Differences in quality of life between women and men in the older population of Spain

Differences in quality of life between women and men in the older population of Spain

ARTICLE IN PRESS Social Science & Medicine 60 (2005) 1229–1240 www.elsevier.com/locate/socscimed Differences in quality of life between women and me...

239KB Sizes 0 Downloads 224 Views

ARTICLE IN PRESS

Social Science & Medicine 60 (2005) 1229–1240 www.elsevier.com/locate/socscimed

Differences in quality of life between women and men in the older population of Spain Pilar Guallar-Castillo´na, A´urea Redondo Sendinoa, Jose´ R. Banegasa, Esther Lo´pez-Garcı´ ab, Fernando Rodrı´ guez-Artalejoa, a

Departamento de Medicina Preventiva y Salud Pu´blica, Facultad de Medicina, Universidad Auto´noma de Madrid, Avenida Arzobispo Morcillo, sn, 28029 Madrid, Spain b Department of Nutrition. Harvard School of Public Health. 665 Huntington Avenue, 02115 Boston, MA, USA Available online 11 September 2004

Abstract The objective of the study was to examine the contribution of sociodemographic factors, lifestyle, social network, chronic morbidity and use of healthcare services to the poorer health-related quality of life (HRQL) of women, as compared to that of men, among the older population of Spain. Data were collected by home-based personal interview and physical examination of 3260 subjects representative of the Spanish non-institutionalized population aged 60 years and over. HRQL was assessed with the SF-36 health questionnaire. Relative differences in HRQL between women and men were summarized using odds ratios of suboptimal health (score o100) on each scale of the SF-36, obtained from logistic regression. The contribution of the variables of interest to the relative differences in HRQL between the sexes was evaluated as the percentage change in the odds ratio before and after adjustment for such variables. The odds ratio of suboptimal health among women versus men was higher than 2 (po0.0001) on all SF-36 scales. Adjustment for sociodemographic variables led to a reduction of 23% (95% confidence limits (CL): 38 to 5%) in the odds ratio on the social functioning scale, while adjustment for lifestyle reduced the odds ratio on the general health and social functioning scales by 45% (95%CL: 64 to 15%) and 29% (95%CL: 42 to 13%), respectively. Adjustment for the social network, chronic morbidity and use of healthcare services variables did not lead to significant changes in the odds ratios on any of the SF-36 scales. In general, the contribution of the study variables to differences in HRQL between the sexes was smaller in the oldest age groups. We conclude that sociodemographic and lifestyle factors may explain a substantial part of the differences between women and men in certain HRQL dimensions. Some of these factors, such as the lower educational level and the higher frequency of sedentary lifestyles and obesity among women, are potentially modifiable. r 2004 Elsevier Ltd. All rights reserved. Keywords: Inequalities; Gender differences; Quality of life; Elderly; Spain

Introduction

Corresponding author. Tel: +34-91-497-5444; fax: +34-91-

497-5353. E-mail address: [email protected] (F. Rodrı´ guez-Artalejo).

Health-related quality of life (HRQL) provides a subjective overview of the state of health of individuals. Worse HRQL is associated with higher mortality (Ries, Kaplan, Limbreg, & Prewitt, 1995) and a greater use of healthcare services (Conelli, Philbrick, Smith, Kaiser &

0277-9536/$ - see front matter r 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.socscimed.2004.07.003

ARTICLE IN PRESS 1230

P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

Wymer, 1989; Siu, Reuben, Ouslander & Osterweil, 1993). Women tend to report a poorer HRQL than men, both in selected samples of subjects (Mele´ndez Herna´ndez, Montero Herrero, Jime´nez Sa´nchez & Blanco Montagut, 2001; Walters, Munro & Brazier, 2001) and in the general population (Alonso, Regidor, Barrio, Prieto, Rodrı´ guez, & De la Fuente, 1998; Azpiazu et al., 2003; Hopman et al., 2000; Loge & Kaasa, 1998; Lo´pez Garcı´ a et al., 2003b; Scott, Tobias, Sarfati & Haslett, 1999; Sullivan & Karlsson, 1998). Although there is a substantial amount of literature on possible explanatory factors of the greater morbidity and disability, and worse subjective health reported by women (Rohlfs, Borrell & Fonseca, 2000; Ruiz & Verbrugge, 1997), very few studies have specifically addressed the possible determinants of the differences in HRQL between men and women. Moreover, such studies are particularly infrequent among samples representative of the older adult population (Arber & Cooper, 1999; Arber & Ginn, 1993; Dahl & Birkelund, 1997). The determinants of the differences in health between women and men may differ with the measure of health used (Macintyre, Hunt & Sweeting, 1996) and, consequently, the results obtained on subjective health or disability may not apply to HRQL. Similarly, factors that explain differences in health status may vary across the life cycle (Macintyre et al., 1996). The study of such factors in older adults is particularly relevant. First, because in this population segment, whose size is progressively growing, health needs are much greater than among the young; second, because the predominance of women over men increases with age and HRQL is worse among the former. Furthermore, differences in HRQL between women and men may change with population’s cultural values and degree of economic development, which differ between southern European countries and those in the north of Europe and North America. Cultural values may influence not only the meaning, interpretation, knowledge and potential determinants of health and disease, but also the manner of reporting them. Moreover, cultural values and degree of economic development influence women’s incorporation to paid work and fulfillment of their social role, with possible effects on differences in HRQL between sexes (Annandale & Hunt, 2001). Lastly, theories explaining the differences in health between women and men include strictly biological factors (genes, anatomy, hormones, reproductive history, etc.), factors stemming from women’s social role (social network and support, non-paid work at home, etc.) and mixed factors that are a combination of the previous two (health-related lifestyles, use of healthcare services, mental health disorders, etc.) (Dahl & Birkelund, 1997). The contribution of these types of factors to differences in HRQL between women and

