Daily Illness Characteristics and

Transcription

1 Daily Illness Charateristis and Health Care Deisions of Older People Tom Hikey Hiroko Akiyama University of Mihigan William Rakowski Brown University Although investigations of health are deision making typially deal with patterns of health servie use, inreasing attention has foused on lay- and self-are ations in response to illness symptoms. This study examined the health are ations of a ommunity sample of 142 older adults, who reorded illness symptoms and orresponding health are ations in daily health diaries for a 14-day period. Self-treatment and no-ation deisions were found to be the most frequent response to illness symptoms. Professional-are deisions were assoiated with greater health are need, suh as multiple symptoms and inreased pain. Lay-are deisions were signifiantly related to symptoms of shorter duration. Women were also more likely than men to self-treat their illness symptoms. Results suggest that older people deal with a greater number of reurrent hroni symptoms than previously thought and that they make most treatment deisions without onsulting their dotors or other health are providers. This investigation undersores the importane of a prospetive diary methodology for studying the daily omplexities of hroni illness experienes and for validating and onduting useful interventions. Muh of what is known about how older people deal with illness is based on studies of their interations with the health are system and on analyses of treatment-seeking outomes (Dean, 1986a). In fat, a widely used explanatory model for health behavior ontinues to demonstrate the strong relationship between health need and the use of medial servies (Andersen & Newman, 1973). However, it is inreasingly evident that professional are is only one of many hoies that individuals make in response to illness AUTHORS NOTE: This work was supported in part by Grant No. AG4696 from the National Institute on Aging to the third author. We are grateful to Dr. Mara Julius, University of Mihigan, for the opportunity to ollet these data during the ourse of her ongoing ommunity health survey. Address orrespondene to Tom Hikey, University of Mihigan Shool of Publi Health, 142 Washington Heights, Ann Arbor, MI The Journal of Applied Gerontology, Vol. 1 No. 2, June &opy; 1991 The Southern Gerontologial Soiety 169

2 17 symptoms (Dean, Hikey, & Holstein, 1986). The proess of making deisions about one s health is a very personal one, often beginning with some sort of ative self-are or, simply, the passive self-treatment of taking no ation at all (Ford, 1986). From a historial perspetive, the emphasis on medial treatment is omparatively reent; self-are and informal measures have always been the first, if not the most frequent, reourse of ation in the fae of illness (Dean, 1986b). Although few studies fous on how older adults deal with reurrent hroni symptoms on a day-to-day basis, it is likely that self-are and nonmedial treatment are quite prevalent. Following a lifetime of experiene with their own health and illnesses, older people may have greater onfidene in the effiay of self-are over professional treatment for their hroni illnesses (whih may or may not get better anyway). Their familiarity with symptom patterns and, perhaps, even the fear of losing their &dquo;redibility&dquo; with health professionals and family members over seemingly small omplaints may be more important fators in determining the steps they take to alleviate their symptoms or to redue the probability of needing further treatment (Brody & Kleban, 1981). Traditional theoretial frameworks for understanding health behavior are thought to be less appliable in late life beause of the nature of hroni illness, the potential influene of various psyhosoial fators and life experienes, and the wide array of possible outome behaviors. The most prominent models of health are behavior have foused on health beliefs and pereptions (Rosenstok, 1974), and on predisposing, enabling, and need fators (Andersen & Newman, 1973). For the most part, however, these fators aount for only a small amount of the variane in studies of older adults that have inluded a range of outome measures. In her review, Dean (1984) found that general health beliefs have only a limited influene on self-are behavior and deisions to seek professional treatment. Speifi health beliefs or attitudes regarding personal responsibility or ontrol over health are more likely to have a greater influene on health are deisions - espeially in the area of self-are (Dean, 1989). Suh beliefs, however, often interat with long-standing pereptions of the harateristis of the illness threat, or with beliefs about the potential effiay of treatment or the likely results of ignoring the symptoms. In fat, when variables related to illness harateristis or experienes are introdued, they have been found to be more important (Mehani, 1979). Thus many studies suggest the importane of symptom-related fators, suh as seriousness and likelihood of reurrene, and expeted treatment outomes in determining daily health are deisions (e.g., Berkanovi, Telesky, & Reeder, 1981; Tanner, Cokerham, & Spaeth, 1983).

