Gender Differences in Self-Rated Health and Mortality Association: Role of Pain-inducing Musculoskeletal Disorders

Abstract

Aims:

We aimed to examine whether pain-inducing musculoskeletal disorders might explain the gender difference in the strength of the self-rated health (SRH)-mortality association.

Methods:

We pooled data from two National Health and Nutrition Examination Surveys (conducted in 1998 and 2001), which were linked to national mortality data of Korea using 13-digit unique personal identification numbers. There were 9,912 study participants, and 456 deaths were recorded (average length of follow-up = 5.7 years). Using a checklist of chronic conditions, the prevalences of major pain-causing musculoskeletal disorders (arthritis, sciatica, and herniated intervertebral disc) were obtained.

Results:

The relative risk (RR) of mortality for the poor and very poor categories of SRH tended to be greater in men than in women. Compared with those without musculoskeletal disorders, the RR for those with musculoskeletal disorders was similar in men but lower in women. Women had a greater prevalence of musculoskeletal disorders than men, and women with musculoskeletal disorders reported poorer health than did men with these disorders. In individuals without major musculoskeletal disorders, the strength of the SRH-mortality association was reduced in men but increased in women. Similar patterns in RRs for SRH by gender were observed when sociodemographic characteristics (education and marital status), number of severe chronic illnesses, and health behaviors (cigarette smoking, alcohol consumption, and regular physical exercise) were additionally adjusted for.

Conclusions:

Nonfatal musculoskeletal disorders may explain gender differences in the SRH-mortality association. Larger prospective studies in different cultural settings may help advance our understanding of the role of pain and pain-inducing musculoskeletal disorders in explaining gender differences in the SRH-mortality association.

Introduction

Self-rated health (SRH) has often been used as an important measure of health status, based on epidemiological studies showing a close relation between SRH and mortality.^1
–3 The strength of the SRH-mortality association has been reported to vary with gender, however. In their review of 27 studies on the SRH-mortality association, Idler and Benyamini¹ indicated that the ability of SRH to predict mortality was greater in men than in women. Although exceptions have been reported,^4,5 additional researchers have suggested that the SRH-mortality association is less strong in women than in men.^3,6

–9 Based on these findings, the need for closer examination of the effects of gender on the SRH-mortality association has been suggested.^8,10,11 For example, Idler et al. have urged researchers to systematically examine potential gender differences in the effects of SRH on mortality.^{8 p.882}

Several explanations for the gender difference have been proposed,^6,11

–14 including the suggestion that the SRH-mortality relationship in women could be lowered because women may experience more pain-inducing musculoskeletal diseases than men, leading to poorly perceived health but usually not affecting survival.^6,9,15 This hypothesis is plausible because it is now generally accepted that marked gender differences in pain reporting, pain tolerance, and pain perception exist.^16
–18 Epidemiological studies indicate that women report more musculoskeletal pain experiences than do men.^18,19 Experimental pain induction studies showed that women exhibit reduced thresholds and tolerances to a wide range of noxious stimuli,^16,17 and a prior study showed that women were more likely to include nonfatal conditions in their global health judgments and that this contributed to the weaker relationship between SRH and mortality in women.⁶ Apart from this work, no systematic analysis has specifically examined the hypothesis that pain-inducing musculoskeletal disorders may contribute to gender differences in the SRH-mortality association.

The purpose of this study is to examine the SRH-mortality association in men and women with or without musculoskeletal disorders causing severe pain.

Materials and Methods

Study subjects

This study was approved by the Institutional Review Board of the Asan Medical Center, Seoul, Korea. In this study, we used pooled mortality follow-up data of a nationally representative sample of Korean individuals, available from the 1998 and 2001 National Health and Nutrition Examination Surveys (NHANES) conducted by the Korea Institute for Health and Social Affairs. Information was collected from a stratified multistage probability sample of Korean households that represented the civilian, noninstitutionalized population. The surveys were divided into four parts: (1) health interview survey, (2) health examination survey, (3) health behavior survey, and (4) nutrition survey. The response rates of the health examination survey were 86.5% in the 1998 NHANES and 77.3% in the 2001 NHANES.^20,21 Information on physician-diagnosed chronic illnesses, including musculoskeletal disorders, was obtained from the health interview survey results. Health examination survey data included recording of 13-digit personal identification numbers to allow linkage to mortality data. The SRH parameter was obtained from the health behavior survey. Data from these three surveys were individually linked. The survey questionnaires on SRH, health behaviors, and other demographic variables in the 1998 and 2001 surveys were similar. A prior study pooled data in the two surveys.²² Each participant in the health examination surveys was interviewed about SRH and health behaviors, and the head of the family provided the information about participants' educational attainment, marital status, and chronic illnesses (severe chronic illness and musculoskeletal disorders). Information on age was based on the 13-digit personal identification numbers. Additional details about study design and methods are provided elsewhere.^20,21 A total of 11,969 men and women aged ≥30 years participated in the health examination surveys of the 1998 and 2001 NHANES. Of these, 10,437 reported valid 13-digit personal identification numbers, which were linked to data on mortality from the National Statistical Office of Korea (NSO). We excluded respondents lacking an SRH variable (n = 525, 5.0%). Of the remaining 9,912 respondents aged ≥30 years (95.0%), 456 (341 from the 1998 NHANES and 115 from the 2001 NHANES) died through December 2005.

