This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Research Protocols, is properly cited. The complete bibliographic information, a link to the original publication on https://www.researchprotocols.org, as well as this copyright and license information must be included.
Digital health technologies using mobile apps and wearable devices are a promising approach to the investigation of substance use in the real world and for the analysis of predictive factors or harms from substance use. Moreover, consecutive repeated data collection enables the development of predictive algorithms for substance use by machine learning methods.
We developed a new self-monitoring mobile app to record daily substance use, triggers, and cravings. Additionally, a wearable activity tracker (Fitbit) was used to collect objective biological and behavioral data before, during, and after substance use. This study aims to describe a model using machine learning methods to determine substance use.
This study is an ongoing observational study using a Fitbit and a self-monitoring app. Participants of this study were people with health risks due to alcohol or methamphetamine use. They were required to record their daily substance use and related factors on the self-monitoring app and to always wear a Fitbit for 8 weeks, which collected the following data: (1) heart rate per minute, (2) sleep duration per day, (3) sleep stages per day, (4) the number of steps per day, and (5) the amount of physical activity per day. Fitbit data will first be visualized for data analysis to confirm typical Fitbit data patterns for individual users. Next, machine learning and statistical analysis methods will be performed to create a detection model for substance use based on the combined Fitbit and self-monitoring data. The model will be tested based on 5-fold cross-validation, and further preprocessing and machine learning methods will be conducted based on the preliminary results. The usability and feasibility of this approach will also be evaluated.
Enrollment for the trial began in September 2020, and the data collection finished in April 2021. In total, 13 people with methamphetamine use disorder and 36 with alcohol problems participated in this study. The severity of methamphetamine or alcohol use disorder assessed by the Drug Abuse Screening Test-10 or the Alcohol Use Disorders Identification Test-10 was moderate to severe. The anticipated results of this study include understanding the physiological and behavioral data before, during, and after alcohol or methamphetamine use and identifying individual patterns of behavior.
Real-time data on daily life among people with substance use problems were collected in this study. This new approach to data collection might be helpful because of its high confidentiality and convenience. The findings of this study will provide data to support the development of interventions to reduce alcohol and methamphetamine use and associated negative consequences.
DERR1-10.2196/44275
Substance use disorder (SUD) is a major public health concern worldwide. The harmful use of alcohol is responsible for 5.1% of the global burden of disease (7.1% for males and 2.2% for females) [
Advances in technology have the potential to reduce barriers to treatment services in this field. Broadly known as digital health or mobile health (mHealth), research in these areas uses mobile and computer software apps and wearable biosensor devices to understand and treat health conditions better [
Although EMA has advantages in assessing substance use accurately and repeatedly, compliance and missing data are still challenges [
One of the solutions to the challenges of EMA might be the use of a wearable activity tracker (eg, Fitbit and Apple Watch), which can automatically collect participants’ data. Although these products are not specifically made for the data collection on substance use, researchers can repeatedly gather various objective data, such as behavioral data (eg, steps and sleep) and biological data (eg, heart rate, blood pressure, and degree of stress). These devices have been used in observational studies to collect accurate real-time data [
If a wearable activity tracker and a self-monitoring tool that records variables related to substance use are combined in a study, researchers can collect behavioral and biological data with minimal reliance on participant input. Moreover, researchers can analyze the data collected by a wearable activity tracker in many ways because these data could be potential predictive factors of substance use or subsequent outcomes following substance use. Daily self-monitored behavioral data along with momentary biological data collected by wearable activity trackers have the potential to bridge important data gaps.
This paper aims to introduce a procedure for implementing wearable tracker technologies as a method of EMA. We describe the methods to analyze data on substance use and psychological data using a self-monitoring app and objective behavioral and biological data using a wearable activity tracker (Fitbit). We will develop a predictive model of substance use by analyzing combined data using machine learning methods.
Fitbit (a registered trademark of Fitbit LLC) provides devices that offer a range of wearable activity trackers that can automatically track and display a users’ daily activity, exercise, and sleep data in real time. Fitbit is designed for use on both Android and iOS smartphones. Its popularity has grown consistently over the years, selling approximately 16 million devices in over 100+ countries in 2019, with an active user community of over 29 million users [
Two self-monitoring mobile apps were newly developed for use in this study, depending on whether the participant had mainly alcohol or drug-related problems. Self-monitoring is a widespread approach in mHealth apps in the field of substance use [
This study is an ongoing observational study using a Fitbit and a self-monitoring app (
Content included in the self-monitoring apps. API: application programming interface; DB: database.