3 171 Using a different approah Holtzman, Akiyama, and Maxwell (1986) reported that older people are strongly influened by their pereptions of treatment effiay. They ompared older persons beliefs about the most appropriate ways to treat ommon symptoms with their atual responses to the same symptoms. They found that pereptions of seriousness did not orrespond to personal health behavior. Older persons regarded many more symptoms as serious and requiring professional treatment than their own personal behavior refleted. They onluded that self-are deisions were made on the basis of their potential for effetiveness, if not the pereption that professional treatment was unlikely to be any more effetive. The study reported here foused on the relationship of professionaltreatment deisions and lay-are deisions to speifi beliefs about illness and treatment. However, in addition to looking at how illness symptoms influene daily health are ations, we were interested in the extent to whih older patients are involved diretly in their own are on a day-to-day basis. Health are deisions were examined in a ommunity sample of older persons who reorded their reations to symptoms on a daily basis over a 2-week period. We were partiularly interested in three issues: (a) how relatively healthy adults deal with the daily reurrene of hroni illness symptoms; (b) what illness harateristis and personal fators might affet the type of health are deisions they make; and () the usefulness of a daily health diary for assessing symptom experienes and health are deisions in an older population. This artile extends an earlier summary of the overall study (Rakowski, Julius, Hikey, Verbrugge, & Halter, 1988) by reporting more speifially about how older people respond to illness symptoms on a daily basis and how their health are deisions are influened by gender and age, health-related attitudes, and by the harateristis of their illness symptoms. Unlike most other studies in this area, whih have relied on retrospetive interview data, our investigation made use of a prospetive diary methodology to inrease the likelihood that respondents would reord most of their daily illness symptoms and health are ations as they ourred. It has been suggested that health diaries are a more effiient way of olleting an abundane of data about hroni illness episodes (Verbrugge, 198). For older persons likely to experiene simultaneous illness problems, health diaries not only enhane the potential for apturing more relevant data, but also make it easier to trak the relationship between multiple illnesses and various health are ations and deisions. Beause older people are often relutant to share information about their illness symptoms with others (e.g., Brody & Kleban, 1981), the diary method represents a less threatening way to ollet suh information.

4 172 Methods Partiipants and Proedure Symptom experienes and health are deisions were based on daily health diaries maintained for a 2-week period by 142 older adults (82 women and 6 men) between the ages of 62 and 94. Refleting the population omposition of their Detroit suburban ommunity, they were predominantly Jewish (45%) and almost exlusively White. Partiipants were part of an original random sample of 243 noninstitutionalized older adults who agreed to partiipate in the third wave of a longitudinal epidemiologi survey. Of the original group, 18 had died, 22 ould not be loated, and 31 delined to be interviewed, resulting in interviews with 172 (7.8%) of the original sample. Of this group, 2 delined to partiipate in the health diary part of the study, and 1 who agreed, failed to return the diaries. Although those who returned the diaries (82.6% of those interviewed) tended to be somewhat younger than those who delined to partiipate, there were no differenes in other basi demographi and personal harateristis, or in health status and number of reported illnesses. Respondents maintained a daily health log that ontained a list of 36 possible symptoms and a orresponding list of 22 health ations or responses to illness symptoms (f. Rakowski et al., 1988, p. 282). Symptoms were numbered so that they ould be learly linked to health are ations/responses. The symptom list was organized by organ systems, although not ategorized as suh in the diary itself. The list of health ations inluded several possible responses within four general approahes to treatment (using mediations, seeking professional are, self-initiated responses of an informal nature, and taking no ation). Both lists were based on standard measures used in other health interview surveys, as well as on earlier experienes with the diary method (Verbrugge, 198). Partiipants reorded their illness symptoms eah day, indiating the speifi ations they took in response to eah symptom by writing the symptom number next to the ation taken. To maximize partiipation for 14 onseutive days, the response format was designed to be ompleted as easily and quikly as possible. Partiipants were asked to indiate only those ations related to symptoms, that is, any other daily health are or health maintenane ativities, inluding the use of mediations on a regular basis, were not reorded unless in diret response to a speifi symptom. Also, in those few ases where more than one response was reorded, the more ative or formal deision (e.g., &dquo;sheduled appointment&dquo;) was used on the basis that it better