Self-rated health

One question on SRH, based on the report from each participant, was included in this study (How would you rate your health as compared to that of others your age?), with a 5-point Likert scale answer category ranging from 1 = very good to 5 = very poor. Because of the small number of deaths that occurred in the very good SRH group and the similar mortality risks for very good, good, and fair SRH groups (Table 2), analyses of the combined categories of very good, good, and fair were also conducted.

Pain-inducing musculoskeletal disorders

Using a checklist of 37 (in 1998) and 53 (in 2001) chronic conditions, the head of a family reported each participant's chronic illnesses that had been diagnosed by physicians during the past 12 months. In this chronic disease checklist, we identified arthritis, sciatica, and herniated intervertebral disc as major musculoskeletal disorders and used these conditions as markers for pain-inducing musculoskeletal disorders.

Covariates

Variables on sociodemographic characteristics (education and marital status), other health conditions (severe chronic illness), and health behaviors (cigarette smoking, alcohol consumption, and regular physical exercise) were also considered in our analysis. Education level was determined by education level completed and classified as elementary school or less, middle school, high school, and college or over. Current marital status was grouped as married or unmarried. Using the same checklist of chronic conditions used for identification of musculoskeletal disorders, we identified respondents who had suffered any severe chronic diseases (cancer, diabetes, stroke, ischemic heart disease, other heart disease, chronic obstructive pulmonary disease, chronic liver disease, chronic renal disease, dementia, and pulmonary tuberculosis). The number of severe chronic diseases was used in the analysis. Three major health behaviors were also considered. Cigarette smoking status was grouped as one of four categories: never smoker, former smoker, irregular smoker, and daily smoker. Alcohol consumption was categorized as one of five groups: nondrinker, former drinker, minimal drinker, regular drinker, and frequent drinker. Regular physical exercise was assessed by asking if respondents had exercised regularly in the past month, with the response choices being yes and no.

Mortality

The outcome variable for this study was all-cause mortality. It is a legal requirement that all deaths of Koreans be reported to the NSO, and death certification data among Korean adults are known to be complete.²³ Dates of death were obtained from NSO death certificate data.

Statistical analysis

Unpaired Student's t test and chi-square test were used to compare simple differences in continuous variables and categorical variables, respectively, between men and women (Table 1). Cox's proportional hazard model was used to estimate relative risks (RRs) and associated 95% confidence intervals (CIs) of all-cause mortality according to SRH and major musculoskeletal disorders, adjusted for survey year (1998 and 2001) and age (10-year age groups). Adjustment of continuous age and age square terms produced similar results reported here. To take into account household clustering, Cox's regression models with robust standard errors that accounted for the clustering between household members were used but produced very similar results. Hazard ratios (HR) by Cox regression analysis were used as approximations of RR. Separate analyses were performed for men and women. We also analyzed gender differences in proportions of SRH categories and prevalences of musculoskeletal disorders. Logistic regression models were used to estimate odds of reporting SRH or musculoskeletal disorders in both genders, after adjusting for age. We next examined gender differences in the proportions of subjects with musculoskeletal disorders by categories of SRH. Finally, we present RRs for groups with or without musculoskeletal disorders among men and women. Gender differences in the magnitudes of RRs were checked by examining p values for interaction between gender and SRH variables (ordinal variables) in the Cox's regression model estimating mortality risk. All statistical analyses were performed using SAS statistical software (SAS Institute, Cary, NC), and a p value of 0.05 was considered statistically significant.

Table 1.

