Symposium on Psychiatry in Internal Medicine
Improving Patient Compliance A Guide for Physicians
Daryl Matthews, M.D.,* and Ralph Hingson, Sc.D.**
Patient compliance and lack of it has received widespread attention in recent years. If anyone conclusion can be reached from the spate of studies published on this topic, it is that physicians simply cannot expect all patients or, perhaps, even a majority of patients to adhere to the medical and therapeutic regimens they prescribe. Moreover, it is difficult for physicians to distinguish which patients are compliant and which are not. 5,7 Difficulty in distinguishing compliant from noncompliant patients can pose a serious obstacle to patient care. Unless physicians can determine the extent to which a patient is following a specific therapy, they cannot fully assess the utility of that therapy for the patient. On a broader scale, the value of specific treatments for a number of diseases may be questioned if the degree of patient compliance has not been assessed. The physician can no longer afford to simply take compliance for granted. Unless the physician takes the few extra minutes necessary to promote the highest level of patient compliance, his or her efforts, as well as those of other health professionals involved in the development and delivery of the regimen to the patient, may be jeopardized.
Factors Which Influence Compliance In the past few years a great deal of study has been devoted to this topic. The bulk of this research has consisted of descriptive comparisons of compliant and noncompliant patients. From this research we have learned a great deal about the influence of characteristics and attitudes of the patient, characteristics of the treatment regimen, and the physician-patient interaction on compliance. A number of articles review this literature in depth. 14, [8,25,27 Briefly summarized, this research informs us, first, that noncompliance can pose a problem regardless of the severity of a patient's disorder in terms of painfulness, disability, or threat to life. While non"'Assistant Professor, Department of Socio-Medical Sciences and Community Medicine and Division of Psychiatry, Boston University School of Medicine; Assistant Visiting Physician (Psychiatry), University Hospital, Boston, Massachusetts ':"'Assistant Professor, Department of Socio-Medical Sciences and Community Medicine, Boston University School of Medicine, Boston, Massachusetts Medical Clinics of North America-Vo!. 61, No. 4, July 1977
879
880
DARYL MATTHEWS AND RALPH HINGSON
compliance is particularly high when the disorder being treated is asynlptomatic or psychiatric in nature,14 numerous studies have shown that, in general, the severity of a disorder as measured by evaluation by a physician, prior hospitalizations, or prior diagnoses is unrelated to compliance. While one would think that people with more severe problems would be more motivated to comply, such disorders often require more complex, difficult regimens. Persons with severe disorders may find compliance difficult because of limitations imposed on them by their illnesses, and may also be discouraged by the failure of their previous efforts to prevent or cure their ailments. Just the fact that a person's illness is serious, painful, or even life threatening does not ensure a high level of compliance. Second, the patient's demographic characteristics do not consistently predict whether or not he will be compliant. Studies which report no relation between patient compliance and social class, age, sex, education, occupation, income, and marital status outnumber those which do almost three to one. 14 ,27 Surprising as such a finding may seem, the lack of association between demographic characteristics and compliance has been observed for a variety of disorders, in a variety of medical settings, using a variety of measurement techniques. 1R Though this does not mean that such demographic characteristics as social class never predict compliance, it does mean that noncompliance can be a problem with any patient population, regardless of social class level, age composition, or racial and ethnic background. Third, while common sense tells us that knowledge about one's illness and understanding how to follow the regimen are necessary for a patient to comply, they are not always sufficient. Graphically illustrating this point is a study by Bergman and Werner in which 59 families of children having streptococcal pharyngitis were followed. When questioned at home, 95 per cent of respondents correctly recalled the appropriate regimen. By the sixth day of a 10 day course of treatment, however, pill counts and urine assays revealed that only 30 per cent of the patients had received the prescribed penicillin. 4 More recently, Sackett et al. 31 conducted a randomized controlled trial to increase patient compliance with antihypertension regimens among more than 230 men. Although the experimental group receiving intensive instruction did show greater increases in knowledge about hypertension than a control group, the experimental group was no more compliant than the control group. In fact, measurements of patient knowledge at intake and 6 month follow-up in both the intervention and control groups were not related to compliance. 31 More consistently predictive of compliance are the nature of the regimen offered to patients, the beliefs patients hold about their illness and treatment, and certain features of the interaction between patients and their physicians. Some treatment plans are more difficult for patients to follow. The more complex a regimen, e.g., the greater the number of medications and the more times per day that medications must be taken, the less
IMPROVING PATIENT COMPLIANCE
881
likely is full compliance. Compliance tends to diminish as the duration of treatment increases over time. Regimens which require marked changes in behavior, especially abandoning established behaviors like smoking or drinking or certain dietary habits, are less likely to be followed than regimens which ask patients to initiate new behaviors that do not disrupt their normal routines. Regimens which produce adverse side-effects for the patient are also less likely to be followed. A number of studies have rather consistently demonstrated that one's beliefs about one's illness and its treatment can influence compliance independently of knowledge about the illness and treatment, the actual nature of the illness and treatment, or the characteristics of the regimen. These studies have explored the impact of the beliefs outlined in the "Health Belief Model" originally proposed to explain preventive health behaviors such as obtaining chest x-rays or immunizations. 30 Applied to compliance, this model suggests that patients (a) who feel susceptible to problems or complications because of their illness, susceptible to further attacks of the illness itself, (b) who believe that their illness could pose severe consequences for their health and daily functioning, (c) who feel that the proposed treatment plan will be highly effective in treating their illness, and (d) who do not foresee such major obstacles to compliance as adverse drug effects, cost of the regimen, or perceived lack of safety of a medication, will be more likely to follow regimens offered to them than patients not holding these beliefs. While very few studies have systematically explored all the major features of the model, a number of studies have examined individual component beliefs. 2.13 •17 In the most thorough review ofliterature on the health belief model as it applies to compliance to date, Becker concluded: "Although no single effort has provided (or could provide) convincing confirmation of all Health Belief Model Variables, most studies have produced internally consistent findings in the predicted direction which taken together yield relatively strong support for this conceptual model of compliance behavior. "3 Characteristics of the interaction between patients and physicians also seem to influence whether patients will adhere to their doctor's recommendations. In general, the more satisfied patients are with their physicians and the settings in which care is provided, the more likely they will be to comply. Many factors have been found to undermine the physician-patient interaction and thereby reduce compliance. We will mention only a few here. When physicians underestimate patient knowledge and ability to understand information about their disorders, they give their patients less information about their illness and treatment. Their patients, in turn, are less likely to actively inquire about their illness and treatment and are less likely to follow prescribed regimens. 29 Patients are also less likely to comply: (1) if they feel they are not held in adequate esteem by their physician;lO (2) if their physician actively seeks information from them without providing feedback about either why that information is being gathered or on the patient's condition;9 (3) iftension emerges during a physician-patient interaction and is not addressed or resolved;9 (4)
882
DARYL MATTHEWS AND RALPH HINGSON
if patients feel that their expectations are not being met or if they feel their physician is not behaving in a friendly manner. 12,22 Interactions between physician and patient are, understandably, often laden with emotion. The patient may be in pain, may be apprehensive about his or her health and the consequences of the illness, may find the setting of the office, clinic or hospital complex, impersonal, or frightening. Moreover, as most physicians quickly learn, patients bring many different motivations to their medical visits including the desire to reduce pain, to cure disease, to prevent illness, to discuss family or other life problems, to seek company and compassion, to confirm they are sick, or to learn they are well. Deciphering a patient's motivations and attempting to meet patient expectations in a potentially emotionally charged interaction is a difficult task which changes with each new patient. Though the pitfalls in physician-patient interaction can take varied and sometimes unpredictable forms, most research in the area leads to the conclusion that the physician-patient interaction can be an important determinant of compliance.
