- Review previous knowledge of the associations among musculoskeletal diagnoses (MSD), pain, and common mental disorders (CMD).
- Summarize the new findings on the association between number of pain locations and CMD on the risk of disability pension, based on a Swedish twin survey.
- Discuss the implications for strategies to prevent disability pension.
Musculoskeletal diagnoses (MSDs) and common mental disorders are the two major groups of chronic conditions adding to the global burden of disease among those are still at working-age.1,2 An important factor underlying MSDs is pain, which, often comorbid with common mental disorders,3,4 may additionally worsen work ability.5 Furthermore, among pain conditions, low back pain constitutes one of the major public health problems today.6 To get the perspective, low back and neck pain form the fourth leading cause of the burden of disease measured as loss of disability-adjusted life years, whereas depressive disorders are at the 11th place.7
Many MSDs include pain, but pain can also appear as a sole symptom, without clear underlying medical condition, and episodes of pain will in most cases eventually resolve. Chronic pain, which affects over 20% of the adult population, is the most common cause of severe, long-term, physical disability and consequently a substantial burden to both individuals and societies.1 The number of pain locations, that is, pain in various locations of the body, may add to the risk of losing work capacity or alternatively imply a dose–response of pain.8,9
Comorbidity adds some complexity, as pain can have a different association with common mental disorders. Previous studies show that the prevalence of common mental disorders, including depressive disorders, anxiety, and unipolar depressive disorder, is increasing.10 In Europe, every third person who has a chronic pain condition has also been diagnosed with depression.11 Hence, it remains to be clarified whether common mental disorders are a cause or a consequence of pain. Some indications exist that depression and anxiety may predict pain,12,13 whereas depression can also be a consequence of pain,14 as it has been shown to be a part of the natural course of pain.15 On the basis of earlier studies, we know that back pain and depression are strongly interlinked.16,17 Furthermore, sick leave due to depression or back pain is a strong risk factor for DP due to MSD and common mental disorders.5 However, more studies are still needed to elaborate the role of common mental disorders, pain, and their mutual associations in the process to DP.
Epidemiological studies of pain, common mental disorders, and DP can be extended by analyzing the associations among twins. Even though a DP always requires an objective medical ground, the risk of DP may also be affected by other factors, such as the patient's social background. Genetics is known to play a moderate role in pain, common mental disorders, and DP, suggesting that familial confounding (ie, genetics and social background mainly in childhood) is likely when studying the association between pain and common mental disorders with future DP.18–25 Previous twin cohort studies9,26–29 have shown that pain and common mental disorders have a direct effect on the risk for DP, meaning that the association was still present after genetic factors were controlled for. Twin studies provide a possibility to control for genetics and shared environment (familial factors) through a cotwin control method. The cotwin control design includes twins discordant for DP (one twin granted DP but not the other) and provides a great possibility to ascertain whether the associations between pain, common mental disorders, and DP remain after taking unmeasured predisposing familial factors into account. Furthermore, the cotwin control design where same-sexed twins are perfectly matched for age and sex adds to the knowledge based on the epidemiological studies of nontwin populations.
This study aimed to examine whether there is a higher risk of all-cause DP, DP due to MSD, or DP due to mental diagnoses regarding number of pain locations and common mental disorders (anxiety and/or depression). Another aim was to investigate whether the associations were influenced by familial factors.
Data for this study were available from a prospective twin cohort study, the Swedish Twin project Of Disability pension and Sickness absence (STODS), which includes survey and national register data.30 From STODS, we included twins who participated in the Screening Across the Lifespan Twin study (SALT) telephone interview conducted by the Swedish Twin Registry between January 1998 and March 2003.31 Then, we restricted the data to those who were less than 65 years old and not on old-age retirement or DP at the time of the SALT interview. The final sample consisted of 27,165 twin individuals (51% women) born in 1935 to 1958, with baseline data from the survey, of whom 2640 were monozygotic and 3506 dizygotic complete twin pairs and the remaining 14,873 were individuals without a cotwin.
