A greatly increased level and duration of suppression of HIV replication has become possible with the use of protease inhibitors in combination with nucleoside analogues as antiretroviral therapy [1,2]. The long-term effectiveness of protease inhibitors, and other antiretroviral medications, is dependent upon strict adherence to the prescribed regimen, since HIV resistance to these drugs can develop with subtherapeutic doses [3,4]. Consequences of poor adherence include not only diminished outcome for the patient, but also the public health threat of multidrug-resistant HIV, and widespread transmission of drug-resistant virus, similar to that seen with multidrug-resistant tuberculosis .
The aim of this article is to review the scientific literature on drug adherence in studies based on methodologic criteria and among patient populations relevant to HIV infection. We reviewed the adherence literature on patients with HIV infection and on two other groups that may have relevance to the adherence issues faced by HIV-infected individuals: patients with chronic illness and patients with mental illness. Chronic illness was chosen because it is well-established that HIV infection has a median incubation period of 10 years or more , with multiple comorbid complications, and complex treatment regimens. Mental illness was chosen because psychiatric diagnoses are common in patients who acquire HIV [7–9]. We divided factors associated with decreased adherence into four broad categories: demographic characteristics, psychosocial/ behavioral characteristics, clinical aspects, and health-care administration and delivery.
Much of the literature use the words ‘compliance’ and ‘adherence’ interchangeably, although adherence has come to be considered the less pejorative term. Although they are inherently related, several articles reviewed described individual definitions of the terms.
Adherence relates to the extent to which the patient follows a prescribed regimen [10–12]. For example, a patient may take his/her medication 80% of the time. Conversely, Young et al.  reported adherence in terms of a default rate from follow-up, or significant deviation from a prescribed regimen. The specific characteristics of what the adherence or default rate is measuring depends on the prescribed regimen; for example, it may apply to medication-taking behaviors, dietary intake or physical therapy.
Compliance, on the other hand, is an overall evaluation of adherence. Those patients who adhere to their prescribed medical regimen 80% of the time may be considered to show high compliance, whereas those patients who demonstrate 50% adherence may be demonstrating moderate or low compliance. The level of compliance may vary between patients depending on the type of medicine prescribed, the type of illness, and the accompanying level of symptoms. Evaluation of compliance should also consider the threshold of therapy necessary for effectiveness. For example, with triple combination antiretroviral therapy, it is necessary to take a high proportion (90–100%) of doses to maintain suppression of viral replication [13,14]. In contrast, prophylactic therapy for Pneumocystis carinii pneumonia may be forgiving of an adherence of less than 90–100% [15–17].
Although it has been well demonstrated that certain demographic and HIV risk groups have poorer access to therapy [e.g., African-Americans [18,19] and injecting drug users (IDU) ] few data are available on the influence of these factors on adherence among HIV-infected patients. Demographic characteristics have been associated with decreased adherence in a variety of disorders, including chronic illness, mental illness, and amongst the elderly.
Adherence increases with age, except in the most elderly (those aged over 75 years) . Often, the very elderly have more complex medical regimens and more comorbidities, such as vision, hearing, or memory impairment. In comparison, studies of medication adherence in patients aged under 75 years with chronic or mental illness showed a positive correlation between age and adherence [10,11,22,23]. In one prospective evaluation of patients with HIV that addressed this issue, age did not correlate with adherence . However, HIV-infected individuals are typically no older than 50 years . Thus, risks associated with old age may not be a major issue.
In several studies of patients with HIV infection, chronic illness, mental illness, and the elderly, male gender was associated with decreased adherence [10,11,26–28]. However, other studies have not found a significant association between men and decreased adherence. Notably, in one study of HIV-infected patients, it was observed that women were missing a higher percentage of clinic visits than men. As a result, free on-site child care was provided, and adherence with scheduled visits improved dramatically . In situations where women are responsible for care of the household and children, immediate concerns or practical barriers may adversely affect adherence among women more than among men. As rates of HIV infection are increasing among women [29,30], the extent to which these barriers can be alleviated may have a significant impact on treatment outcome.