The Compliance-Oriented History As with other problems in medicine, the importance of history-taking in addressing the problem of noncompliance cannot be overstated. The extension of the medical history to include information on the patient's health-related beliefs can be of use to the practitioner in two distinct ways: (1) An understanding ofthese beliefs on the part of the practitioner can help to predict to what degree a patient is at risk for poor compliance. (2) Investigating the issues dictated by the Health Belief Model can contribute to the development of a relationship with the patient most likely to foster compliance. A systematic exploration of these beliefs can offer the physician more compliance-related information than is obtained in any other facet of the medical encounter. The model implies that the following areas merit attention: BELIEFS ABOUT THE ILLNESS ITSELF. These include beliefs about the seriousness of the disorder and, if applicable, beliefs about the likelihood perceived by the patient of suffering future episodes. Such beliefs can be ascertained by the physician by posing such questions as, "What is it that worries you the most about having developed high blood pressure?" "How likely do you think it is that your pain will return?" "What do you think is causing the problem you describe?" "Have you known anyone with a rash such as yours? How did things turn out for that person?" Questions such as these can identify misconceptions on the part of the patient which may be predictive of poor compliance. Such misconceptions, it has been found, may involve either under or overestimates of the seriousness of the disorder and the likelihood of recurrent episodes. 24 Depending, of course, on the disorder in question, it may be beliefs about the symptom rather than about the illness itself which must be investigated. Most women, for example, would be fearful of cancer upon detecting a breast mass. Even when the diagnosis of a benign lesion is made, the fear of cancer may remain. In the case of symptoms with
IMPROVING PATIENT COMPLIANCE
883
considerable meaning to patients, questions about perceived seriousness and susceptibility may reasonably be asked early in the visit, at the time when historical material is customarily elicited. With other disorders, including those which present asymptomatic ally, the practitioner must bear in mind that the relevant beliefs about the illness may only surface once the patient has been informed of the diagnosis. This requires that the questioning of the patient not be confined, as is often the case, to the early part of the visit. Rather, it must be reintroduced at various times throughout the encounter as information is provided to the patient. BELIEFS ABOUT THE TREATMENT PLAN. The Health Belief Model implies that patients undertake something like a "cost-benefit analysis" in deciding whether or not to follow medical instructions. Those patients who perceive many or serious costs associated with the treatment plan, while anticipating little benefit, are at risk for poor compliance. The costs of adhering to a therapeutic regimen may be more than just financial. They may also involve giving up familiar behavior patterns or acquiring new ones, the involvemept required of family members, the anticipated adverse effects of treatment, and the abandonment of other remedies from which the patient may have been deriving real or placebo benefit. A practitioner can assess such perceived costs by asking such questions as "What have you been doing for the pain on your own?" "Can you think of any problems you might have in taking these pills an hour before meals?" "Is there anything that worries you about having these monthly blood tests?" "Are you concerned about any particular side-effects of the medication ?" The patient's perception of the benefits of the proposed treatment includes notions about the efficacy of the regimen as well as ideas about what problems he or she might be spared if treatment is carried out as recommended. These beliefs may be discerned by asking, "How effective do you think the injections will be for you?" and "What do you think would happen if you forgot to take your medicine one night?" In implementing the suggestions offered here, the physician must recognize that patients' beliefs can change markedly from visit to visit. It is unusual for an individual to employ his or her physician as the sole source of health information. Friends, relatives, books, magazine articles, and other health professionals may be consulted between the time the diagnosis is made and the time treatment is completed. The information or misinformation obtained from such sources can have as much or more influence on compliance than what has been explained to the patient by the physician. In order to ensure optimal compliance with an ongoing treatment plan, the physician would be well advised to repeatedly assess the status of the beliefs considered above over the course of the several visits which may be necessary to manage a particular episode of illness. In addition to the more specific interventions which will be discussed below, there is reason to suspect that the mere process of taking a compliance-oriented history can do much to contribute to improved compliance. The use of the type of questions advocated here communicates to the patient that the physician is interested in the patient's own percep-
884
DARYL MATTHEWS AND RALPH HINGSON
tion of the illness and the patient's own appraisal of the treatment plan. This may serve to enhance the feeling on the part of the patient that the physician understands the complaint and has taken a friendly interest. Insofar as these feelings contribute toward increased patient satisfaction with the encounter, the likelihood of good compliance is increased. 12 Similarly, the taking of a history such as is being suggested here may contribute toward the patient's perceiving respect on the part of the physician, another perception which appears to be related to compliance. 1o Finally, having access to the information yielded by a complianceoriented history may mitigate against the physician's underestimating the level of the patient's information about the disorder, another aspect of the physician-patient relationship associated with poor compliance. 29
Interventions to Improve Compliance Once the physician has explored the patient's perceptions of his or her illness and treatment, several different steps can be taken to encourage the patient to follow the prescribed regimen. ':' The information elicited during the compliance-oriented history can be most useful in this regard. First, providing the patient with a clear description of the condition, how it is treated, and the implications of the illness for the patient when treated and untreated probably constitutes the most frequently tried way to promote compliance. While useful, merely presenting information does not always insure compliance. One recent study has further suggested that not merely describing a regimen but also discussing the specific purpose and function of the prescribed medications will enhance compliance. 21 To be sure that the patient understands what is said, the physician can ask the patient to repeat key parts of what he or she has been told, especially how to follow the prescribed regimen. Clearly written or typed instructions are also useful. 8 Whenever presenting new information to a patient, the physician should ask if there are any questions. Patients will generally hold some beliefs, correct or incorrect, about their illness before contacting their physician. Together with these old beliefs, which will have been held for varying lengths of time and with varying degrees ofintensity, new beliefs will emerge during the course of treatment. If the physician feels that the patient's beliefs are appropriate, he or she should attempt to reinforce them. This may be done by indicating to the patient that he or she seems to have a good understanding of the situation and by asking the patient to telephone if any problems emerge during the course of treatment. Alternately, asking the patient to write down questions or concerns as they come up may help the patient remember them at follow-up visits. A much more difficult situation exists when the patient holds erroneous beliefs about his or her illness or treatment plan. Several options are available to change such beliefs. Perhaps the most over-used strat"'Some of the material outlined in this section was presented by Dr. Marshall Becker at a lecture at Boston University School of Medicine, May 3, 1976.