5 173 refleted the endpoint of a deision proess. Prior to using the health diary, it was pretested with a separate group of elderly persons to ensure that the symptom/ation lists were omprehensive and relatively easy to omplete (see Rakowski et al., 1988, for a full desription of the diary). Partiipants were introdued to the health diary during the ourse of a ommunity health survey, at whih time basi information was olleted about their health status. The purpose of the diary was explained following the health interview, and interested respondents were given an opportunity to omplete a sample day with supervision. A follow-up telephone all was made after 1 week to determine whether there were any problems in ompleting the diaries; after a seond week, partiipants returned the diaries by mail. Respondents were reontated by telephone in the event of inomplete or unlear entries. Dependent variables. Two illness behavior indies were onstruted by ombining various symptom responses heked in the diaries. Although all responses were self-initiated and, therefore, a form of self-are, response ategories were worded to indiate learly the hoie between &dquo;medial help&dquo; and &dquo;on my own.&dquo; Informal lay are inluded self/lay-are ations (e.g., &dquo;stayed in bed,&dquo; &dquo;hanged diet,&dquo; &dquo;ut down on ativities,&dquo; &dquo;talked with someone for advie,&dquo; and so on) and the taking of nonpresribed or over-theounter mediations. Formal professional are resulted from ombining various ations related to seeking professional treatment (e.g., &dquo;alled for advie from physiian/nurse/dentist&dquo; &dquo;went to emergeny room,&dquo; &dquo;visited medial/dental offie,&dquo; &dquo;sheduled appointment,&dquo; and so on) and using presription mediations. These two indies of health are deision making were the major outome measures for the analyses we report. The index of eah of the two types of illness behavior indiated the ratio of the number of ation responses in a ertain behavior ategory to the total number of ation responses an individual reported during the 2-week period. It was possible for ratio sores to range from % to 1% beause a few respondents reported no ations in response to symptoms, and others who experiened only a few illness symptoms in the 14 days, may have employed only a single type of health are ation in response to all of their symptoms. Thus an individual who reported only a few symptoms during the 14-day period and who took only self-are ations in response to them, had a ratio sore of % for professional are and 1% for self-are ations. Independent variables. The health interviews, onduted before the partiipants ompleted the diaries, provided additional information about personal harateristis and life outlook, and urrent health status and health

6 174 attitudes. Age and gender are the only demographi variables reported here, based on earlier analyses whih found other demographi variables to be less important (Rakowski et al., 1988). Life outlook was based on three different measures: the Philadelphia Geriatri Center Morale Sale, a future-orientation sale drawn from the work of William Rakowski and a Cantril-ladder rating of urrent quality of life. Health status was measured by a single question: &dquo;overall, how healthy would you say you are now?&dquo; Health attitudes inluded lous ofhealth ontrol, whih ontained seven items worded to reflet personal versus other ontrol over health (alpha =.78) ; pereived interferene of illness with daily life, whih inluded three items indiating resistane to letting illness interfere with daily ativities (alpha.54); and = onern or sensitivity about one s health, also assessed with three items measuring the degree of onern that the respondent felt about urrent health status (alpha.44). This = measure resulted from a fator analysis of 17 items drawn from the Rand Health Insurane Study and various studies of the Health Belief Model. A more detailed desription of the measures and their reliability an be found in an earlier artile (Rakowski, Julius, Hikey, & Halter, 1987). In addition to these variables, four dihotomous indiators assessed illness harateristis or the pereived need to take ation in response to symptoms. Symptom days with pain differentiated those who reported pain along with illness symptoms from those who reported no pain on symptom days. Average number of symptoms distinguished single-symptom days from days on whih multiple symptoms were experiened. Symptom duration indiated the average length of symptom episodes based on the number of onseutive days on whih the same symptom was reported; this variable was divided into &dquo;fewer than 3 days&dquo; and &dquo;3 days or longer,&dquo; on the basis of suggestions from liniians regarding how long people are likely to self-treat illness symptoms. Health now was based on a self-reported above average or good health (good) versus average or less than average health (poor). Results The 1 most frequently reported illness symptoms were identified initially, followed by an analysis of symptom patterns and how the respondents dealt with their illness episodes on a daily basis. We then analyzed the relationship of the type of health are deision with the personal and illness harateristis of the respondents using both bivariate and multivariate methods. By holding &dquo;all else equal,&dquo; the multiple regression analysis was intended to demonstrate the relative strength of the various individual pre-