Lower socioeconomic status (SES) has been shown to be another contributor to decreased adherence [11,21,26–28]. The components of lower SES associated with decreased adherence are unstable or poor housing, low income, low level of education [12,23,24] and lack of medical insurance , the latter factor particularly important in the United States. Low income and lack of insurance may prevent patients from easily accessing health care, in terms of purchasing medications, transportation, or child care. Poor housing may either contribute directly to lower adherence by preoccupying the patient with more immediate concerns, or be an indirect indicator of low social support and educational status. These risk factors may also apply to IDU. In a cohort study of HIV-infected IDU, unstable or poor housing, low income, and lack of insurance were associated with lower adherence to zidovudine treatment . One study found that low adherence in urban indigent patients with HIV was associated with limited health resources and access to health assistance until significant immunosuppression occurred . Medical assistance and other public sector programs may need to provide resources earlier in the disease course to maximize compliance.
The association of race with adherence appears to vary with different illnesses. In a study of adherence to hypertension treatment in an inner-city population, race was not found to be associated with adherence . Similarly, in a study of systemic lupus erythematosus, there was no association between race and adherence when controlling for SES and psychosocial factors . However, two studies of HIV-infected individuals showed that blacks were significantly more likely to be non-adherent, independently of injecting drug use and educational status [24,25]. The authors suggested that the reasons for this may be multifactorial, such as differences in health beliefs or social support structures . In the United States in 1994, African American men were becoming HIV-infected at a rate four times higher than white men, and African American women were becoming HIV-infected at a rate nine times higher than white women. AIDS is now the leading cause of death among African-American men and women aged 25–44 years [25,29,30]. Interventions aimed at increasing compliance among minorities may have to take into account cultural and social network issues independent of SES, level of education, and drug use.
Presence of psychiatric illness is commonly associated with decreased adherence in patients with mental illness, elderly status, and HIV infection [10,12,23,31,34–36]. Other psychological factors associated with lower adherence among the mentally ill are hostility, guilt, anxiety, paranoia, and grandiosity [10,35]. In a prospective study of HIV-infected individuals, compliant patients (defined as ≥ 80% adherence) had significantly less depression than non-compliant patients . In a large study of zidovudine compliance, only 52.6% of patients diagnosed with psychiatric illness had good compliance, compared with 81% compliance in those without psychiatric diagnosis (P < 0.001) . Major depression is the most commonly cited psychiatric problem among HIV-infected individuals, with prevalence rates ranging from 17 to 30% . It has been suggested that active psychiatric intervention should be an important component in the effective management of HIV-infected patients .
Patient attitudes and beliefs
Negative attitudes about medications or illness may also interfere with patient adherence. Amongst the mentally ill, reasons cited for not taking medications were fear of addiction and the belief that medication use was a sign of weakness . The patient's perception of how beneficial treatment would be in affecting illness outcome was also associated with compliance. Amongst HIV-infected patients, attitudes and beliefs related to decreased adherence included the patient's acceptance/perception of disease, and perceived lack of benefit . Specifically, scepticism about zidovudine (perceived lack of benefit) and denial of necessity of treatment have been associated with decreased adherence [23,37,38]. Conversely, the belief that zidovudine prolongs life has been strongly associated with increased adherence [40,41]. Perceived lack of benefit and lack of necessity for treatment can also be risk factors among patients with chronic illness.
Poor social relationships and activities have been associated with lower adherence in several studies. In the mentally ill, lack of involvement by family and friends, social isolation, and living alone, were found to be risk factors . In comparison, a broader array of social activities had a positive effect on adherence . Living alone and lack of support have also been associated with increased risk of non-adherence in the elderly. Kissinger et al.  speculated that HIV infection is a potential cause of social isolation. Thus, social isolation may be a risk factor for decreased compliance in HIV-infected individuals. However, in one study, higher perceived social support was not found to be significantly associated with adherence .
