Medical and pharmaceutical advances have made it possible for people living with HIV (PLWH) to manage the disease and reduce the risk of infecting others, but many patients don’t adhere to their prescribed medication regimen. CU College of Nursing Associate Professor Paul Cook, PhD, seeks to understand medication-taking behaviors in PLWH who are prescribed antiretroviral therapy (ART).

In a study published in the July/August 2017 issue of Nursing Research, Cook and his co-authors recorded their findings based on data from technology-enhanced medication bottles that time-stamp when opened. The researchers also used smartphones to send daily surveys to the participants that evaluated the six predictors of adherence considered in the study: thoughts, mood, stress, coping, social support, and treatment motivation. “Even behaviors that are usually habitual, like taking medication, can change based on day-to-day variability in mood, motivation, or other factors that serve as barriers or facilitators,” Cook, Schmiege, Star, Carrington, & Bradley-Springer wrote (2017, p. 276).¹

To make findings that would be considered valuable beyond Colorado, the researchers studied a diverse sample population. “Here [in Colorado] most people living with HIV are White men who have sex with men (MSM). That's what the HIV epidemic looked like in its early days. But now the largest numbers of new diagnoses are among minority men and women,” says Cook. “Young Latino and African-American MSM are at the highest risk, but heterosexual contact is the route of infection in 24% of cases, and 19% of all new diagnoses are among women.² In order to have a sample that provides meaningful information for the U.S. HIV epidemic as a whole, we needed to recruit more carefully in Colorado where the population of PLWH doesn't completely resemble the national picture.”

After the survey data was compared with the time-stamps from the medication bottles, the results were not completely what had been expected. “The most surprising thing in the current study was the direct relationship between mood and adherence. We expected that worse mood would lead to people being tired, demoralized, etc. and that they would be nonadherent as a result. The reverse was actually true: When people felt happier, they were less likely to take their medication,” Cook says. “It's important to note that we weren't looking at clinical depression, just mood being a little better or a little worse on any given day. People who are more seriously depressed might still have adherence problems. But in our study, feeling a little bit down actually made you more adherent.”

So what implications do this study’s findings have for clinicians who want to ensure their patients are adherent to prescribed ART? “This I don't think we know yet,” says Cook. “There's more research to be done before we have solid advice for clinicians about how to help. I think we can confidently say (a) that motivation goes up and down, and is a good predictor of actual adherence behavior going up and down, and (b) that mood is probably also important to monitor.”

One warning Cook offers for clinicians is not to consider decreased viral load as a true indicator of patient adherence. In this study, 91% of the patients were virally suppressed, but their average adherence after the first few weeks of the study was only 74%, which is well below the level at which patients are expected to have therapeutic results.

“Clinicians are likely to assume that patients with good viral load outcomes are automatically adherent — some reviewers have even asked me ‘Why are you doing this study when the patients' viral load shows that they are adherent?’ — but that doesn't seem to be the case,” Cook says. “Clinicians need to think about the process of adherence separate from the outcome of viral suppression, and consider whether even the ‘good’ patients may have some lapses that put them at risk for treatment failure.”

The purpose of the study was not to determine who takes medication and who does not, but rather to understand the circumstances that can lead to nonadherence. “Overall, I think it highlights the need to look at health behaviors in the context where they actually occur, rather than relying on retrospective measures,” says Cook. “New technologies are giving us important new windows into how patients actually take care of their health in the course of their everyday lives.”

Cook plans to use technology for subsequent studies that will build upon the variables, such as mood, which were found to be important in his previous work: “Some additions will be data collection on symptom experiences like fatigue and daily activity levels to get at the question of whether mood is associated with more or less behavioral activation, and alcohol or drug use, which is an important risk factor for nonadherence. I am hoping to include FitBits in that one as a behavioral measure of activity, heart rate, and sleep, along with self-report surveys and adherence data.”

1. Cook, P.F., Schmiege, S.J., Starr. W., Carrington, J.M., Bradley-Springer, L. (2017). Prospective State and Trait Predictors of Daily Medication Adherence Behavior in HIV. Nursing Research, 66(4):275-285. doi: 10.1097/NNR.0000000000000216.
2. HIV in the United States: At A Glance. (2017, 9 June). Retrieved from https://www.cdc.gov/hiv/statistics/overview/ataglance.html