28 Health, Behavior, and Aging Ilene C Siegler Lori A Bastian Hayden B Bosworth Duke University Health psychology has always been sensitive to age as an important construct because the distributions of diseases by age are not random and are important in determining the psychological impact of different diseases Epidemiology, on the other hand, studies age as a prominent risk factor for disease Both are important in understanding the set of associations in health, behavior, and aging The psychology of adult development and aging looks at persons aging normally, some with and some without specific diseases, to examine the ways disease influences the aging process Health psychology studies individuals with specific physical illnesses and seeks to understand how the aging process might modify the impact of that disease (Siegler & Vitaliano, 1998) The psychology of adult development and aging and health psychology are two subdisciplines of psychology that have multidisciplinary partners The multidisciplinary aspects of studying aging are part of gerontology and limited to studying primarily the elderly, whereas the medical aspects of aging are studied as a postgraduate branch of medicine called geriatrics (Hazzard, Bierman, Blass, Ettinger, & Halter, 1994; Maddox et al., 1995) Behavioral medicine is a multidisciplinary approach to understanding problems in health psychology that interact with the same problems in psychosomatic medicine (Blechman & Brownell, 1998; Matthews, in press) Handbooks are common in the psychology of adult development and aging In each of the Handbooks there has been a “health psychology” chapter (Deeg, Kardaun, & Fozard, 1996; Eisdorfer & Wilkie, 1977; M F Elias, J W Elias, & P K Elias, 1990; Siegler & Costa, 1985) As a group, they provide excellent reviews of the relevant literature that need not be repeated here As part of a set of master lectures on adult development and aging, Siegler (1989) was given the “health psychology” assignment and tried to conceptualize the intersection of health, behavior, and aging as developmental health psychology This chapter reflects an updating of that initial effort and focuses on emergent findings in the past 10 years and aims to be illustrative rather than exhaustive Understanding the issues in health, behavior, and aging first requires discussing what has been a central question in the field: What is normal aging? Second, some important methodological ideas are reviewed from the psychology of adult development that will be useful in health psychology The chapter then takes up the issue of a life-span developmental versus “phase” theory view of a developmental health psychology It then considers some new data and thinking about women's and men's health in middle and later life, and reviews the findings from some recent empirical studies that show the excellent results from the synergy of developmental and health psychology WHAT IS NORMAL AGING AND HOW IS IT DIFFERENT FROM DISEASE? This question drove the initial longitudinal studies of normal aging, such as the Duke Longitudinal Study (Busse et al., 1985) Shock's (see Shock et al., 1984) initial observations were essentially correct, that some, albeit rare, individuals -469- age without the typical declines (as was shown by a careful testing of participants in the Baltimore Longitudinal Study of Aging, BLSA) and normal aging itself is a relatively benign set of processes (Williams, 1994) This is not to say that older persons not have health problems or that the probability of health problems does not increase with age But rather, when they do, they can be attributed to a particular disease process rather than just the passage of time The resulting problems then could be considered the fault of a chronic disease process that is not rapidly fatal, but treated, and remains as a companion for the rest of life Thus, it seems “unfair” to blame aging It used to be that normal aging was accepted and disease treated However, current cohorts raised on such slogans as “Better living through chemistry” are trying to treat normal aging as well A table taken from a recent NIA publication summarizes what the Baltimore Longitudinal Study of Aging (BLSA) teaches about normal aging (NIA, 1996) Definitions of normal aging are a moving target and change as a function of individual's risk modification behavior and effective treatments Thus, Table 28.1 is true today; but future research may point out changes for future cohorts As new cohorts age, they may well be different from current middleaged and elderly persons As is discussed later, psychology of aging has spent considerable time understanding these shifting patterns of aging METHODOLOGICAL CONCERNS FROM PSYCHOLOGY OF AGING Psychology of aging has made major contributions in the explication of the meanings of age, period, and cohort and important factors in understanding development (Baltes, 1968; Schaie, 1965; Schaie & Herzog, 1985) Here, the focus is on the implications of developmental designs for health psychology Period/Time Effects The definition of a period, or time effect, is a societal or cultural change that may occur between two measurements that present plausible alternative rival explanations for the outcome of a study (Baltes, Reese, & Nesselroade, 1988; Schaie, 1977) In order to specifically describe a period, or time effect, in health and disease, these effects are described in terms of new diagnostic tools or medical therapies that present plausible