Form of medication
In their review of adherence in ambulatory patients with schizophrenia, Young et al.  presented data summarizing the observed ranges of default rates with oral medications in 28 studies. Default rates for oral medications ranged from 10 to 76%; injected medication default rates ranged from 14 to 36%, but in general were considerably lower than rates for oral medication. Due to the simplifying nature of the form of the medication, the depot form of medication seems to have a higher rate of adherence. A higher rate of adherence to depot medication has also been shown by comparing use of oral contraceptives with levonorgestrel implants (a sustained-release contraceptive that is implanted subcutaneously) . In addition to reducing dosing frequency, depot preparations are also more likely to be supervised. Like depot injections, oral controlled-release drugs, such as procainamide for cardiovascular use, can decrease side-effects and prolong dosing intervals, increasing patient compliance . Using drugs with longer half-lives can also reduce dosing and prolong dosing intervals . Although depot forms are not likely to be an option in HIV disease, combined pill forms, longer half-life drugs (e.g., single daily dose), or long-acting controlled-release forms may become important strategies in improving compliance with antiretroviral therapy.
Amongst the mentally ill, specific side-effects are associated with decreased adherence [10,36]. In schizophrenic patients, an initial dysphoric response to thiothixene, or extrapyramidal side-effects related to fluphenazine decanoate have been associated with medication cessation or reduced dosage . In a small subset of mentally ill patients, discontinuance of medication occurred as a result of fear of side-effects . Undesirable side-effects were also a risk factor for non-adherence in chronically ill and elderly patients, such as insulin cessation in diabetic patients . In patients with HIV infection, major side-effects leading to decreased adherence and treatment cessation are leukopenia, anemia, transfusion, and gastrointestinal upset [31,40,41,45], and one reason for refusal of zidovudine is fear of side-effects [31,41]. It is possible that combination antiretroviral regimens will result in a greater risk of adverse effects. Given that the prognosis of AIDS is so poor in untreated patients, patients with HIV may tolerate side-effects better than patients with less severe chronic diseases such as hypertension or diabetes.
Complexity of prescribed regimen
As the complexity of the prescribed regimen increases, so do rates of non-adherence [10,28,40]. Moreover, as a prescribed regimen becomes more complex, it also becomes more inconvenient and difficult to incorporate into daily living. A medication timer may be one possible solution, and has been shown to increase compliance in a study of HIV-infected patients . Nevertheless, in patients who are less motivated, who have less social reinforcement, or who live disordered lives (e.g., IDU, alcoholics), regimen complexity may prove to be an insurmountable barrier.
Health-care administration and delivery
Patient non-adherence has been found to be significantly associated with the patient not knowing the correct dose of medication or that chronic medications have to be taken continuously [27,28,36,39]. In a sample of elderly adults, 50% of the non-adherent group reported skipping a prescribed drug because they did not believe they needed it . A group of non-adherent mentally ill patients claimed remission of symptoms as a reason for medication cessation . Although this may not be of concern for patients with chronic or acute illness and physical symptoms, a study of coronary artery disease in the elderly has suggested that patients may be more adherent to medications that relieve their symptoms (i.e., they may be more likely to take their antihypertensives than their prophylactic aspirin). Conversely, instead of stopping medications because they are feeling better, some patients abandon their medications due to delay of clinical response, or failure to provide a ‘cure’ within a limited period . If these results extend to patients with HIV infection, it becomes important to explain to patients the necessity of continued medication during asymptomatic periods. In the absence of symptom cues, informing patients of their HIV viral load and CD4 count during chronic antiretroviral therapy may assist adherence.
Patients' lack of knowledge of their diagnosis and the expected course of their illness or treatment has also been associated with decreased adherence . One study found that patients who learned the names of their medications were more adherent than those who did not know the medication names .