A national social security DP program covers all citizens in Sweden. All above age 16 who have income from work or unemployment benefits are entitled to sickness benefits when they are incapable to work. Eligibility for both sickness absence and DP benefits requires a medically confirmed disorder or injury, and for DP that the disorder permanently reduces work capacity by at least 25%. All citizens between 19 and 64 years of age are covered by the DP program. Evaluation of work incapacity always includes the assessment of functional capacity, occupational skills, education, work tasks, and work history. A thorough assessment of level of work incapacity is done by the Social Insurance Agency as a basis to grant DP that covers about 65% of lost income and is a taxable income. In Sweden, it is also possible to have part- or full-time DP based on the ordinary working hours of the applicant. In case of part-time employment (eg, 50% = 4 hours/day), full-time DP grant means that person will be granted DP for 4 hours/day. If a part-time DP would be granted, for example, 50%, then there would be work 50% of ordinary part-time, that is, 2 hours/day.
In this study, data on DP included date and diagnosis according to the International Classification of Diseases version 10.32 DP data were obtained from the National Social Insurance Agency Micro-Data for Analysis of the Social Insurance System (MiDAS), and DP diagnoses were categorized into all-cause DP (including all diagnoses except mental and MSD), and into two main diagnosis groups including DP due mental diagnoses (F00-F99), and DP due to MSD (M00-M99). Date of SALT response (between January 1, 1998, and March 31, 2003) was the beginning of the follow-up, which continued until the age of 65, old-age retirement (if retired before the age of 65), emigration (data available from Statistics Sweden),33 death (data available from the National Board of Health and Welfare), or up to December 31, 2013, whichever occurred first. The unique 10-digit Swedish identification number of each twin individual was used for linking data from the different national registers together.
Data on the main exposures of interest, number of pain locations, common mental disorders (major depression and anxiety), and covariates were available from the SALT31 interview. Up to five pain locations were evaluated using a set of three questions. The first question queried about the presence of back, shoulder, or neck pain with “Have you been suffering from pain in [then came the list]?” with response alternatives yes/no. Then another question of back pain was stated for sciatica with a dichotomous response option (yes/no). To further define the pain, a third question included was related to having low back pain (yes/no). All these pain questions were for lifetime pain. On the basis of the three-way cross-tabulations, these three questions were not mutually exclusive, hence we constructed a combined variable on number of pain locations. Those reporting no pain in any of the locations nor sciatica or low back problem were coded into “0.” Then, all the other categories were constructed summing up the pain locations (1 to 5 pain locations, in which 5 would be pain in back, low back, sciatica, neck, and shoulders).
For common mental disorders, major depression was evaluated for lifetime major depression using the WHO Composite International Diagnostic Interview procedure as described in detail earlier.34,35 Major depression was considered if the response “In your lifetime, have you ever had two weeks or more when you nearly every day felt sad, low, depressed?” was yes, and additionally at least four of eight additional symptoms were identified: lost interest (in general/in hobbies), weight change, trouble falling asleep, feeling tired, trouble concentrating, and/or thoughts about death. Current or previous history of anxiety was also measured according to DSM-IV.36 Anxiety were queried by a question “Have you had an episode lasting at least a month, in which you felt worried and anxious most of the time?”, and at least one of five additional symptoms: feeling tense, irritable, tired, trouble sleeping, and the combination of being restless and on the edge.
