 1 2 Yet, in our study the female exess was limited to shorter sikness absene spells. In previous studies gender differenes have been found in both self-ertified and medially onfirmed sikness absene spells but with regard to the length of the absene their results have been inonsistent. The reasons for these disrepanies are unlear. In ontrast to our finding of no gender differenes in sikness absene spells longer than 60 days, Swedish studies using the same ut-off point have reported onsiderably more sikness absene among women. Our finding of weaker gender differenes for longer absene spells is onsistent with studies showing that gender differenes tend to be smaller in more serious and longstanding health problems. Physial funtioning and self-reported diagnosed diseases explained gender differenes in sikness absene spells of all lengths. These were the only fators that explained some of the gender differenes also in self-ertified absene spells, but their effets were stronger in medially onfirmed absene spells. This is plausible as longer sikness absene spells are likely to better indiate hroni morbidity, while shorter absene spells are likely to reflet minor transient morbidity and auses other than health. 40 Among the work-related fators, physial work demands and work fatigue explained gender differenes in medially 4 Oup Environ Med 2007;0:1 6. doi: /oem om33910 Module 1 Oupational and Environmental Mediine 14/11/07 14:48:43

5 Original artile onfirmed sikness absene spells. In ontrast, other working onditions did not explain any of the female exess in sikness absene. In Finland systemati nationwide reords are kept only for sikness absene spells longer than two weeks. 36 Suh sikness absene spells have inreased over the last deade and this has to a great extent been beause of mental reasons. The inrease has been larger among women, for whom mental diagnoses are more important auses for sikness absene than for men. Job strain and other psyhosoial working onditions might be assumed to have ontributed to this inrease as well as to the prevailing gender differenes in sikness absene. However, our study did not find any support for suh fators in explaining the female exess in sikness absene. Women report their work to be mentally strenuous slightly more often than men, but there have been no marked hanges over time. 41 Various fators relating to family and private life have been hypothesised to ontribute to the female exess in sikness absene. This has been a persistent assumption even if the evidene about the assoiations of these fators with sikness is inonsistent. In our study none of the family-related fators explained gender differenes in sikness absene. However, our study inluded only year-old employees. In younger age groups with more demanding family responsibilities the effets of family-related fators might be stronger. Previous studies have also shown that in younger age-groups a large amount of women s sikness absene is aused by pregnany-related problems, but this is unlikely to have muh influene in the age groups studied here. Poor physial funtioning, diagnosed diseases, physial work demands and work fatigue explained gender differenes in sikness absene beause they were more ommon among women, but not beause women were more responsive to these problems. Women have been found to report more health problems and work fatigue than men, but the finding of more physially demanding work among women is against the ommonly held belief of men working in physially heavier jobs. This finding may be beause our study was restrited to muniipal employees. In a nationwide study from Finland men and women reported similar levels of physial work demands. 41 In partiular, in shorter sikness absene spells, where the gender differenes were larger, subjetive assessment and responsiveness to signs of illness ould be assumed to ontribute to gender differenes. However, when we speifially tested whether the four fators had different effets on sikness absene in women and men we found no gender differenes. Methodologial onsiderations The strengths of this study inlude a relatively large sample, register-based data on sikness absene, and a prospetive study design. However, information on the health status measures, working onditions and family-related fators were based on self-reports, and if they were inadequately reported this would redue their explanatory power. Some of the explanatory fators were measured by single-item questions whih may have redued their measurement validity. Self-reports of working onditions ould also be affeted by one s health status. Our study inluded a large number of potential explanatory fators for gender differenes in sikness absene. However, we did not have measures of minor transient morbidity and symptoms whih ould relate espeially to shorter absene spells. We examined interations between gender and the potential explanatory fators only for those four fators that were found to explain gender differenes in sikness absene. We did not want to test all possible interations beause a large number of tests would inrease the risk of hane findings. Theoretially, if one explanatory fator had a positive effet in one gender and a negative one in the other, the interations ould be important even if the main effets are not. We heked the assoiations of all explanatory fators with sikness absene separately in women and men. All assoiations