7 175 ditors. On the other hand, beause our intent was to ompare professionalare ations with self/lay-are ations in a way useful for pratitioners, the bivariate approah provided an opportunity to identify all possible influenes on these two health are behavior outomes without ruling out anything. Symptom frequeny. Altogether, 128 respondents reported 696 illness episodes enompassing 2,91 daily symptoms and ations during the 2-week period. Illness episodes were defined in terms of onseutive days on whih the same symptom was reported. Of the 142 respondents who ompleted the diaries, 14 reported no symptoms during that period. The respondents reported a total of 582 (2%) ations based on professional-treatment responses and 1,28 (44%) ations based on self/lay-are ations in response to the symptoms. No ations were taken for 1,48 (36%) symptoms reported. These figures show that, on a day-to-day basis, the respondents took twie as many ations based on lay-are deisions as they did on professional-are deisions. In Table 1 we present the 1 most frequently reported symptoms and the different types of ations taken in response to them. The high inidene of musuloskeletal symptoms is onsistent with other reports that arthritisrelated pains represent the most frequent hroni omplaints of the elderly. In general, arthritis-related symptoms and allergy symptoms were treated by both professional and lay are, partiularly by presribed and/or over-theounter mediines. Headahe and ough were also largely treated by mediation. By ontrast, the respondents did not seek either professional are or take mediines for fatigue or lak of energy, one of the more prominent symptoms. The &dquo;no ation&dquo; ategory inluded no ations and responses that did not fit any other ategory. However, beause the number of unlear responses were almost negligible for most symptoms, we should onsider that the perentages in this olumn are more indiative of no ation. Most people did not take any ations when they had ringing in ears, shortness of breath, and pain, weakness or numbness in fae, arm or leg. The issue of whether to inlude &dquo;no ation&dquo; as a form of self-are is no ation in a lear and deliberate somewhat ontroversial. Some have suggested that by taking response to illness symptoms, people are making self-are hoie (Dean, 1989; Haug, Wykle, & Namazi, 1989). However, one ould just as easily argue that no onsious deision is involved in doing &dquo;nothing&dquo; about various illness symptoms, that many people give little thought to their symptoms for various reasons. Beause the first approah tends to obsure the differenes between deliberate ations involving informal and lay are and seeking professional treatment, &dquo;no ation&dquo; has been exluded from the analyses reported in Tables 2 and 3.

8 E ä. E.. mi- U) s a en d te a tm as II) II) a as M as d a II) oe U) M 4 8 as s To II) I as o E ä. EN E E û) 2 T II) m !%.5 a N O t E 7an an E O LLJ... z

9 M.r.r i o a m o m T5 x -C < II) v w II) 16. S<a m o II) II) 1 C4 2 m 9 L6 q v po q v I?- 177

10 m M ) a 2 Il U t T5 w x o u U) dm s < m 6.L.C Q. l i E Q m m 178 To.s ~5 a4 t4 d m 9 ui j V *a XQ v ~o..

11 179 Illness harateristis and responses. In Table 2 we summarize the relationship between illness harateristis and health are deisions. The preditable assoiation was found between greater health needs and medial treatment seeking. More speifially, a greater perentage of symptom responses were professional-are deisions in the presene of pain, multiple symptoms, and/or symptoms of longer duration, and by persons in &dquo;poor self-pereived health.&dquo; Shorter symptom duration was the only illness harateristi related to lay-are deisions. Overall, lay-are ations were still made more frequently for all types of symptoms than were professional-are deisions. However, professional-are ations were more likely to be influened by the harateristis and duration of the symptoms, as well as by overall pereptions of health status. When examining the relationship of demographi harateristis to health are deisions (Table 3), we found that women were more likely than men to respond to their illness symptoms with lay-are ations. There were, however, no signifiant gender differenes in professional-are deisions. Attitudinal fators were based on Likert-type-saled items in whih &dquo;above average&dquo; and &dquo;good&dquo; were onsidered positive, and &dquo;average,&dquo; &dquo;below average,&dquo; and &dquo;poor&dquo; were labeled negative attitudes for purposes of this report. Among the attitudinal harateristis, the level of onern or sensitivity about health and the lous of health ontrol were signifiantly assoiated with professional-are deisions. Less onern or sensitivity about health and a weaker sense of health ontrol appear to be linked to professional-are deisions. Also, those who had poor future outlook tended to respond to illness symptoms by ations based on professional-are deisions. These attitudinal harateristis, however, were not signifiantly related to lay-are deisions. Just as in Table 2, the data in Table 3 are presented in the ontext of a ontrast between professional- and lay-are deisions. Overall, lay-are deisions were more ommon than professional-are deisions. Finally, the illness behavior indies were regressed separately against the 1 1 demographi, attitudinal, and illness harateristis to determine the relative importane of those variables in health are deisions. The multiple regression analyses onfirmed the overall signifiane of these fators in professionalare ations (R2 =.27; p =.4) and self-/lay-are ations (R2 =.2; p =.17). Among the individual preditors in the multiple regression equations, health onern was the only fator signifiantly assoiated with both professional-are (p =.35) and lay-are ations (p =.1). People who are more sensitive or vigilant about their health are more likely to do something in response to illness symptoms. Thus, by ontrolling for the various objetive indiators of illness need (e.g., number/duration of symptoms, pain, and so on), a personal orientation to health behavior emerged as important.