IMPROVING PATIENT COMPLIANCE
885
egy is fear arousal. So many health care providers have attempted to scare their patients into complying that the literature about this approach is particularly extensive. Review articles on the use of feararousal have reached the same fundamental conclusion: fear arousal is not consistently successful and, in some situations, may actually hinder compliance. 19,23 Fear arousal technqiues can falter for several reasons. 20 Patients exposed to a fear arousal message may perceive the threatened outcome as improbable, inapplicable, unimportant, or remote in time. Patients advised, for example, to stop smoking because of its adverse health consequences can and do deny or rationalize the advice away. Moreover, even if a fear appeal succeeds in arousing sufficient anxiety, it may fail to produce the intended behavioral change because the physician's recommendation may not be regarded as very effective. In some instances, if the patient was already anxious about his or her disorder, use of scare techniques might actually make thinking about one's illness and treatment so painful that any action on the issue is avoided. If the patient is already deeply anxious about the illness, a far more helpful approach would be one which attempted to allay fears by pointing out the promises held by treatment. Although appeals to other emotions may not have been given the same attention as fear arousal, they may actually prove more effective in altering beliefs and behavior. For example, appeals to love, joy, or parental responsibility can be powerfully motivating in some individuals. Pointing out any inconsistencies that may exist in the structure of the patient's stated belief may also be helpful. Often a patient's ambivalence about an illness or its treatment needs to be addressed before he or she will comply. A patient may be reluctant to follow a particular regimen which is consciously or unconciously in conflict with other desired goals, be they work-related, recreational, or family-oriented. Another alternative is to appeal to sources of information in which the patient has or can develop considerable trust. It may be difficult, for example, for a seriously ill patient to fully trust the physician who has never experienced the problems associated with having the disorder in question. In such situations, it is frequently useful to have the patient speak with other individuals having similar conditions who have benefited from the suggested therapy. Though not as technically knowledgeable, such individuals may appear more credible to the patient than the physician. The use of social pressure and social support can also be of help to patients. Such strategies have long been used with behavioral problems such as obesity, smoking, and alcohol abuse. More recently, Avery et al. 1 examined the use of such techniques with asthmatic patients. Comparing asthmatic patients placed in discussion groups focusing on ways the patients could prevent asthma attacks with controls drawn from the same emergency room population, they found that subsequent visits to the emergency room by patients in the discussion groups during the next 4 months were but half the number made by controls. 1 Patients with similar problems need not be the only resource for this kind of strategy.
886
DARYL MATTHEWS AND RALPH HINGSON
In many circumstances family members, employers, and friends can be of invaluable assistance in helping patients follow their regimens and informing the physician if the patient appears to be having difficulty. Finally, other direct incentives can be identified and employed to encourage compliance. For example, many patients dislike being told forthrightly what to do or not to do concerning health matters. To the extent that decision-making and follow-up responsibility can be shared with them, their compliance may be improved. In a recent experimental study, Haynes et al. 16 divided into two groups a sample of patients who, at 6 months after the initiation of antihypertension therapy, were not compliant. The experimental group was given instructions about how to take and record home blood pressures and chart their medication consumption. With each patient, ways were discussed to tailor his regimen into habitual daily activities. Six months later, 80 per cent of those in the experimental group were compliant, twice the percentage of compliant control group patients. 16 Involving patients in a decision-making capacity and teaching them how to observe the impact of their illness seems to have provided incentives for compliance which these patients had previously not experienced. A further set of interventions to improve compliance revolve around the treatment plan itself. One general principle of compliance-oriented prescribing is to simplify the regimen as much as possible. This may be done in a variety of ways. One of the most consistently successful strategies in this regard is the use, when available, of long-acting, injectable pharmaceuticals. The use of such preparations has been found to enhance both compliance and clinical outcome in streptococcal pharyngitis, rheumatic fever prophylaxis, schizophrenia, and tuberculosis.15 A related approach involves the scheduling of oral medication. In general, the physician would be well advised to avoid divided doses of medication when once a day administration would be equally effective. Although in many cases the possibility of implementing this strategy hinges on product availability, the increased attention to the problem of noncompliance on the part of the pharmaceutical industry has resulted in the recent introduction of a variety of products designed for once a day oral administration. These can often be used to good advantage in those patients for whom noncompliance is suspected to be a problem. The principle of keeping down the number of pills per day applies particularly in the case of multiple drug regimens. This principle dictates the avoidance on the part of the physician of routine prescription of such medications as vitamin supplements, bowel preparations, and tranquilizing drugs. The addition of such drugs may increase both errors in medication and noncompliance. Similarly, though they have been criticized on other grounds, the prescribing of combination drugs rather than several individual agents may improve compliance. Consideration should be given to the use of such preparations once the appropriate ratio of the drugs involved has been determined by separate administration. There are further ways in which the physician can simplify the treatment plan. Questioning the patient about his daily routine and
887
IMPROVING PATIENT COMPLIANCE
suggesting times for medication that involve the least disruption of that routine has been found to be beneficial. When behavior changes that are more difficult to comply with are involved, such as dietary habits, exercise habits, or smoking, the principles of be havi or modification 26 suggest the introduction of such changes gradually over the course of several visits rather than all at once. In these difficult areas, by reinforcing compliance as it occurs and only then adding a new task, changes in behavior will come more easily. The occurrence of side-effects has been found to hinder compliance. 6 If certain side-effects are common with the medication to be prescribed, the physician should indicate this in advance to the patient, adding, where appropriate, that the side-effect may be unpleasant but is not serious. It is also important to keep in mind that the monetary cost of following a regimen may interfere with compliance. The physician alert to the problem of noncompliance will investigate or delegate the investigation of the patient's health insurance coverage to ensure that he or she receives all due benefits. Prescribing by generic name can significantly reduce the cost of medication. Further, the patient should be informed that prescription prices can vary widely from pharmacy to pharmacy and that a pharmacy offering the best price on one product may not offer the best price on others. Patients should therefore be encouraged to request price information on a prescription before having it filled and be supported in the notion that there is no onus to comparative shopping for pharmaceuticals. After initial formulation of a treatment plan and presentation to the patient, physician follow-up can prove most important in promoting good compliance. The use of mailed appointment reminder cards and followup telephone calls have been found to be successful in encouraging patients to keep appointments with health care providers. ll ,28 Similarly, there is evidence that home visits by nonprofessional workers to help arrange transportation and babysitting and to schedule convenient appointment times can improve clinic attendance. Patients receiving written instructions concerning the proposed regimen have been found, as has been noted, to be more compliant than control groups.
