The Behavior of the Health Care Consumer: a Selective Review

Lawrence H. Wortzel, Boston University
ABSTRACT - Consumer behavior analysis and findings are an important input to the design of health care marketing programs. This paper is an attempt to present a framework for understanding consumer health care behavior, and to present selected findings. The paper concentrates on primary demand aspects and focuses on three types of physician visits: preventive, diagnostic, and therapeutic. A model is presented to predict behavior for preventive and diagnostic situations, and behavior in therapeutic situations is described and analyzed.
[ to cite ]:
Lawrence H. Wortzel (1976) ,"The Behavior of the Health Care Consumer: a Selective Review", in NA - Advances in Consumer Research Volume 03, eds. Beverlee B. Anderson, Cincinnati, OH : Association for Consumer Research, Pages: 295-301.

Advances in Consumer Research Volume 3, 1976      Pages 295-301

THE BEHAVIOR OF THE HEALTH CARE CONSUMER: A SELECTIVE REVIEW

Lawrence H. Wortzel, Boston University

ABSTRACT -

Consumer behavior analysis and findings are an important input to the design of health care marketing programs. This paper is an attempt to present a framework for understanding consumer health care behavior, and to present selected findings. The paper concentrates on primary demand aspects and focuses on three types of physician visits: preventive, diagnostic, and therapeutic. A model is presented to predict behavior for preventive and diagnostic situations, and behavior in therapeutic situations is described and analyzed.

INTRODUCTION

For a variety of reasons ranging from the purely humanitarian to the purely economic, marketing has become an accepted activity in many health care institutions and settings. The functions that marketing is expected to serve in the health care field are not dissimilar to the functions marketing is expected to fulfill in the commercial sector of the economy. And it is not surprising that the marketing problems faced by many health care institutions are similar to those faced by firms in the commercial sector. Perhaps some examples will make this point more clearly.

Beale and Schroeder (1973) describe the marketing efforts undertaken by a new urban health center. The marketing task was simply to register potential users with the center. An adequate level of registration was required in order to justify a full staff for the center. Registrants did not have to pay either a registration or a maintenance fee nor did they take on any other contractual obligation; registration merely indicated intent to use the center's services. The center was promoted primarily through direct mail advertising.

A Government technical assistance publication (U.S. Department of Health, Education, and Welfare, 1973) describes the marketing tasks involved in HMO marketing in similar terms, referring to media advertising and to personal selling efforts directed both toward organizations and toward individual potential subscribers within those organizations. Kotler (1975) lists health institutions including hospitals, HMO's and communities that could undertake health care marketing efforts. He goes on to describe the variety of marketing efforts that these organizations and institutions could undertake, including attracting patients, attracting donations of blood or money, attracting physicians, attracting contributions to medical causes, and encouraging behavior likely to be productive of better health. Invariably, these examples discuss the use of the usual marketing tools.

Virtually all marketers working in the commercial sector of the economy recognize that the basis of a marketing program is an understanding of the relevant aspects of consumer behavior. The emergence of a consumer behavior literature, consisting of both conceptual work and empirical findings organized and reported so that it is both accessible and useful to commercial marketers is a testimonial to this recognition. Moreover, the way in which this body of literature has developed has facilitated the conduct of additional research because of the relative ease with which a researcher can identify gaps in knowledge. Health care marketers also recognize the importance of understanding consumer behavior. There is a growing body of literature relating to health care that can certainly be labeled "consumer behavior" literature. As yet, however, this literature has not been organized so that its concepts and findings can be easily used by health care marketers in building their marketing programs. And, possibly, the content and dimensions of this literature may not be well enough known among potential researchers in health care consumer behavior. This paper, therefore, will attempt to do two jobs: (1) to review some of the content and dimensions of the literature relating to health care consumer behavior, and (2) to organize this literature, and to develop a focus for it that will make somewhat easier the tasks of applying it to marketing problems and of identifying potentially worthwhile research projects. The review to follow is a very selective one; no attempt has been made to cover all of the literature. Its major purpose is to give interested health care marketers and researchers some guides and some entry points for a possibly more intensive search. The attempt at organization is an effort to cast the literature into a simple straight forward, marketing oriented and managerially useful framework.