men depends on two elements: (a) the effect of each on health, something that may vary with sex (e.g., whereas tobacco and alcohol have a greater influence on men’s health, sedentary lifestyle and obesity have a greater influence on women’s health) (Denton & Walters, 1999); and (b) the frequency and distribution of such factors in each sex. Both elements may vary with the country, culture, age and calendar time (Hunt, 2002; Wiggins et al., 2002). Accordingly, this study examines the contribution of sociodemographic factors, lifestyle, social network, chronic morbidity and use of healthcare services to the poorer HRQL of women, as compared to that of men, in the older population of Spain; it also ascertains whether the contribution of such factors changed with age. To our knowledge this is the first study of its type conducted in a European Mediterranean country.

Materials and methods Study design and subjects This was a cross-sectional survey covering a sample of 4000 subjects representative of the non-institutionalized Spanish population aged 60 years and over. The study was approved by the Clinical Research Ethics Committee of the ‘‘La Paz’’ University Hospital in Madrid. Study subjects were selected through probabilistic multistage cluster sampling. Firstly, clusters were stratified by region of residence and size of town. Thereafter, census sections were selected at random in each cluster, followed by individual households where information was then obtained from residents. Data were collected on a total of 450 census sections in Spain, with subjects being selected in two sex and three age (60–69 y, 70–79 y, 80 y and over) strata. Individuals aged 80 years and over were oversampled to assure a large enough number of subjects for a meaningful analysis. Subjects were replaced for interviews only after 10 failed visits by the interviewer or because of subject’s incapacity, death, institutionalization or refusal to participate. There was an overall study response rate of 71%. Reasons for non-response were ‘impossible to locate after several attempts’ (17%), ‘refused to be interviewed’ (6%) and the rest of motives (6%). Given the study’s sample design, subjects were assigned a weighting coefficient according to their sex, age, region and size of town of residence, which allowed for reconstructing the characteristics of the Spanish population in the analysis. Study variables Information was collected from October 2000 to February 2001 by home-based personal interview using

ARTICLE IN PRESS P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

a structured questionnaire, followed by a physical examination to measure blood pressure and anthropometric variables. In all cases, informed consent was obtained from subjects or cohabiting next-of-kin. Of the 4000 subjects surveyed, 3260 (81.32%) furnished complete information on all study variables. Compared to the subjects who provided complete information on the study variables, those who did not were more frequently women (67% versus 54%) and older (a mean of 73.6 versus 71.6 years). Interviewers underwent standardized training to administer the questionnaire, and take blood pressure readings and anthropometric measurements. HRQL was measured using the Spanish version of the SF-36 questionnaire (Alonso, Prieto & Anto´, 1995), which was included at the start of the survey. This questionnaire is made up of 36 items, which assess the following eight HRQL dimensions or scales: physical functioning, role-physical, body pain, general health, vitality, social functioning, role-emotional and mental health. Physical functioning, role-physical and body pain reflect the physical component of health; social functioning, role-emotional and mental health cover the psycho-social aspects; and vitality and general health give an overall idea of subjective health, and are associated with both the physical and mental aspects Among the subjects that answered the SF-36 questionnaire, the percentage of persons who completed all the items on each of the scales was very high. In the case of the vitality scale, which was the one that registered the lowest response, 95.8% of subjects answered all the items. The SF-36 allows for assigning values to items in cases where individuals answer more than half the items comprising a scale. In this study, a value was imputed to only 1.5% of subjects. When asked as to the degree to which they understood the SF-36 questions, 95.8% of subjects rated it very high or fairly high. The Spanish version of the SF-36 has previously been used in the elderly (Alonso et al., 1998; Ferrer & Alonso, 1998; Mele´ndez Herna´ndez et al., 2001), and has shown good reproducibility and validity for measuring HRQL (Alonso et al., 1995). Aside from sex, information was collected on the following sociodemographic variables: age, size of town of residence, marital status, head-of-family status, educational level, employment status and last occupation of the head of the family. Lifestyle variables on which information was obtained were: tobacco use, physical activity and alcohol consumption. Moderate drinkers were deemed to be men who consumed p30 g and women who consumed p20 g of alcohol daily; heavy drinkers were those who exceeded the limits of moderate consumption in each sex. In addition, weight and height were measured using standardized procedures (Lo´pez Garcı´ a et al., 2003a). Body mass index (BMI) was calculated as weight