SUMMARY Compliance can be problematic for any patient, irrespective of demographic factors and the illness involved. The most consistent predictors of compliance have been found to be certain health-related beliefs on the part of the patient, various characteristics of the regimen, and some aspects of the interaction between physician and patient. The health-related beliefs implicated in compliance are perceptions of the severity of the illness, susceptibility to future attacks or complications, and views of the costs and benefits associated with following the treatment plan. We believe that if physicians explore such patient beliefs, they will thereby (1) improve the quality of the physician-patient interac-
888
DARYL MATTHEWS AND RALPH HINGSON
tion and (2) learn much which can help them decide how best to promote good compliance in that particular patient. It is important to keep in mind that the strategy or strategies selected to improve compliance must be based on the beliefs and needs of the individual patient as determined in the compliance-oriented history; there is no universally effective technique. Possible interventions to promote compliance include fully informing the patient about the illness and the regimen, the use of emotional appeals, pointing out inconsistencies in the patient's belief system, the use of social supports, and the exposure of the patient to other sources of information. Further, the regimen can be simplified and tailored to the patient's individual activity pattern and careful follow-up by a variety of means can be provided. If the physician recalls that patient beliefs can change over time and continues to investigate them over the course of treatment, modifying his or her approach to the patient accordingly, good compliance will be a likely outcome.
REFERENCES 1. Avery, C. et al.: Reducing emergency room visits of asthmatics: an experiment in patient education. Testimony, President's Committee on Health Education, Pittsburgh, 1972. 2. Becker, M. et al.: Predicting mothers' compliance with pediatric medical regimens. J. Pediat., 81 :843, 1972. 3. Becker, M.: Sociobehavioral determinants of compliance. In Sackett, D., and Haynes, R., (eds.) Compliance With Therapeutic Regimens. Baltimore, Johns Hopkins University Press, 1976. 4. Bergman, A. B., and Werner, R. J.: Failure of children to receive penicillin by mouth. New Eng. J. Med., 268:1334, 1963. 5. Berkowitz, N.: Patient follow-through in the outpatient department. Nursing Research, 19:312, 1970. 6. Blackwell, B.: Patient compliance. New Eng. J. Med., 289:249, 1973. 7. Caron, H., and Roth, H.: Patient cooperation with medical regimens. J.A.M.A., 203 :922, 1968. 8. Colcher, 1., and Bass, J.: Penicillin treatment of streptococcal pharyngitis. J.A.M.A., 222:657,1972. 9. Davis, M.: Variations in patient compliance with doctor's advice: an empirical analysis of patterns of communication. Amer. J. Publ. Health, 58:274, 1968. 10. Elling, R.: Patient participation in a pediatric program. J. Health Human Behavior, 1 : 183, 1960. 11. Fletcher, S., et al.: Improving emergency-room patient follow-up in a metropolitan teaching hospital. New Eng. J. Med., 291 :385, 1974. 12. Francis, V., et al.: Gaps in doctor-patient communication. New Eng. J. Med., 280:535, 1969. 13. Gordis, L., et al.: Why patients don't follow medical advice: a study of children on long term anti-streptococcal prophylaxis. J. Pediat., 75:957, 1969. 14. Haynes, R.: A critical review of the determinants of patient compliance with therapeutic regimens. In Sackett and Haynes, eds.: Compliance With Therapeutic Regimens. Baltimore, Johns Hopkins University Press, 1976. 15. Haynes, R.: Strategies for improving compliance: a methodologic analysis and review. In Sackett and Haynes, eds.: Compliance With Therapeutic Regimens. Baltimore, Johns Hopkins University Press, 1976. 16. Haynes, R., et al.: Improvement of medication compliance in uncontrolled hypertension. Lancet, 1 :1265, 1976. 17. Heinzelman, F.: Factors in prophylaxis behavior in treating rheumatic fever. J. Health Human Behavior, 3 :73, 1962. 18. Hingson, R.: The physician's problems in identifying non-compliant patients. In Barofsky, 1., ed.: Medication Compliance-A Behavioral Management Approach. Charles B. Slack Co., in press. 19. Higbee, K.: Fifteen years of fear arousal: research on threat appeals. Psychological Bull., 72:426, 1969.
IMPROVING PATIENT COMPLIANCE
889
20. Hovland, C., et al.: Communication and Persuasion. New Haven, Yale University Press, 1953. 21. Hulka, B., et al.: Communication, compliance and concordance between physicians and patients with prescribed medications. Amer. J. Pub!. Health, 66:847,1976. 22. Korsch, B., et al.: Gaps in doctor/patient communication.!. Doctor/patient interaction and patient satisfaction. Pediatrics, 42 :855, 1968. 23. Leventhal, H.: Effect of fear communication in the acceptance of preventive health practices. Bul!. New York Acad. Med., 41 :11, 1965. 24. Ley, P., and Spelman, M.: Communication in an outpatient setting. Brit. J. Soc. Clin. Psycho!., 4 :115, 1965. 25. Marston, M.: Compliance with medical regimens: a review of the literature. Nursing Research, 19 :312, 1970. 26. Meyer, V., and Chesser, E.: Behaviour Therapy in Clinical Psychiatry. Baltimore, Penguin Books, 1970. 27. Mitchell, J.: Compliance with medical regimens: an annotated bibliography. Health Education Monograph, 2 :75, 1974. 28. Nazarian, L., et a!.: Effect of a mailed appointment reminder on appointment keeping. Pediatrics, 53 :349, 1974. 29. Pratt, L., et al.: Physician views on the level of medical information among patients. Amer. J. Pub!. Health, 47:1277, 1957. 30. Rosenstock, 1.: Why people use health services. Milbank Memorial Fund Quarterly, 44 :94, 1966. 31. Sackett, D., et al.: A randomized clinical trial of strategies for improving medication compliance in primary hypertension. Lancet, 1 :1205, 1975. Department of Socio-Medical Sciences and Community Medicine Boston University School of Medicine 80 East Concord Street Boston, Massachusetts 02118
Improving Patient Compliance A Guide for Physicians
Daryl Matthews, M.D.,* and Ralph Hingson, Sc.D.**
Patient compliance and lack of it has received widespread attention in recent years. If anyone conclusion can be reached from the spate of studies published on this topic, it is that physicians simply cannot expect all patients or, perhaps, even a majority of patients to adhere to the medical and therapeutic regimens they prescribe. Moreover, it is difficult for physicians to distinguish which patients are compliant and which are not. 5,7 Difficulty in distinguishing compliant from noncompliant patients can pose a serious obstacle to patient care. Unless physicians can determine the extent to which a patient is following a specific therapy, they cannot fully assess the utility of that therapy for the patient. On a broader scale, the value of specific treatments for a number of diseases may be questioned if the degree of patient compliance has not been assessed. The physician can no longer afford to simply take compliance for granted. Unless the physician takes the few extra minutes necessary to promote the highest level of patient compliance, his or her efforts, as well as those of other health professionals involved in the development and delivery of the regimen to the patient, may be jeopardized.