A FRAMEWORK FOR HEALTH CARE CONSUMER BEHAVIOR

Exhibit 1 presents schematically the framework that will be used in organizing this paper. The words and concepts are familiar enough and appear, at least in part, in either oral or written publications; is by no means wholly original. Essentially, the framework is a 3 x 3 matrix in which the columns are types of medical care visits and the rows are consumer decisions. Thus, consumer can make a visit for preventive reasons (a polio shot), or for diagnosis (a "pap" smear), or for treatment (in response, say, to a pain in the neck). Similarly, the consumer has both a primary and a selective demand choice to make, both of which may be affected by the consumer's previous experience with visit types and with providers.

Each cell of the matrix encloses the factors that determine behavior with respect to each combination. These factors can be classified into four broad groups: people specific (such as ethnic background, social class, age, sex, life style); promotion specific (whatever an institution does to obtain and satisfy patients) "product service" specific (whatever it is that the consumer perceives is being provided); and institution specific ( such as type of provider and location of the provider in relation to the consumer).

Since this paper had to conform to limitation in space, it has been necessary to concentrate the review on only certain cells of the matrix, and on only selected factors within each cell. In order to conform most closely to the theme of the session in which this paper will appear, the paper will concentrate on primary demand and on people specific factors.

EXHIBIT 1

HEALTH CARE CONSUMER DECISIONS

THE DECISION TO SEEK CARE: PREVENTIVE AND DIAGNOSTIC VISITS

Kasl and Cobb (1965) provide rather good descriptions of preventive and diagnostic behavior, defining each respectively as "any activity undertaken by a person who believes himself to be healthy for the purpose of preventing disease or detecting disease in any asymptomatic state." A conceptual model called the Health Belief model designed to explain demand for these two types of visits has been proposed by Rosenstock (1966). The Health Belief model consists of two dimensions or consumer states. One is defined as "the psychological state of readiness to take specific action" and the other is defined as "the extent to which a particular course of action is believed, on the whole, to be beneficial in reducing the threat." Both cognitive and "emotional" psychological elements are viewed as affecting the individual's psychological readiness. These include perceived susceptibility to a disease or condition and its perceived seriousness if contracted. Rosenstock presents hypotheses indicating the action consumers will take given each state combination. The action hypotheses are, however, somewhat confusing and somewhat incomplete. Essentially, Rosenstock states that, when readiness is low, intense stimuli will be needed if action is to result. He also states that when readiness is high, even a slight stimulus will result in some action. But, because actions that may be effective in reducing the threat of illness may also be perceived as causing significant emotional, physical or financial discomfort, the consumer may simply avoid taking such action. The consumer might instead try to remove himself psychologically by engaging in some "activity" that he may be able to pretend is threat reducing, but which in actuality is not. Or, the consumer may experience an increase in fear, and as a result become incapable of taking reasoned action.

Actually, the Rosenstock model is basically an extension and elaboration of the classic Janis and Feshbach (1953) study of fear arousing communications, but allowing for more diffuse responses to a state of high readiness. The model does not appear to be an unreasonable representation, but it does seem less precise than one would ideally like, because of the variety of responses it predicts. Ideally, one would want to know the circumstances under which the consumer will take the threat reducing action, or will indulge in psychological removal, or will take unreasoned action based on fear. Perhaps a conceptual scheme originally proposed to deal with situations such as those described by Janis and Feshbach (1953) can be of some assistance.