1231

in kilograms divided by the square of the height in meters, and subjects were classified in three groups: low and normal weight (o25 kg/m2), overweight (25–29.9 kg/m2), and obese (X30 kg/m2). Blood pressure was determined in a standardized manner (Banegas et al., 2002), with individuals deemed hypertensive where systolic blood pressure was X140 mm Hg, diastolic blood pressure was X90 mm Hg or on current antihypertensive drug treatment. Social network was assessed by whether subjects lived alone or not, saw family members daily/almost daily or more seldom, and saw friends and neighbors daily/ almost daily or more seldom. Health-related variables included self-reported chronic diseases diagnosed by the physician, and use of healthcare services. Among the former, data were obtained on asthma and chronic bronchitis, ischemic heart disease, cerebrovascular disease, arthritis, cataracts without treatment, diabetes mellitus, Parkinson’s disease, cancer (any site), and depression with need for treatment. The number of medications taken by subjects was also obtained. Data were likewise collected on use of healthcare services, specifically: the frequency with which they had visited their physician in the previous 2 weeks; the frequency of home visits by the physician in the previous 2 weeks; and admission to hospital in the year preceding the interview.

Data analysis Results obtained for scales of the SF-36 questionnaire receive a numerical score, which, after being coded, is standardized and ranked on a scale of 0–100. The higher the score the better the state of health (Medical Outcomes Trust, 1996). For each SF-36 scale a dichotomous variable was defined, with a value of 1 being assigned to cases with score 100 and a value of 0 to cases where score was o100. The relative differences in HRQL between women and men were summarized using odds ratios (OR) of suboptimal health (score o100) on each scale of the SF-36, obtained through logistic regression models in which HRQL was the dependent and sex the independent variable. First, crude OR was obtained; then, models were adjusted for groups of variables (sociodemographic, lifestyle, social network, and healthrelated). All independent variables were modeled as dummies. Because age is conceptually a very important variable, first we present results adjusted solely for age (in five year groups) and, then, for all sociodemographic variables including age. The respective contribution of the groups of variables to the relative differences in the frequency of suboptimal health between women and men were evaluated as the percentage change (PC) in the OR of the variable

ARTICLE IN PRESS 1232

P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

‘‘sex’’, before and after adjustment for such groups of variables (PC OR=ORadjustedORcrude/ORcrude  100). Analyses were performed with the SAS package version 8.02 (2001).

Results The study sample comprised 1768 (54.2%) women and 1492 (45.8%) men, with a mean age of 72.2 and 70.8 years, respectively. Table 1 shows the distribution by sex of the variables of interest organized into four groups, i.e., sociodemographic, lifestyle, social network and health-related. Differences between women and men (po0.05) were observed for most of the variables, except for size of town of residence and use of healthcare services. A total of 80.8% of men, versus 45.6% of women, were married. A greater percentage of men than women were head of the family (94% versus 27%) had a higher educational level and were in paid work (33.1% versus 13.3%). As regards social network, a greater percentage of women than men lived alone and saw family members daily (25.3% versus 8.6%, and 55.7% versus 51.8%, respectively). However, men tended to see friends and neighbors daily more frequently than did women (87.6% versus 82.8%). In terms of lifestyles, women led more sedentary lives and were obese more frequently than men, yet there were nevertheless more overweight men than women. Alcohol and tobacco consumption were higher among men. Women took more medications, used services from primary care physicians more frequently, and presented with a higher number of chronic diseases than did men. On all scales of the SF-36 questionnaire, the percentage of women with maximum score was lower than that of men (Table 2). The odds ratio of suboptimal health among women versus men was higher than 2 (po0.0001) on all SF-36 scales (Table 2). Adjustment for age did not produce a significant change in odds ratios for any of the SF-36 scales (graph a in Fig. 1). Further adjustment for the rest of sociodemographic variables led to a statistically significant reduction of 22.98% (95% confidence limits (CL): 37.67 to 4.80%) in the odds ratio on the social functioning scale (graph b in Fig. 1), while adjustment for lifestyle reduced the odds ratio on the general health and social functioning scales by 44.76% (95%CL: 63.73 to –15.91%) and 29.15% (95%CL: 42.23 to –13.08%), respectively (graph c in Fig. 1). Adjustment for the social network and health-related variables did not change materially the odds ratios on any of the HRQL scales (graphs d and e in Fig. 1). Simultaneous adjustment for sociodemographic variables and lifestyle did not modify the odds ratio substantially on any of the HRQL scales beyond that observed after adjustment for lifestyle

(graph f in Fig. 1). Adjustment for the four types of variables led to a reduction of 25.06% (95%CL: 43.41 to –0.02%) in the odds ratio on the role-physical scale (graph g in Fig. 1). Among the sociodemographic variables, adjusting solely for head-of-family status or educational level reduced—though not significantly (p40.05)—the odds ratio of suboptimal health of women versus that of men on the social functioning scale. In the case of the lifestyle variables, adjusting solely for BMI or physical activity reduced the odds ratio on the general health and social functioning scales but, once again, the reduction failed to attain statistical significance (data not shown). The contribution of the study variables to differences in HRQL between women and men appears to decline at the most advanced ages (Fig. 2). In persons aged 60–69 years, simultaneous adjustment for sociodemographic, lifestyle, social network and health-related variables led to a significant (po0.05) and substantial (over 40%) reduction in the odds ratios on the physical functioning, general health, vitality, social functioning and mental health scales. Among persons aged 70–79 years, adjustment for the four types of variables reduced the odds ratios solely on the general health and social functioning (po0.05) scales, and among those aged 80 years and over, it reduced the odds ratios on the role-physical scale alone (po0.05).