12 18 Disussion These findings are onsistent with Dean (1986b) and others who have reported that self-are and lay treatment are the first, if not the most frequent reourse of ation in the fae of illness. However, this study goes beyond many earlier ones by aggregating the response patterns to illness symptoms based on daily reords of behavior. The results suggest that self-are deisions are far more important than previously thought and should be given more prominene when haraterizing the health are behavior of the elderly and in the onsideration of various health are interventions. For the most part, if a partiular symptom is neither painful nor long lasting, and older people are onfident of their own judgments about their health, they are less treatment deisions. likely to make professional The multivariate analyses provided useful onfirmations of the overall importane of various personal, attitudinal, and illness harateristis in determining responses to illness symptoms on a daily basis. It is likely that there is an interation among these variables that must be onsidered arefully in the design of additional studies. However, at this stage in the development of oneptual models for understanding the health behaviors of older people, it is important not to rule out something that might be potentially useful in explaining how older people respond to illness symptoms. Moreover, from the perspetive of the pratitioner, it is helpful to know all of the fators that motivate older people to take various ations on behalf of their own health. Therefore, the results of the bivariate analyses should be given areful onsideration. Although self-are predominane in the treatment of hroni illness episodes is a somewhat preditable finding, this study provides more empirial validation than what has been available from previous investigations. The health diary aptures onsiderable data about illness episodes and health are ations resulting in a more thorough desription of the daily experienes of illness and their onsequenes. Beause the diary methodology has rarely been used in studies of older people, the study reported here validates this prospetive approah for olleting important information from this age group-espeially when there are problems with reall and a need for more detailed data. An additional value of the diary approah is its potential for providing a substantial data base for treatment intervention and program development. The daily health diaries reviewed here also reveal an interesting piture of the kinds of routine problems dealt with by relatively healthy older people on a day-to-day basis that have not been identified in other studies. One

13 181 again, this is important information for the pratitioner. For example, arthritis pains, weakness, and fatigue are even more frequent omplaints than previously thought, aounting for over 6% of all symptoms reported. The diary method seems to be espeially useful in highlighting the frequeny of these symptoms, suggesting their potential impat on the overall quality of the daily lives of older people. That these data were olleted from a fairly healthy sample of older people should lend even more signifiane to this finding. The health diary data also indiated heavy use of presribed and over-theounter mediations by this group of older people. These data probably underreported mediation usage, beause the diary direted partiipants to reord only those mediations taken in response to speifi symptoms, and not their use of other mediations taken on a daily basis for the prevention of illness symptoms (e.g., antihypertensives). This is onsistent with other studies that suggest that older persons who are not limited finanially or otherwise in their aess to mediations, are more likely to be heavy users (Anderson & Cartwright, 1986; Eve, 1986; Ostrom, Hammarlund, Christensen, Plein, & Kethley, 1985). However, our investigation provided more speifi information about whih symptoms are more likely to be treated with mediations than what has been learned from previous interview studies. For example, the deision to use some type of mediation was made about 5% of the time for symptoms related to joint and musle pains, headahes, oughs, and allergies. Whether partiipants were using reently presribed mediations or merely self-treating with presription mediines they had on hand from earlier illnesses ould not be determined. In addition to fousing on how older adults deal with the daily reurrene of hroni illness symptoms, we were interested in how various harateristis of their illness symptoms might affet the type of health are deisions they make. Consistent with most of the literature, health need fators were again found to be assoiated with professional-treatment deisions (Andersen & Newman, 1973; Berkanovi et al., 1981; Ford, 1986; Tanner et al., 1983; Wolinsky et al., 1983). Greater pain, symptoms of a longer duration and/or a serious nature, and more negative self-pereived health were assoiated with professional-are deisions. Health need fators were muh less important in determining self-are responses to illness, suggesting the influene of other fators, as well as the need for a oneptual framework to aount for multiple, and possible interative, preditors of self-are behavior. In addition to health needs, this study showed that personal onerns about how muh ontrol older persons have over their health and how it affets their future outlook and other aspets of their lives were also assoiated with their professional-are deisions. Although the literature is equivoal regarding the influene of lous of ontrol on health are deisions (Dean, 1989),