Number of Men and Women and Associated Proportion (%) According to Sociodemographic Characteristics, Severe Chronic Illness, and Health Behaviors

Characteristic	Men	Women	p value
No. of subjects	4,539 (100.0)	5,373 (100.0)
Person-years	25,614	30,448
No. of deaths	269 (5.9)	187 (3.5)	<0.0001
Age, mean (standard deviation)	48.4 (13.0)	48.9 (13.9)	0.054
Education
College or over	2,071 (45.6)	1,498 (27.9)	<0.0001
High school	1,153 (25.4)	1,360 (25.3)
Middle school	819 (18.0)	1,116 (20.8)
Elementary or less	494 (10.9)	1,397 (26.0)
Missing	2 (0.0)	2 (0.0)
Current marital status			<0.0001
Married	4,115 (90.7)	4,194 (78.1)
Unmarried	424 (9.3)	1,179 (21.9)
No. of severe chronic illness			0.958
0	3,812 (84.0)	4,515 (84.0)
1	638 (14.1)	757 (14.1)
≥2	89 (2.0)	101 (1.9)
Cigarette smoking			<0.0001
Never smoker	809 (17.8)	4,961 (92.3)
Former smoker	914 (20.1)	101 (1.9)
Irregular smoker	152 (3.3)	43 (0.8)
Daily smoker	2,663 (58.7)	268 (5.0)
Missing	1 (0.0)	0 (0.0)
Alcohol consumption			<0.0001
Nondrinker	581 (12.8)	2,470 (46.0)
Former drinker	333 (7.3)	138 (2.6)
Minimal drinker	601 (13.2)	1,419 (26.4)
Regular drinker	1,550 (34.1)	1,153 (21.5)
Frequent drinker	1,474 (32.5)	193 (3.6)
Regular physical exercise			<0.0001
Yes	1,224 (27.0)	1,088 (20.2)
No	3,313 (73.0)	4,283 (79.7)
Missing	2 (0.0)	2 (0.0)

Results

As shown in Table 1, the study cohort contained 56,082 person-years of follow-up (25,614 person-years for men and 30,448 person-years for women). A total of 37,583 and 18,479 person-years were from the 1998 and 2001 NHANES data, respectively. The average length of follow-up was 5.7 years (6.9 years for the 1998 NHANES and 4.1 for the 2001 NHANES). Of 456 total deaths, 59.0% (n = 269) were in men. Mean age difference between men and women was not statistically significant (p = 0.054); however, the proportion of participants aged ≥65 years was 13.7% (n = 624) in men and 16.5% (n = 886) in women. Men were more educated than women (p < 0.0001), and less men than women were unmarried (p < 0.0001). The prevalence of severe chronic illness was similar in both genders (p = 0.958). However, significant gender differences in health behaviors were observed: more men than women engaged in cigarette smoking, alcohol consumption, and regular physical exercise (all p < 0.0001).

As shown in Table 2, for both genders, mortality risks for the good and fair SRH categories were similar to those for the very good category in both genders. However, there were significant differences in RRs for the poor and very poor categories compared with those for the very good category. For the poor and very poor categories of SRH, the RRs tended to be greater in men than in women, but these gender differences were not significant. The p values for the interaction of gender with five-category and three-category SRH ordinal variables were 0.268 and 0.229, respectively. When we adjusted for education, marital status, and number of severe chronic illnesses (or plus three health behavior variables), we found similar patterns: RRs for poor and very poor SRH categories tended to be greater in men than women.

Table 2.

Number of Subjects and Relative Risk (RR) ^a of Mortality by Self-Rated Health Category among 4,539 Men and 5,373 Women: 1998–2001 National Health and Nutrition Examination Surveys of Korea

	No. of subjects (%)	No. of deaths	RR (95% CI)	RR (95% CI)
Men
Total	4,539 (100.0)	269
Very good	286 (6.3)	10	1.00 (reference)
Good	1,720 (37.9)	68	1.08 (0.55-2.09)	1.00 (reference)
Fair	1,672 (36.8)	58	1.06 (0.54-2.07)
Poor	707 (15.6)	92	2.34 (1.21-4.50)	2.20 (1.68-2.88)
Very poor	154 (3.4)	41	4.68 (2.33-9.41)	4.41 (3.08-6.31)
Women
Total	5,373 (100.0)	187
Very good	199 (3.7)	3	1.00 (reference)
Good	1,745 (32.5)	40	1.15 (0.35-3.71)	1.00 (reference)
Fair	1,768 (32.9)	27	1.16 (0.35-3.82)
Poor	1,363 (25.4)	81	2.17 (0.68-6.88)	1.89 (1.37-2.62)
Very poor	298 (5.5)	36	3.34 (1.03-10.90)	2.93 (1.94-4.40)

Adjustments were made for survey year (1998 and 2001) and age (10-year age groups).