Data from Statistics Sweden at the time of SALT included years of education, marital status (dichotomized into cohabiting or living alone), and children living at home. We also used data on age at the interview occasion and sex as covariates. Body mass index (BMI) was calculated using self-reported weight and height information (kg/m2). Self-rated health was assessed using the question: “How would you rate your general health status?” responses were collapsed into three alternatives: excellent/good, moderate, and fairly poor/poor.37
Furthermore, the use of tobacco products was measured in four categories: never used (never tried, only tried; the reference category), occasional user, regular user, and past user as has been done before.28 Use of any type of alcohol (beer, wine, and spirits) was enquired and the sex-specific amount of alcohol consumption was calculated. The alcohol use was then categorized into abstainers, light users, moderate users, and heavy users as reported in detail before.38 Annual leisure-time physical activity was assessed by “If you consider the physical activity you take during your leisure time, which of these seven options fit you the best if you look at the year as a whole?” The response alternatives were collapsed to very much/a lot, quite a lot, not much, very little, and hardly ever/almost never. In addition, headache and use of painkillers (ie, analgesics) with dichotomous yes/no responses were used in the analyses.
Occurrence of diseases was surveyed for 53 health problems and conditions that were classified for their life-threatening severity as has been described in detail earlier.37,39 Then, a combination of number and severity of diseases was created with eight categories as has been done before.37,39
The study was approved by the Regional Ethical Review Board in Stockholm (2007/524–31).
The differences between DP diagnostic groups and those without DP for main exposures and covariates were analyzed using Chi-square tests for categorical variables and by t tests for continuous variables. Cox proportional hazards regression models were used to calculate hazard ratios (HRs) with 95% confidence intervals (95% CIs) for all-cause DP, DP due to mental diagnoses, or DP due to MSD as the outcomes, while follow-up time was in days. The log-log curves for the proportional hazards’ assumption were graphically tested for the categories of pain locations and common mental disorders and shown to be acceptably parallel. As pain locations and common mental disorders were suspected to differ in their association to DP due to MSD versus mental diagnoses, hence, different strategies for censoring were tested. To see the effect of censoring, we excluded the competing DP diagnostic group from the data to estimate the risks for each DP diagnostic group. As risk estimates did not differ depending on analytical strategy, we included all the competing DP diagnostic groups. First, we ran Cox proportional hazard regressions models adjusted for age, and by providing men and women with their own baseline hazards to adjust for sex. To account for the interdependency in twin pairs, the models included clustering on twin pair identity.
Then, the models were adjusted for several covariates. The model, including education as a proxy for socioeconomic status, family situation indicated by marital status and number of children, and BMI and tobacco consumption for health behaviors, was chosen as a main result for this study based on the relative importance of the covariates in the model. Furthermore, we conducted additional models to account for the effects of various health factors, including the effect of the number and severity of diseases, headache, use of painkillers (yes/no), alcohol consumption, physical activity, and self-rated health based on their known role not only in pain, common mental disorders, and/or DP, but also in the mutual association. As these additional covariates played only a minor role, we have chosen not to present the results.
As a sensitivity analysis, we restricted the sample and excluded those who had cancer, rheumatism in lower back or in lower limbs, or arthrosis in knee or hip (n = 4066), leaving 2424 cases with all-cause DP, 900 cases with DP due to MSD, and 552 cases with DP due to mental diagnoses. In this restricted sample, the point estimates for pain locations or common mental disorders were not affected (data not shown). Hence, we chose to keep the full sample in order to retain power.
Then, we ran a sensitivity analysis for the number of pain locations. For this study, we used the number of pain locations (zero to five locations) constructed from three separate questions, which were not mutually exclusive in our survey. As we also had a single question with response alternatives zero to three pain locations, we wanted to see if that would impact the associations. We ran all the same models for this, but as the results were similar for direction and magnitude as for zero to five pain locations, we chose to present the results for the latter. In addition, we tested the interaction for common mental disorders and pain locations and that was nonsignificant for any of the DP diagnosis groups.