14 182 there is some onsisteny in the overall pattern reported here. People with negative health experienes tend to rely more on professional assistane. A lak of self-onfidene in one s ability to deal effetively with illness symptoms ould similarly lead to a greater reliane on professional treatment for symptoms that might be dealt with as effetively with some form of self-treatment. Not surprisingly, symptoms of shorter duration were most often self-treated. Gender was also assoiated with self-treatment deisions. As reported earlier (Rakowski et al., 1988), women were far more likely than men to take a more ative role in their health are on a daily basis. Although there were no gender differenes in professional-are deisions (presumably for more serious illnesses), women are more likely to initiate nonmedial self-are ations. This finding is onsistent with what is known about other patterns of health behavior among men and women (Dean, 1989). Men are more likely to ignore many symptoms until they are serious enough to seek professional are. Women have higher reported morbidity and are more frequent users of health servies; typially, they are also experiened aretakers of sik family members and in a position to know more about the treatment of various symptoms. The traditional soialization of males may result in ignoring minor symptoms and the early stages of illness whereas women are more likely to &dquo;do something&dquo; (Akiyama, Hikey, & Rakowski, 1987). Lay-are deisions were not explained by any other demographi, attitudinal, or need fators. In fat, suh deisions did not appear to follow a preditable pattern other than that nonmedial self-are was the most frequent response to illness - an important finding. Lay-are deisions seem to be typial responses to more routine as well as more serious sysmptoms of illness. The absene of a pattern should not be surprising if lay-are deisions are the result of people doing &dquo;what works best&dquo; in eah situation (Holtzman et al., 1986). As indiated at the outset, no single oneptual model seems appliable to the wide range of health are deisions that older people make in response to daily illness symptoms. Generalized health beliefs are less important in determining self-are and professional treatment deisions than are speifi pereptions about the nature of the symptom, one s previous experiene with it, and the pereived effiay of various treatment options. Conlusion The extensive and reurrent illness symptoms reported by the relatively healthy older population studied here are indiative of the hroni nature of

15 183 their health problems in late life. Studying a larger and more representative ross-setion of the older population should only enhane this basi finding. We also identified important attitudinal and demographi fators, as well as the harateristis of illness symptoms. Suh fators were found to be espeially important influenes on professional treatment deisions. The health diary method provides a useful way to identify the number and type of illness symptoms that tend to our daily and to examine the proess of making treatment deisions. This methodology has the potential for olleting a onsiderable amount of useful information about the daily experienes of hroni illnesses and their impat on the quality of older people s lives. For example, the diary method need not be limited to symptom responses. It ould be used to identify what people do for their overall health on a daily basis, inluding routine preventive praties, health maintenane behavior, and the use of mediations on a regular basis. As suh, it an provide important information for planning health are interventions. Thus the study reported here provides additional insight into the hoies that people make between self-treatment and professional treatment, as well as a better understanding of how to use a prospetive diary methodology for studying hroni illness symptoms. Further researh is needed to advane our understanding of the deisions that older people make about their health are on a day-to-day basis in late life. Faed with the reurrent symptoms of various hroni illnesses, suh deisions are likely to be influened in an interative fashion by the severity and duration of symptoms and pain, the pereived effiay of various professional and lay treatments, and a number of other personal and situational fators. Referenes Akiyama, H., Hikey, T., & Rakowski, W. (1987, Otober). Daily response to illness symptoms among the elderly. Paper presented at the 115th Annual Meeting of the Amerian Publi Health Assoiation, New Orleans, LA. Andersen, R., & Newman, J. F. (1973). Soietal and individual determinants of medial are utilization in the United States. Milbank Memorial Fund Quarterly, 51, Anderson, R., & Cartwright, A. (1986). The use of mediines by older people. In K. Dean, T. Hikey, & B. E. Holstein (Eds.), Self-are and health in old age (pp ). London: Croom Helm. Berkanovi, E., Telesky, C., & Reeder, S. (1981). Strutural and soial psyhologial fators in the deision to seek medial are for symptoms. Medial Care, 24, Brody, E. M., & Kleban, M. H. (1981). Physial and mental health symptoms of older people: Who do they tell? Journal of the Amerian Geriatris Soiety, 29, Dean, K. (1984). The influene of health beliefs on lifestyle: What do we know? European Monograph of Health Eduation Researh, 6,