Bauer with Cox (1963) also working from findings such as Janis and Feshbach's, propose that communications can have two dimensions, an "emotional" or drive-arousing dimension, and a "rational," or drive-reducing dimension. They argue that the effect of a communication of a given strength of drive-arousing and/or drive-reducing content depends on the exact state of the recipient prior to receiving the communication. Exhibit 2 presents their argument diagrammatically. [There is no intention here to slight the most interesting model proposed by Zaltman and Vertinsky (1971). It is simply that the Zaltman-Vertinsky model was designed with an emphasis on less developed countries and is somewhat more elaborate than is necessary for the purpose of this paper.]

EXHIBIT 2

EXTENDED FORMULATION

The diagram is taken directly from Bauer with Cox (1963), but the labeling is an extension of the Rosenstock Health Belief Model, and the formulation to follow may include an idea or two of my own.

A line of unknown elope (here 45 degrees is taken arbitrarily) divides the diagram into an action and an inaction zone. This line indicates that the more severe the threat (severity is the product of perceived susceptibility and perceived seriousness), the more "sure" the outcome of a threat reducing action must be in relation to its cost before the consumer will take that action. The exact probability of success required in order for a consumer to take any given action at any given level of readiness is a function also of the perceived cost of the action; at any given level of readiness, higher probabilities of success are required for high "cost" actions to be undertaken.

The conditions under which each of the alternative actions proposed by Rosenstock will be taken can now be hypothesized somewhat more precisely for consumers who are in a state of high readiness. These consumer actions are predicted on the basis of perceived subjective probability of success (SpS) and cost (C) of the action and they do not have to be predicated on reactions to a given communication. Specifically,

a. Favorable (the "medically correct") action will result most readily when (SpS) is high and (C) is low.

b. Inaction will result when all alternatives are perceived as low in both (SpS) and (C). The tendency toward inaction increases as (C) increases while (SpS) remains constant.

c. An action that is perceived as threat-reducing but which in actuality is not will result when that action is perceived to have a lower (C) attached to it than to other alternative actions and when the additional (C) attached to "better" alternatives does not appear to be balanced by an increased (SpS).

d. The consumer's response is simply increased fear when the seemingly "best" alternative has an unacceptably high (C) and/or a low (SpS). In this condition there is no alternative the consumer can pretend is threat reducing. [It should be noted that in very extreme cases, e.g. where a consumer has been diagnosed as having an irreversibly fatal disease, it is likely that any and every possible action, regardless of its(C) or (SpS) might be undertaken.]

It should be a worthwhile exercise to look at soma studies that report on consumer behavior with respect to preventive and therapeutic visits to see whether the hypotheses Just stated seem to hold up. The first of such studies antedates the Rosenstock model but it is of interest both because it, itself, is a compilation and analysis of several other studies, and because of its subject matter: poliomyelitis vaccination. Polio vaccination is purely preventive medicine. At the time the study to be reported here was compiled, Salk vaccine had been proven to be a highly effective polio preventive, and the vaccination procedure was a relatively easy one for the patient to undergo. Thus, the treatment had a high (SpS) and a low (C). Moreover, a minor epidemic of paralytic polio had just occurred, indicating that the study population might be expected to have been in a reasonably high state of readiness.

Rosenstock, Derryberry, and Carriger (1959) analyzed six studies on acceptance of polio vaccination which they considered methodologically sound. Their conclusions were that "a basic determinant of the decision.., is the extent to which the individual believes that he is susceptible.., or ... to which the parent believes ...his child is..." There seems to have been little question among consumers that the Salk vaccine was effective and could be administered with little discomfort to the patient. Non-vaccinatees, then, simply seemed to be consumers to whom polio was of little significance; data are cited which show that vaccination rates were higher in populations that (1) demonstrated stronger beliefs in the seriousness of the disease, or (2) were exposed to some form of "promotion" designed to raise the salience of the disease (thus increasing consumer's state of readiness).

There is no indication from any of the findings that raising the salience of polio had any effect other than to increase the consumer's propensity to undergo vaccination.