Table 2 also shows relatively poor levels of SRH in women compared with men. A total of 6.3% (286 of 4,539) and 3.4% (154 of 4,539) of men reported their health as very good and very poor, respectively; the corresponding percentages in women were 3.7% (199 of 5,373) and 5.5% (298 of 5,373). When we examined the likelihood of reporting very good or very poor health status, by adjusting for age in the logistic models, gender differences were statistically significant (p < 0.0001 for both very good and very poor health).

Table 3 shows gender differences in the proportions of subjects with musculoskeletal disorders. Of a total of 4,539 men, 14.3% (n = 647) reported musculoskeletal disorders, whereas 28.0% (1,505 of 5,373) of women reported such disorders. Gender differences in the prevalence of each musculoskeletal disorder were significant for arthritis (7.3% of men vs. 18.7% of women) and sciatica (5.4% of men vs. 11.6% of women), but the gender differences were not so great in the prevalence of herniated intervertebral disc (3.3% of men vs. 4.1% of women). However, the gender difference in the likelihood of reporting herniated intervertebral disc after adjusting for age was significant (p = 0.016).

Table 3.

Number of Subjects and Relative Risk (RR) ^a of Mortality by Major Musculoskeletal Disorders among 4,539 Men and 5,373 Women: 1998–2001 National Health and Nutrition Examination Surveys of Korea

	Men		Women
	No. of subjects (No. of deaths)	RR (95% CI)	No. of subjects (No. of deaths)	RR (95% CI)
All musculoskeletal disorders
No	3,892 (203)	1.00 (reference)	3,868 (112)	1.00 (reference)
Yes	647 (66)	1.03 (0.77-1.36)	1,505 (75)	0.68 (0.50-0.91)
Arthritis
No	4,207 (235)	1.00 (reference)	4,370 (139)	1.00 (reference)
Yes	332 (34)	0.90 (0.63-1.30)	1,003 (48)	0.65 (0.46-0.90)
Sciatica
No	4,296 (233)	1.00 (reference)	4,749 (148)	1.00 (reference)
Yes	243 (36)	1.47 (1.03-2.10)	624 (39)	0.70 (0.49-1.00)
Herniated intervertebral disc
No	4,391 (260)	1.00 (reference)	5,151 (184)	1.00 (reference)
Yes	148 (9)	0.82 (0.42-1.59)	222 (3)	0.41 (0.13-1.29)

Adjustments were made for survey year (1998 and 2001) and age (10-year age groups).

Table 3 shows RRs for mortality by major musculoskeletal disorders in men and women and reveals gender differences in the association between musculoskeletal disorders and mortality. There was a similar mortality risk (RR = 1.03, 95% CI 0.77-1.36) for men with musculoskeletal disorders compared with those without, whereas the mortality risk among women with musculoskeletal disorders was significantly lower (RR = 0.68, 95% CI 0.50-0.91) than in women without musculoskeletal disorders. These reduced mortality risks were found in women with arthritis, sciatica, or herniated intervertebral disc. In men with sciatica, however, there was a significantly greater mortality risk compared with men without sciatica.

As shown in Table 4, subjects with major musculoskeletal disorders reported poorer health than did those without such disorders. For example, 8.4% of men with musculoskeletal disorders rated their health as very poor, whereas only 2.6% of men without musculoskeletal disorders reported such a rating. Similar results were found in women.

Table 4.

Numbers and Proportion (%) of Subjects with/without Major Musculoskeletal Disorders by Gender and Categories of Self-Rated Health: 1998–2001 National Health and Nutrition Examination Surveys of Korea