16 184 Dean, K. (1986a). Self-are behaviour: Impliations for aging. In K. Dean, T. Hikey, & B. E. Holstein (Eds.), Self-are and health in old age (pp ). London: Croom Helm. Dean, K. (1986b). Lay are in illness. Soial Siene and Mediine, 22, Dean, K. (1989). Coneptual, theoretial and methodologial issues in self-are researh. Soial Siene and Mediine, 29, Dean, K., Hikey, T., & Holstein, B. E. (Eds.). (1986). Self-are and health in old age. London: Croom Helm. Eve, S. B. (1986). Self-mediation among older adults in the United States. In K. Dean, T. Hikey, & B. E. Holstein (Eds.), Self-are and health in old age (pp ). London: Croom Helm. Ford, G. (1986). Illness behaviour in the elderly. In K. Dean, T. Hikey, & Self-are and health in old age (pp ). London: Croom Helm. Haug, M. R., Wykle, M. L., & Namazi, K. H. (1989). Self-are among older adults. Soial Siene and Mediine, 29, Holtzman, J. M., Akiyama, H., & Maxwell, A. J. (1986). Symptoms and self-are in old age. Journal of Applied Gerontology, 5, Mehani, D. (1979). Correlates of physiian utilization: Why do major multivariate studies of physiian utilization find trivial psyhosoial and organizational effets? Journal of Health and Soial Behavior, 29, Ostrom, J. R., Hammarlund, E. R., Christensen, D. B., Plein, J. B., & Kethley, A. J. (1985). Mediation usage in an elderly population. Medial Care, 23, Rakowski, W., Julius, M., Hikey, T., & Halter, J. (1987). Correlates of preventive health behavior in late life. Researh on Aging, 9, Rakowski, W., Julius, M., Hikey, T., Verbrugge, L. M., & Halter, J. B. (1988). Daily symptoms and behavioral responses: Results of a health diary with older adults. Medial Care, 26, Rosenstok, I. M. (1974). The Health Belief Model and preventive health behavior. Health Eduation Monographs, 2, Tanner, J. L., Cokerham, W. C., & Spaeth, J. L. (1983). Prediting physiian utilization. Medial Care, 21, Verbrugge, L. M. (198). Health diaries. Medial Care, 18, Wolinsky, F. D., Coe, R. M., Miller, D. K., Prendergast, J. M., Creel, M. J., & Chavez, M. N. (1983). Health servies utilization among the noninstitutionalized elderly. Journal of Health and Soial Behavior, 24, B. E. Holstein (Eds.), Tom Hikey, Dr.P.H., is a professor in the Department of Health Behavior and Health Eduation at the University ofmihigan Shool of Publi Health, and a faulty assoiate at the Institute of Gerontology in Ann Arbor, Mihigan. Hiroko Akiyama, Ph.D., is an assistant researh sientist at the Survey Researh Center of the University of Mihigan Institute for Soial Researh and an adjunt leturer at the Shool of Soial Work in Ann Arbor, Mihigan. William Rakowski, Ph.D., is the Division Head of Disease Prevention/Health Promotion at the Center for Gerontology and Health Care Researh and an assoiate professor in the Department of Community Health Program in Mediine at Brown University in Providene, Rhode Island.