Discussion Among the older population of Spain, women have a substantially worse HRQL than men, on both the physical and mental scales. The study variables had an impact on gender differences in HRQL only for a minority of dimensions. However, sociodemographic factors, such as not being head of the family and having a lower educational level, and lifestyle-related variables, such as a higher BMI and less physical activity, may partly explain the worse score registered by women on the general health and social functioning scales. The contribution of the study factors to differences in HRQL between women and men seems to decline with age. These findings are relevant because some of the variables that can explain such differences in HRQL are potentially modifiable; moreover, our results suggest that the effect of the intervention would be greater if it is targeted at the younger strata within the population aged 60 years and over. Earlier publications have examined the contribution of sociodemographic variables to differences in HRQL between sexes. Like us, Arber and Cooper (1999) did not observe differences in health between women and men aged 60 years and over in terms of marital status, last held occupation and income level. On the other hand, HRQL is worse in older adults with a lower educational

ARTICLE IN PRESS P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

1233

Table 1 Sociodemographic, lifestyle, social network and health-related variables, in men and women Variables

Men (n ¼ 1492) %

Women (n ¼ 1768) %

P

SOCIODEMOGRAPHIC Age (years) 60–64 65–69 70–74 75–79 80 and over

25.87 23.17 22.54 14.66 13.76

25.23 16.37 21.50 16.18 20.74

o0.0001

Size of town of residence (population) Less than 50.000 50.000 or over Marital status: married Head-of-family status

53.72 46.28 80.81 93.96

52.06 47.94 45.59 26.82

0.34

Educational level No studies Primary school Secondary school University

44.61 36.69 13.21 5.49

56.96 34.48 6.08 2.48

o0.0001

Head of family’s work status Paid employment Unemployed Retired/housewife

33.09 1.86 65.05

13.29 0.87 85.84

o0.0001

Head of family’s last held occupation Self-employed worker Non-manual salaried worker Manual salaried worker

26.03 21.48 52.49

29.86 19.76 50.38

0.048

LIFESTYLE Physical activity Sedentary Occasional Regular

34.69 61.32 3.99

50.09 47.69 2.21

o0.0001

Body mass index (kg/m2) (%) o25 25–29.9 X30

19.22 48.83 31.96

18.23 39.17 42.59

o0.0001

Tobacco consumption Never smoked Ex-smoker Current smoker

28.73 50.06 21.21

94.57 3.42 2.01

o0.0001

Alcohol consumption Never drinker Ex-drinker Moderate drinker Heavy drinker Hypertension

24.71 18.91 37.66 18.71 66.20

70.30 6.17 20.74 2.74 69.66

o0.0001

SOCIAL NETWORK Living alone Seeing family members daily or almost daily Seeing friends or neighbors daily or almost daily

8.62 51.83 87.62

25.30 55.74 82.75

o0.0001 0.0356 0.0001

HEALTH-RELATED Medical drug consumption 0 1–2 3 or over Medical consultation 4=1 per month

15.96 42.27 41.77 37.05

12.23 35.51 52.27 42.32

o0.0001

o0.0001 o0.0001

0.035

0.003

ARTICLE IN PRESS 1234

P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

Table 1 (continued ) Variables

Men (n ¼ 1492) %

Women (n ¼ 1768) %

P

Medical home visit 4=1 per year Admission to hospital last year Number of chronic diseases 0 1 2 3 or over

9.73 18.29

15.59 17.93

o0.0001 0.7964 o0.0001

11.77 32.66 33.49 22.08

7.08 26.18 36.67 30.06

Table 2 Percentage of subjects (95% confidence limits) with maximum score (100) on each SF-36 questionnaire scale, in men and women. Crude odds ratios of suboptimal health (scoreo100) among women versus men, on each SF-36 questionnaire scale. Men (n ¼ 1492)

Physical functioning (PF) Role-physical (RP) Body pain (BP) General health (GH) Vitality (VT) Social functioning (SF) Role-emotional (RE) Mental health (MH)

Women (n ¼ 1768)

%

95%CL

%

95%CL

15.15 75.33 46.05 7.92 9.18 61.26 87.53 12.87

13.39–17.09 73.05–77.49 43.50–48.61 6.61–9.42 7.79–10.79 58.73–63.73 85.72–89.14 11.23–14.70