Participants were outpatients with alcohol or methamphetamine use disorders from 2 psychiatric hospitals in Japan. A psychiatrist specializing in SUD screened potential participants based on inclusion and exclusion criteria, and patients were referred to the study if applicable. General residents with alcohol problems were recruited through a research company. Considering realistic circumstances at each facility and the approximate number of participants required to adequately perform preliminary model testing using machine learning, 20 people with methamphetamine use problems and 40 people with alcohol problems were recruited for this feasibility study. Inclusion criteria were as follows: (1) alcohol consumption with health risks in the past year (Alcohol Use Disorders Identification Test [AUDIT]: 8-19 points) [
The following data were measured while wearing the Fitbit for 8 weeks: (1) heart rate per minute, (2) sleep duration per day, (3) sleep stages per day, (4) the number of steps per day, and (5) the amount of physical activity per day. Data were synced from Fitbit to the Fitbit app every time it was connected to the participant’s smartphone. These data were retrieved from the Fitbit servers daily through an application programming interface and stored per user in a secure database created for this study. If there were no records for over 3 days in the past week, an email reminder was sent to the participant by the researcher.
The following data were measured daily through the self-monitoring app for 8 weeks (
Framework for data collection using Fitbit and self-monitoring app.
Surveys were conducted at baseline, 4 weeks, and 8 weeks from the start of participation (
Overview of measurements for each survey period.
Baseline | 4 weeks | 8 weeks | ||
|
||||
|
Amount of alcohol (grams of pure alcohol) or drug consumed per day in the past 7 days | ✓ | ✓ | ✓ |
|
Number of drinking or drug-use days during the observational period | ✓ | ✓ | ✓ |
|
University of Rhode Island Change Assessment Scale | ✓ | ✓ | ✓ |
|
General Health Questionnaire | ✓ | ✓ | ✓ |
|
Short Form-8 | ✓ | ✓ | ✓ |
|
||||
|
System Usability Scale |
|
✓ |
|
|
Client Satisfaction Questionnaire |
|
✓ |
|
|
Adverse events concerning physical problems or discomfort using the app |
|
✓ |
|
|
||||
|
Gender, age, last education, employment status, and marital status | ✓ |
|
|
|
Primary substance (alcohol or drug) | ✓ |
|
|
|
Severity of substance use disorder (AUDITa or DASTb-10) | ✓ |
|
|
|
Treatment history | ✓ |
aAUDIT: Alcohol Use Disorders Identification Test.
bDAST: Drug Abuse Screening Test.
The amount of alcohol or drug consumed per day (grams of pure alcohol or drug equivalent) in the past 7 days and the number of drinking or drug-use days during the observational period were assessed in this section. Mental health was measured by the General Health Questionnaire abbreviated version with a range of 0-30 points, validated in a Japanese population sample with a cutoff of over 6 points indicating mental distress [
The System Usability Scale (SUS), which consists of a total of 10 items, was used to assess the ease of use of the app [
Gender, age, last education, employment status, and marital status were asked as general demographic data. The following were also asked concerning the participants’ treatment history: the substance primarily responsible for the problem (alcohol, methamphetamine, marijuana, dangerous drugs, prescription drugs, over-the-counter drugs, organic solvents, or other drugs); the severity of SUD (AUDIT or DAST-10); the frequency of primary substance use in the past year; frequency of primary substance use in the past month (in days); the age of first alcohol or drug use; the age when first recognized an alcohol or drug use problem; the age when first sought medical care for an alcohol or drug use problem; social resources (medical institutions, mental health welfare centers, public health centers, or private addiction rehabilitation facilities); and the presence of comorbid psychiatric disorders and diagnosis.
Fitbit’s application programming interfaces capture data for heart rate per minute, sleep duration per day, sleep stages per day, the number of steps per day, and the amount of activity per day (hereafter referred to as “Fitbit data”). We finished collecting the data in April 2021 and are preparing to conduct data analysis. In order to determine substance use based on the Fitbit data, 3 steps were applied, as shown in
Planned future data analysis. HR: heart rate.
First, the Fitbit data will be visualized to confirm typical Fitbit data patterns for individual users. There are many patterns to detect drug or alcohol use; for example, the heart rate may remain high while drinking alcohol. Accordingly, data visualization is beneficial for confirming such patterns. If, for example, we can detect a typical pattern as mentioned above, signals other than that pattern may be considered noise for this analysis. Subsequently, noise removal methods will be applied to the Fitbit data. For example, average heart rate or heart rate variability may also be influenced by exercise or walking rather than drug or alcohol use. Therefore, we will use step counts or activity amounts on Fitbit data to specifically filter out such movement or workout periods. Other possible methods for noise removal, in terms of signal processing, may include signal smoothing methods, for example, moving average and Gaussian filter. Methods will be chosen to retain typical patterns but remove the other signals.