Health-care practitioner–patient communication
Patient education as to the correct use of medications relies primarily on health-care providers. Even when patients have received proper instruction, it may be difficult to remember to take medications as prescribed. In these situations, the practitioner can work with the patient to incorporate the individual drug regimen into a daily schedule. Several strategies have been suggested: timed pill dispensers, alarm clocks, and having someone else to act as a reminder.
In patients with mental illness, specifically targeted education about illness treatment and medications has improved adherence . In a study of hypertensive patients, adherence increased proportionally with the number of times the patient talked to a doctor about hypertension . Lack of a primary care physician also carried an increased risk of decreased adherence . However, one study of adherence with zidovudine in HIV-infected individuals, which found that IDU were more likely to refuse zidovudine therapy, suggests that physicians may view IDU negatively or insist less on the importance of antiretroviral treatment in case of refusal [31,47].
Several of these studies have suggested more active involvement by health-care providers in providing support and educating patients about treatment benefits. Methods may include closer supervision by scheduling more visits or calling patients periodically. Unfortunately, supervised therapy such as that used for treatment of tuberculosis, is less likely to be useful in HIV infection, because of the typical need to ingest medication more than once daily.
Extrinsic barriers to treatment
Health-care administration and delivery often involves physical barriers to patient adherence, including cost, lack of transportation, lack of child care, severe illness, place of treatment, and lack of primary care physician [21,26–28]. Cost of treatment was repeatedly cited as a risk of decreased compliance in all three patient groups [28–30]. Nevertheless, even when cost of health care was subsidized, other costs adversely affected adherence. In a group of diabetic patients, although insulin cost was subsidized at the hospital, 50% of patients who stopped taking insulin did so because of lack of financial resources for transportation to a pharmacy . In a group of HIV-infected patients, it was found that more severely immunocompromised patients were more likely to miss their visits. The investigators suggested that severely immunocompromised patients may have an increased need for the provision of hospice and home-based care . In this same study, it was found that providing free on-site child care greatly improved compliance with scheduled appointments.
Possible remedies to the barriers may include community-based or individual care. A study of appointment compliance in a community-based clinic serving minority and low income communities found that adherence was high (85% of initial scheduled appointments kept), and was even higher (90%) when considering rescheduled appointments . The comparison was with a hospital-based ambulatory care network serving the same population, which had an appointment adherence rate of 60%. Reasons for missed appointments were conflicts with work schedules, lack of child care, no transportation, family illness, and hospitalization. Thus, community-based clinics may help solve many of these problems. In HIV-infected patients, there is a need for community-based primary care facilities that offer a combination of therapeutic options, not just for HIV care alone, but also for drug use, gynecologic screening, sexually transmitted disease treatment, mental health, and general health care.
Our review has focused on demographic and psychosocial/behavioral characteristics, the medication regimen, and health-care administration and delivery as characteristics potentially relevant to adherence in the HIV-infected patient. These factors are summarized in Table 1. Clearly, there is an overlap between the characteristics studied and these variables cannot therefore be considered in isolation. Strategies to improve adherence in many HIV-infected patients will need to focus on the combined psychosocial and health-care access problems of the poor, as well as the special barriers to women and IDU. Many of the suggested solutions to problems with adherence in non-HIV-infected patients may be applicable in patients with HIV infection. In particular, strategies specific to women, the poor, and IDU with HIV are likely to be necessary, including expansion of availability of drug treatment, primary care provision, and the possible introduction of simpler medication regimens in the future.