Studies of behavior in situations that have diagnostic as well as preventive overtones, however, appear to elicit consumer behavior that is somewhat more difficult to interpret. Fink, Shapiro and Roester (1972) describe an effort to increase participation in a breast cancer detection program. The effort was an experiment conducted over a four year period among a sample of female members aged 40-64 of the Health Insurance Plan of greater New York. Members in the experimental group were first contacted by a mailing which informed them of the initiation of a breast cancer detection study and asked them to make an appointment to participate in a screening examination. Non-respondents were sent a second letter reemphasizing the importance of the exam, and were further contacted by telephone if they did not respond to this second letter. Later contacts were made with each participant at annual reexamination time.

If one assumes that the contents of the communications were actually as described, the description indicates a communication that appears to be essentially drive reducing (more specifically, reducing the cost, or (C) term) since it appears simply to offer a convenient service. Thus, the communication should have been most effective with consumers who were already in a state of readiness. Study findings indicate that this was, in fact, the case. Participants in the first screening tended to be younger, to be more highly educated, more favorably disposed toward the concept of screening, and to express more concern about the possibility of having breast cancer than did non-participants.

Similar factors distinguish participants who completed all four exams from those who did not. Education, income and occupation were all positively correlated with completion. Both previous use of medical services and previously having had a polio vaccination were also positively related to completion. There was also a positive relationship between respondents' self-reported state of health and their completion of the four screening exams.

Women who started the screening program but did not take all four examinations showed a significantly higher vote of agreement with the statement that "Physical examinations just make you worry; it's like looking for trouble" than did women who did take all four examinations. This finding would also be consistent with the model if, for these women, the examination itself could have been a drive-arousing rather than a drive-reducing experience. It is not difficult to imagine such a possibility; an uncommunicative or seemingly unsympathetic or uncaring physician can easily raise a patient's anxiety level, with the result that the patient will become a "drop out."

Given the high level of publicity relating smoking and disease, cigarette smokers ought to be in a perpetually high state of readiness with respect to medical problems. The risks involved in smoking might indicate more frequent exams. But, the extended model would predict that cigarette smokers should be less frequent users of preventive and diagnostic medical services than should non-smokers. This is because the diseases smokers contact as a result of their smoking are relatively incurable, and it's not unreasonable to expect that smokers know it.

Oakes et al (1974) studied participation in health examinations among a sample of Kaiser-Permanente Medical Care program members. Members prepay and are entitled to free physical examinations. A mailed questionnaire was used to elicit rates of participation in physical examinations. The study included procedures to adjust for non-respondents. Data were analyzed controlling for socio-economic status and for cigarette smoking. Among males, current smokers were less likely to have had any kind of health examination over the previous five years than were either ex-smokers or non-smokers. The difference was most pronounced among men of higher social class (who are likely to be best informed about the consequences). Among women, the differences between smokers and non-smokers were much less pronounced. It does not seem unreasonable to explain this finding by noting that women have other occasions, principally gynecological, for having regular physical examinations.

Predicting use of a center offering free, comprehensive medical care to a low income population for whom medical care had previously been inaccessible and un-affordable should also be an interesting exercise for the model. If such a center is well-promoted, it might even succeed in raising consumer's readiness status to too high a level. The result might be overuse of the center, even in circumstances where the consumer's (SOS) is relatively low. This is because certain dimensions of the consumer's (C) are also low: if the center is both free and easily accessible. Given that these (C) dimensions are low in every circumstance, overuse might be predicted specifically in those situations where the expected treatment is not uncomfortable.

A "before-and-during" study of consumer's attitudes toward the use of various medical care services in such a population was conducted among residents of a somewhat isolated Boston-area low income housing project (Bellin & Geiger, 1972). Two waves of interviews were conducted, both before and two years after the initiation of a comprehensive community health center in the project. Project residents were involved in the design and operation of the center; the center was heavily promoted and much talked about, and residents could use the center's facilities without charge.

Comparison of the results of the two studies showed that attitudes toward asymptomatic medical checkups become significantly more favorable and that the proportion of project residents who reported having had a general physical during the past 12 months rose from 17% to 59%.