7.36 60.25 28.99 3.84 3.85 43.66 74.23 4.07

6.21–8.70 57.92–62.53 26.89–31.18 3.02–4.88 3.02–4.88 41.33–46.01 72.11–76.25 3.22–5.13

Odds ratio

2.26 2.01 2.09 2.16 2.54 2.04 2.45 3.47

 po0.0001.

level (Regidor et al., 1999). This, coupled with women’s lower educational level, is consistent with our results on the role of educational level modestly explaining the poorer HRQL of women. In our study, lifestyle may account for a part of the differences between women and men on several SF-36 scales. These results receive external support from other studies in which regular physical activity has been associated with a better perception of health among men (Ross & Bird, 1994) while excess weight has been linked with a poorer HRQL, particularly among women (Yan et al., 2004). On the other hand, tobacco smoking and alcohol consumption is much more infrequent among elderly Spanish women than among men, making it unlikely that these variables could explain a substantial proportion of the gender differences in HRQL. However, this observation may not apply to future generations of older adults, as consumption of both products is rather similar in women and men aged 16–45 years (Ministerio de Sanidad y Consumo, 2003). As regards social support, this could be an important determinant of HRQL in both sexes, though its association with HRQL is stronger among women than men (Denton & Walters, 1999). However, we have found no evidence of an appreciable contribution of social network to differences in HRQL between men and women.

There is evidence of the relationship between healthrelated variables—such as the number of chronic diseases and use of healthcare services—and worse HRQL. Specifically, suffering a greater number of chronic diseases and visiting the doctor more frequently are associated with a worse perception of health among women (Damia´n, Ruigo´mez, Pastor, & Martı´ n-Moreno, 1999), though the association declines with age ( Seculi et al., 2001). Moreover, for any given perceived state of health, women are more likely to seek medical advice than men (Fernandez, Schiaffino, Rajmil, Badia, & Segura, 1999). Consequently, chronic morbidity and use of healthcare services would be expected to explain part of the differences in HRQL between men and women. To our surprise, we did not observe a clear contribution of physical disorders to differences in HRQL between the sexes, nor did other researchers (Linzer et al., 1996). It should be noted that although women took more medications, used services from primary care physicians more frequently, and presented with a higher number of chronic diseases than did men, differences were not large. Future studies should also consider gathering information on the severity of chronic disease besides their number. Certain methodological characteristics of our study must be commented upon. First, the cross-sectional design does not allow for a causal interpretation of the

60

40

40

20

20

0

OR 2.09

OR 2.01

OR 2.16

OR 2.45

OR 2.04

OR 2.54

0

OR 3.47

-20

OR 2.45 OR 1.74

OR 2.17

OR 1.78

-20

OR 3.03

*OR 1.57

OR 1.68 OR 1.46

-40

-40

-60

-60

-80

-80 -100

PF

RP

BP

GH

VT

SF

RE

PF

MH

(a) Adjustment for lifestyle.

60

GH

VT

SF

RE

MH

Adjustment for social network.

40 20

20 OR=2.54

0

0

OR=2.04 OR=1.88

-20

%

%

BP

60

40

OR=1.61 OR=1.89

-40

OR=2.19

OR=2.04

OR=1.95

OR=2.54

OR=2.07

OR=2.37

OR=1.93

OR=3.34

-20

* OR=1.45 -40

* OR=1.19

-60

-60

-80

-80 -100

-100

(c)

RP

(b)

PF

RP

BP

GH

VT

SF

RE

MH

(d)

PF

RP

BP

GH

VT

SF

RE

MH

*p<0.05 for percentage change in odds ratio. PF: physical functioning. RP: role-physical. BP: bodily pain. GH: general health. VT: vitality. SF: social functioning. RE: role-emotional. MH: mental health. OR: Odds ratios adjusted for groups of variables

1235

Fig. 1. Percentage change (95% confidence limits) in the odds ratio of suboptimal health among women versus men on each SF-36 questionnaire scale, after adjustment for groups of variables. The point estimate of the adjusted odds ratio is also included.

ARTICLE IN PRESS

-100

P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

%

OR 2.26

%

Adjustment for sociodemographic variables.

Adjustment for age.

60

1236

Adjustment for health-related variables.

60 40

40

20

20 0

0 OR=1.87

OR=2.26

OR=1.83

OR 2.44

OR=3.20

OR=2.22

OR=1.89

%

OR=1.93

-20

OR=1.88

OR=1.77

-20

OR 2.13

OR=1.61

OR 2.51

OR=1.64

-40

-40

-60

-60

-80

-80

* OR 1.28 * OR=0.94

PF

RP

BP

GH

VT

SF

RE

PF

MH

(e)

RP

BP

GH

VT

SF

RE

MH

(f ) Adjustment for all types of variables**.

60 40 20

%

0

OR=2.37 OR=1.77

-20

OR=1.95

* OR=1.51

OR=1.55

OR=2.39

* OR=1.21

-40 -60

*OR=0.85

-80 -100 PF

RP

BP

GH

VT

SF

RE

MH

(g)

*p<0.05 for percentage change in odds ratio.**Variables in table 1.PF: physical functioning. RP: role-physical. BP: bodily pain. GH: general health. VT: vitality. SF: social functioning. RE: role-emotional. MH: mental health. OR: Odds ratios adjusted for groups of variables Fig. 1. (Continued)

ARTICLE IN PRESS

-100

-100

P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

%

Adjustment for sociodemographic variables and lifestyles.

