Stats-of-1 at Brown University
Last updated: Apr 29, 2023
Thanks for coming and giving such an energetic talk. The turnout was one of the best of the semester, if not the best, so it seems that we have a lot of interested students.
It meant a lot to me to get your invitation to give a talk—to get to share ideas and warm camaraderie (out of the frosty air) with gregarious and engaging (and wicked smart!) biostatisticians working on n-of-1 AND causal inference. Y’all are amazing, and doing excellent work! I am super excited to see Brown Biostats continue to grow.
Other faculty and staff I met included Ani Eloyan, Anarina Murillo, Roee Gutman, Joseph Hogan, Zhijin Jean Wu, Roberta De Vito, Arman Oganisian, Stavroula Chrysanthopoulou, and Lourdez Caballeros. All were so warm and friendly, and we connected over my work—both in n-of-1 and in DEI (diversity, equity, and inclusion). (My DEI work is with the American Statistical Association Justice, Equity, Diversity, and Inclusion (JEDI) Outreach Group, where I currently chair the Professional Development Committee.)
More on my talk:
Talk Title: Using Wearables and Apps to Characterize Your Own Recurring Average Treatment Effects
Abstract: Temporally dense single-person “small data” have become widely available thanks to mobile apps (e.g., that provide patient-reported outcomes) and wearable sensors. Many caregivers and self-trackers want to use these intensive longitudinal data to help a specific person change their behavior to achieve desired health outcomes. Ideally, this involves discerning possible causes from correlations using that person’s own observational time series data. In paper one, we estimate within-individual average treatment effects of sleep duration on physical activity, and vice-versa. We introduce the model-twin randomization (MoTR; “motor”) and propensity score twin (PSTn; “piston”) methods for analyzing Fitbit sensor data. MoTR is a Monte Carlo implementation of the g-formula (i.e., standardization, back-door adjustment); PSTn implements propensity score inverse probability weighting. They estimate idiographic stable recurring effects, as done in n-of-1 trials and single case experimental designs. We characterize and apply both methods to the two authors’ own data, and compare our approaches to standard methods (with possible confounding) to show how to use causal inference to make truly personalized recommendations for health behavior change. In paper two, we apply MoTR to the three authors, thereby providing a guide for using MoTR to investigate your own recurring health conditions—and demonstrating how any suggested effects can differ greatly from those of other individuals.
I covered the first of these three papers:
- Main paper: Daza EJ, Schneider L. Model-Twin Randomization (MoTR): A Monte Carlo Method for Estimating the Within-Individual Average Treatment Effect Using Wearable Sensors. arXiv preprint arXiv:2208.00739. 2022 Aug 2. arxiv.org/abs/2208.00739
- Application: Matias I, Daza EJ, Wac K. What possibly affects nighttime heart rate? Conclusions from N-of-1 observational data. Digital Health. 2022 Aug 24. journals.sagepub.com/doi/10.1177/20552076221120725
- Underlying framework: Daza EJ. Causal analysis of self-tracked time series data using a counterfactual framework for N-of-1 trials. Methods of information in medicine. 2018 Feb;57(Suppl 1):e10. thieme-connect.de/products/ejournals/abstract/10.3414/ME16-02-0044 (LaTeX pre-print with cleaner equations and identical content here)
These can also be found here: https://statsof1.org/resources/#causal-inference–causality
The slides are here. Please request access, and tell me how you intend to use them. In the spirit of “trust but verify”, please also include a link to your professional bio or related work (e.g., LinkedIn profile, ResearchGate profile, Google Scholar profile, professional website, or official institutional bio page). Thanks!