exept negative family-related life events in absene spells over 60 days were in the same diretion for men and women. However, the number of men and women with two or more negative life events was very small and the interation with gender was not statistially signifiant (p = 0.65). These analyses provide further support for the finding that gender differenes in sikness absene are not explained by different effets of these fators in women and men. In Finland and elsewhere the labour market is strongly segregated into male and female oupations. Beause of this segregation work tasks and working onditions between men and women may differ substantially, making it ompliated to firmly establish the role of working onditions in the explanation of gender differenes in sikness absene. Previous studies have approahed this question indiretly, and found that men working in speifi female-dominated oupations have even more sikness absene than women, whih would suggest that the nature of work rather than gender is the reason for sikness absene However, a more reent study examining a broad range of oupations did not verify these previous findings. 46 In another study, eliminating the effets of oupation and workplae strengthened gender differenes in sikness absene, also suggesting that diretly work-related fators do not explain the gender differenes in sikness absene. 5 We examined the explanations for gender differenes in sikness absene diretly by adjusting for a large number of general indiators of working onditions. However, it is possible that there are oupationspeifi exposures that are not fully grasped by these general indiators. Main messages Women had more self-ertified sikness absene spells than men, but in lengthening medially onfirmed spells gender differenes were smaller or non-existent. Poor physial funtioning, diagnosed diseases, physial work demands and work fatigue explained the female exess in sikness absene, whereas psyhosoial working onditions and family-related fators had no effets. These fators explained gender differenes in sikness absene beause they were more ommon among women, not beause women were more responsive to these problems. Poliy impliations Gender differenes were smaller in longer sikness absene spells. Suh sikness absene spells are likely to better reflet hroni morbidity whereas for shorter absene spells the explanations are likely to be more varied. Reasons for shorter sikness absene spells should be better understood as they aount for a major part of all absene days. Oup Environ Med 2007;0:1 6. doi: /oem om33910 Module 1 Oupational and Environmental Mediine 14/11/07 14:48:43

6 Original artile CONCLUSION Women had more sikness absene than men but the differenes were gradually weaker for longer absene spells. The overall gender differenes in sikness absene are therefore due to relatively short absene spells being more ommon among women. The examined health status measures, working onditions and family-related fators explained more of the gender differenes in longer than in shorter sikness absene spells. Gender differenes in longer sikness absene spells are mainly due to heavier burden of ill-health and to a lesser extent higher physial work demands among women. We found no support for greater vulnerability to health- and work-related problems among women as reasons for sikness absene. Part of the gender differenes in short sikness absene spells remained unexplained by the fators studied here. Further explanatory fators for short sikness absene spells, suh as the prevalene of minor symptoms and differenes in illness behaviour, need to be onsidered in future studies. Funding: The study was supported by the Aademy of Finland (#204894, # and #210435) and the Finnish Work Environment Fund (#106066). Competing interests: None. REFERENCES 1. Allebek P, Mastekaasa A. Swedish Counil on Tehnology Assessment in Health Care (SBU). Chapter 5. Risk fators for sik leave general studies. Sand J Publi Health 2004;32(Suppl 63): Barnby T, Erolani M, Treble J. Sikness absene: an international omparison. Eon J 2002;112: Gimeno D, Benavides FG, Benah J, et al. Distribution of sikness absene in the European Union ountries. Oup Environ Med 2004;61: Melhior M, Niedhammer I, Berkman LF, et al. Do psyhosoial work fators and soial relations exert independent effets on sikness absene? A six year prospetive study of the GAZEL ohort. J Epidemiol Community Health 2003;57: Mastekaasa A, Dale-Olsen H. Do women or men have the less healthy jobs? an analysis of gender differenes in sikness absene. European Soiologial Review 2000;16: Feeney A, North F, Head J, et al. Soioeonomi and sex differentials in reason for sikness absene from the Whitehall II Study. Oup Environ Med 1998;55: Kivimäki M, Head J, Ferrie JE, et al. Sikness absene as a global measure of health: evidene from mortality in the Whitehall II prospetive ohort study. BMJ 2003;327: Kivimäki M, Forma P, Wikström J, et al. Sikness absene as a risk marker of future disability pension: the 10-town study. J Epidemiol Community Health 2004;58: Vahtera J, Pentti J, Kivimäki M. Sikness absene as a preditor of mortality among male and female employees. J Epidemiol Community Health 2004;58: Gijsbers van Wijk CMT, Kolk AM. Sex differenes in physial symptoms: the ontribution of symptom pereption theory. So Si Med 1997;45: MDonough P, Walters V. Gender and health: reassessing patterns and explanations. So