JMIR Publications

ABSTRACT

Background:

There is increasing interest to reuse person-generated wearable device data for research purposes, which raises concerns about data quality. However, the literature on data quality challenges, specifically for person-generated wearable device data, is sparse.

Objective:

The objective of this study is to systematically review the literature on factors affecting quality of person-generated wearable device data and identify challenges associated with their secondary uses.

Methods:

We searched PubMed, ACM, IEEE, and Google Scholar with search terms related to wearable device and data quality. Using PRISMA guidelines, we reviewed the papers to find factors affecting the quality of wearable device data. We annotated those papers and categorized semantically similar factors. If any data quality challenges were mentioned in the study, we extracted those contents as well.

Results:

Twenty-six papers were included. We identified 3 high-level factors: device and technical, user-related, and data governance factors. Device and technical factors include problems with hardware, software, connectivity; user-related factors include device non-wear and user error; and data governance factors include lack of standardization and data accessibility issues. The identified factors potentially can lead to data quality problems such as incomplete, inaccurate, and heterogeneous data.

Conclusions:

Our study identifies potential data quality challenges that could occur when analyzing wearable device data for research and 3 major contributing factors for these challenges. As poor data quality can compromise the reliability and accuracy of research results, further investigation is warranted on how to address data quality challenges facing wearable devices.