In a new paper, my colleagues and I used Fitbit fitness trackers to look at patterns of exercise among people with HIV. It might come as no surprise that the people in our study exercised more on some days than on others. We measured exercise in two different ways, total number of steps taken and number of active minutes (defined based on Fitbit's algorithm as burning at least 3x the individual's normal number of calories). The day-to-day variability was greater on the total-steps measure than on the active-minutes measure, because some people never had much of the higher level of activity at all. Fitbits aren't research-grade devices or ones that you would use for training if you were a professional athlete, but they are pretty good at picking up this type of difference between sedentary behavior and at least some level of activity.
The higher level of physical activity is particularly important for people with a chronic disease like HIV, as I have written about previously; it helps to prevent cardiovascular complications, diabetes, and other problems of "premature aging" that otherwise tend to occur. Even for this higher level of activity, about 30% of the variability in exercise was between individual days within a specific person -- in other words, even the very active people weren't that way very consistently.
Most important from the perspective of Two Minds Theory, we also looked at variables that predicted the days on which people would be more versus less active. These were things that could be measured at multiple points in time for the same person, like sleep (sensor data collected via the Fitbit), stress (both a daily survey and a sensor measure based on heart rate variability), and various survey-based measures of fatigue, self-efficacy, mood, coping behaviors, environmental barriers, and social support. Most of the same questions that predicted active minutes also predicted total steps per day, but there were fewer predictors for active minutes and it's also the more important behavior to predict, so I will focus on those results here.
One other helpful feature of this study, besides the use of sensors to get more objective data, is that we applied time-lagged statistics to determine what someone's experiences today might mean for their behavior tomorrow. This one-day lag helps to establish causation, because the cause comes before the effect in time. That's not quite as good as a randomized trial where researchers create a condition for one group and not for another, but it's much better than the usual correlational design where people just report on things that happen together at the same time point.
Interestingly, sleep didn't matter at all for exercise the next day, counter to what we might have expected about the effects of being tired on someone's tendency to be active or not. However, a more subjective measure of fatigue did predict exercise the next day -- so when someone felt tired they were less likely to exercise, independent of how well they actually slept. For stress, the pattern went the other way: Subjective stress reported on a daily survey wasn't a good predictor of next-day exercise, but a physiological measure of stress using heart rate variability did predict exercise, in the direction of more stress predicting less exercise and vice-versa. The self-reported barriers to exercise mattered, including things like alcohol and drug use as well as feeling sick or experiencing medication side effects. And perhaps uniquely for our population of people with HIV, feeling discriminated against or stigmatized by other people led to less exercise, while feeling positively supported by other people had no effect.
One final finding was a bit surprising to us, which was the direction in which coping behavior affected exercise the next day. Coping is often divided into two subtypes, approach and avoidance, with the assumption that confronting your problems head-on using approach coping is usually the healthy way to go. Avoidance coping includes things like trying watching TV, distracting yourself, or otherwise trying not to think about your problems, and it's generally considered unhealthy. In this study, however, people who had less approach coping (or for the total-steps measure of exercise more avoidance coping) were actually more likely to exercise the next day. Why might that be? One explanation we considered was that we just had a poor measure of coping. But another possibility is that exercise itself might serve as a form of avoidance coping -- trying to "get out of your head and into your body" for a while. If that's the case, it might be a place where something that seems psychologically unhealthy actually turns out to be healthy for you overall! This finding is worth some further research: It would be interesting to know, for example, if people who cope with their problems through exercise end up having a better outcome or are more likely to succeed in approach coping later on, once they have cleared their heads through exercise.
Overall, this study confirmed some important predictions of Two Minds Theory by showing that day-by-day experiences can affect our everyday behaviors. The inclusion of sensor data such as heart rate variability showed that sometimes the influences on behavior are completely outside our awareness, at the level of the Intuitive mind (recall that people couldn't necessarily tell you their stress level in a way that predicted exercise, but their sensor-measured stress level did predict their behavior). And in another set of analyses where we looked at Narrative-level measures of some of the same concepts (stress, mood, sleep, etc.) we failed to find such strong effects. The day-by-day level of experience is once again what we found to be most useful in predicting behavior, above and beyond what we could account for by asking people about their "average" or "typical" experiences.
Comments
Post a Comment