Zusammenfassung
To summarize, the usage of the instrumental variable allowed us to gauge the unconfounded
effect of meal logging on weight loss. Based on the results, we conclude that app-based
meal logging in itself is conducive to weight loss in a blended-care setting. As the
data suggests a linear relationship (more meal logs, more weight loss), prompting
patients to log more meals is a promising intervention strategy to drive weight loss.
In general, more research is needed to identify (and subsequently increase) the genuine
effects of app-based behaviour change techniques, such as self-monitoring. The resulting
evidence will be pivotal in designing scalable, blended-care weight loss interventions
that make optimal use of dietitians’ scarce resources, while relying on app-based
care for behaviour change techniques that prove efficacious when delivered digitally.
Einleitung
App-based meal logging – along with other self-monitoring techniques aiming at behaviour
change – has proven to be beneficial for weight-loss.
Material und Methodik
To identify the genuine effect of meal logging, we analysed the data of 2856 obese
patients (mean age: 50 years (STD: 13.64); mean start body-mass-index: 37,5 (STD:
6,25); mean percentual weight loss after 12 weeks of weight loss intervention: −3.66
(STD: 4,08)) in a retrospective real-world-data analysis. Patients participated in
blended-care weight loss interventions at a specialized nutritional care provider,
combining dietitian support and app-based care (e. g. self-monitoring via photo-based
meal logging). To estimate the ‘causal’ treatment effect of meal logs on weight loss
and thus to eliminate confounders, we used an instrumental variable, quantifying the
volume of assigned meal logging tasks by a dietitian.
Ergebnisse
Variations across our instrumental variable yielded a response in weight loss that
is smaller than the observed correlation between logging frequency and weight loss.
However, logging meals once a day on its own had a positive effect on weight loss
(p=0.05) with an expectation of additional ~0.2% weight loss relative to the start
weight for each increment of the daily logging rate (e. g. an increase of 0.6% for
3 meals logged per day).