alternative rival explanations for the outcome of a study The introduction of the prostate-specific antigen (PSA) test in 1987 is an example of the effects of period/time effects accounting for age-related changes in detecting prostate cancer With the increased use of the PSA test as a diagnostic tool for prostate cancer, there is now an increasing number of prostate cancers being diagnosed among older adults that would not have been diagnosed based on previous diagnostic techniques (Amling et al., 1998) Because prostate cancer prevalence rates increase with age, researchers studying longitudinally the relation between age and onset of prostate cancer would have to account for the introduction of this relatively new diagnostic tool Information is becoming made available quicker and improvements in diagnostic techniques and treatments are increasing in frequency This is a benefit for the population, but it makes research more difficult, particularly if an intervention is in process and new information or changes in medical procedure is being made and available to the public at-large Subsequently, health psychologists are going to have to become more aware and flexible in the way they deal with these increasing period/time effects An example of researchers adapting to historical or period effects has been the ongoing Women's Health Initiative study, in which information on estrogen use is being collected (Matthews et al., 1997) Since the design of this large observational and intervention study in the late 1980s (Roussow et al., 1995), there has been an increased number of data to suggest that the influence of estrogen -470is related to long-term benefits, such as preventing or delaying osteoporosis, heart disease, and Alzheimer's disease (Jacobs & Hillard, 1996) More recently, researchers have identified newer selective estrogen receptor modulators (i.e., Raloxifene and lower doses of conjugated estrogen (0.3 mg) that may be as effective as the more traditiona10–625 mg Premarin but with fewer side effects (Delmas et al., 1997; Bastian, Couchman, Nanda, & Siegler, 1998; Genant et al., 1997) The increase in awareness of the possible association of estrogen and diseases, such as Alzheimer's disease, and the addition of newer therapies may have altered women's perception and experiences with hormonal replacement medication and influence study observations Human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) information is being made readily available, which is quickly influencing treatment outcomes AIDS and aging were once mutually exclusive conditions: The AIDS epidemic began in young adults and they died before they had time to age (Justice & Whalen, 1996) Much has happened since 1982, when the first Centers for Disease Control (CDC) definition of AIDS was published Better therapies, specifically protease inhibitors, have changed HIV infection to be a chronic condition (Flexner, 1998; Hogg et al., 1998) Given the current success of these therapies, more HIV infected persons should be expected across the life span It will be necessary to learn how to adjust management for specific populations, such as the elderly More may be learned about the immune function and aging while treating HIV infection and AIDS in various age groups As researchers move from studying diseases cross- sectionally and move toward examining the disease path and as they continue to apply this acquired information to interventions, they will need to better understand period and time effects and the ways they may influence ongoing studies Is the Age at Disease Onset Important? An area often neglected in the study of aging and disease is whether the development of a particular disease in young adults has the same etiology or recommendations for treatment, as among older adults For example, a disease that is influenced by age of onset is prostate cancer The risk of prostate cancer increases faster with age than any other form of cancer (National Cancer Institute, 1991) After age 50, both mortality and incidence rates from prostate cancer increase almost exponentially Ninety-five percent of cases of prostate cancer are diagnosed in men between age 45 and age 89, with a median age at diagnosis of age 72 (Winkelstein & Ernster, 1979) There is widespread clinical impression among physicians treating prostate cancer that the disease is more virulent and rapidly progressive in younger men (Meikle & Smith, 1990) and as a result treatment options vary based on the age of onset More invasive procedures are used for later stages at earlier age of onset than at comparable stages at later age of onset Depression is another example in which age of onset is an important factor in understanding the etiology and subsequent treatment Depression in the elderly is a serious medical condition that is underdiagnosed and undertreated (NIH, 1992) A common approach to characterizing depression has been to study its risk factors and presentations according to age of first onset This method dichotomizes depression into early-onset depression and late-onset depression, generally using the range from age 50 to 60 as a cutoff point (Steffens, Hays, George, Krishnan, & Blazer, 1996) Late-onset depression is more frequently associated with structural brain changes and cerebrovascular disease, and early-onset depression seems to be more influenced by family and genetic factors (Coffey, Figiel, Djang, & Weiner, 1990; Figiel et al., 1991) Clinically, patients with late-onset