Gender, psychosocial factors and the use of medical services: a longitudinal analysis

https://doi.org/10.1016/S0277-9536(98)00440-7Get rights and content

Abstract

Many researchers have reported gender differences in levels of reported symptoms, morbidity, mortality and medical care utilization, but the debate continues about the underlying causes of these differences. Some have argued that women use more medical services because they are more sensitive to symptoms and interested in health, while others believe that women's greater service utilization arises from the fact that women experience more morbidities than do men. To date, these questions have not been studied prospectively. Using data from a household interview survey carried out in 1970–1971 and linked to 22 years of health services utilization records, we explored the effects of gender, self-reported health status, mental and physical symptom levels, health knowledge, illness behaviors and health concerns and interest on the long-term use of health services. After controlling for the aforementioned factors, female gender remained an independent predictor of higher utilization over the 22-year period studied, and psychosocial and health factors measured at the initial interview predicted service use even 19–22 years later. Controlling for factors identified as likely causes of gender-related differences in healthcare utilization, gender remains an important predictor of medical care use before and after removing sex-specific utilization. In addition, the consistent predictive ability of attitudinal and behavioral factors, combined with the finding that health knowledge did not predict utilization, indicates that efforts to help patients assess their service needs should target the attitudinal and behavioral factors that vary with gender, rather than health-related knowledge alone.

Introduction

Many researchers note that women use more medical services and report more morbidities than do men and, paradoxically, that they live longer (Nathanson, 1975, Verbrugge, 1985, Verbrugge, 1988; Macintyre, 1996). These findings have generated considerable debate about how gender-based differences in socialization, health knowledge, sensitivity to and reporting of symptoms, and health status affect health services use. Verbrugge (1988) argued that women have higher incidences of less severe acute and chronic conditions than men, while men show excess morbidity for life-threatening conditions. Other researchers, however, suggest that gendered patterns of illness may be more complex than Verbrugge maintained, varying considerably during the lifecourse Mutran and Ferraro, 1988, Macintyre et al., 1996.

Various social psychological reasons for women's greater service use have also been posited. Verbrugge, 1985, Gijsbers van Wijk et al., 1991 suggested that women are more sensitive to symptoms, while Verbrugge argued in addition that women recall symptoms more readily than men and that they may be more willing to report their health problems. Several investigators have begun to test these hypotheses, with mixed results. Macintyre (1993) found that men were more likely than women to overrate symptoms of the common cold compared to ratings made by a trained clinical observer. Marshall and Funch (1986) found that women with symptoms of rectal cancer were more likely than men to delay careseeking and, in research on gender and mental health, Tousignant et al. (1987) reported that women were no more likely than men to confide their mental-health symptoms. In contrast, a study of gender differences in the experience of headache (Celetano et al., 1990) found evidence for gender-based differences in headache physiology as well as in related illness orientations and helpseeking behaviors. Lastly, in a study exploring physical symptoms among military personnel, Bishop (1984) found that women reported having experienced twice as many symptoms as men at the initial retrospective interview, but that no gender differences emerged in prospective symptom diaries of those with comparable duties. In addition, Bishop did find differences in reporting between administrative and combat units, with combat units reporting twice as many symptoms. These results support contentions that social circumstances play a more critical role in symptom reporting than gender per se and the pattern of greater retrospective reporting among women argues for the Verbrugge (1985) assertion that some of the gender differences in reported morbidities result from women's better recall of the symptoms they experience.

Some researchers also speculate that women's more frequent health services use relates to their acceptance of helpseeking, compliance with treatment regimens, and willingness to adopt the sick-role (Verbrugge, 1985). In support of this, women have been shown to be more willing than men to use preventive services, to seek help for medical problems and to adopt illness behaviors Nathanson, 1977, Verbrugge, 1985. Conversely, Gove (1984) has argued that gender differences in morbidity result from differential health status caused by the requirements and strains of women's roles. He posits that such roles increase morbidity because they result in poor mental health and prevent women from caring for themselves when they are ill (via adoption of sick-role behaviors).