Si Med 2001;52: Lahelma E, Martikainen P, Rahkonen O, et al. Gender differenes in illhealth in Finland: patterns, magnitude and hange. So Si Med 1999;48: Maintyre S, Hunt K, Sweeting H. Gender differenes in health: are things really as simple as they seem? So Si Med 1996;42: Lund T, Labriola M, Christensen KB, et al. Physial work environment risk fators for long term sikness absene: prospetive findings among a ohort of 5357 employees in Denmark. BMJ 2006;332: Niedhammer I, Bugel I, Goldberg M, et al. Psyhosoial fators at work and sikness absene in the Gazel ohort: a prospetive study. Oup Environ Med 1998;55: Krantz G, Lundberg U. Workload, work stress, and sikness absene in Swedish male and female white-ollar employees. Sand J Publi Health 2006;34: Väänänen A, Toppinen-Tanner S, Kalimo R, et al. Job harateristis, physial and psyhologial symptoms, and soial support as anteedents of sikness absene among men and women in the private industrial setor. So Si Med 2003;57: Chandola T, Martikainen P, Bartley M, et al. Does onflit between home and work explain the effet of multiple roles on mental health? A omparative study of Finland, Japan, and the UK. Int J Epidemiol 2004;33: Winter T, Roos E, Rahkonen O, et al. Work-family onflits and self-rated health among middle-aged muniipal employees in Finland. Int J Behav Med 2006;13: Mastekaasa A. Parenthood, gender and sikness absene. So Si Med 2000;50: Akerlind I, Alexanderson K, Hensing G, et al. Sex differenes in sikness absene in relation to parental status. Sand J So Med 1996;24: Bratberg E, Dahl SA, Risa AE. The double burden : do ombinations of areer and family obligations inrease sikness absene among women? European Soiologial Review 2002;18: Jansen NW, Kant IJ, van Amelsvoort LG, et al. Work-family onflit as a risk fator for sikness absene. Oup Environ Med 2006;63: Mastekaasa A, Olsen KM. Gender, absenteeism, and job harateristis. Work and Oupations 1998;25: Kivimaki M, Vahtera J, Thomson L, et al. Psyhosoial fators prediting employee sikness absene during eonomi deline. J Appl Psyhol 1997;82: North F, Syme SL, Feeney A, et al. Explaining soioeonomi differenes in sikness absene: the Whitehall II Study. BMJ 1993;306: Lahelma E, Martikainen P, Rahkonen O, et al. Oupational lass inequalities aross key domains of health: results from the Helsinki Health Study. Eur J Publi Health 2005;15: Laaksonen M, Aittomäki A, Lallukka T, et al. Register-based study among employees showed small non-partiipation bias in health surveys and hek-ups. J Clin Epidemiol 2007 (in press). 29. Martikainen P, Laaksonen M, Piha K, et al. Does survey non-response bias the assoiation between oupational soial lass and health? Sand J Publi Health Ware JE, Kosinski M, Keller SD. SF-36 physial and mental omponent summary measures: a user s manual. Boston, MA: The Health Institute, New England Medial Center, Karasek RA. Job Content Questionnaire and user s guide. Revision 1.1. Lowell, MA: University of Massahusetts, Department of Work Environment, Kalimo R, Olkkonen M, Toppinen S. Ihminen kehittyvässä tuotannossa: Tutkimus- ja kehityshanke teollisuudessa. Työ ja ihminen 1993;7: Sarason IG, Sarason BR, Shearin EN, et al. A brief measure of soial support: pratial and theoretial impliations. J So Pers Relat 1987;4: Gardner W, Mulvey EP, Shaw EC. Regression analyses of ounts and rates: Poisson, overdispersed Poisson, and negative binomial models. Psyhol Bull 1995;118: Piha K, Martikainen P, Rahkonen O, et al. Trends in soioeonomi differenes in sikness absene among Finnish muniipal employees Sand J Publi Health 2007;35: Soial Insurane Institution. Health Insurane Statistis 2005 (In Finnish). Helsinki: Soial Insurane Institution, Ylitalo M. Palkansaajien sairauspoissaolot. In: Lehto AM, Sutela H, Miettinen A, eds. Kaikilla mausteilla. Artikkeleita työolotutkimuksesta. Researh Reports No 244. Helsinki: Statistis Finland, Lidwall U, Marklund S. What is healthy work for women and men? A ase-ontrol study of gender- and setor-speifi effets of psyho-soial working onditions on long-term sikness absene. Work 2006;27: Stenbek M, Persson G. Chapter 10: Working life, work environment and health. In: Persson G, Danielsson M, Rosen M, et al, eds. Health in Sweden: the National Publi Health Report Sand J Publi Health 2006;34(Suppl 67): Marmot M, Feeney A, Shipley M, et al. Sikness absene as a measure of health status and funtioning: from the UK Whitehall II study. J Epidemiol Community Health 1995;49: Lehto AM, Sutela H. Threats and opportunities. Findings of Finnish quality of work life surveys Helsinki: Statistis Finland, Alexanderson K, Sydsjo A, Hensing G, et al. Impat of pregnany on gender differenes in sikness absene. Sand J So Med 1996;24: Alexanderson K, Hensing G, Leijon M, et al. Pregnany related sikness absene in a Swedish ounty, J Epidemiol Community Health 1994;48: Evans O, Steptoe A. The ontribution of gender-role orientation, work fators and home stressors to psyhologial well-being and sikness absene in male- and female-dominated oupational groups. So Si Med 2002;54: Hensing G, Alexanderson K, Akerlind I, et al. Sik-leave due to minor psyhiatri morbidity: role of sex integration. So Psyhiatry Psyhiatr Epidemiol 1995;30: Mastekaasa A. Sikness absene in female- and male-dominated oupations and workplaes. So Si Med 2005;60: Oup Environ Med 2007;0:1 6. doi: /oem om33910 Module 1 Oupational and Environmental Mediine 14/11/07 14:48:46