depression show more loss of interest, less pathological guilt, more psychosis, and more generalized anxiety (Krishnan, Hays, Tupler, George, & Blazer, 1995) There is also evidence that late-onset depressives may be more refractory to antidepressant treatment than patients with early-onset depression (Hi&e, Scott, Wilhelm, & Brodaty, 1997) and suffer higher mortality rates (Philbert, Richards, Lynch, & Winokur, 1997) Age of Disease Onset by Gender Interaction Not only is age of disease onset an important issue to consider when examining the relation between aging, disease, and behavior, but the consideration of how gender interacts with age of disease onset is just as important This is particularly the case for coronary heart disease because clinical manifestations of coronary heart disease (CHD) occurs among women on average 10 years later than for men, with the occurrence of myocardial infarction being almost 20 years later (Wenger, 1995) An example of the importance of considering gender in terms of age of onset interaction is provided in the Framingham study, which is a 30 year followup examining a number of specific risk factors for CHD by age group and gender Results from the study indicate that the majority of significant associations between risk factors and CHD apparent in younger men and women remain significant in older age groups, but not consistently in both sexes Systolic blood pressure and vital capacity, for example, both demonstrate strong risk associations for CHD in younger men and women and the association increases for older men, but decreases for older women (age 65–94) The effects of diastolic blood pressure are strong risk associations for CHD in younger adults and the effects decrease in older adults, particularly among older women The risk association for serum cholesterol, cigarettes, relative weight decreases the risk of CHD in older adults, whereas blood glucose increases the risk of CHD (Harris, Cook, Kannel, & Goldman, 1988) Another gender interaction is diabetes At all ages, but especially in premenopausal women, diabetes mellitus is a far more powerful risk factor of CHD for women than for men (Barrett-Connor, Cohn, Wingard, & Edelstein, 1991) Adult onset diabetes, in the Nurses' Health Study was associated with a three- to sevenfold increased risk of a cardiovascular event, with this risk amplified by associated risk factors (Manson et al., 199 I) The mechanisms imparting risk among diabetic women are uncertain, but may include lipid abnormalities, hypertension, fibrinogen abnormalities, and the upper body obesity syndrome, all of which are common -471concomitants of diabetes mellitus Further, after age 45, women are twice as likely as men to develop diabetes (Wenger, 1995) Usefulness of a Life Span Perspective In the psychology of adult development and aging, it is almost a matter of catechism that development needs to be understood in a life span perspective Is this true in developmental health psychology as well? Two areas of research that have gotten attention in the past 10 years are the tremendous increase in the numbers of centenarians and the role of hormone replacement therapy (HRT) as an antidote to aging for women Both research areas argue for lifespan development within particular phases, not as a continuous process Centenarians and the role of health and aging Centenarian status is not easy to predict from earlier in the life cycle It sounds trite to say that one must live until 80 or 90 to get to be 100; but it is not at all obvious from a group of very old people, who will be the rare person to get all the way to 100 This not only makes them very hard to locate and to study, but also to describe In studies that require cognitive testing, centenarians are generally seen as expert survivors (Poon, Johnson, & Martin, 1997) When the full population of living centenarians is studied, the variances are extreme in both physical health and cognitive functioning (Forette, 1997) HRT and the Logic of Estrogen-Related Disease Until recently, women's health research has mainly focused on reproduction and cancers unique to women Given that the incidence of chronic diseases in women increase after menopause (a marker of midlife), it is logical to use the terminology pre- and postmenopausal to describe adult women's health The postmenopausal period extends beyond age 50 and can be divided into three additional phases based on the incidence of chronic diseases As shown in Table 28.2, women's health can be described in four phases: Phase 1-premenopause; Phase 2-postmenopause (age 50–64) with the development of diseases such as breast cancer; Phase 3-postmenopause (age 65–79) with the development of diseases such as heart disease; and Phase 4-postmenopause (age 80 and up) with the long-term development of diseases such as osteoporosis and Alzheimer's dementia The terminology “estrogen-related diseases” can be used to organize and describe diseases associated with estrogen Although some of these diseases, such as osteoporosis, heart disease, endometrial cancer, and breast cancer are well established to be associated with estrogen in women (Col et al., 1997; Colditz et al., 1995; Grady et al., 1992; Grady, Gebretsadik, Kerlikowske, Ernster, & Petitti, 1995; Newcomb & Storer, 1995), other diseases like Alzheimer's dementia and colon cancer have been reported to be associated with estrogen but are generally considered more controversial (Kawas et al., 1997; Nanda, Bastian, Hasselblad, & Simal, 