	All musculoskeletal disorders		Arthritis		Sciatica		Herniated intervertebral disc
Categories of self-rated health	No (%)	Yes (%)	No (%)	Yes (%)	No (%)	Yes (%)	No (%)	Yes (%)
Men
Total	3,892 (100)	647 (100)	4,207 (100)	332 (100)	4,296 (100)	243 (100)	4,391 (100)	148 (100)
Very good	271 (7.0)	15 (2.3)	279 (6.6)	7 (2.1)	283 (6.6)	3 (1.2)	281 (6.4)	5 (3.4)
Good	1,545 (39.7)	175 (27.1)	1,645 (39.1)	75 (22.6)	1,640 (38.2)	80 (32.9)	1,692 (38.5)	28 (18.9)
Fair	1,454 (37.4)	218 (33.7)	1,559 (37.1)	113 (34.0)	1,589 (37.0)	83 (34.2)	1,623 (37.0)	49 (33.1)
Poor	522 (13.4)	185 (28.6)	602 (14.3)	105 (31.6)	644 (15.0)	63 (25.9)	658 (15.0)	49 (33.1)
Very poor	100 (2.6)	54 (8.4)	122 (2.9)	32 (9.6)	140 (3.3)	14 (5.8)	137 (3.1)	17 (11.5)
Women
Total	3,868 (100)	1,505 (100)	4,370 (100)	1,003 (100)	4,749 (100)	624 (100)	5,151 (100)	222 (100)
Very good	185 (4.8)	14 (0.9)	189 (4.3)	10 (1.0)	195 (4.1)	4 (0.6)	198 (3.8)	1 (0.5)
Good	1,469 (38.0)	276 (18.3)	1,566 (35.8)	179 (17.9)	1,628 (34.3)	117 (18.8)	1,725 (33.5)	20 (9.0)
Fair	1,391 (36.0)	377 (25.1)	1,528 (35.0)	240 (23.9)	1,613 (34.0)	155 (24.8)	1,715 (33.3)	53 (23.9)
Poor	690 (17.8)	673 (44.7)	907 (20.8)	456 (45.5)	1,078 (22.7)	285 (45.7)	1,249 (24.3)	114 (51.4)
Very poor	133 (3.4)	165 (11.0)	180 (4.1)	118 (11.8)	235 (5.0)	63 (10.1)	264 (5.1)	34 (15.3)

Table 4 also shows gender differences in the distribution of SRH categories by musculoskeletal disorders. Among participants with musculoskeletal disorders, women tended to report poorer health than did men. For example, 37.0% (28.6% and 8.4%) of men with musculoskeletal disorders reported their health as poor or very poor, but the proportion of women with musculoskeletal disorders reporting poor or very poor health was 55.7% (44.7% and 11.0%) among female participants. However, gender differences in the distribution of SRH categories were not so great among those without musculoskeletal disorders. In those without such disorders, a total of 16.0% (13.4% and 2.6%) of men and 21.2% (17.8% and 3.4%) of women reported poor or very poor health. Among subjects without musculoskeletal disorders, the gender difference in the proportion of those reporting good or very good health was not marked (46.7% in men vs. 42.8% in women). Among participants who reported musculoskeletal disorders, the gender difference in the prevalence of reporting good or very good SRH was evident (29.4% in men vs. 19.2% in women). Similar patterns were seen for each musculoskeletal disorder examined in this study.

Table 5 shows RRs of mortality by SRH categories in groups with or without major musculoskeletal disorders. Based on the comparison of RRs before and after exclusion of musculoskeletal disorders (comparison of RRs in Tables 2 and 5), RRs in men without musculoskeletal disorders tended to decrease (e.g., 4.41 in Table 2 to 3.99 in Table 5 for the very poor category), whereas RRs in women without such disorders showed an increasing tendency (e.g., 2.93 in Table 2 to 4.35 in Table 5 for the very poor category). Contrary to the findings shown in Table 2, Table 5 also indicates that the RR for each category of SRH among those without musculoskeletal disorders tends to be greater in women than in men, although the interaction between the ordinal SRH variable and gender was not significant. The p values for the interaction between gender and SRH ordinal variables (five-category and three-category) in the Cox's regression model predicting mortality were 0.546 and 0.693, respectively. Meanwhile, in the group with musculoskeletal disorders, there was a more marked gender difference in the RR by SRH category. For example, the interaction of gender variable with gender and five-category or three-category SRH ordinal variable was marginally significant (p = 0.055 and p = 0.081, respectively). When we included sociodemographic characteristics (education and marital status) and numbers of severe chronic illnesses or added health behaviors (cigarette smoking, alcohol consumption, and regular physical exercise) as well as these sociodemographic characteristics and numbers of severe chronic illnesses, similar patterns in the RRs of SRH categories by genders were observed (data not shown but available upon request).

Table 5.