1. Goebel FD: Combination therapy from a clinician's perspective
. J Acquir Immune Defic Syndr
2. Vella S: HIV pathogenesis and treatment strategies
. J Acquir Immune Defic Syndr
3. Cinatl Jr J, Cinatl J, Rabenau H, Doerr HW, Weber B: Failure of antiretroviral therapy: role of viral and cellular factors
4. Moutouh L, Corbeil J, Richman DD: Recombination leads to the rapid emergence of HIV-1 dually resistant mutants under selective drug pressure
. Proc Natl Acad Sci USA
5. Edlin BR, Toakrs JI, Grieco MH, et al.
: An outbreak of multidrug-resistant tuberculosis among hospitalized patients with the acquired immunodeficiency syndrome
. N Engl J Med
6. Munoz A, Wang, M, Bass S, et al.
: Acquired immunodeficiency syndrome (AIDS)-free time after human immunodeficiency virus type 1 (HIV-1) seroconversion in homosexual men
. Am J Epidemiol
7. Lyketsos C, Hanson AL, Fishman M, et al.
: Mania early and late in the course of HIV
. Am J Psychiatry
8. Harris MJ, Jeste DV, Gleghorn A, et al.
: New-onset psychosis in HIV-infected patients
. J Clin Psychiatry
9. Neugebauer R, Rabkin JG, Williams JB, et al.
: Bereavement reactions among homosexual men experiencing multiple losses in the AIDS epidemic
. Am J Psychiatry
10. Young J, Howard Z, Shepler L: Medication noncompliance in schizophrenia: codification and update
. Bull Am Acad Psychiatry Law
11. Daniels D, Rene A, Daniels V: Race: an explanation of patient compliance – fact or fiction?J Natl Med Assoc
12. Coons S, Sheahan S, Martin S, Hendricks J, Robbins C, Johnson J: Predictors of medication noncompliance in a sample of older adults
. Clin Therapeut
13. Myers MW, Montaner JG: A randomized, double-blinded comparative trial of the effects of zidovudine, didanosine and nevirapine combinations in antiviral naive, AIDS-free, HIV-infected patients with CD4 counts 200–600/mm3
. XI International Conference on AIDS
. Vancouver, July 1996 [abstract MoB294].
14. Gulick RM, Mellors J, Havlir D, et al.
: Potent and sustained antiretroviral activity of indinavir (IDV), zidovudine (ZDV), and lamivudine (3TC)
. XI International Conference on AIDS
. Vancouver, July 1996 [abstract ThB931].
15. Hughes WT, Rivera GK, Schell MJ, Thornton D, Lott L: Successful intermittent chemoprophylaxis forPneumocysits cariniipneumonitis
. N Engl J Med
16. Ruskin J, LaRiviere M: Low-dose cotrimoxazole for prevention of Pneumocysitis carinii pneumonia in human immunodeficiency virus disease
17. Stein DS, Stevens RC, Terry D, et al.
: Use of low-dose trimethoprim–sulfamethoxazole thrice weekly for primary and secondary prophylaxis ofPneumocystic cariniiin human immunodeficiency virus-infected patients
. Antimicrob Agents Chemother
18. Moore RD, Stanton D, Gopalan R, Chaisson RE: Racial differences in the use of drug therapy for HIV disease in an urban community
. N Engl J Med
19. Graham NMH, Jacobson LP, Kuo V, Chmiel JH, Morgenstern H, Zucconi SL: Access to therapy in the multicenter AIDS cohort study
. J Clin Epidemiol
20. Solomon L, Vlahov D, Astemborski J, Galai N, Graham NMH, Nelson KE: Factors associated with initiation of zidovudine in a cohort of injection drug users
. J Drug Issues
21. Fedder DO: Drug use in the elderly: issues of noncompliance
. Drug Intell Clin Pharm
22. Shea S, Misra D, Ehrlich M, Field L, Francis C: Correlates of nonadherence to hypertension treatment in an inner-city minority population
. Am J Public Health
23. Draine J, Solomon P: Explaining attitudes toward medication compliance among a seriously mentally ill population
. J Nerv Ment Dis
24. Singh N, Squier C, Hayes P, et al.
: Determinants of compliance in patients with HIV: prospective assessment with implications for enhancing compliance
. 34th Interscience Conference on Antimicrobial Agents and Chemotherapy
. Orlando, October 1994 [abstract I170].