Factors Which Influence Compliance In the past few years a great deal of study has been devoted to this topic. The bulk of this research has consisted of descriptive comparisons of compliant and noncompliant patients. From this research we have learned a great deal about the influence of characteristics and attitudes of the patient, characteristics of the treatment regimen, and the physician-patient interaction on compliance. A number of articles review this literature in depth. 14, [8,25,27 Briefly summarized, this research informs us, first, that noncompliance can pose a problem regardless of the severity of a patient's disorder in terms of painfulness, disability, or threat to life. While non"'Assistant Professor, Department of Socio-Medical Sciences and Community Medicine and Division of Psychiatry, Boston University School of Medicine; Assistant Visiting Physician (Psychiatry), University Hospital, Boston, Massachusetts ':"'Assistant Professor, Department of Socio-Medical Sciences and Community Medicine, Boston University School of Medicine, Boston, Massachusetts Medical Clinics of North America-Vo!. 61, No. 4, July 1977
879
880
DARYL MATTHEWS AND RALPH HINGSON
compliance is particularly high when the disorder being treated is asynlptomatic or psychiatric in nature,14 numerous studies have shown that, in general, the severity of a disorder as measured by evaluation by a physician, prior hospitalizations, or prior diagnoses is unrelated to compliance. While one would think that people with more severe problems would be more motivated to comply, such disorders often require more complex, difficult regimens. Persons with severe disorders may find compliance difficult because of limitations imposed on them by their illnesses, and may also be discouraged by the failure of their previous efforts to prevent or cure their ailments. Just the fact that a person's illness is serious, painful, or even life threatening does not ensure a high level of compliance. Second, the patient's demographic characteristics do not consistently predict whether or not he will be compliant. Studies which report no relation between patient compliance and social class, age, sex, education, occupation, income, and marital status outnumber those which do almost three to one. 14 ,27 Surprising as such a finding may seem, the lack of association between demographic characteristics and compliance has been observed for a variety of disorders, in a variety of medical settings, using a variety of measurement techniques. 1R Though this does not mean that such demographic characteristics as social class never predict compliance, it does mean that noncompliance can be a problem with any patient population, regardless of social class level, age composition, or racial and ethnic background. Third, while common sense tells us that knowledge about one's illness and understanding how to follow the regimen are necessary for a patient to comply, they are not always sufficient. Graphically illustrating this point is a study by Bergman and Werner in which 59 families of children having streptococcal pharyngitis were followed. When questioned at home, 95 per cent of respondents correctly recalled the appropriate regimen. By the sixth day of a 10 day course of treatment, however, pill counts and urine assays revealed that only 30 per cent of the patients had received the prescribed penicillin. 4 More recently, Sackett et al. 31 conducted a randomized controlled trial to increase patient compliance with antihypertension regimens among more than 230 men. Although the experimental group receiving intensive instruction did show greater increases in knowledge about hypertension than a control group, the experimental group was no more compliant than the control group. In fact, measurements of patient knowledge at intake and 6 month follow-up in both the intervention and control groups were not related to compliance. 31 More consistently predictive of compliance are the nature of the regimen offered to patients, the beliefs patients hold about their illness and treatment, and certain features of the interaction between patients and their physicians. Some treatment plans are more difficult for patients to follow. The more complex a regimen, e.g., the greater the number of medications and the more times per day that medications must be taken, the less
IMPROVING PATIENT COMPLIANCE
881
likely is full compliance. Compliance tends to diminish as the duration of treatment increases over time. Regimens which require marked changes in behavior, especially abandoning established behaviors like smoking or drinking or certain dietary habits, are less likely to be followed than regimens which ask patients to initiate new behaviors that do not disrupt their normal routines. Regimens which produce adverse side-effects for the patient are also less likely to be followed. A number of studies have rather consistently demonstrated that one's beliefs about one's illness and its treatment can influence compliance independently of knowledge about the illness and treatment, the actual nature of the illness and treatment, or the characteristics of the regimen. These studies have explored the impact of the beliefs outlined in the "Health Belief Model" originally proposed to explain preventive health behaviors such as obtaining chest x-rays or immunizations. 30 Applied to compliance, this model suggests that patients (a) who feel susceptible to problems or complications because of their illness, susceptible to further attacks of the illness itself, (b) who believe that their illness could pose severe consequences for their health and daily functioning, (c) who feel that the proposed treatment plan will be highly effective in treating their illness, and (d) who do not foresee such major obstacles to compliance as adverse drug effects, cost of the regimen, or perceived lack of safety of a medication, will be more likely to follow regimens offered to them than patients not holding these beliefs. While very few studies have systematically explored all the major features of the model, a number of studies have examined individual component beliefs. 2.13 •17 In the most thorough review ofliterature on the health belief model as it applies to compliance to date, Becker concluded: "Although no single effort has provided (or could provide) convincing confirmation of all Health Belief Model Variables, most studies have produced internally consistent findings in the predicted direction which taken together yield relatively strong support for this conceptual model of compliance behavior. "3 Characteristics of the interaction between patients and physicians also seem to influence whether patients will adhere to their doctor's recommendations. In general, the more satisfied patients are with their physicians and the settings in which care is provided, the more likely they will be to comply. Many factors have been found to undermine the physician-patient interaction and thereby reduce compliance. We will mention only a few here. When physicians underestimate patient knowledge and ability to understand information about their disorders, they give their patients less information about their illness and treatment. Their patients, in turn, are less likely to actively inquire about their illness and treatment and are less likely to follow prescribed regimens. 29 Patients are also less likely to comply: (1) if they feel they are not held in adequate esteem by their physician;lO (2) if their physician actively seeks information from them without providing feedback about either why that information is being gathered or on the patient's condition;9 (3) iftension emerges during a physician-patient interaction and is not addressed or resolved;9 (4)
882
DARYL MATTHEWS AND RALPH HINGSON
if patients feel that their expectations are not being met or if they feel their physician is not behaving in a friendly manner. 12,22 Interactions between physician and patient are, understandably, often laden with emotion. The patient may be in pain, may be apprehensive about his or her health and the consequences of the illness, may find the setting of the office, clinic or hospital complex, impersonal, or frightening. Moreover, as most physicians quickly learn, patients bring many different motivations to their medical visits including the desire to reduce pain, to cure disease, to prevent illness, to discuss family or other life problems, to seek company and compassion, to confirm they are sick, or to learn they are well. Deciphering a patient's motivations and attempting to meet patient expectations in a potentially emotionally charged interaction is a difficult task which changes with each new patient. Though the pitfalls in physician-patient interaction can take varied and sometimes unpredictable forms, most research in the area leads to the conclusion that the physician-patient interaction can be an important determinant of compliance.