In a final step, we restricted the sample to same-sex discordant twin pairs to conduct conditional Cox proportional hazard regression models. The conditional model utilizes a cotwin control design where one twin in a pair had been granted DP, while the twin sibling had not been granted DP during the follow-up; hence, the follow-up times were evaluated for both twins in a pair. In the cotwin control design, familial confounding can be evaluated, as twins in pair share genes [100% in monozygotic (MZ) and on average 50% of their segregating genes in dizygotic (DZ) twin pairs], and social background that is home and family environment mainly in childhood. The familial confounding is suggested if the association is found in the analyses of the whole cohort, but not in the analyses of the discordant twin pairs. If the association is present both in the whole cohort and discordant twin pairs, then there is none or only minor familial confounding, that is, the association is most likely due to nonfamilial environmental factors. In the final sample, 1029 were same-sex discordant twin pairs with all-cause DP, 484 with DP due to MSD, and 250 with mental diagnoses (number of MZ and DZ twins separately are summarized in Tables 1–3).
All statistical analyses were conducted using Stata version 12.1 (Stata Corporation, College Station, Texas).
During the mean follow-up of 8 years, DP due to any cause was granted to 12% of the final sample (3290 individuals), whereas 5% got DP due to MSD (1338 individuals) and 3% DP due to mental diagnoses (683 individuals). More than half were women, 61% of those DP due to all causes, 65% in DP due to MSD, and 69% in DP due to mental diagnoses, whereas among those without DP, there were no sex differences (Table 4).
First, we investigated the separate associations between pain locations and common mental disorders and DP due to all causes (Table 1). The age and sex-adjusted model indicated a trend toward a dose–response effect of number of pain locations on the risk of DP due to all causes. Also, common mental disorders were significantly associated with risk of DP due to all causes. All these associations remained while mutually adjusting for pain locations and common mental disorders as well as adjusting for several covariates, including the familial ones.
Furthermore, we stratified the analyses by anxiety and depression to see whether they would affect the associations of pain locations with all-cause DP, but that had no effect (Table 5).
The number of pain locations showed a dose–response effect for DP due to MSD (Table 2). The associations and the dose–response effect retained in all tested models, also among discordant twins pairs. Instead, the effect of anxiety attenuated to statistical nonsignificance among discordant twin pairs or when depression was controlled for. Third, we assessed the effect of baseline common mental disorders on the associations between pain locations and DP due to MSD and detected no effect (Table 5).
For DP due to mental diagnoses, the number of pain locations showed a weak trend toward the more pain locations, the higher risk (Table 3). Adjusting for common mental disorders or various covariates slightly attenuated the associations, but the magnitude and direction remained. In the analyses of discordant twins, the associations between pain locations and DP due to mental diagnoses attenuated into statistical nonsignificance.
Common mental disorders were important predictors for DP due to mental diagnoses in all models. Controlling for baseline common mental disorders attenuated the associations between pain locations and DP due to mental diagnoses into statistical nonsignificance (Table 5).
This population-based study of over 27,000 Swedish twins with comprehensive survey and register data and with more than 8 years of follow-up aimed to investigate whether the risk of all-cause DP, or DP due to MSD, or DP due to mental diagnoses would differ regarding the number of pain locations and how common mental disorders affect the association. The number of pain locations (from none to five locations) indicated a dose–response effect on the risk of DP in all studied diagnosis groups not affected by coexisting common mental disorders. This adds to the existing knowledge based on population-based samples regarding associations between sickness absence due to common mental disorders or back pain and later DP5 or even assessing synergistic effect between common mental disorders and back pain.17 The dose–response effects are in line with the findings based on the Finnish Twin Cohort, both with results over two time points, but also in regard with independency of various covariates.8
Access to twin data allowed us to investigate whether the associations were influenced by familial factors. The cotwin-control design provided us with a powerful tool to analyze discordant twin pairs, that is, individuals within a twin pair who differed in DP, pain, and common mental disorders by the optimal matching of cases and controls, as they were same-sexed twin pairs. This is a unique feature exploring the role of genetics and shared (mainly childhood) environmental factors on the associations between pain, common mental disorders, and DP due to three diagnosis groups. We detected no effect of familial factors on the association between pain and DP due to all causes. Instead for DP due to MSD, associations with common mental disorders attenuated suggesting familial confounding, as did associations with pain for DP due to mental diagnosis. This points into direction that early life circumstances, genetics, and shared family environment (such as social background) might be important for associations between pain and DP due to mental diagnoses, and for common mental disorders and DP due to MSD.
