Evaluation of an Intervention on Adherence to Highly Active Antiretroviral Therapy (HAART) in HIV Infected Adults

--Background-- The management of HIV has changed dramatically since the introduction of highly active antiretroviral therapy (HAART). HAART consists of a drug regimen that targets the virus at various points in its life cycle and is thus more effective than monotherapy. The drugs that comprise this regimen usually belong to different classes of antiretrovirals. HAART regimens can be complex and difficult to tolerate. They usually consist of multiple pills that have to be taken two to three times a day and may also need to be taken with food, on an empty stomach or with plenty of fluids (depending on the particular drug). In addition, they have multiple side effects that may limit the patient's willingness to take them. Adherance to HAART is of paramount importance. Paterson et al showed that patients should take 95% of their prescribed doses in order to achieve a high likelihood of success and maintaining an undetectable viral load. Poor adherence can lead to serious consequences including failure to prevent viral replication, shortened survival and development of viral resistance. Various methods have been used to measure adherence to HAART. All these methods have their limitations and no gold standard exists with which they can be compared. The easiest methods employ questionnaires, interviews or written documentation (ie diaries). These can be supplemented by a pill identification test. This test asks the patients to examine a board displaying two similar pills for each antiretroviral drug and to identify which they have been taking. Correct scores on the test have been shown to correlate with adherence. The accuracy of self reported adherence depends on the patient's ability to recall and accurately report this behavior. Since patients may have difficulty in recalling their adherence over longer periods of time, one commonly used measure relies on self-report over each of the past 4 days. All methods of self-report overestimate the level of adherence compared with treatment adherence measures. Patients' wish to please their health care provider may be partially responsible for this. Patients who admit being non-adherent by self-report also have poor adherence on other measures. The ability of healthcare professionals to predict adherence to HAART is poor. In one study, the physicians and nurses were wrong 41% of the time when asked which patients were \>80% adherent with HAART. Other methods to monitor adherence are pill counts, pharmacy records, electronic monitoring devices, such as Medication Events Monitoring System (MEMS) and serum drug levels. All these methods have their particular shortcomings. Pill counts are time consuming. Patients may remove pills from their bottles without taking them ("pill dumping"), mix pills or forget their medications when coming to clinic. Pill counts, announced or unannounced may also be viewed as intrusive by many patients. Pharmacy-based measures assume that patients use the same pharmacy or the same payer source to refill their medications. Thus if prescriptions are not refilled on time it indicates non-adherence. Limitations include that patients may get free medication samples from their physicians and that even if the prescriptions are filled on time, it does not mean that the patient took the drugs. Electronic devices are expensive and have been used in the context of research. MEMS cap was such a device which registered every time the bottle cap was removed. Patients may remove more than one dose at a time, open the cap and not take the medication or engage in "pill dumping". Serum drug levels have been used recently in research settings. It is expensive and only reflects the level of the drug after the most recent dose of the medication. A study being conducted currently at the University of Chicago Hospitals is using Q methodology to study factors that influence adherence to HAART. Preliminary results of this study show that patients can be subdivided into five categories based upon responses to questions regarding their experience with HAART, the health system and living life with HIV. Attempting to predict an individual patient's potential adherence to HAART is extremely difficult. Studies have shown conflicting results and are hard to interpret. Factors that are consistently shown to be associated with poor adherence are lower level of education, depression, active drug/alcohol use and younger age. Health care professionals have been trying to improve adherence to HAART since its inception. The newer antiretrovirals are simpler to take, have longer half-lives and combine two or three medicines in one pill. A study by Stone et al showed that poor adherence was associated with medications which needed to be taken on an empty stomach or had to be taken three or more times a day. Campo and colleagues found that the higher the number of antiretroviral medications in the regimen, the less likely it is that patients would be adherent. Interventions to improve adherence are difficult to evaluate. This is due to the complex nature of measurement of adherence itself, and then after the intervention, the effect of the intervention on adherence. Pill boxes, charts of medications with photographs, electronic reminders, pharmacy support have all been studied. It has been shown that most interventions are useful short term but their effect is not durable. Thus reinforcement of the intervention is necessary. --Materials and Methods-- We plan to study if adherance to HAART is improved with an electronic reminder device (watch with an alarm), a pill box and phone calls made by a health care professional. About 200 HIV infected adults will be asked to participate in the study. For inclusion in the study, patients must be at least 18 years of age or older and able to give consent. They must have been on HAART for six weeks. Also patients included in the study must provide the research staff with a working telephone number, where they can be contacted. Patients will be randomized to the intervention or the control groups. All patients will have their HIV viral loads noted at the beginning of the study. This will not incur any additional cost to the patient as HIV viral loads are measured as a standard of care in patients who are on HAART at regular intervals. This measurement reflects the efficacy of HAART and HIV viral loads that are high may reflect poor adherance. All patients will be administered a questionnaire on enrollment. This questionnaire will consist of three parts. The first part will address the demographics of the patients. The second will be a pill identification test. The patients will be asked to recognize the agents that they are taking as part of their HAART regimen from a chart that has photographs of all antiretroviral medications. The third part will be a Q sort questionnaire where the patients will be asked to rank statements regarding their experience with HAART, the healthcare system, living life with HIV etc. This is the same format that is currently being used in the ongoing study at the University of Chicago Hospitals which studies subjective attitudes and adherance to HAART in HIV infected adults. That study has enrolled \~90 subjects. Data from the ongoing study will be used as part of this new study. All participants will be asked to provide the research staff with a working telephone number. Lack of a working telephone number is an exclusion criterion. After the questionnaire, the patients will be randomized into intervention and control groups. The intervention group will be given a wristwatch with an alarm. The patients in this group will receive instructions on how to program the watch so that it reminds them to take their medications. They will be given a pillbox. The pillbox will be able to hold one week's supply of HAART. Also the intervention group will receive a standardized phone call from a physician not directly related in their health care once a month for three months. The phone call will not last more than 15 minutes and will comprise a standard set of questions about their medications. The responses of the patients will be noted. The control group will be observed during the study period. At the end of the study, the HIV viral loads of both groups will be noted again and an attempt will be made to see if the intervention had any effect on adherance with HAART. If there is a difference in adherance between the intervention and the control groups at the end of the study, we think that this intervention is simple and cost effective. It can be easily reproduced at other centers that take care of patients with HIV.