This is certainly not an unreasonable finding, but it really says very little about the efficacy of the model. There is, however, another finding that is of some interest. Respondents in both waves were given a list of conditions or symptoms including several from each of three categories: serious (e.g. severe shortness of breath); mixed (e.g. feeling tired all the time); minor (e.g. sore throat, running nose). They were asked whether a doctor should be seen for each condition or symptom. In the baseline survey, a much larger percentage of respondents said "yes" for the serious conditions than for the minor ones. The percentage for the mixed conditions fell in between the serious and minor percentages. Results of the later survey showed a striking increase in the percentage of respondents who said "yes" for minor conditions (from 30% to 50% for sore throat, runny nose, for example). Again, these results are consistent with the model, and the thrust of the studies so far discussed suggest that the model can be a useful schema for working with preventive and diagnostic visits. Specifically, it suggests, schema for determining the content of promotional efforts that are based on consumers' existing attitude states. The model should be most effective in working with specific preventive and diagnostic problems.

Given any specific situation, straightforward research techniques can be used to establish existing consumer states with respect to (1) each of the readiness dimensions, (2) the perceived (SpS) of the desirable medical action, (3) perceived (SpS)'s of other medical activities that consumers might incorrectly perceive as desirable, and (4) the (C) for each possible medical action. Ideally, these data can be linked to other consumer characteristics, either demographic, or psychographic, or both, so that identifiable market segments can be defined.

These market segments can then he plotted on a graph such as the graph in Exhibit 2. Given the position of each market segment on the graph, it should be feasible (1) to identify the probable direction (up, down, right, left) that market segment would have to be moved in order for the medically favorable action to result and (2) to identify the way in which the content should be balanced for susceptibility, seriousness, (SpS), and (C) so that the medically favorable action would be most likely to result.

It should then be possible to construct specific promotional programs aimed at individual market segments, to test them and to measure the results. Both testing and measurement of results are often accomplishable since the marketer will usually be using either direct mail or other controllable media and/or personal selling efforts. In addition, consumers who take the medically favorable action can be specifically identified and "tied back" to specific promotional efforts and to specific market segments, and therefore to predisposition. Thus, the model's prediction can also be validated.

It is not likely, however, that the process just described will be quite as tidy in actuality as it appears on paper. For one thing, the actual slope of the line dividing the action and inaction zones in the model is unknown. For another, it may be difficult to estimate exactly how strong a specific state of readiness may actually be. Even in the face of these complications, however, consumer behavior in preventive and diagnostic situations is likely to be somewhat more straightforward than behavior in therapeutic situations. It is to these situations that we shall turn.

THE DECISION TO SEEK MEDICAL CARE: THERAPEUTIC VISITS

Kasl and Cobb (1965) also provide a good description of therapeutic behavior, defining it as "any activity undertaken by a person who feels ill, for the purpose of defining the state of his health and of discovering suitable remedy." Any review of studies conducted on consumers' use of medical care services for therapeutic reasons quickly identifies that a visit to the physician may be the end, rather than the beginning of a behavior chain that was initiated when some symptom was experienced. And the behavior chain may end before the physician visit stage is reached.

A landmark study conducted among members of a Bronx, N.Y. prepaid medical plan (Friedson, 1961) discovered the existence of "lay referred systems" that consumers progressed through before they formally consulted a physician. Usually, the subject began the search for medical help with self-diagnosis and some resultant self-treatment. If the self-treatment did not appear successful, in some cases others would be consulted, first members of the household, and then outside lay people. In these cases, advice and diagnosis would be sought from each lay person, and the physician would be consulted only when the lay advisers' prescriptions proved ineffective.