1999; Paganini-Hill & Henderson, 1996; Potter, 1995) Estrogen-related diseases can be used as a shorthand to represent the potential HRT may have on the public health of women A much better understanding of how estrogen is related to these and other diseases can be expected because of anticipated results from the Women's Health Initiative (WHI; Rossouw et al., 1995) Reported to be the largest study of women's health in the world, WHI researchers have compiled a battery of psychological measures in addition to the longitudinal assessment of disease incidence as outcome measures (Matthews et al., 1997) Conversely, there is not a clear marker of midlife in men However, there may be a role for discussing phases of diseases in men's health as well Like estrogens, androgen levels decrease with age and have a broad range of effects on sexual organs and metabolic processes Androgen deficiency in men older than age 65 leads to a decrease in muscle mass, osteoporosis, decrease in sexual activity, and changes in mood and cognitive function (Swerdloff & Wang, 1993) leading to speculation that there may be at least two phases of chronic diseases related to androgen levels in men Whether men over age 65 would benefit from androgen replacement therapy is not known Any potential benefits from this therapy would need to be weighed against the possible adverse effects on the prostate and cardiovascular system EMERGENT FINDINGS ON AGING AND HEALTH/DISEASE The literature has matured in this area as investigators have started to ask important questions about how age interacts with other factors to try to look at the potential mechanisms that relate psychosocial factors to disease outcomes Kop (1997) provided an interesting theoretical statement He argued that psychosocial factors, such as hostility and socioeconomic status are important in understanding the risks of heart disease only under age 55 Whether this turns out to be true requires significantly more empirical verification However, it is consistent with findings reported by House et al (1992), suggesting from survey data that the number of chronic conditions -472reported by individuals stratified by SES and age indicated that the average number of conditions for an older person, age 75+, in the upper social class was the same as for a middle-aged person (around age 45) in the lower social class strata.Jennings et al (1997) studied middle-aged men (age 46– 64) from the Kuopio study to ask empirical questions about the role of age, disease (hypertension), and medication on cardiovascular reactivity and found that there are no simple answers All three of the factors have effects of reactivity Thus, the role of age in studies of health psychology needs to be determined paradigm by paradigm, and disease by disease.Kaplan (1992) reported on data from the Alameda county study to ask if risk factor modification undertaken in later life has an impact on mortality-the answer is a definite yes When those over age 70 in the Alameda county study were followed for 17 years, current smoking, physical inactivity, consuming more than 45 drinks per month, being more than 10% underweight or 30% overweight, and having a low social network index were associated with mortality, whereas marital status, race, and SES were not The Role of Health-Related Quality of Life Clinicians and policymakers are recognizing the importance of measuring health-related quality of life to inform patient management and policy decisions, but researchers have been a little slower to examine this outcome An understanding of what determines good health outcomes is highly valued by patients and is necessary in order to maintain function and, therefore, improve health-related quality of life (Stewart et al., 1989) One reason why researchers have chosen not to consider health-related quality of life as an outcome measure is that this concept misleadingly suggests an abstract philosophical approach, whereas most approaches used in the medical contexts not attempt to include more general notions such as life satisfaction or living standards and instead concentrate on aspects of personal experience that might be directly related to health (Fitzpatrick et al., 1992) Nevertheless, despite the proliferation of instruments and the theoretical literature devoted to the measurement of health-related quality of life, no unified approach has been devised for its measurement, and little agreement has been attained on what it means (Bergner, 1989; Gill & Feinstein, 1994; Spilker, Molinek, Johnson, Simpson, & Tilson, 1990).Although, health-related quality of life has not been clearly defined, it is important to begin to consider outcomes other than mortality and morbidity, particularly as life expectancy continues to increase and chronic diseases are becoming more prevalent Health-related quality of life is important for measuring the impact of chronic disease (Patrick & Erickson, 1993) It will be interesting to see if Baby Boomers have the same patterns as earlier generations Physiologic and clinical measures provide information to clinicians, but they often correlate poorly with functional capacity and well- being (Guyatt, Feeny, & Patrick, 1993) For example, in patients with chronic heart and lung disease, exercise capacity in the laboratory is only weakly related to exercise capacity in daily life (Guyatt et al., 1985) Another example of health-related quality of life instruments improving assessment is that these