Gove's contention is consistent with other researchers who suggest that women's greater experience of illness arises, at least in part, as a function of their social and nurturant roles, generally lower social status and the strains that arise out of these social positions (Woods and Hulka, 1979, Bishop, 1984, Gove, 1984, Verbrugge, 1986, Verbrugge, 1988, Barnett et al., 1991, Kandrack et al., 1991, Anson et al., 1993, Hibbard and Pope, 1993, Popay et al., 1993). Previous examinations of some of the data used in this study support these contentions; they indicate that women's socialization and caregiving responsibilities produce heightened attention to symptoms (Hibbard and Pope, 1987). Such attention may enhance recall, reporting and knowledge about appropriate responses to symptoms and lead to increases in medical service use. Such an interpretation is consistent with the hypothesis that socialization and roles that require caring for and protecting the health of family members are associated with greater interest in health, also previously supported in this sample (Hibbard and Pope, 1987).

To date, no one has prospectively tested the relationships between these psychosocial factors and health service utilization. For this paper, we linked 22 years of health service utilization records (1970–1991) to demographic and psychosocial data from an extensive household interview survey carried out in 1970–1971. Controlling for key demographic variables, we investigated the effects of psychosocial, behavioral, perceptual/attitudinal factors and baseline health status as mediators of gender-based differences in long-term utilization. We included the following in our model: (1) differences in knowledge about appropriate responses to symptoms, (2) tendency to adopt illness behaviors, (3) concerns about health, (4) interest in health, (5) self-reported health status and (6) mental and physical health symptom levels.

Section snippets

Study population

The study population was drawn from the membership of the Northwest Region of Kaiser Permanente (KPNW), the largest nonprofit health maintenance organization in the United States. The Region's membership is generally representative of the northwest Oregon and southwest Washington population from which it is drawn (Greenlick et al., 1988). In 1970–1971, a random sample of 2603 adult members from 1599 subscribing families with at least two years of enrollment participated in an extensive

Symptoms, health status, health concerns, and the sick-role

Table 1 shows analyses of variance testing for gender differences in mental and physical health symptom levels, self-reported health status, health concerns, health interest and illness behaviors. Although most of the mean gender differences we found were small, the majority were statistically significant. Women reported more symptoms than men, although self-reported global health status did not differ. In addition, whereas women did report more interest in health as expected, they did not

Discussion

This study yielded three major sets of findings: (1) baseline self-reported health and symptom levels, as well as psychosocial and behavioral factors, are stable predictors of medical care use over long periods, (2) gender continues to predict utilization after controlling for these key factors, and (3) removing sex-specific utilization from the analyses does not significantly affect the resulting models, although it does reduce gender's predictive ability to some extent.

In our models, the

Directions for future research

The results of this project indicate that unidentified gender-related factors, in addition to those we studied, affect medical care utilization, and that our understanding of the processes through which gender alters healthcare utilization remains incomplete. In addition, although we know there are gender-related patterns of utilization, we still do not understand whether gender directly or indirectly affects the appropriateness of service use. We believe that future research should explore how

Limitations of the study

A limitation of this study results from our comparison of physician and respondent recommendations in creating the RSS. It is possible that difference scores were affected by changes in physician knowledge and/or practice, since physician ratings were completed roughly 25 years after those of respondents. However, it is also true that physician recommendations for most of the conditions we surveyed (e.g. common cold, headaches, skin rashes) have changed little (if at all) over the period of

Acknowledgements

Dr. Green gratefully acknowledges the support of a National Service Research Award from the Agency for Health Care Policy and Research (#T32 HS 00069-01). In addition, the authors are indebted to research assistant Carol Sullivan for her invaluable help retrieving, linking and helping us interpret the medical record data used in this project. We would also like to thank the nine physicians who provided their recommendations for symptom-related responses.

References (27)

  • D.D Celetano et al.

    Gender differences in the experience of headache

    Social Science and Medicine

    (1990)
  • O.D Duncan

    A socioeconomic index for all occupations

  • C.M.T Gijsbers van Wijk et al.

    Symptom sensitivity and sex differences: a review of health surveys in the United States and the Netherlands

    Women and Health

    (1991)
  • Cited by (358)

    View all citing articles on Scopus
    View full text