I'm a collaborator on a new article in which we looked at patients' adherence to a different kind of treatment than in my usual medication studies -- continuous glucose monitoring technology, or CGM. The lead investigator on this project, Dr. Laurel Messer, is a former CU Nursing PhD student who has previously published with me on Two Minds Theory. CGM devices are used by people with diabetes -- in this case, children and adolescents with type I diabetes -- to monitor ups and downs of their blood sugar in real time. Patients can use the information to more effectively time doses of insulin in order to keep their blood sugar in the desired range more of the time. The real-time sensor data generated by CGM devices is ideal for studying questions about how Intuitive-level thinking might affect self-management behavior and health outcomes in the context of everyday life.
In this study, which used Two Minds Theory as a conceptual framework, 88 adolescents and young adults (ranging from 14 to 23 years old) completed daily surveys for about a week, and also provided CGM data collected by their device during the same timeframe. Our outcome measure was time-in-range (TIR), which represents the percentage of CGM measurements that were within normal limits for blood sugar -- i.e., times when the participant wasn't experiencing either hyper- or hypoglycemia. We used within-person statistical models to test the ability of each item on the daily survey to predict that person's individual TIR results on that specific day. Out of 25 items on the survey, we were able to drill down to just 7 that were particularly good predictors of good blood sugar control.
Not surprisingly, one of the best predictors of someone's overall glucose control was their actual glucose level at the time the survey was completed -- in other words, good control at a particular point in time is a marker for good control over the course of the whole day. More interestingly, people had better glucose control when they answered more affirmatively on the questions "I am planning on managing my diabetes today" and "I want to manage my diabetes today." Both of these items, which were excellent predictors of TIR, come from a measure of motivation that I have used in previous studies of treatment adherence in people with HIV. In those studies, people's in-the-moment motivation emerged as one of the strongest predictors of their behavior. I was pleased to see very similar results in a very different population (adolescents vs. adults) who were engaged in a very different behavior (using a CGM device vs. taking pills) for a very different health condition (type I diabetes vs. HIV). It's also worth noting that the motivation items emerged out of a pool of possible predictors in a purely empirical analysis (LASSO regression), not because we thought they were theoretically the most important. This study builds on prior work suggesting that motivation is a particularly important factor to consider in understanding people's everyday health behaviors based on real-time data.
The other four items included lack of symptoms as a barrier to diabetes self-management: The question "do you feel like skipping diabetes self-management activities because you feel fine?" predicted worse TIR results. A second question also showed that feeling well presented a barrier to self-management: Better scores on the perceived health item from the SF-36 also predicted lower TIR. This useful survey question -- "in general, would you say your health right now is ..." [Poor/Fair/Good/Very Good/Excellent]? -- also been found to predict mortality in older patient populations independent of someone's actual health condition, but it's a somewhat novel finding that it also predicted worse health outcomes in younger patients here. Additionally, TIR was predicted by a question about perceived need for help -- "do you think you could use some extra support for your diabetes management today?" This finding suggests that people who were aware of a need for help, or perhaps who had lower self-efficacy related to their ability to manage on their own, tended to have worse TIR that day.
Finally, in an item that might be unique to an adolescent population, higher scores on the question "I feel good about who I am" predicted better diabetes self-management. This suggests that in addition to typical self-management predictors like barriers, self-efficacy, and health status, models of self-management in adolescents and young adults might need to include a self-image or social perception variable that accounts for how attuned this group is to what others might think of them. In a prior study, Dr. Messer also found that body image was a key component of whether adolescents and young adults were willing to use a CGM device at all.
Although the 7 survey items in this study together predicted about 17% of participants' day-to-day variability in glucose control, that still leaves a lot of unexplained variability on the table. The survey items were a bit better at predicting whether people said they had achieved their daily self-management goals (29% of variance explained), but there may be more to actual glucose control than just feeling that you are keeping on top of the problem. In future work we are looking at additional predictor variables drawn from the CGM device itself, like whether someone checked their blood sugar during the day, tracked their carbohydrate intake, or delivered an insulin bolus at the appropriate times based on their numbers.
This is a great example of a study that combines biological with behavioral variables in order to understand what's actually happening in someone's experience at the Intuitive-mind level, right at the moment that a health behavior occurs. Our hope is that this type of in-the-moment understanding will lead to better interventions that speak to the Intuitive mind, giving people a nudge at just the time and in just the way that's most helpful to support them in managing their diabetes.
Comments
Post a Comment