60

ARTICLE IN PRESS P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

1237

Persons aged 60-69 years 280 240 200 160

%

120 80 40 OR 2.45

0

OR 1.85

OR 1. 76

-40

* OR 0.74

* OR 1. 35

* OR 0 .9 5

* OR 0. 93

* OR 1.43

-80 -120 -160 PF

RP

BP

GH

VT

SF

RE

MH

Persons aged 70-79 years 280 240 200 160

%

120 80 40

OR 3.5 7 OR 2.93

0

OR 2.0 8

OR 1.69

OR 3 .42

OR 1.73

-40

* OR 1 .2 7

-80

* OR 0.56

-120 -160

PF

RP

BP

GH

VT

SF

RE

MH

Personsaged 80 years and over 280 240 200 160

%

120 80 40

OR 4.88 OR 2.6 8

0

OR 1.8 9

-40

*OR 1.2 5

-80

OR 1.7 OR 1.2 1

OR 0.6

OR 0.5 9

-120 -160

PF

RP

BP

GH

VT

SF

RE

MH

*p<0.05 for percentage change in odds ratio. PF: physical functioning. RP: role-physical. BP: body pain. GH: general health. VT: vitality. SF: social functioning. RE: role-emotional. MH: mental health. OR: Odds ratios adjusted for groups of variables. Fig. 2. Percentage change (95% confidence limit) in the odds ratio of suboptimal health among women versus men on each SF-36 questionnaire scale, after adjustment for all variables in Table 1, by age groups. The point estimate of the adjusted odds ratio is also included.

ARTICLE IN PRESS 1238

P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

results linking sociodemographic and lifestyle variables to differences in HRQL between women and men. In addition, this design does not reveal whether the observed reduction with age in the contribution of the study factors to gender differences in HRQL is a consequence of age itself or of operating cohort-effects. Second, our study, like the great majority of population-based studies on older adults, does not include institutionalized subjects. Since women are widowed more frequently than men, and since widowhood is an important cause of institutionalization due to the lack of a caregiver (e.g., a spouse) capable of attending to personal needs (Arber & Cooper, 1999), it is likely that there may be a higher degree of institutionalization among women than among men in a relatively good state of health (Damia´n Moreno, 2002). It is therefore feasible that the study may have somewhat overestimated women’s poorer HRQL versus that of men. It can be argued that the factors responsible for the differences in HRQL between women and men vary between the institutionalized and non-institutionalized populations, because gender roles may operate with a lower intensity among institutionalized persons. Were this to be true, the argument would nevertheless have a very limited effect on sociodemographic factors (educational level and head-of-family status) and lifestyle factors (sedentary nature and obesity), precisely those that in our study have displayed a certain contribution to HRQL differences between sexes. Third, data are self-reported. This is the appropriate way of getting information on subjective aspects of health, such as HRQL. It is possible that the concept of health may not be the same in the two sexes, and that differences in health may be due to different ways of perceiving or expressing it (Hibbard & Pope, 1983). However, recent research has compared men’s and women’s answers to a global, commonly used question about chronic illness and to a series of more specific prompts (Macintyre, Ford & Hunt, 1999). This research found: no gender differences in the initial reporting of conditions; that men reported a higher proportion of their conditions in response to the initial global question and, lastly, no evidence that women were more likely to report ‘‘trivial’’ and mental health conditions in response to the initial question. This supports the view that the women’s poorer HRQL observed in our study is, in good extent, real. Furthermore, there is evidence of the reliability of the reporting of lifestyles (Bowling et al., 1993; Colditz et al., 1986;), chronic diseases (Bush, Miller, Golden & Hale, 1989; Harlow & Linet, 1989), and use of healthcare services in the time period covered by our questions (Roberts, Bergstralh, Schmidt & Jacobsen, 1996). Last, despite the great number of variables included in our study, no consideration has been given to some factors that are specifically linked to social structure and

the role of each sex, and that might partly explain gender-related differences in state of health (Annandale & Hunt, 2001; Arber & Khlat, 2002; Rohlfs et al., 2000). Among these are paid and non-paid work, duration of the work-day (full versus half-day), family demands or burden, possession of material resources, and income level. Similarly no information was included on mental disorders, which are more frequent among women and which may thus account for their poorer HRQL (Linzer et al., 1996). Hence, future studies should include such variables, in addition to those included in our study, to explain a greater proportion of the differences in HRQL between women and men.

Acknowledgements This work was funded, in part, by a grant (Dossier 24/02) from the Instituto de la Mujer, Ministerio de Trabajo y Asuntos Sociales and by ISCIII (Red de Centros RCESP C03/09). During this study, Esther Lopez-Garcı´ a was the recipient of a Fulbright fellowship from Secretarı´a de Estado de Educacio´n y Universidades, Ministerio de Educacio´n y Cultura de Espan˜a, y el Fondo Social Europeo.