The design of this study was built on the known interlinkage of back pain and depression.16,17 In addition, we acknowledged the assumption that common mental disorders might have a different role for pain, in which anxiety was assumed as a predictor of pain, and depression linked with the course of pain15 or perhaps being a consequence of pain.14 We tested these assumptions by assessing pain locations, and depression and/or anxiety in two ways. First, we adjusted for them mutually, and then controlled for mutual associations at the baseline. Another underlying assumption was that the associations might differ depending on DP diagnosis groups. Results pointed into the same direction: the number of pain locations had a dose–response effect on DP due to all causes, mental diagnoses, and MSD. In addition, we detected a trend toward having several pain locations may have an influential role on DP due to mental diagnoses. This might reflect the fact that common mental disorders include some degree of cognitive, emotional, social, and physical inhibition, which may consequently influence functional capacity negatively. For example, those with common mental disorders might suffer from fatigue, which could have an additional effect on work capacity. The other way around and in regards of work ability, such effects might be exacerbated by pain. As we also detected an effect of familial confounding on the associations between pain locations and DP due to mental diagnosis, one can speculate that the shared neurobiology and genetics in depression and pain may play a role.40,41
This study had several strengths, including the comprehensive survey data for more than 27,000 individuals, comprehensive national register data with date and diagnosis for DP, and censoring with no loss to follow-up. The incidence of DP was at the same level as has been in earlier studies of the Nordic countries.42–45 We also had the possibility to control for several known and influential factors on the associations between the main factors of interest (pain and common mental disorders) and DP. Among these, the access to twin data with control for familial confounding merits mentioning, as our finding that familial confounding affected the associations between pain and DP due to mental or MSD, but not DP due to all cause. This was an expected finding, as pain, common mental disorders, and DP all have a moderate genetic component.18–25 Considering primary health care, or occupational health, one may assume that not only to prevent pain or common mental disorders, but also DP, may require early and tailored actions and interventions due to the influence of familial factors.
Any study carries some pitfalls that may affect the interpretation of results. In our study, despite the large sample, we may have lacked power in the discordant pair analysis, at least for MZ and DZ twins separately. Hence, further studies with even larger samples would be needed to confirm these findings. Another aspect to keep in mind while assessing our results is the fact that both pain locations and common mental disorders were assessed at different time periods and based on self-reports 10 to 15 years ago. This may affect the observed associations, but some confidence remains, as these results are in line with those published in Sweden earlier and based on register data of sickness absences with diagnoses.17 Despite the comprehensive survey data, we did not have information on physical job demands that should be kept in mind while interpreting the results. One further limitation may be related to the Nordic welfare model and comprehensive register data on DP, which may affect generalizability to other countries. As behavior is one of the important aspects affecting work while having pain and common mental disorders present, one should note that we were not able to control the effects of pain beliefs, self-determination, willingness to work despite pain, or functional restoration programs. These should be accounted in further studies.
The number of pain locations show a dose–response effect on the risk of DP due to any cause, mental diagnoses, and MSD. Common mental disorders also predict DP due to any diagnosis group. Comorbidity of pain and common mental disorders do not add to the risk of DP, but the risk attenuates for DP due to mental diagnoses. The effect of familial confounding on the associations with DP due mental and MSD suggests that in strategies to prevent DP, early signs of pain or common mental disorders should be taken into consideration.
We acknowledge The Swedish Twin Registry for access to data.
1. Lim SS, Vos T, Flaxman AD, et al. A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet
2. March L, Smith EU, Hoy DG, et al. Burden of disability due to musculoskeletal (MSK) disorders. Best Pract Res Clin Rheumatol
3. Schatzberg AF. Mayer EA, Bushnell MC. Chronic pain
and depression. Functional Pain Syndromes: Presentation, Pathophysiology
. Seattle: IASP Press; 2009. 215–226.