The Compliance-Oriented History As with other problems in medicine, the importance of history-taking in addressing the problem of noncompliance cannot be overstated. The extension of the medical history to include information on the patient's health-related beliefs can be of use to the practitioner in two distinct ways: (1) An understanding ofthese beliefs on the part of the practitioner can help to predict to what degree a patient is at risk for poor compliance. (2) Investigating the issues dictated by the Health Belief Model can contribute to the development of a relationship with the patient most likely to foster compliance. A systematic exploration of these beliefs can offer the physician more compliance-related information than is obtained in any other facet of the medical encounter. The model implies that the following areas merit attention: BELIEFS ABOUT THE ILLNESS ITSELF. These include beliefs about the seriousness of the disorder and, if applicable, beliefs about the likelihood perceived by the patient of suffering future episodes. Such beliefs can be ascertained by the physician by posing such questions as, "What is it that worries you the most about having developed high blood pressure?" "How likely do you think it is that your pain will return?" "What do you think is causing the problem you describe?" "Have you known anyone with a rash such as yours? How did things turn out for that person?" Questions such as these can identify misconceptions on the part of the patient which may be predictive of poor compliance. Such misconceptions, it has been found, may involve either under or overestimates of the seriousness of the disorder and the likelihood of recurrent episodes. 24 Depending, of course, on the disorder in question, it may be beliefs about the symptom rather than about the illness itself which must be investigated. Most women, for example, would be fearful of cancer upon detecting a breast mass. Even when the diagnosis of a benign lesion is made, the fear of cancer may remain. In the case of symptoms with
IMPROVING PATIENT COMPLIANCE
883
considerable meaning to patients, questions about perceived seriousness and susceptibility may reasonably be asked early in the visit, at the time when historical material is customarily elicited. With other disorders, including those which present asymptomatic ally, the practitioner must bear in mind that the relevant beliefs about the illness may only surface once the patient has been informed of the diagnosis. This requires that the questioning of the patient not be confined, as is often the case, to the early part of the visit. Rather, it must be reintroduced at various times throughout the encounter as information is provided to the patient. BELIEFS ABOUT THE TREATMENT PLAN. The Health Belief Model implies that patients undertake something like a "cost-benefit analysis" in deciding whether or not to follow medical instructions. Those patients who perceive many or serious costs associated with the treatment plan, while anticipating little benefit, are at risk for poor compliance. The costs of adhering to a therapeutic regimen may be more than just financial. They may also involve giving up familiar behavior patterns or acquiring new ones, the involvemept required of family members, the anticipated adverse effects of treatment, and the abandonment of other remedies from which the patient may have been deriving real or placebo benefit. A practitioner can assess such perceived costs by asking such questions as "What have you been doing for the pain on your own?" "Can you think of any problems you might have in taking these pills an hour before meals?" "Is there anything that worries you about having these monthly blood tests?" "Are you concerned about any particular side-effects of the medication ?" The patient's perception of the benefits of the proposed treatment includes notions about the efficacy of the regimen as well as ideas about what problems he or she might be spared if treatment is carried out as recommended. These beliefs may be discerned by asking, "How effective do you think the injections will be for you?" and "What do you think would happen if you forgot to take your medicine one night?" In implementing the suggestions offered here, the physician must recognize that patients' beliefs can change markedly from visit to visit. It is unusual for an individual to employ his or her physician as the sole source of health information. Friends, relatives, books, magazine articles, and other health professionals may be consulted between the time the diagnosis is made and the time treatment is completed. The information or misinformation obtained from such sources can have as much or more influence on compliance than what has been explained to the patient by the physician. In order to ensure optimal compliance with an ongoing treatment plan, the physician would be well advised to repeatedly assess the status of the beliefs considered above over the course of the several visits which may be necessary to manage a particular episode of illness. In addition to the more specific interventions which will be discussed below, there is reason to suspect that the mere process of taking a compliance-oriented history can do much to contribute to improved compliance. The use of the type of questions advocated here communicates to the patient that the physician is interested in the patient's own percep-
884
DARYL MATTHEWS AND RALPH HINGSON
tion of the illness and the patient's own appraisal of the treatment plan. This may serve to enhance the feeling on the part of the patient that the physician understands the complaint and has taken a friendly interest. Insofar as these feelings contribute toward increased patient satisfaction with the encounter, the likelihood of good compliance is increased. 12 Similarly, the taking of a history such as is being suggested here may contribute toward the patient's perceiving respect on the part of the physician, another perception which appears to be related to compliance. 1o Finally, having access to the information yielded by a complianceoriented history may mitigate against the physician's underestimating the level of the patient's information about the disorder, another aspect of the physician-patient relationship associated with poor compliance. 29
Interventions to Improve Compliance Once the physician has explored the patient's perceptions of his or her illness and treatment, several different steps can be taken to encourage the patient to follow the prescribed regimen. ':' The information elicited during the compliance-oriented history can be most useful in this regard. First, providing the patient with a clear description of the condition, how it is treated, and the implications of the illness for the patient when treated and untreated probably constitutes the most frequently tried way to promote compliance. While useful, merely presenting information does not always insure compliance. One recent study has further suggested that not merely describing a regimen but also discussing the specific purpose and function of the prescribed medications will enhance compliance. 21 To be sure that the patient understands what is said, the physician can ask the patient to repeat key parts of what he or she has been told, especially how to follow the prescribed regimen. Clearly written or typed instructions are also useful. 8 Whenever presenting new information to a patient, the physician should ask if there are any questions. Patients will generally hold some beliefs, correct or incorrect, about their illness before contacting their physician. Together with these old beliefs, which will have been held for varying lengths of time and with varying degrees ofintensity, new beliefs will emerge during the course of treatment. If the physician feels that the patient's beliefs are appropriate, he or she should attempt to reinforce them. This may be done by indicating to the patient that he or she seems to have a good understanding of the situation and by asking the patient to telephone if any problems emerge during the course of treatment. Alternately, asking the patient to write down questions or concerns as they come up may help the patient remember them at follow-up visits. A much more difficult situation exists when the patient holds erroneous beliefs about his or her illness or treatment plan. Several options are available to change such beliefs. Perhaps the most over-used strat"'Some of the material outlined in this section was presented by Dr. Marshall Becker at a lecture at Boston University School of Medicine, May 3, 1976.