References Alonso, J., Prieto, L., & Anto´, J. (1995). La versio´n espan˜ola del SF-36 Health Survey (Cuestionario de Salud SF-36): un instrumento para la medida de los resultados clı´ nicos. Medicina Clı´nica, 104, 771–776. Alonso, J., Regidor, E., Barrio, G., Prieto, L., Rodrı´ guez, C., & De la Fuente, L. (1998). Valores poblacionales de referencia del Cuestionario de Salud SF-36. Medicina Clı´nica, 111, 410–416. Annandale, E., & Hunt, K. (2001). Gender inequalities in health, Buckingham: Open University Press. Arber, S., & Cooper, H. (1999). Gender differences in health in later life: the new paradox? Social Science & Medicine, 48, 61–76. Arber, S., & Ginn, J. (1993). Gender and inequalities in health in later life. Social Science & Medicine, 36, 33–46. Arber, S., & Khlat, M. (2002). Introduction to ‘social and economic patterning of women’s health in a changing world’. Social Science & Medicine, 54, 643–647. Azpiazu, M., Cruz, A., Villagrasa, J., Abanades, J., Garcı´ a, N., & A´lvarez de Mon, C. (2003). Calidad de vida en mayores de 65 an˜os no institucionalizados de dos a´reas sanitarias de Madrid. Atencio´n Primaria, 32, 285–294. Banegas, J. R., Rodrı´ guez-Artalejo, F., Ruilope, L. M., Graciani, M. A., Luque, M., & de la Cruz-Troca, J. J., et al. (2002). Hypertension magnitude and management in the elderly population of Spain. Journal of Hypertension, 20, 2157–2164. Bowling, S. J., Morrill, B. D., Nafziger, A. N., Jenkins, P. L., Lewis, C., & Pearson, T. P. (1993). Validity of cardiovascular

ARTICLE IN PRESS P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240 disease risk factors assessed by telephone survey: the behavioral risk factor survey. Journal of Clinical Epidemiology, 46, 561–571. Bush, T. L., Miller, S. R., Golden, A. L., & Hale, W. E. (1989). Self-report and medical record report agreement of selected medical conditions in the elderly. American Journal of Public Health, 79, 1554–1556. Colditz, G. A., Martin, P., Stampfer, M. J., Willett, W. C., Sampson, L., & Rosner, B. etal. (1986). Validation of questionnaire information on risk factors and disease outcomes in a prospective cohort study of women. American Journal of Epidemiology, 123, 894–900. Conelli, J. E., Philbrick, J. T., Smith, G. R., Kaiser, D. L., & Wymer, A. (1989). Health perceptions of primary care patients and the influence on health care utilization. Medical Care, 27, S99–S109. Dahl, E., & Birkelund, G. E. (1997). Health inequalities in later life in a social democratic welfare state. Social Science & Medicine, 44, 871–881. Damia´n Moreno, F.J. (2002). The health of the older adults in Madrid. Doctoral Dissertation, Universidad Auto´noma de Madrid, Madrid. Damia´n, J., Ruigo´mez, A., Pastor, V., & Martı´ n-Moreno, J. M. (1999). Determinants of self assessed health among Spanish older people living at home. Journal of Epidemiology and Community Health, 53, 412–416. Denton, M., & Walters, V. (1999). Gender differences in structural and behavioral determinants of health: an analysis of the social production of health. Social Science & Medicine, 48, 1221–1235. Fernandez, E., Schiaffino, A., Rajmil, L., Badia, X., & Segura, A. (1999). Gender inequalities in health and health care services use in Catalonia (Spain). Journal of Epidemiology and Community Health, 53, 218–222. Ferrer, M., & Alonso, J. (1998). The use of the Short Form (SF)-36 questionnaire for older adults. Age and Ageing, 27, 755–756. Harlow, S. S. D., & Linet, M. S. (1989). Agreement between questionnaire data and medical records. The evidence for accuracy of recall. American Journal of Epidemiology, 129, 233–248. Hibbard, J. H., & Pope, C. R. (1983). Gender roles, illness orientation and use of medical services. Social Science & Medicine, 17, 129–137. Hopman, W. M., Towheed, T., Anastassiades, T., Tenenhouse, A., Poliquin, S., & Berger, C., et al. (2000). Canadian normative data for the SF-36 health survey. Canadian Multicentre Osteoporosis Study Research Group. Canadian Medical Association Journal, 163, 265–271. Hunt, K. (2002). A generation apart? Gender-related experiences and health in women in early and late mid-life. Social Science & Medicine, 54, 663–676. Linzer, M., Spitzer, R., Kroenke, K., Williams, J. B., Hahn, S., & Brody, D., et al. (1996). Gender, quality of life, and mental disorders in primary care: results from the PRIMEMD 1000 study. American Journal of Medicine, 101, 526–533. Loge, J. H., & Kaasa, S. (1998). Short form 36 (SF-36) health survey: normative data from the general Norwegian population. Scandinavian Journal of Social Medicine, 26, 250–258.