These lay referral systems, however, appeared to differ based on the social class of the patient, because patients of upper versus lower social classes evidenced different views of their bodies and bodily processes, and of illness. Essentially, lower class people tended to be apprehensive about illness and to be relatively ignorant with respect to their bodies and bodily functions. They also evidenced considerable ignorance about illness, and about types of medical treatment. They tended to view the physician as something of a mystic, whose skills were arcane, incomprehensible, and above questioning. The upper class patient, on the other hand had a much more realistic and detached view of his or her body and its functions, and evidenced much more sophistication with respect to illness. This patient tended to view the physician much more as a colleague, and to take a somewhat detached but realistic and concerned interest in the treatment the physician was providing.

The lower social class referral system tended to be localized to kin and neighbors, selected on the basis of family relationship and physical proximity. The upper class lay referral system did not depend on kin or neighbors; advisers were selected on the basis of perceived knowledge, rather than on the basis of family or neighborhood. A summary view of these social class differences might describe lower social class patients as unwilling to try the system except as a last resort, but blindly trusting the system once they entered it. It might describe upper social class patients as much more willing to try the system, but also much less trusting of the system once they enter it.

Suchman (1965) conducted an extensive study, also among New York residents, that supported many of Friedson's (1961) findings. He hypothesized that an individual's health behavior resulted from that individual's medical orientation, and that the individual's medical orientation was a consequence of his or her socio demographic characteristics and thus the concomitant community orientations. His findings identified social class linked community systems that paralleled the "lay referral" systems of Friedson (1961). Lower social class systems tended to be parochial, while higher class systems tended to be more cosmopolitan. Lower social classes also evidenced "lower knowledge about disease and higher skepticism toward medical care..." which again supports Friedson's (1961) findings. The thrust of these findings would suggest that social class membership plays an important role in determining the therapeutic aspect of medical care utilization; however, Suchman's data are not strong where the link between medical orientation and behavior must be made. At least one, more recent study fails to replicate all of the Friedson (1961) and Suchman (1965) findings.

Reeder and Berkanovic (1973) conducted a study in the Los Angeles area in 1970 using many of Suchman's variables. They did not find a significant relationship between social class and medical orientations. These researchers cited all of the usual reasons for the failure to replicate, the most compelling of which seem to be time and place. The late 1960's, and California may well have bred a lower social class population that is more aware of good medical care and the medical care system. As the Bellin and Geiger (1972) study cited earlier suggests, perhaps people can learn to use the medical care system for therapeutic reasons and there may now be other determinants of its use such as heightened feelings of entitlement, that are more important than social class.

In any event, the conclusion of a review of several studies conducted among different populations (Zola, 1972) and concerned with different medical problems should not be surprising. This review concluded that a substantial number of medical complaints are not brought to physicians for treatment (at least they are not purposefully and specifically brought to physicians). However, as Zola (1972) forcefully points out, this does not mean that these complaints go completely untreated. Self-medication (or medication prescribed by a friend or pharmacist), a visit to a paramedical (podiatrist, osteopath etc.) and "informal" medical consultation (while the physician is treating another family member, for example) are among the many ways in which some of these medical conditions actually do get treated.

Whether such behavior is functional or not depends both on the point of view of the analyst and on the specific medical conditions involved. For example, if the analyst is a physician in a fee-for-service practice, great concern might be expressed that any complaint may be going untreated. A physician in an organization offering a prepayment plan, however, could consider the situation highly functional, depending on the specific medical condition that may be going without physician treatment. Both, however, might be interested in shaping behavior, although toward somewhat different ends. Thus, it should be useful to explore (1) what complaints tend to go without treatment, and, (2) whether there are systematic differences among consumers with respect to what conditions do or do not get treated.