25. Centers for Disease Control and Prevention: HIV/AIDS Surveillance Report.
June 1996. Atlanta: CDC; 1996.
26. Kissinger P, Cohen D, Brandon W, Rice J, Morse A, Clark R: Compliance with public sector HIV medical care
. J Natl Med Assoc
27. Musey V, Lee J, Crawford R, Klatka M, McAdams D, Phillips L: Diabetes in urban African-Americans. I. Cessation of insulin therapy is the major precipitating cause of diabetic ketoacidosis
. Diabetes Care
28. Wallsten S, Sullivan R, Hanlon J, Blazer D, Tyrey M, Westlund R: Medication taking behaviors in the high- and low-functioning elderly: MacArthur field studies of successful aging
. Ann Pharmacother
29. Centers for Disease Control and Prevention: Update: mortality attributable to HIV infection among persons 25–44 years – United States, 1994
30. Centers for Disease Control and Prevention: HIV testing among women aged 18–44 years – United States, 1991 and 1993
31. Broers B, Morabia A, Hirschel B: A cohort study of drug users' compliance with zidovudine treatment
. Arch Intern Med
32. Francis H, Armstrong C, Chandler K, Grinage Y, McCree C: Nonbiologic constraints on early HIV intervention programs
. Third Conference on Retroviruses and Opportunistic Infections
. Washington, DC, January–February 1996 [abstract 276].
33. Karlson EW, Daltroy LH, Lew RA, et al.
: The relationship of socioeconomic status, race, and modifiable risk factors to outcome in patients with systemic lupus erythematosus
. Arthritis Rheum
34. Carney R, Freedland K, Eisen S, Rich M, Jaffe A: Major depression and medication adherence in elderly patients with coronary artery disase
. Health Psychol
35. Pugh R: An association between hostility and poor adherence to treatment in patients suffering from depression
. Br J Med Psychol
36. Johnson DW: Depression: treatment compliance in general practice
. Acta Psychiatr Scand Suppl
37. Fernandez F, Ruiz P: Psychiatric aspects of HIV disease
. South Med J
38. John JK, Kumar TM: Psychiatric challenges in HIV management
. XIII International Conference on AIDS
. Amsterdam, July 1992 [abstract PuB7263].
39. Youssef F: Adherence to therapy in psychiatric patients: an empirical investigation
. Int J Nurs Stud
40. Samet J, Libman H, Steger K, et al.
: Compliance with zidovudine therapy in patients infected with human immunodeficiency virus, type 1: a cross-sectional study in a municipal hospital clinic
. Am J Med
41. Samuels J, Hendrix J, Hilton M, Marantz P, Sloan V, Small C: Zidovudine therapy in an inner city population
. J Acquir Immune Defic Syndr
42. Glantz S, Schaff E, Campbell-Heider N, Glantz JC, Bartlett M: Contraceptive implant use among inner city teens
. J Adolesc Health
43. Arnold RJ, Kaniecki DJ: Selection of oral controlled-release drugs: a critical decision for the Physician
. South Med J
44. Croteau D, Bergeron MG, LeBel M: Pharmacokinetic advantages of erythromycin estolate over ethylsuccinate as determined by high-pressure liquid chromatography
. Antimicrob Agents Chemother
45. Easterbrook P, Keruly J, Creagh-Kirk T, Richman D, Chaisson R, Moore R: Racial and ethnic differences in outcome in zidovudine-treated patients with advanced HIV disease
46. Hulka BS, Cassell JC, Kupper LL, Burdette JA: Communication, compliance, and concordance between physicians and patients with prescribed medications
. Am J Publ Health
47. Gerbert B, Maguire BT, Bleeker T, et al.
: Primary care physicians and AIDS: attitudinal and structural barriers to care
48. Norris B, Davis I, Leibel J, Laqueur P, Morgan A, Paroski P: Evaluation of compliance rates in a clinic serving minority and low income communities
. VI International Conference on AIDS
. San Francisco, June 1990 [abstract SD769].