Second, the Fitbit data will be analyzed by using statistical analysis methods in order to predict drug or alcohol use on a relatively long-term basis, such as within a day or a week. By using self-monitoring data, we can separate the Fitbit data into data reflecting substance use and data from normal activity. Some descriptive statistical methods, for example, averaging, variance, self-correlation, or frequency-domain analysis, will be applied to each parameter in the separated data sets to determine features that include the days a user has declared drug or alcohol use. In this analysis, survey data containing the participants’ relation to drug or alcohol use, along with psychological and behavioral data points, may also be added as complements through the descriptive statistics. After transforming the data into its statistics, we would like to apply clustering methods to provide a quantitative evaluation of typical patterns reflecting drug or alcohol use.
There are 2 approaches to determining an optimal detection method for drug or alcohol use. If we can successfully obtain distinctive features by using the above statistical analysis, we will accordingly construct a detection model based on it. However, constructing such a model may be insufficient to isolate use with high accuracy and precision. Therefore, data analysis using both machine learning and statistical analysis methods will be performed in order to create a model that solves the problem as indicated above, based on the combined Fitbit and self-monitoring data. The primary candidate for machine learning methods will be Long Short-Term Memory. Long Short-Term Memory is considered one of the most suitable methods for analyzing time series data, which applies to this study. Meanwhile, other deep learning methods will also be considered to fit the data. The deep learning model will be carefully selected after we perform the statistical analyses, and then this model will be tested based on 5-fold cross-validation. Partitioning data depends on the degree of temporal precision with which drugs and alcohol consumption is estimated. Further preprocessing and machine learning methods will be conducted based on the preliminary results of the validation.
This study was approved by the Ethics Committee of the Faculty of Medicine and Graduate School of Medicine of Tokyo Medical and Dental University (M2020-189) and the Institutional Review Boards of each recruiting hospital. Data collected from the self-monitoring app were automatically stored in a protected database. Fitbit data were archived from the Fitbit server and preprocessed on a cloud-based, protected database without personal information attached. These data were managed by Humanome Lab, Inc. Both self-monitoring data and Fitbit data were then matched using participants’ ID numbers. Patients or participants provided their written informed consent in person before participating in this study.
Enrollment for the trial began in September 2020, and the data collection was completed in April 2021. In total, 13 people with methamphetamine use disorder and 36 with alcohol problems participated in this study.
There are 2 main anticipated results in this study. The first is to gain a better understanding of physiological and behavioral data before, during, and after alcohol or methamphetamine use. For example, an extended period of sleep deprivation may lead to a heightened sense of craving, and alcohol or methamphetamine use may increase. On the other hand, sleep quality may decrease due to several days of drinking.
The second anticipated result is to develop more personalized predictive models for alcohol or methamphetamine use. Data gathered through Fitbit are valuable in that it provides longitudinal and sequential data compared to periodic visits to hospitals. Although Fitbit is not a medical device, these data offer a new perspective into the daily life of participants with alcohol or methamphetamine problems. For example, individual participants may drink at night on workdays when stressed, and their heart rate increases to 110 beats per minute during the day.
In the future, understanding physiological and behavioral data and identifying individual patterns of behavior may lead to more effective personalized interventions that support preventive behaviors such as promoting eating or sleeping as a method of harm reduction when there is the detection of continuous levels of high stress and accelerated heart rate. Limitations of previous EMA and wearable tracker studies are still prevalent in this study, as compliance and missing data are still predictable risks. However, along with traditional deterrence methods such as data visualization, rewards, and reminders, this study uses consecutive biological data collected through Fitbit and an extended study period of 8 weeks to supplement missing data.
Demographic characteristics of participants.
