Development and Evaluation of an Ecological Momentary Assessment–Based mHealth Platform for Monitoring HIV PrEP Adherence

In the world of public health, knowing what people actually do is often harder than knowing what they say they do. When researchers ask people to remember their habits over weeks or months, the answers are often clouded by forgetfulness or a desire to look good. This is particularly true for medication, where taking a pill every day or at a specific time before an event requires consistent effort. To solve this, scientists have developed a method called ecological momentary assessment. Instead of asking for a memory of the past, this approach asks people to report their actions in real time, right as they happen, usually through a smartphone. The goal is to catch behavior as it unfolds, reducing the errors that come from trying to recall days or weeks later.

This real-time approach is vital for preventing HIV through a medication called pre-exposure prophylaxis, or PrEP. When taken correctly, PrEP is highly effective at stopping the virus, but its power depends entirely on the user taking the pills as prescribed. There are two main ways to take it: a daily regimen, where one pill is taken every day, and an event-driven regimen, where pills are taken only around the time of sexual activity. For the event-driven group, timing is everything, and missing a dose can leave a person unprotected. Because so many people in China now get PrEP through online platforms, where doctors cannot see them in person, finding a way to track whether they are actually taking their medication has become a critical challenge.

A team of researchers from Tsinghua University and various health centers in China set out to test a new digital tool designed to solve this problem. They built a mobile health platform that uses the real-time reporting method to track how well people are sticking to their PrEP schedules. To see if this digital tool was actually accurate, they did not just trust the users' words. Instead, they compared the app's data against a biological gold standard: a dried blood spot test. This test measures the actual amount of drug in a person's blood, providing an objective record of whether the medication was truly taken. They also compared the app's results to traditional questionnaires, where users filled out surveys every few weeks about their past behavior.

The study followed 342 people who were already using PrEP through an online service. The participants were divided into two groups: those taking the daily dose and those taking the event-driven dose. Over a period of eighteen weeks, the researchers checked in with them at six, twelve, and eighteen weeks. The app prompted users to log their medication intake immediately after taking it or after a sexual event, while the questionnaires asked them to look back over the previous six weeks. At each check-in, the participants also mailed in a small blood sample from a finger prick to be analyzed for drug levels.

The results revealed a clear difference between how people remembered their habits and how they actually behaved. For those on the daily regimen, the traditional questionnaires suggested that nearly everyone was taking their medication perfectly, with adherence rates hovering around 80 to 83 percent. The app, however, recorded lower rates, showing that only about 56 to 69 percent of daily users were taking their pills as scheduled. When the researchers looked at the blood tests, the app's numbers were much closer to reality. The app correctly identified that a person was taking their medication about 65 to 86 percent of the time, matching the biological evidence well. The questionnaires, by contrast, often claimed a person was taking their pills when the blood test showed they were not, leading to a high number of false alarms about good behavior.

The picture was slightly different for the event-driven users, who take pills only when needed. Here, the app was very good at spotting people who were not taking their medication correctly. If the app said a person was not adhering, the blood test almost always confirmed it. This means the tool rarely gave false reassurance to people who were actually missing doses. However, the app was less good at catching everyone who was taking their medication correctly; it missed some adherent users that the questionnaires caught. Despite this, the researchers found that for the event-driven group, the app was still more reliable than the surveys at identifying who was truly at risk because it avoided the common mistake of assuming people were following the rules when they were not.

The study suggests that using a smartphone to track medication in the moment offers a more honest picture than asking people to remember their past weeks. While the app did not catch every single instance of perfect behavior, it was far better at identifying when people were falling short. This is a crucial distinction for public health, because knowing who needs help is more important than assuming everyone is doing fine. The researchers noted that the tool worked well for both daily and event-driven users, though it was particularly valuable for the event-driven group, whose complex schedules are difficult to track with older methods.

This work provides strong evidence that digital tools can serve as a reliable way to monitor health behaviors in the real world. By combining immediate reporting with biological verification, the study shows that we can move beyond guesswork and self-reporting to understand how people truly manage their health. The findings support the idea that these mobile platforms can be scaled up to help millions of people, offering a practical way to ensure that life-saving medications are taken when they are needed most.