Ludwig and Gibson (1969) studied the extent to which medical care was sought among a population of applicants for social security disability benefits who claimed to be in "poor" health or worse. He divided this population into two groups, those who had "seen a doctor or visited a clinic or hospital" within two months (users), and those who had not done so for six months or more (non-users). Data were collected for both groups on their recognition of symptoms, situation factors affecting the ease or difficulty with which medical attention could be obtained, and faith in the medical system. The results of the study are quite interesting. The type and number of symptoms respondents reported had no significant effect on their use of medical care; respondents with any specific symptom were no more or no less likely to have sought medical care than were respondents with any other specific symptom. Respondents reporting 5 or more symptoms were no more likely to report having sought medical care than were respondents with 4 or fewer symptoms. Both faith in the medical care and the ease of obtaining medical attention, however, affected the use of medical care. Among respondents who had a positive medical-scientific orientation (as measured by a three-item index), just 7% had not sought medical care in the past 6 months, while 32% of respondents with negative orientations had not done so.

There was a strong positive relationship between income and use of medical care: 37% of very low income respondents versus just 10% of somewhat higher income respondents had not sought care. Recent welfare contact inhibited medical care seeking: respondents who had recent contact with welfare agencies were twice as likely not to have sought medical care. Among this study population, therefore, it was not a question of what symptoms went untreated as much as a question of what beliefs and circumstances facilitated or inhibited the seeking of care. It should also be noted that this study deals with an essentially a lower social class population, yet both positive and negative medical-scientific orientations were found. In addition, some other variables appeared to be at least as important as respondents' medical-scientific orientations in predicting their use of the medical care system.

Hackett, Cassem and Raker (1973) provide strong further reinforcement for the finding that it is not primarily the presence or absence of symptoms that causes consumers to use medical services. They surveyed both in-and-out patients visiting the tumor unit at the Massachusetts General Hospital during the period 1968-70. Their survey collected data on delay (the time between each patient's first awareness of a sign or symptom and the first physician visit). Also collected were data on each patient's reason for first visiting the tumor unit, a self-description of their conditions (after diagnosis), a self-rating of delay time, the extent to which they worried about their health, family cancer history, and social class.

The researchers found considerable evidence of delay among study respondents. Just 20% had consulted a physician within a week of becoming aware of the first symptom or sign, and 66% had waited at least a month. The least amount of delay was found among patients whose reason for visiting the tumor unit was that their tumors were discovered during a routine physical examination. Those whose reason was worry about their symptoms showed the next to least amount of delay those who came in because their symptoms were incapacitating had delayed still more, while those who visit the tumor unit because of pain had delayed the longest. Delayers also tended to deny their condition, often calling it a "tumor," while those who were more prompt in seeking treatment were more likely than delayers to refer to their condition as "cancer." Those who worried about their health delayed somewhat, but not significantly more than those who did not. Respondents who reported a family history of cancer delayed somewhat longer than those who did not report such a history, and those who claimed they generally procrastinated in seeing physicians were more likely to have delayed in this particular instance. Patients higher in social class tended to delay less than lower social class patients; however, this finding may simply be related to a greater propensity of people in higher social classes to have regular physical exams.

It is difficult to disagree with these researchers' conclusion that the patients who delayed did so knowingly and purposefully. It is also difficult not to look at the findings from this study in light of the model presented in Exhibit 2. Obviously, the presence of signs or symptoms indicates a consumer in a high state of readiness, which suggests that the way to encourage a physician visit is through drive reduction. The group who visited the tumor unit with the least amount of delay were those whose signs or symptoms were discovered during a routine physical examination. These are also the patients who were the most likely to have been exposed to a drive reducing communication, since the physician could have explained the relationship between early treatment and success. Unfortunately, this study did not measure respondents' perceptions of the likelihood of successful treatment.

EMERGING GENERALIZATIONS

This paper has attempted to present some findings from a selected group of studies covering a range of consumers and medical situations. In general, it would appear that medical care consumer behavior is shaped more by psychographic than by demographic factors. Where demographics appear to predict behavior, it may well be that these demographic factors correlate closely with associated psychographic factors which are more likely to be causal. Such a generalization, of course, is consonant with expectations derived from the study of consumer behavior in the commercial sector.

