Background Medication adherence is a global public health challenge, as only approximately 50% of people adhere to their medication regimens. Medication reminders have shown promising results in terms of promoting medication adherence. However, practical mechanisms to determine whether a medication has been taken or not, once people are reminded, remain elusive. Emerging smartwatch technology may more objectively, unobtrusively, and automatically detect medication taking than currently available methods. Objective This study aimed to examine the feasibility of detecting natural medication-taking gestures using smartwatches. Methods A convenience sample (N=28) was recruited using the snowball sampling method. During data collection, each participant recorded at least 5 protocol-guided (scripted) medication-taking events and at least 10 natural instances of medication-taking events per day for 5 days. Using a smartwatch, the accelerometer data were recorded for each session at a sampling rate of 25 Hz. The raw recordings were scrutinized by a team member to validate the accuracy of the self-reports. The validated data were used to train an artificial neural network (ANN) to detect a medication-taking event. The training and testing data included previously recorded accelerometer data from smoking, eating, and jogging activities in addition to the medication-taking data recorded in this study. The accuracy of the model to identify medication taking was evaluated by comparing the ANN’s output with the actual output. Results Most (n=20, 71%) of the 28 study participants were college students and aged 20 to 56 years. Most individuals were Asian (n=12, 43%) or White (n=12, 43%), single (n=24, 86%), and right-hand dominant (n=23, 82%). In total, 2800 medication-taking gestures (n=1400, 50% natural plus n=1400, 50% scripted gestures) were used to train the network. During the testing session, 560 natural medication-taking events that were not previously presented to the ANN were used to assess the network. The accuracy, precision, and recall were calculated to confirm the performance of the network. The trained ANN exhibited an average true-positive and true-negative performance of 96.5% and 94.5%, respectively. The network exhibited <5% error in the incorrect classification of medication-taking gestures. Conclusions Smartwatch technology may provide an accurate, nonintrusive means of monitoring complex human behaviors such as natural medication-taking gestures. Future research is warranted to evaluate the efficacy of using modern sensing devices and machine learning algorithms to monitor medication-taking behavior and improve medication adherence.

BACKGROUND Medication adherence is a complex human behavior associated with chronic condition self-management. Medication adherence is a global public health challenge, as only about 50% of people adhere to their medication regimes. Smartphone apps and reminders have shown promising results in promoting medication adherence. However, practical mechanisms to determine whether a medication has been taken or not, once people are reminded, have been elusive. Emerging smartwatch technology may more objectively, unobtrusively, and automatically detect the medication-taking than currently available methods. OBJECTIVE This study aimed to examine the feasibility of detecting natural medication-taking gestures using smartwatches. METHODS Recruited participants (N=28) ranged in age (20 to 60 years) and comprised 57.0% males and 43.0% females. The majority were college students (71.4%), single (86.0%), and working at least part-time (61.0%). The sample represented racial diversity with 4% African American, 43% Asian, 43% White, and 10% reported two or more races1. Most participants were right-hand dominant (82%), while only 1 participant (4%) was ambidextrous. During data collection, each participant recorded at least five protocol-guided (scripted) medication-taking events (sMTE) and at least ten natural instances of medication-taking events (nMTE) per day for 5 days. Using a smartwatch, the accelerometer data was recorded for each session at 25Hz of sampling rate. The raw recordings were scrutinized by a team member to validate the accuracy of self-reports. The validated data were used to train an Artificial Neural Network (ANN) to detect a medication-taking activity. The training and testing data included previously recorded accelerometer data from smoking, eating, and jogging activities in addition to the medication-taking data recorded in this work. The accuracy of the model to identify medication-taking was evaluated by comparing the ANN’s output to the actual output. RESULTS In total, 2,800 medication-taking gestures (1400 natural plus 1400 scripted gestures) were used to train the network. During the testing session, 560 nMTE events that were not previously presented to the ANN were used to assess the network. Various metrics, such as accuracy, precision, and recall, were calculated to confirm the performance of the network. The trained ANN exhibited an average True-Positive performance of 96.5% and an average True-Negative performance of 94.5%. The network exhibited less than 5% error in incorrect classification of the medication-taking gestures. CONCLUSIONS Smartwatch technology can provide an accurate, non-intrusive means of monitoring human behaviors such as natural medication-taking gestures. The use of machine learning algorithms combined with modern sensing devices may significantly improve medication adherence and monitoring.