IMPROVING PATIENT COMPLIANCE
885
egy is fear arousal. So many health care providers have attempted to scare their patients into complying that the literature about this approach is particularly extensive. Review articles on the use of feararousal have reached the same fundamental conclusion: fear arousal is not consistently successful and, in some situations, may actually hinder compliance. 19,23 Fear arousal technqiues can falter for several reasons. 20 Patients exposed to a fear arousal message may perceive the threatened outcome as improbable, inapplicable, unimportant, or remote in time. Patients advised, for example, to stop smoking because of its adverse health consequences can and do deny or rationalize the advice away. Moreover, even if a fear appeal succeeds in arousing sufficient anxiety, it may fail to produce the intended behavioral change because the physician's recommendation may not be regarded as very effective. In some instances, if the patient was already anxious about his or her disorder, use of scare techniques might actually make thinking about one's illness and treatment so painful that any action on the issue is avoided. If the patient is already deeply anxious about the illness, a far more helpful approach would be one which attempted to allay fears by pointing out the promises held by treatment. Although appeals to other emotions may not have been given the same attention as fear arousal, they may actually prove more effective in altering beliefs and behavior. For example, appeals to love, joy, or parental responsibility can be powerfully motivating in some individuals. Pointing out any inconsistencies that may exist in the structure of the patient's stated belief may also be helpful. Often a patient's ambivalence about an illness or its treatment needs to be addressed before he or she will comply. A patient may be reluctant to follow a particular regimen which is consciously or unconciously in conflict with other desired goals, be they work-related, recreational, or family-oriented. Another alternative is to appeal to sources of information in which the patient has or can develop considerable trust. It may be difficult, for example, for a seriously ill patient to fully trust the physician who has never experienced the problems associated with having the disorder in question. In such situations, it is frequently useful to have the patient speak with other individuals having similar conditions who have benefited from the suggested therapy. Though not as technically knowledgeable, such individuals may appear more credible to the patient than the physician. The use of social pressure and social support can also be of help to patients. Such strategies have long been used with behavioral problems such as obesity, smoking, and alcohol abuse. More recently, Avery et al. 1 examined the use of such techniques with asthmatic patients. Comparing asthmatic patients placed in discussion groups focusing on ways the patients could prevent asthma attacks with controls drawn from the same emergency room population, they found that subsequent visits to the emergency room by patients in the discussion groups during the next 4 months were but half the number made by controls. 1 Patients with similar problems need not be the only resource for this kind of strategy.
886
DARYL MATTHEWS AND RALPH HINGSON
In many circumstances family members, employers, and friends can be of invaluable assistance in helping patients follow their regimens and informing the physician if the patient appears to be having difficulty. Finally, other direct incentives can be identified and employed to encourage compliance. For example, many patients dislike being told forthrightly what to do or not to do concerning health matters. To the extent that decision-making and follow-up responsibility can be shared with them, their compliance may be improved. In a recent experimental study, Haynes et al. 16 divided into two groups a sample of patients who, at 6 months after the initiation of antihypertension therapy, were not compliant. The experimental group was given instructions about how to take and record home blood pressures and chart their medication consumption. With each patient, ways were discussed to tailor his regimen into habitual daily activities. Six months later, 80 per cent of those in the experimental group were compliant, twice the percentage of compliant control group patients. 16 Involving patients in a decision-making capacity and teaching them how to observe the impact of their illness seems to have provided incentives for compliance which these patients had previously not experienced. A further set of interventions to improve compliance revolve around the treatment plan itself. One general principle of compliance-oriented prescribing is to simplify the regimen as much as possible. This may be done in a variety of ways. One of the most consistently successful strategies in this regard is the use, when available, of long-acting, injectable pharmaceuticals. The use of such preparations has been found to enhance both compliance and clinical outcome in streptococcal pharyngitis, rheumatic fever prophylaxis, schizophrenia, and tuberculosis.15 A related approach involves the scheduling of oral medication. In general, the physician would be well advised to avoid divided doses of medication when once a day administration would be equally effective. Although in many cases the possibility of implementing this strategy hinges on product availability, the increased attention to the problem of noncompliance on the part of the pharmaceutical industry has resulted in the recent introduction of a variety of products designed for once a day oral administration. These can often be used to good advantage in those patients for whom noncompliance is suspected to be a problem. The principle of keeping down the number of pills per day applies particularly in the case of multiple drug regimens. This principle dictates the avoidance on the part of the physician of routine prescription of such medications as vitamin supplements, bowel preparations, and tranquilizing drugs. The addition of such drugs may increase both errors in medication and noncompliance. Similarly, though they have been criticized on other grounds, the prescribing of combination drugs rather than several individual agents may improve compliance. Consideration should be given to the use of such preparations once the appropriate ratio of the drugs involved has been determined by separate administration. There are further ways in which the physician can simplify the treatment plan. Questioning the patient about his daily routine and
887
IMPROVING PATIENT COMPLIANCE
suggesting times for medication that involve the least disruption of that routine has been found to be beneficial. When behavior changes that are more difficult to comply with are involved, such as dietary habits, exercise habits, or smoking, the principles of be havi or modification 26 suggest the introduction of such changes gradually over the course of several visits rather than all at once. In these difficult areas, by reinforcing compliance as it occurs and only then adding a new task, changes in behavior will come more easily. The occurrence of side-effects has been found to hinder compliance. 6 If certain side-effects are common with the medication to be prescribed, the physician should indicate this in advance to the patient, adding, where appropriate, that the side-effect may be unpleasant but is not serious. It is also important to keep in mind that the monetary cost of following a regimen may interfere with compliance. The physician alert to the problem of noncompliance will investigate or delegate the investigation of the patient's health insurance coverage to ensure that he or she receives all due benefits. Prescribing by generic name can significantly reduce the cost of medication. Further, the patient should be informed that prescription prices can vary widely from pharmacy to pharmacy and that a pharmacy offering the best price on one product may not offer the best price on others. Patients should therefore be encouraged to request price information on a prescription before having it filled and be supported in the notion that there is no onus to comparative shopping for pharmaceuticals. After initial formulation of a treatment plan and presentation to the patient, physician follow-up can prove most important in promoting good compliance. The use of mailed appointment reminder cards and followup telephone calls have been found to be successful in encouraging patients to keep appointments with health care providers. ll ,28 Similarly, there is evidence that home visits by nonprofessional workers to help arrange transportation and babysitting and to schedule convenient appointment times can improve clinic attendance. Patients receiving written instructions concerning the proposed regimen have been found, as has been noted, to be more compliant than control groups.