1239

Lo´pez Garcı´ a, E., Banegas Banegas, J. R., Gutie´rrez-Fisac, J. L., Graciani Pe´rez-Regadera, A., Dı´ ez Gan˜a´n, L., & Rodrı´ guez Artalejo, F. (2003a). The relationship between body weight and health-related quality of life among the elderly in Spain. International Journal of Obesity and Related Metabolic Disorders, 27, 701–709. Lo´pez Garcı´ a, E., Banegas, J. R., Graciani Pe´rez-Regadera, A., Gutie´rrez-Fisac, J. L., Alonso, J., & Rodrı´ guez Artalejo, F. (2003b). Valores de referencia de la versio´n espan˜ola del Cuestionario de Salud SF-36 en poblacio´n adulta de ma´s de 60 an˜os. Medicina Clinica, 120, 568–573. Macintyre, S., Ford, G., & Hunt, K. (1999). Do women ‘overreport’ morbidity? Men’s and women’s responses to structured prompting on a standard question on long standing illness. Social Science & Medicine, 48, 89–98. Macintyre, S., Hunt, K., & Sweeting, H. (1996). Gender differences in health: are things really as simple as they seem. Social Science & Medicine, 42, 617–624. Medical Outcomes Trust (1996). Scoring the SF-36 Health Survey. Spanish version, Boston, MA: Medical Outcomes Trust. Mele´ndez Herna´ndez, M., Montero Herrero, R., Jime´nez Sa´nchez, C., & Blanco Montagut, L. E. (2001). Autopercepcio´n de salud en ancianos no institucionalizados. Aten Primaria, 28, 32–91. Ministerio de Sanidad y Consumo (2003). Encuesta Nacional de Salud de Espan˜a, 2001, Madrid: Ministerio de Sanidad y Consumo. Regidor, E., Barrio, G., De La Fuente, L., Domingo, A., Rodriguez, C., & Alonso, J. (1999). Association between educational level and health related quality of life in Spanish adults. Journal of Epidemiology and Community Health, 53, 75–82. Ries, A. L., Kaplan, R. M., Limbreg, T. M., & Prewitt, L. M. (1995). Effects of pulmonary rehabilitation on physiologic and psychosocial outcomes in patients with chronic obstructive pulmonary disease. Annals of Internal Medicine, 122, 823–832. Roberts, R. O., Bergstralh, E., Schmidt, L., & Jacobsen, S. J. (1996). Comparison of self-reported and medical record health care utilization measures. Journal of Clinical Epidemiology, 49, 989–995. Rohlfs, I., Borrell, C., & Fonseca, M. C. (2000). Ge´nero, desigualdades y salud pu´blica: conocimientos y desconocimientos. Gaceta Sanitaria, 14(Suppl 3), 60–71. Ross, C., & Bird, C. (1994). Sex stratification and health lifestyle: consequences for men’s and women’s perceived health. Journal of Health and Social Behavior, 35, 161–178. Ruiz, M. T., & Verbrugge, L. M. (1997). A two way view of gender bias in medicine. Journal of Epidemiology and Community Health, 51, 106–109. SAS/STAT package version 8.02. (2001). SAS/STATguide for personal computers, version 8.2. Cary, NC: SAS Institute. Scott, K. M., Tobias, M. I., Sarfati, D., & Haslett, S. J. (1999). SF-36 health survey reliability, validity and norms for New Zealand. Australian and New Zealand Journal of Public Health, 23, 401–406. Seculi, E., Fuste, J., Brugulat, P., Junca, S., Rue´, M., & Guille´n, M. (2001). Percepcio´n del estado de salud en varones y

ARTICLE IN PRESS 1240

P. Guallar-Castillo´n et al. / Social Science & Medicine 60 (2005) 1229–1240

mujeres en las u´ltimas etapas de la vida. Gaceta Sanitaria, 15, 217–223. Siu, A. L., Reuben, D. B., Ouslander, J. B., & Osterweil, D. (1993). Using multidimensional health measures in older persons to identify the risk of hospitalization and skilled nursing placement. Quality of Life Research, 2, 253–261. Sullivan, M., & Karlsson, J. (1998). The Swedish SF-36 Health Survey III. Evaluation of criterion-based validity: results from normative population. Journal of Clinical Epidemiology, 51, 1105–1113.

Walters, S. J., Munro, J. F., & Brazier, J. E. (2001). Using the SF-36 with older adults: a cross-sectional community-based survey. Age and Ageing, 30, 337–343. Wiggins, R. D., Joshi, H., Bartley, M., Gleave, S., Lynch, K., & Cullis, A. (2002). Place and personal circumstances in a multilevel analysis of women’s long-term illness. Social Science & Medicine, 54, 827–838. Yan, L. L., Daviglus, M. L., Liu, K., Pirzada, A., Garside, D. B., Schiffer, L., Dyer, A. R., & Greenland, P. (2004). BMI and health-related quality of life in adults 65 years and older. Obesity Research, 12, 69–76.