Number of Subjects and Relative Risk (RR) ^a of Mortality among Men and Women with/without Major Musculoskeletal Disorders by Categories of Self-Rated Health: 1998–2001 National Health and Nutrition Examination Surveys of Korea

	Men			Women
Categories of self-rated health	No. of subjects (No. of deaths)	RR (95% CI)	RR (95% CI)	No. of subjects (No. of deaths)	RR (95% CI)	RR (95% CI)
Those without musculoskeletal disorders
Total	3,892 (203)			3,868 (112)
Very good	271 (10)	1.00 (reference)	1.00 (reference)	185 (3)	1.00 (reference)	1.00 (reference)
Good	1,545 (59)	1.11 (0.57-2.17)		1,469 (28)	1.00 (0.30-3.28)
Fair	1,454 (47)	1.03 (0.52-2.05)		1,391 (21)	1.44 (0.43-4.87)
Poor	522 (62)	2.27 (1.16-4.44)	2.13 (1.55-2.91)	690 (37)	2.55 (0.78-8.29)	2.25 (1.46-3.46)
Very poor	100 (25)	4.26 (2.03-8.93)	3.99 (2.56-6.22)	133 (23)	4.87 (1.46-16.29)	4.35 (2.63-7.19)
Those with musculoskeletal disorders
Total	647 (66)	1.00 (reference)		1,505 (75)	1.00 (reference)
Very good	15 (0)		1.00 (reference)	14 (0)		1.00 (reference)
Good	175 (9)			276 (12)
Fair	218 (11)	1.38 (0.57-3.36)		377 (6)	0.55 (0.21-1.46)
Poor	185 (30)	3.07 (1.42-6.63)	2.59 (1.44-4.68)	673 (44)	1.66 (0.87-3.15)	2.12 (1.22-3.68)
Very poor	54 (16)	6.77 (2.90-15.77)	5.71 (2.89-11.27)	165 (13)	1.83 (0.83-4.03)	2.34 (1.14-4.79)

Adjustments were made for survey year (1998 and 2001) and age (10-year age groups).

Discussion

In this study, we examined whether major pain-causing musculoskeletal disorders could explain the gender difference in the association (measured by RR) between SRH and mortality. We found that the SRH-mortality association tended to decrease in men but increase in women. The RR for the very poor SRH group compared with the very good/good/fair SRH group was 4.41 in all male participants (Table 2) but showed a decreasing tendency (to 3.99) in male subjects without musculoskeletal disorders (Table 5). The RR for the very poor SRH category in all women was 2.93 (Table 2) but tended to increase to 4.35 when women with musculoskeletal disorders were excluded (Table 5). These patterns were also found when we further adjusted for education, marital status, and numbers of severe chronic illnesses or added cigarette smoking, alcohol consumption, and regular physical exercise as covariates. Although the interaction between ordinal SRH variable and gender was not significant in any group either before or after those without musculoskeletal disorders were excluded, we consider this gender difference in the SRH-mortality association to be an important finding, suggesting that pain-causing musculoskeletal disorders may explain gender differences in the SRH-mortality association.

Several mechanisms can be proposed to explain the effect of musculoskeletal disorders on gender differences in the SRH-mortality association. If pain-inducing musculoskeletal disorders reduce the SRH-mortality association in both genders (i.e., yielding a less clear SRH-mortality association among those with musculoskeletal disorders), the increased prevalence of pain-causing musculoskeletal diseases in females could be responsible for the smaller RRs by SRH categories in women. Otherwise, musculoskeletal disorders may affect the SRH-mortality association differently in men and women. This seems to be the case in our analysis. The SRH-mortality association was smaller in women with musculoskeletal disorders than in women without such disorders. The mortality risk for the very poor SRH group compared with the very good/good/fair SRH group was 4.35 in women without musculoskeletal disorders but 2.34 in women with such disorders (Table 5). This was not true for men, however. The RR for the very poor SRH group was 3.99 in men without musculoskeletal disorders but 5.71 in men with such disorders (Table 5).

An interesting question is why the SRH-mortality association was reduced in women with pain-inducing musculoskeletal disorders but not in men. Considering the SRH-mortality association by musculoskeletal disorders (Table 5), gender differences in the prevalence of pain-inducing musculoskeletal disorders may not explain gender differences in the SRH-mortality association. Rather, the reason for a reduced SRH-mortality association in women may be differences in pain perception between genders. Many studies have reported that women show greater perceptual responses to various painful stimuli than do men.^16,17,24 This observed gender difference in pain perception has been attributed to biological (e.g., resting blood pressure),²⁵ psychological (e.g., mood states),²⁶ and social factors (e.g., social learning and gender role expectation).²⁷ Our analysis in Table 4 showed that women tended to report poorer SRH than men when they had pain-inducing musculoskeletal disorders. This suggests that women consider pain a more important constituent of SRH than do men. This perception among women may obscure mortality prediction abilities of other health conditions that are more directly related to mortality. Health conditions that are less strongly related to mortality have been proposed to be more important constituents of SRH in women than in men.^9,13 Deeg and Kriegsman¹³ found that the most notable gender difference in the predictive ability of mortality was observed when study subjects were asked to rate their health compared with others of the same age. In our study, the question about SRH was on perceived health compared with age peers. The data suggest that women's rating of health status may be disproportionately influenced by chronic, painful, but less fatal conditions among their female age peers, whereas men's ratings of health are better predictors of mortality risk because men base their SRH estimates on fatal health conditions.¹¹