SUMMARY Compliance can be problematic for any patient, irrespective of demographic factors and the illness involved. The most consistent predictors of compliance have been found to be certain health-related beliefs on the part of the patient, various characteristics of the regimen, and some aspects of the interaction between physician and patient. The health-related beliefs implicated in compliance are perceptions of the severity of the illness, susceptibility to future attacks or complications, and views of the costs and benefits associated with following the treatment plan. We believe that if physicians explore such patient beliefs, they will thereby (1) improve the quality of the physician-patient interac-
888
DARYL MATTHEWS AND RALPH HINGSON
tion and (2) learn much which can help them decide how best to promote good compliance in that particular patient. It is important to keep in mind that the strategy or strategies selected to improve compliance must be based on the beliefs and needs of the individual patient as determined in the compliance-oriented history; there is no universally effective technique. Possible interventions to promote compliance include fully informing the patient about the illness and the regimen, the use of emotional appeals, pointing out inconsistencies in the patient's belief system, the use of social supports, and the exposure of the patient to other sources of information. Further, the regimen can be simplified and tailored to the patient's individual activity pattern and careful follow-up by a variety of means can be provided. If the physician recalls that patient beliefs can change over time and continues to investigate them over the course of treatment, modifying his or her approach to the patient accordingly, good compliance will be a likely outcome.
REFERENCES 1. Avery, C. et al.: Reducing emergency room visits of asthmatics: an experiment in patient education. Testimony, President's Committee on Health Education, Pittsburgh, 1972. 2. Becker, M. et al.: Predicting mothers' compliance with pediatric medical regimens. J. Pediat., 81 :843, 1972. 3. Becker, M.: Sociobehavioral determinants of compliance. In Sackett, D., and Haynes, R., (eds.) Compliance With Therapeutic Regimens. Baltimore, Johns Hopkins University Press, 1976. 4. Bergman, A. B., and Werner, R. J.: Failure of children to receive penicillin by mouth. New Eng. J. Med., 268:1334, 1963. 5. Berkowitz, N.: Patient follow-through in the outpatient department. Nursing Research, 19:312, 1970. 6. Blackwell, B.: Patient compliance. New Eng. J. Med., 289:249, 1973. 7. Caron, H., and Roth, H.: Patient cooperation with medical regimens. J.A.M.A., 203 :922, 1968. 8. Colcher, 1., and Bass, J.: Penicillin treatment of streptococcal pharyngitis. J.A.M.A., 222:657,1972. 9. Davis, M.: Variations in patient compliance with doctor's advice: an empirical analysis of patterns of communication. Amer. J. Publ. Health, 58:274, 1968. 10. Elling, R.: Patient participation in a pediatric program. J. Health Human Behavior, 1 : 183, 1960. 11. Fletcher, S., et al.: Improving emergency-room patient follow-up in a metropolitan teaching hospital. New Eng. J. Med., 291 :385, 1974. 12. Francis, V., et al.: Gaps in doctor-patient communication. New Eng. J. Med., 280:535, 1969. 13. Gordis, L., et al.: Why patients don't follow medical advice: a study of children on long term anti-streptococcal prophylaxis. J. Pediat., 75:957, 1969. 14. Haynes, R.: A critical review of the determinants of patient compliance with therapeutic regimens. In Sackett and Haynes, eds.: Compliance With Therapeutic Regimens. Baltimore, Johns Hopkins University Press, 1976. 15. Haynes, R.: Strategies for improving compliance: a methodologic analysis and review. In Sackett and Haynes, eds.: Compliance With Therapeutic Regimens. Baltimore, Johns Hopkins University Press, 1976. 16. Haynes, R., et al.: Improvement of medication compliance in uncontrolled hypertension. Lancet, 1 :1265, 1976. 17. Heinzelman, F.: Factors in prophylaxis behavior in treating rheumatic fever. J. Health Human Behavior, 3 :73, 1962. 18. Hingson, R.: The physician's problems in identifying non-compliant patients. In Barofsky, 1., ed.: Medication Compliance-A Behavioral Management Approach. Charles B. Slack Co., in press. 19. Higbee, K.: Fifteen years of fear arousal: research on threat appeals. Psychological Bull., 72:426, 1969.
IMPROVING PATIENT COMPLIANCE
889
20. Hovland, C., et al.: Communication and Persuasion. New Haven, Yale University Press, 1953. 21. Hulka, B., et al.: Communication, compliance and concordance between physicians and patients with prescribed medications. Amer. J. Pub!. Health, 66:847,1976. 22. Korsch, B., et al.: Gaps in doctor/patient communication.!. Doctor/patient interaction and patient satisfaction. Pediatrics, 42 :855, 1968. 23. Leventhal, H.: Effect of fear communication in the acceptance of preventive health practices. Bul!. New York Acad. Med., 41 :11, 1965. 24. Ley, P., and Spelman, M.: Communication in an outpatient setting. Brit. J. Soc. Clin. Psycho!., 4 :115, 1965. 25. Marston, M.: Compliance with medical regimens: a review of the literature. Nursing Research, 19 :312, 1970. 26. Meyer, V., and Chesser, E.: Behaviour Therapy in Clinical Psychiatry. Baltimore, Penguin Books, 1970. 27. Mitchell, J.: Compliance with medical regimens: an annotated bibliography. Health Education Monograph, 2 :75, 1974. 28. Nazarian, L., et a!.: Effect of a mailed appointment reminder on appointment keeping. Pediatrics, 53 :349, 1974. 29. Pratt, L., et al.: Physician views on the level of medical information among patients. Amer. J. Pub!. Health, 47:1277, 1957. 30. Rosenstock, 1.: Why people use health services. Milbank Memorial Fund Quarterly, 44 :94, 1966. 31. Sackett, D., et al.: A randomized clinical trial of strategies for improving medication compliance in primary hypertension. Lancet, 1 :1205, 1975. Department of Socio-Medical Sciences and Community Medicine Boston University School of Medicine 80 East Concord Street Boston, Massachusetts 02118