However, the thrust of these findings may also be at cross currents with commercial sector findings that point to an increasing regard for the role of situation specific factors in shaping behavior. With respect to health care behavior, we might be more inclined to suspect a basic orientation toward medicine and medical care that is significant in predicting behavior in any given situation. This basic orientation might predict most effectively where there are no institutional or other barriers to action. That is to say, a given consumer's behavior toward a polio shot versus a cancer screening examination, versus a symptom for example might be more uniform than that same consumer's behavior in purchasing a detergent versus a set of tires. However, such a speculation must be made with caution if one's concern is with producing behavior change.

The studies reviewed also suggest that behavior can be changed in specific health care situations. It would be worthwhile to know the extent to which changing the consumer's behavior in any given situation has a carryover effect in terms of changed behavior toward other health care situations. Such knowledge could be extremely useful, for example, in finding the most efficient entry points to produce the most pervasive health behavioral change.

REFERENCES

Raymond A. Bauer with Donald F. Cox, "Rational Vs. Emotional Communications: A New Approach," in L. Arons and M. May, eds. Television and Human Behavior, Appleton-Century-Crofts, New York, 1963.

Eunice C. Beale and Steven A. Schroeder, "Marketing for an Urban Health Center," Health Services Reports, Volume 8, 1973 pp. 84-88.

Seymour S. Bellin and H. Jack Geiger, "The Impact of a Neighborhood Health Center on Patients' Behavior and Attitudes Relating to Health Care," Medical Care, Volume 10, No. 3, May-June 1972 pp. 224-239.

Bionetics Research Laboratories, Inc., Marketing of Health Maintenance Organization Services, U.S. Dept. of H.E.W., Washington, D.C., 1973.

Raymond Fink, Sam Shapiro, and Ruth Roester, "Impact of Efforts to Increase Participation in Repetitive Screening for Early Breast Cancer Detection," American Journal of Public Health, March 1972, pp. 328-336.

Eliot Friedson, Patients' View of Medical Practice, Russell Sage Foundation, New York, 1961.

Thomas P. Hackett, N. H. Cassem, and John W. Raker, "Patient Delay in Cancer," The New England Journal of Medicine, Volume 289, No. 1, pp. 14-20.

Irving L. Janis and Sidney Feshbach, "Effects of Fear Arousing Communications," J. Abnormal & Social Psychology, 48, 1953, pp. 78-92.

Stanislav V. Kasl and Sidney Cobb, Health Behavior, Illness Behavior, and Sick Role Behavior, ISR, University of Michigan, Ann Arbor, 1965.

Philip Kotler, Marketing for Nonprofit Organizations, Prentice Hall, Inc., N.J., 1975.

Edward G. Ludwig and Geoffrey Gibson, "Self Perception of Sickness and The Seeking of Medical Care," Journal of Health and Social Behavior, II, July 1969, pp. 125-133.

Thomas W. Oakes et al, "Health Service Utilization by Smokers and Nonsmokers," Medical Care, Volume 12, No. 11, November, 1974, pp. 958-966.

Leo G. Reeder and Emil Berkanovic, "Sociological Concomitants of Health Orientations: A Partial Replication of Suchman," Journal of Health & Social Behavior 14, June 1973, pp. 134-143.

Irwin M. Rosenstock, Mayhew Derryberry, and Barbara K. Carriger, Public Health Reports, Volume 74, No. 2, February 1959, pp. 98-102.

Irwing M. Rosenstock, "Why People Use Health Services," Milbank Memorial Fund Quarterly, Volume 44, 1966, pp. 94-127.

Edward A. Suchman, "Social Patterns of Illness and Medical Care," Journal of Health and Human Behavior, 6, Spring 1965, pp. 2-16.

Gerald Zaltman and Ilan Vertinsky," Health Service Marketing: A Suggested Model," Journal of Marketing, Volume 35, July 1971, pp. 19-27.

I. K. Zola, "Studying the Decision to See a Doctor," Adv. Psychosom. Med., Volume 8, 1972 pp. 216-236.

---------------------------------------