4. Naliboff BD, Rhudy JL. Mayer EA, Bushnell MC. Anxiety in functional pain
disorders. Functional Pain Syndromes: Presentation and Pathophysiology
. Seattle: IASP Press; 2009. 185–213.
5. Dorner TE, Alexanderson K, Svedberg P, Ropponen A, Stein KV, Mittendorfer-Rutz E. Sickness absence due to back pain
or depressive episode and the risk of all-cause and diagnosis-specific disability pension: a Swedish cohort study
of 4,823,069 individuals. Eur J Pain
6. Försäkringskassan. Social Insurance in Figures 2014. Stockholm: Försäkringskassan; 2014.
7. Murray CJ, Barber RM, Foreman KJ, et al. Global, regional, and national disability-adjusted life years (DALYs) for 306 diseases and injuries and healthy life expectancy (HALE) for 188 countries, 1990–2013: quantifying the epidemiological transition. Lancet
8. Ropponen A, Svedberg P, Kalso E, Koskenvuo M, Silventoinen K, Kaprio J. A prospective twin cohort study
of disability pensions due to musculoskeletal diagnoses in relation to stability and change in pain
9. Ropponen A, Silventoinen K, Svedberg P, et al. Health-related risk factors for disability pensions due to musculoskeletal diagnoses: a 30-year Finnish twin cohort study
. Scand J Public Health
10. Mathers CD, Loncar D. Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med
11. Breivik H, Collett B, Ventafridda V, Cohen R, Gallacher D. Survey of chronic pain
in Europe: prevalence, impact on daily life, and treatment. Eur J Pain
12. Bair MJ, Robinson RL, Katon W, Kroenke K. Depression and pain
comorbidity: a literature review. Arch Intern Med
13. Niles AN, O’Donovan A. Comparing anxiety and depression to obesity and smoking as predictors of major medical illnesses and somatic symptoms. Health Psychol
14. Knaster P, Estlander AM, Karlsson H, Kaprio J, Kalso E. Temperament traits and chronic pain
: the association of harm avoidance and pain
-related anxiety. PLoS One
15. Vranceanu A-M, Barsky A, Ring D. Psychosocial aspects of disabling musculoskeletal pain
. J Bone Joint Surg Am
16. Carstensen J, Andersson D, André M, Engström S, Magnusson H, Borgquist LA. How does comorbidity influence healthcare costs? A population-based cross-sectional study of depression, back pain
and osteoarthritis. BMJ Open
17. Dorner TE, Alexanderson K, Svedberg P, Tinghog P, Ropponen A, Mittendorfer-Rutz E. Synergistic effect between back pain
and common mental disorders
and the risk of future disability pension: a nationwide study from Sweden. Psychol Med
18. Fejer R, Hartvigsen J, Kyvik KO. Heritability of neck pain
: a population-based study of 33 794 Danish twins. Rheumatology
19. Hartvigsen J, Nielsen J, Kyvik KOHM, et al. Heritability of spinal pain
and consequences of spinal pain
: a comprehensive genetic epidemiologic analysis using a population-based sample of 15,328 twins ages 20–71 years. Arthritis Care Res
20. Battie MC, Videman T, Levalahti E, Gill K, Kaprio J. Heritability of low back pain
and the role of disc degeneration. Pain
21. Harkonmäki K, Silventoinen K, Levälahti E, et al. The genetic liability to disability retirement: a 30-year follow-up study of 24,000 Finnish twins. PLoS One
22. Narusyte J, Ropponen A, Silventoinen K, et al. Genetic liability to disability pension in women and men: a prospective population-based twin study. PLoS One
23. Young EE, Lariviere WR, Belfer I. Genetic basis of pain
variability: recent advances. J Med Genet
24. Otte C, Gold SM, Penninx BW, et al. Major depressive disorder. Nat Rev Dis Primers
25. Gottschalk MG, Domschke K. Genetics of generalized anxiety disorder and related traits. Dialogues Clin Neurosci
26. Markkula R, Kalso E, Huunan-Seppala A, et al. The burden of symptoms predicts early retirement: a twin cohort study
