# Mobile device data for the study of miscarriage and its causes

## TL;DR

This paper shows how data from a pregnancy-tracking app can help study miscarriage and its causes, including social and environmental factors.

## Contribution

The novel use of mobile app data to detect and study miscarriage patterns, including upstream factors previously hard to study.

## Key findings

- Mobile app data accurately detects miscarriage and replicates known risk patterns.
- App data can expand research into social and environmental factors linked to miscarriage.
- The approach offers a scalable method for studying pregnancy complications.

## Abstract

Miscarriage is a common pregnancy complication that is difficult to study. Using information on 580,000 US pregnancies from a menstrual- and pregnancy-tracking app, we demonstrate that mobile device data can detect miscarriage and replicate patterns of miscarriage risk found in high-quality Norwegian administrative data and in US clinical cohorts. App data may greatly expand research on miscarriage, including the contributions of “upstream” social, economic, and environmental factors that have eluded study.