The results of the present study show that mortality risks among women with musculoskeletal disorders were smaller than those among women without such disorders, whereas the mortality risks were similar for men with and without musculoskeletal disorders (Table 3). Most musculoskeletal disorders are not fatal, so gender differences in mortality patterns by musculoskeletal disorders may arise because of gender differences in other characteristics of those with musculoskeletal disorders. For example, our additional analyses (data not presented) show gender differences in the association between musculoskeletal disorders and severe chronic diseases. After adjusting for survey year and age, the odds of having a severe chronic disease were greater in men with musculoskeletal disorders than in men without such disorders (odds ratio [OR] = 1.25, 95% CI 1.01-1.55), whereas the odds ratio for women was not significant (OR = 0.99, 95% CI 0.83-1.17). This indicates that men who report musculoskeletal disorders are more likely to have another severe chronic disorder, whereas women who report musculoskeletal disorders are not. A prior study showed that the SRH-mortality association was strong in respondents with circulatory system disease, but the association was not strong among healthy respondents.²⁸

This study showed that mortality risks for very good, good, and fair SRH categories were similar in both men and women. Although many previous studies have shown a graded mortality risk by SRH category,^1
–3 several studies have shown similar mortality risks in excellent and good SRH categories.^8,29,30 No explanation for the observed similar risk of mortality has been proposed. Investigations into the similar mortality risk patterns among favorable SRH categories need to be conducted.

This study should be viewed in light of its limitations. First, the overall number of deaths in this study (n = 456) was not smaller than those in previous cohort studies³ examining the SRH-mortality association but was not sufficiently high to show a significant gender difference in the SRH-mortality association. Considering this limitation, a larger prospective study should be performed to examine patterns in and causes of gender differences. Second, we identified physician-diagnosed major musculoskeletal disorders by a self-reported survey but had no accurate information on the severity and duration of pain caused by those disorders. Future studies using this information may help determine the contribution of nonfatal pain to the gender difference in the SRH-mortality association.

Conclusions

This study provides evidence of a gender difference in the SRH-mortality association in a representative South Korean population sample. The findings show that pain-causing musculoskeletal diseases may explain the gender difference in the SRH-mortality association. Considering the increasing incidence of musculoskeletal disorders, especially in women,³¹ the role of musculoskeletal disorders could become greater in measuring SRH and explaining the SRH-mortality association. Given that cultural background is known to have a gender-specific effect on SRH⁹ and that considerable differences in the response to and expression of pain have been reported among different cultures,³² more studies in different cultural settings are warranted to advance our understanding of the role of pain-inducing musculoskeletal disorders in explaining gender differences in the SRH-mortality association.

Footnotes

Disclosure Statement

The authors have no conflicts of interest to report.

References

Idler

, Benyamini

. Self-rated health and mortality: A review of twenty-seven community studies. J Health Soc Behav, 1997; 38:21–37.

Benyamini

, Idler

. Community studies reporting association between self-rated health and mortality. Res Aging, 1999; 21:392–401.

DeSalvo

, Bloser

, Reynolds

, He

, Muntner

. Mortality prediction with a single general self-rated health question: A meta-analysis. J Gen Intern Med, 2005; 20:267–275.

Miilunpalo

, Vuori

, Oja

, Pasanen

, Urponen

. Self-rated health status as a health measure: The predictive value of self-reported health status on the use of physician services and on mortality in the working-age population. J Clin Epidemiol, 1997; 50:517–528.

Strawbridge

, Wallhagen

. Self-rated health and mortality over three decades. Results from a time-dependent covariate analysis. Res Aging, 1999; 21:402–416.

Benyamini

, Leventhal

. Gender differences in processing information for making self-assessments of health. Psychosom Med, 2000; 62:354–364.

Hays

, Shoenfeld

, Blazer

, Gold

. Global self-ratings of health and mortality: Hazard in the North Carolina Piedmont. J Clin Epidemiol, 1996; 49:969–979.

Idler

, Russell

, Davis

. Survival, functional limitations, and self-rated health in the NHANES I Epidemiologic Follow-up Study, 1992. First National Health and Nutrition Examination Survey. Am J Epidemiol, 2000; 152:874–883.