on fibromyalgia-associated symptoms. Eur J Pain
27. Pietikäinen S, Silventoinen K, Svedberg P, et al. Health-related and sociodemographic risk factors for disability pension due to low back disorders: a 30-year prospective Finnish twin cohort study
. J Occup Environ Med
28. Ropponen A, Narusyte J, Alexanderson K, Svedberg P. Stability and change in health behaviours as predictors for disability pension: a prospective cohort study
of Swedish twins. BMC Public Health
29. Narusyte J, Ropponen A, Alexanderson K, Svedberg P. Genetic and environmental influences on disability pension due to mental diagnoses: limited importance of major depression, generalized anxiety, and chronic fatigue. Twin Res Hum Genet
30. Svedberg P, Ropponen A, Lichtenstein P, Alexanderson K. Are self-report of disability pension and long-term sickness absence accurate? Comparisons of self-reported interview data with national register data in a Swedish twin cohort. BMC Public Health
31. Lichtenstein P, De faire U, Floderus B, Svartengren M, Svedberg P, Pedersen NL. The Swedish Twin Registry: a unique resource for clinical, epidemiological and genetic studies. J Intern Med
32. World Health Organization. ICD-10, the ICD-10 Classification of Mental and Behavioural Disorders: Diagnostic Criteria for Research Version: 2016. Geneva: World Health Organization; 1993.
33. Ludvigsson JF, Svedberg P, Olen O, Bruze G, Neovius M. The longitudinal integrated database for health insurance and labour market studies (LISA) and its use in medical research. Eur J Epidemiol
34. Kessler RC, Ustun TB. The World Mental Health (WMH) Survey Initiative Version of the World Health Organization (WHO) Composite International Diagnostic Interview (CIDI). Int J Methods Psychiatr Res
35. Ropponen A, Narusyte J, Mather L, Mittendorfer-Rutz E, Akerstedt T, Svedberg P. Night work as a risk factor for future cause-specific disability pension: a prospective twin cohort study
in Sweden. Chronobiol Int
36. Bell CC. DSM-IV: diagnostic and statistical manual of mental disorders. JAMA
37. Svedberg P, Bardage C, Sandin S, Pedersen NL. A prospective study of health, life-style and psychosocial predictors of self-rated health. Eur J Epidemiol
38. Ropponen A, Svedberg P. Single and additive effects of health behaviours on the risk for disability pensions among Swedish twins. Eur J Public Health
39. Gold CH, Malmberg B, McClearn GE, Pedersen NL, Berg S. Gender and health: a study of older unlike-sex twins. J Gerontol B Psychol Sci Soc Sci
40. Reichborn-Kjennerud T, Stoltenberg C, Tambs K, et al. Back-neck pain
and symptoms of anxiety and depression: a population-based twin study. Psychol Med
41. Goesling J, Clauw DJ, Hassett AL. Pain
and depression: an integrative review of neurobiological and psychological factors. Curr Psychiatry Rep
42. Jansson C, Alexanderson K. Sickness absence due to musculoskeletal diagnoses and risk of diagnosis-specific disability pension: a nationwide Swedish prospective cohort study
43. Bjorkenstam C, Alexanderson K, Bjorkenstam E, Lindholm C, Mittendorfer-Rutz E. Diagnosis-specific disability pension and risk of all-cause and cause-specific mortality: a cohort study
of 4.9 million inhabitants in Sweden. BMC Public Health
44. Samuelsson A, Alexanderson K, Ropponen A, Lichtenstein P, Svedberg P. Incidence of disability pension and associations with socio-demographic factors in a Swedish twin cohort. Soc Psychiatry Psychiatr Epidemiol
45. Ropponen A, Silventoinen K, Koskenvuo M, Svedberg P, Kaprio J. Stability and change of body mass index as a predictor of disability pension. Scand J Public Health