Drug use disorder (N=13) | Alcohol use disorder (N=36) | ||||
Sex (male %), mean (SD) | 12 (92.3) | 18 (50) | |||
Age (years), mean (SD) | 46.9 (9) | 45.5 (10.7) | |||
|
|||||
|
Middle school | 3 (23.1) | 1 (2.8) | ||
|
Secondary school | 4 (30.8) | 6 (16.7) | ||
|
College | 1 (7.7) | 6 (16.7) | ||
|
University | 5 (38.5) | 23 (63.9) | ||
|
|||||
|
Employed | 5 (38.5) | 26 (72.2) | ||
|
Unemployed | 5 (38.5) | 3 (2.8) | ||
|
Other | 3 (23.1) | 7 (19.4) | ||
|
|||||
|
Married | 11 (84.6) | 10 (27.8) | ||
|
Single | 2 (15.4) | 22 (27.8) | ||
|
Divorced | 0 (0) | 4 (27.8) | ||
Age at first drug or alcohol use (years), mean (SD) | 21.2 (6.4) | 18.8 (2) | |||
Age at first hospital encounter (years), mean (SD) | 39.9 (9.3) | —a | |||
|
|||||
|
Score, mean (SD) | 6.9 (1.4) | — | ||
|
Moderate level (3-5; %), mean (SD) | 3 (23.1) | — | ||
|
Substantial level (6-8; %), mean (SD) | 8 (61.5) | — | ||
|
Severe level (9-10; %), mean (SD) | 2 (15.4) | — | ||
|
|||||
|
Score, mean (SD) | — | 13.8 (5) | ||
|
Low-risk consumption (0-7; %), mean (SD) | — | 4 (11.1) | ||
|
Hazardous consumption (8-14; %), mean (SD) | — | 17 (47.2) | ||
|
Alcohol dependence (15-40; %), mean (SD) | — | 15 (41.7) | ||
|
|||||
|
Did not use | 35.9 (17.3) | — | ||
|
Used primary drug | 10.4 (14.9) | — | ||
|
Used secondary drug | 0.3 (0.6) | — | ||
|
Did not answer | 2.2 (5.5) | — | ||
|
No record | 7.2 (12.3) | — | ||
Days of drinking, mean (SD) | — | 33.1 (17.8) | |||
Drinking amount per day (g), mean (SD) | — | 75.6 (65.5) | |||
|
|||||
|
Good | 20.5 (11.3) | 31 (13.9) | ||
|
Not good | 16.8 (13.2) | 4.3 (6.3) | ||
|
Angry | 2.8 (3.5) | 2 (3.1) | ||
|
Sad | 5.5 (5.5) | 3.1 (4.3) | ||
|
Did not answer | 3.1 (5.6) | 3.8 (7) | ||
|
No record | 7.2 (12.3) | 11.9 (12.8) | ||
Craving (0-10) |
4 (3.1) | 4.6 (1.9) | |||
Resting heart rate (beats per minute) |
72.6 (8) | 65.9 (9.4) | |||
Sleep duration per day (minutes) |
408.3 (192.8) | 386.4 (100) | |||
|
|||||
|
Wake | 57.7 (35.8) | 55.8 (22) | ||
|
Light sleep | 242.9 (128.6) | 223.1 (66.8) | ||
|
Deep sleep | 73.8 (40.8) | 66.4 (25) | ||
|
REMd sleep | 82.1 (49.4) | 88.1 (36.5) | ||
Steps per day |
8884.9 (6450.2) | 10,277 (5481.2) | |||
|
|||||
|
Lightly active | 204.8 (113) | 233.7 (100.5) | ||
|
Fairly active | 34.3 (55.6) | 19.7 (23.4) | ||
|
Very active | 38.6 (78.8) | 28.4 (31.2) |
aNot available.
bDAST: Drug Abuse Screening Test.
cAUDIT: Alcohol Use Disorders Identification Test.
dREM: rapid eye movement.
In this study, we collected data related to substance use and consecutive biological and behavioral information simultaneously using Fitbit and a smartphone app. To the best of our knowledge, this study is the first to collect and analyze real-time data on daily life among people with substance use problems in Japan. Similar to other Asian countries, drug policies in Japan are strict, and there is severe stigmatization against people with substance use problems. The new approach to data collection in this study might be helpful because of its high confidentiality and convenience. We will analyze the data in 2023 and then recruit additional participants based on the results for a more robust analysis.
The use of digital health technologies will be increasingly necessary in the future. In fact, during the COVID-19 pandemic, innovations in cost-effective and user-friendly drug prevention and treatment services, such as internet-based or mobile phone-based services, accelerated to increase accessibility and coverage of services [
Alcohol Use Disorders Identification Test
Client Satisfaction Questionnaire
Drug Abuse Screening Test
ecological momentary assessment
mobile health
substance use disorder
System Usability Scale
University of Rhode Island Change Assessment Scale
We wish to thank Dr Kanae Myoenzono for helping us to prepare for the acquisition of Fitbit data. We also thank Mr Junichiro Matsuda and Mr Gen Kikuchi from KANAME Co, Ltd. for the development of the self-monitoring app and for helping us prepare the data analysis. This work was supported by JSPS KAKENHI grant JP20H03977. This study was also partially funded by Revi Co, Ltd. The funders were not involved in the study design, collection, analysis, interpretation of data, writing of this paper, or decision to submit it for publication.
The data sets in this study are not available because we have not obtained the agreement of participants to disclose raw data or the approval of the ethics committee in each institution for data sharing.
AT, MS, MO, and JS were involved in the study design and protocol. AT, KO, KN, YY, and TM were involved in participant recruitment and management of the study. AT, KO, MS, MO, and JS were involved in the preliminary data consideration and analysis, and drafted the manuscript. All authors critically revised the manuscript and approved the final version.
Authors MS, MO, and JS are employed by Humanome Lab, Inc. All authors declare no other competing interests.