Optimal Recall Period and Response Task for Self-Reported HIV Medication Adherence

Minyi Lu; Steven A. Safren; Paul R. Skolnik; William H. Rogers; William Coady; Helene Hardy; Ira B. Wilson


AIDS and Behavior. 2008;12(1):86-94. 

In This Article

Abstract and Introduction

Self-reported measures of antiretroviral adherence vary greatly in recall time periods and response tasks. To determine which time frame is most accurate, we compared 3-, 7-day, and 1-month self-reports with data from medication event monitoring system (MEMS). To determine which response task is most accurate we compared three different 1-month self-report tasks (frequency, percent, and rating) to MEMS. We analyzed 643 study visits made by 156 participants. Over-reporting (self-report minus MEMS) was significantly less for the 1-month recall period (9%) than for the 3 (17%) or 7-day (14%) periods. Over-reporting was significantly less for the 1-month rating task (3%) than for the 1-month frequency and percent tasks (both 12%). We conclude that 1-month recall periods may be more accurate than 3- or 7-day periods, and that items that ask respondents to rate their adherence may be more accurate than those that ask about frequencies or percents.

Highly active antiretroviral therapy (HAART) has led to a dramatic decline in human immunodeficiency virus (HIV) mortality and morbidity (Hogg et al. 1998; Palella et al. 1998). Both clinical trials and cohort studies have shown that adherence to antiretroviral (ARV) medication is one of the most important factors for the success of HAART (Arnsten et al. 2001; Bangsberg et al. 2001; Chesney et al.1999; Paterson et al. 2000). Poor adherence is associated with treatment failure, resistance, and the transmission of resistant viruses (Bangsberg et al. 2000; Vanhove et al. 1996). Consequently, both clinicians and researchers need accurate adherence measures.

Adherence assessment methods include self-report, pill counts, pharmacy records, biological surrogate markers, and the medication event monitoring system (MEMS), each of which has strengths and weaknesses (Gao et al. 2000). There is still no "gold standard" for accurate measurement of adherence (Grymonpre et al. 1998; Wagner and Rabkin 1999). While MEMS assesses pill bottle opening events, it cannot assess pill ingestion. High cost and logistical complexies also limit its usefulness in clinical care (Bova et al. 2005). Self-report is commonly used in both research and clinical care because of its low cost, ease of administration, and good correlation with other indirect adherence measures such as MEMS and pill count; it is likely that these factors will continue to make it the preferred adherence assessment method in many settings (Simoni et al. 2006). Self-report tends to overestimate adherence (Arnsten et al. 2001; Liu et al. 2001; Wagner et al. 2004). Nonetheless, several observational studies have shown that it correlates well with the plasma HIV RNA, supporting its value in HIV ARV medication adherence studies (Nieuwkerk and Oort 2005; Simoni et al. 2006).

An obstacle to progress in the HIV ARV medication adherence field has been the absence of standardized self-report measures. Researchers have used a variety of recall time frames, including 1, 3, 7 days, and 1 month. Shorter time frames have the potential benefit of more accurate recall. The widely used Aids Clinical Trials Group (ACTG) adherence instruments employs a 4-day recall period (Chesney et al. 2000). However it is not known whether these short time frames produce more accurate reports than longer time frames that may capture more clinically relevant variations in adherence behaviors. In addition, both the item content and response tasks of self-report items have varied widely. Item content has included asking about missed pills, missed doses, missed days, problems following exact dosing instructions, and overall percent adherence; and common response tasks have included frequencies of events, numbers of events, and percentages.

Few studies have compared these different self-report methods using an objective indicator such as MEMS. This study had two main goals. First, we assessed the accuracy of 3-, 7-day, and 1-month self-reported ARV adherence by comparing each to MEMS. Second, to learn more about the accuracy of different response tasks, we compared the performance of three different response tasks—a frequency item, a percent item, and a rating item—each of which used a 1-month time frame. For this comparison we examined bias, linearity, convergent validity, and predictive validity, again using MEMS as a standard.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: