Several recent key studies show promising evidence that support digital device use and passive sensing in the treatment of depression.

Almost everything we do now in modern life generates data - and we can learn from it.

‍

The idea of a ‘quantified self’ first emerged with the widespread use of digital devices that are capable of tracking ongoing activity, health information, and other vital data. Here we can point to the Apple Watch, FitBit, the Oura Ring, our smartphones, and other devices that keep tabs on daily living metrics and patterns, so we can view that data and make improvements.

‍

The term quantified self was first proposed in 2007 by editors Gary Wolf and Kevin Kelly at Wired Magazine as "a collaboration of users and tool makers who share an interest in self-knowledge through self-tracking."

‍

Our phones and wearable technology have added meaningful data to the management of our lives and habits…but can they help in the treatment of depression?

‍

John Torous and colleagues first brought the concept of the quantified self to the study of mental health with the term ‘digital phenotyping’ in 2016 as the “moment-by-moment quantification of the individual-level human phenotype (for example depression) in situ (meaning as it is found, unaltered) using data from personal digital devices” (Torous et al. 2016).

‍

Torous and his colleagues want to better understand psychiatric disorders by extracting biomedical and clinical insights from patient smartphone data. While ‘digital phenotyping’ is the term used most often in published literature, it is synonymous with ‘passive sensing’ or ‘personal sensing’ in the mental health field. Here, we will use the terms interchangeably.

‍

Smartphones and wearables collect data passively, persistently, and objectively. These three attributes make passive sensing a potent complement to subjective patient-reported outcomes in mental health.

‍

Despite years of research, we still don’t have objective biomarkers to diagnose the many subtypes of depression that exist. Osmind and others (including Tom Insel, former head of the NIMH) believe passive sensing will yield objective biomarkers that can characterize depression subtypes, predict response to treatment, and even predict depression relapse.

‍

But how close are we to this reality today, six years after John Torous’ seminal paper?

‍

In a review of 25 digital phenotyping studies of depression in college students published in 2020, John Torous’ group (Melcher et al. 2020) concluded that sample sizes were too small (less than 100 patients per study), and investigators had not analyzed data in a way that could be reproduced, among other issues. In short, passive sensing wasn’t ready for primetime in patient care.

‍

However, 2021 brought us three important studies that provide optimism regarding the role of passive sensing in the treatment of depression.

‍

We will take a look at the first study here, and discuss the other two in future posts.

‍

Last year, Meyerhoff and colleagues from David Mohr’s lab published a paper with a larger sample size of 282 patients (Meyerhoff et al. 2021). With this sample, they replicated a correlation between smartphone GPS features and depression previously observed in an independent sample (Saeb et al. 2015). Accumulating multiple independent replication studies is an essential second step in validating a biomarker for clinical care once an initial significant finding is reported.

‍

Myerhoff showed that changes in GPS features (in this case a reduced number of locations and transitions from home) could predict depression two weeks before the subject noticed any changes in their mood.

‍

The correlation was modest. However, if a few more data streams (in addition to GPS) were made available from smartphones or wearables and added to the classification algorithm, the correlations may become stronger, and relapse prediction could become a reliable clinical tool if these findings can be replicated in sufficiently large data sets.

‍

There are some other points from the Meyerhoff paper that are especially relevant to Osmind practices and their patients.

‍

Meyerhoff found that depression relapse correlated more strongly with GPS features when comorbidities were included as clinical variables. Here, they tried to reduce the heterogeneity underlying depression by making the ‘phenotype’ more specific.

‍

We have an opportunity to further this data, because many Osmind practices treat patients with treatment-resistant depression (TRD). TRD is expected to be a more genetically homogenous subtype and we may be able to find unique data streams that predict response and/or relapse to therapies such as ketamine or TMS in Osmind patients.  

‍

Notably, passively-collected social and behavioral data from wearable devices and smartphones may be less susceptible to the complexities introduced by culture and language differences that can pose barriers when using traditional subjective outcomes surveys.

‍

If you’d like to know more about passive sensing, please reach out to Osmind at research@osmind.org.

‍

In the meantime, here’s a great talk from Dr. Thomas Insel about digital phenotyping as a tool for better mental health outcomes.

‍

Sources:

Torous, J., Kiang, M. V., Lorme, J., & Onnela, J. P. (2016). New Tools for New Research in Psychiatry: A Scalable and Customizable Platform to Empower Data Driven Smartphone Research. JMIR mental health, 3 (2), e16. https://doi.org/10.2196/mental.5165

‍

Meyerhoff J, Liu T, Kording KP, Ungar LH, Kaiser SM, Karr CJ, Mohr DC. Evaluation of Changes in Depression, Anxiety, and Social Anxiety Using Smartphone Sensor Features: Longitudinal Cohort Study. J Med Internet Res. 2021 Sep 3;23(9):e22844. doi: 10.2196/22844. PMID: 34477562; PMCID: PMC8449302.

—

‍

Dr. Alison McInnes, VP, Medical Affairs at Osmind, is a nationally recognized expert in psychiatry, mood and anxiety disorders, ketamine treatment, and psychedelic medicine. She graduated from Columbia College of Physicians and Surgeons and was elected to the Alpha Omega Alpha Medical Society.  She completed a postdoctoral fellowship after a residency at UCSF supported by a Howard Hugh’s Research Fellowship and an NIMH K Award. She was an Associate Professor of Psychiatry at the Mount Sinai School of Medicine for many years and ran a laboratory focused on Psychiatric Genetics funded by an NIMH R01 among other grants including a Future Leaders of Psychiatry Award.

‍

Most recently she founded Kaiser Permanente's ketamine infusion therapy program and continues advocacy work to help safeguard the practice of ketamine therapy. Dr. McInnes currently practices KAP, and leads academic research collaboration at Osmind.

‍

The Evidence Base with Alison McInnes, MD

‍

The Evidence Base brings you relevant research to apply to your practice. This research also drives the development of new mental health treatments by a) influencing insurance payors to increase coverage for therapies, and b) providing data to politicians as they lobby for increased access to therapies and reduce stigma regarding mental health illness. Peer-reviewed articles also help unify mental health practice standards for the best possible patient care, and help guard against potential legal challenges.