Jylha

, Guralnik

, Ferrucci

, Jokela

, Heikkinen

. Is self-rated health comparable across cultures and genders? J Gerontol Soc Sci, 1998; 53B:S144–S152.

10.

Deeg

, Bath

. Self-rated health, gender, and mortality in older persons: Introduction to a special section. Gerontologist, 2003; 43:369–371.

11.

Idler

. Discussion: Gender differences in self-rated health, in mortality, and in the relationship between the two. Gerontologist, 2003; 43:372–375.

12.

Benyamini

, Blumstein

, Lusky

, Modan

. Gender differences in the self-rated health-mortality association: Is it poor self-rated health that predicts mortality or excellent self-rated health that predicts survival? Gerontologist, 2003; 43:396–405.

13.

Deeg

DJH

, Kriegsman

DMW

. Concepts of self-rated health: Specifying the gender difference in mortality risk. Gerontologist, 2003; 43:376–386.

14.

Spiers

, Jagger

, Clarke

, Arthur

. Are gender differences in the relationship between self-rated health and mortality enduring? Results from three birth cohorts in Melton Mowbray, United Kingdom. Gerontologist, 2003; 43:406–411.

15.

Benjamins

, Hummer

, Eberstein

, Nam

. Self-reported health and adult mortality risk: An analysis of cause-specific mortality. Soc Sci Med, 2004; 59:1297–1306.

16.

Chesterton

, Barlas

, Foster

, Baxter

, Wright

. Gender differences in pressure pain threshold in healthy humans. Pain, 2003; 101:259–266.

17.

Riley

3rd , Robinson

, Wise

, Myers

, Fillingim

. Sex differences in the perception of noxious experimental stimuli: A meta-analysis. Pain, 1998; 74:181–187.

18.

Unruh

. Gender variations in clinical pain experience. Pain, 1996; 65:123–167.

19.

Rollman

, Lautenbacher

. Sex differences in musculoskeletal pain. Clin J Pain, 2001; 17:20–24.

20.

Ministry of

Health

. Welfare of Korea (MOHW). Interim report on the 1998 National Health and Examination Nutrition Survey [in Korean] Kwachon: Ministry of Health and Welfare of Korea, 1999.

21.

Ministry of

Health

. Welfare of Korea (MOHW). Interim report on the 2001 National Health and Examination Nutrition Survey [in Korean] Kwachon: Ministry of Health and Welfare of Korea, 2002.

22.

Khang

, Lynch

, Harper

, Yang

, Jung-Choi

, Kim

. The contribution of material, psychosocial, and behavioral factors to explain educational and occupational mortality inequalities in a nationally representative sample of South Koreans: Relative and absolute perspectives. Soc Sci Med, 2009; 68:858–866.

23.

Khang

, Lynch

, Kaplan

. Health inequalities in Korea: Age- and sex-specific educational differences in the 10 leading causes of death. Int J Epidemiol, 2004; 33:299–308.

24.

Fillingim

, Maixner

. Gender differences in the responses to noxious stimuli. Pain Forum, 1995; 4:209–221.

25.

Myers

, Robinson

, Riley

3rd , Sheffield

. Sex, gender, and blood pressure: Contributions to experimental pain report. Psychosom Med, 2001; 63:545–550.

26.

Garofalo

, Lawler

, Robinson

, Morgan

, Kenworthy-Heinige

. The role of mood states underlying sex differences in the perception and tolerance of pain. Pain Pract, 2006; 6:186–196.

27.

Wise

, Price

, Myers

, Heft

, Robinson

. Gender role expectations of pain: Relationship to experimental pain perception. Pain, 2002; 96:335–342.

28.

Idler

, Leventhal

, McLaughlin

, Leventhal

. In sickness but not in health: Self-ratings, identity, and mortality. J Health Soc Behav, 2004; 45:336–356.

29.

Bath

. Differences between older men and women in the self-rated health-mortality relationship. Gerontologist, 2003; 43:387–395.

30.

McCallum

, Shadbolt

, Wang

. Self-rated health and survival: A 7-year follow-up study of Australian elderly. Am J Public Health, 1994; 84:1100–1105.

31.

Kopec

, Rahman

, Sayre

et al. Trends in physician-diagnosed osteoarthritis incidence in an administrative database in British Columbia, Canada, 1996–1997 through 2003–2004. Arthritis Rheum, 2008; 59:929–934.

32.

Nayak

, Shiflett

, Eshun

, Levine

. Culture and gender effects in pain beliefs and the prediction of pain tolerance. Cross-Cultural Res, 2000; 34:135–151.