instruments have been shown to be better than conventional rheumatologic measures as predictors of long-term outcomes in rheumatoid arthritis in terms of both morbidity and mortality (Leigh & Fries, 1991; Wolfe & Cathey, 1991).There is evidence of great individual variation in functional status and well-being that is not accounted for by age or disease condition (Sherbourne, Meredith, &Ware, 1992) The field of health psychology needs to consider health- related quality of life because a commonly observed phenomena that two patients with the same clinical criteria often have dramatically different responses For example, two patients, with the same range of motion and even similar ratings of back pain, may have different role function and emotional well-being Although some patients may continue to work without major depression, others may quit their jobs and have major depression Thus, health-related quality of life is often a better index of the impact that health has on functioning than diagnostic or clinical criteria It, however, does not have the same etiologic significance as a verified diagnosis according to standard criteria CONCLUSIONS Aging has taken on a higher profile given the demographic revolution due to increased longevity (Qualls & Abeles, in press) Interest in aging issues is often due to a concern with the health, disease, and disability of the elderly population (Siegler, in press) As the demographic revolution occurs, it is important to note that the number of adults with diseases will continue to increase (prevalence), but because of improvements in diagnostic tools and treatments, will be able to live longer However, despite the fact that there will be increasing numbers of people with chronic diseases, the incidence of certain diseases continues to decline This demographic revolution, however, is spread unevenly around the world (Murray & Lopez, 1996).The writing of this chapter has lead to three conclusions that summarize current thinking on health, behavior, and aging: Gender and age interact in important ways during adult life and aging This is especially true in terms of diseases such as AIDS, coronary artery disease, and cancers In the last decade, gender differences in the etiology and treatment of diseases have been illuminated These differences have provided the impetus to create a women's health medical specialty Examples have been presented in this chapter of gender differences and similarities across the adult life span Several studies have examined the issue of differences in health care use between men and women Landmark studies identified a gender disparity in the diagnosis and treatment of chest pain and CHD (Tobin et al., 1987) This study has led to further studies demonstrating sex differences in the rates of cardiac catheterization and coronary artery bypass surgery (Ayanian & Epstein, 1991) These studies have -473changed the way medical students are taught about evaluating chest pain and have resulted in the reporting of other gender disparities on such topics as renal transplants (Held, Pauly, Bovbjerg, Newmann, & Salvatierra, 1988) and HIV/AIDS (Bastian et al., 1993) The recognition that there is a need to improve the training of clinicians in women's health, has led to the development of a new interdisciplinary specialty in women's health (Wallis, 1992) One consequence of the focus on women's health has been the widening of the definition of health to include social interactions, domestic issues, mental health, and reproductive function (Litt, 1997) Life span development is theory, which at the present time, is not practical given the state of the science Researchers are able to use data at one phase of the life cycle to inform them about developments in adjacent phases and periods approaching 40 years Survival as a quantity of life measure is different than quality of life measures Although there is a growing emphasis to include healthrelated quality of life as an additional outcome measure other than morbidity and mortality for health-related interventions, and that healthrelated quality of life is a useful discriminator among different population segments, and is an important predictor of health and health behaviors, the field still needs to alleviate a number of measurementrelated issues and difficulties The continued efforts to identify those content areas and response dimensions that likely will provide the best discrimination among populations and the greatest sensitivity to change are essential to enable health psychologists to guide future health policy and resources ACKNOWLEDGMENTS This work was supported by grants ROl AG12458 from National Institute on Aging and ROl HL55356 from National Heart Lung and Blood Institute The third author was supported in part by the Department of Veterans Affairs, Veterans Health Administration, HSR&D Service, Program 824 Funds  ... normal aging (NIA, 1996) Definitions of normal aging are a moving target and change as a function of individual's risk modification behavior and effective treatments Thus, Table 28.1 is true today;... three- to sevenfold increased risk of a cardiovascular event, with this risk amplified by associated risk factors (Manson et al., 199 I) The mechanisms imparting risk among diabetic women are uncertain,... among older women The risk association for serum cholesterol, cigarettes, relative weight decreases the risk of CHD in older adults, whereas blood glucose increases the risk of CHD (Harris, Cook,