|Year : 2019 | Volume
| Issue : 4 | Page : 221-223
Smartphone technology for mental health services
Abhijit R Rozatkar1, Nitin Gupta2
1 Department of Psychiatry, All India Institute of Medical Sciences, Bhopal, Madhya Pradesh, India
2 Department of Psychiatry, Government Medical College and Hospital, Chandigarh, India
|Date of Submission||21-Oct-2019|
|Date of Decision||21-Oct-2019|
|Date of Acceptance||21-Oct-2019|
|Date of Web Publication||15-Nov-2019|
Dr. Abhijit R Rozatkar
Department of Psychiatry, All India Institute of Medical Sciences, Bhopal, Madhya Pradesh
Source of Support: None, Conflict of Interest: None
|How to cite this article:|
Rozatkar AR, Gupta N. Smartphone technology for mental health services. Indian J Soc Psychiatry 2019;35:221-3
It is interesting to note that the 26th National Conference of Indian Association for Social Psychiatry (NCIASP) being organized at Bhubaneshwar, Odisha, from November 29 to December 1, 2019, has the theme of “Social Psychiatry in Digital Age.” The theme itself is a reflection as to how an area like social psychiatry is having to embrace this concept due to the advancements in the way information is being made available, disseminated, accessed, utilized, and interpreted.
Digital technology has significantly impacted the way we interact with each other and our environment. The latest data suggest that nearly 4.5 billion people in the world are currently connected to the internet, and nearly 3.3 billion people are smartphone users. In the Indian context, nearly one-quarter of the population are smartphone users, and in all likelihood, these numbers shall continue to rise. Smartphones are not just telecommunication devices but a tool to organize and navigate our daily lives. Apart from social media applications [apps], numerous apps are available on smartphone platforms such as android and iTunes for self-management, improving thinking skills, skill training, illness management, passive symptom tracking, and data collection.
| Application/usage|| |
The assessment of psychiatric disorders involves evaluation of mood, thinking, sociality, biological patterns, and other domains. To achieve a diagnosis, clinicians must rely on patients' self-reports of mental health symptoms. Since smartphone is a personal device, it provides an opportunity to quantify contextual human behavior rather than a clinic-based assessment. Smartphones with their embedded sensors can be used to log patient activity, generating what is called as digital phenotyping (or personal sensing). Digital phenotyping has been defined as “moment-by-moment quantification of the individual-level human phenotype in situ using data from personal digital devices.” The source of data for digital phenotyping can be (a) sensors in smartphone and connected wearables, (b) content on social media platform, and (c) electronic health records. Much of these data are now passively collected rather than actively collecting this information through surveys, thereby removing the linguistic and cultural barriers in information gathering. This differentiates digital phenotyping from the Ecological Momentary Assessment approach, where self-reporting of behavior and symptoms, is done on a digital device by actively logging the experience as close to the event as possible.
Smartphone with wearable devices can provide data not only of spatial location Global Positioning System (GPS) and activity (accelerometer) but also can be used for biological data such as heart rate and rhythm, blood pressure, respiratory rate, and sweating. Spatial location data are helpful in understanding regularity in job/education, confinement to home, frequency of visits to shopping areas or alcohol-serving joints, etc., Accelerometer data can be used to identify energy expenditure in physical activity and identification of medication-induced side effects such as tremors. Data from wearable devices connected to smartphones can augment information related to autonomic hyperarousal. Inference can also be drawn on sleep pattern by smartphone usage, although it may overestimate the sleep duration and fail to identify sleep disturbance. Social networks and social dynamics can be understood from call logs, text logs, and social media activity. Alterations on communication logs can be expected in affective disorders regarding number of communications made, responded, and time spent on communication.
Data obtained from these sensors (raw data) can be further processed to identify changes in other behavior domains. Changes in facial expression (obtained from smartphone camera), changes in vocal patterns (obtained from smartphone microphone), and self-disclosure content from social media can be analyzed by methods such as natural language processing (NLP), image processing, and artificial intelligence. This development is akin to developments in the field of genetic sequencing and psychiatric brain imaging, where statistical methods were used for clinical correlation of genetic and imaging data.
| Application in Mental Illnesses|| |
A recent study identified increasing trend in research related to digital phenotyping in the past 15 years. Of the 1831 articles identified between 2004 and 2017, many have focused on autism spectrum disorders and depression, while serious mental disorders such as schizophrenia, suicide, substance use, and bipolar disorder are recently gaining attention. The clinical applications of digital phenotyping include screening, identifying, treatment monitoring, relapse prevention, and early interventions in the range of psychiatric disorders. Sano et al., showed that by measuring skin conductance and temperature, wearable sensors could accurate indentify stress in students. Smartphone-based data have been used to identify individuals at risk for depression and anxiety. The identification of such at-risk and undiagnosed individuals provides an opportunity of early identification and treatment. Using Linguistic Inquiry and Word count and regression analysis, O'Dea et al. identified that twitter posts strongly concerning suicide were characterized by a higher word count, increased use of first-person pronouns, and more references to death. Online screening of social media content, although has privacy concerns, could possibly identify such unanticipated suicidal behaviors. Smartphone data can also help identify drinking episodes in youth which can then trigger prevention intervention., Smartphone-based apps have also been used to identify symptoms of opioid intoxication.
The principal author of this article is associated with the Smartphone Health Assessment for Relapse Prevention (SHARP) study which is an international digital health study that leverages the smartphone app learn, assess, manage, and prevent (LAMP). LAMP is an open-source smartphone platform built by the digital psychiatry division at Beth Israel Deaconess Medical Center, Boston, in collaboration with an approved vendor that can sync with a smartwatch and capture physiology data into the app. It also offers a range of unique games, assessments, and tools designed to measure cognition. By collectively analyzing activity (via GPS), steps/sleep/exercise (via accelerometer in the phone and on the smartwatch), cognition (via on-screen neuropsychological assessments), and symptoms (via surveys), it can create a personalized summary of how the user experiences his/her mental illness. Preliminary evidence has been collected for relapse prediction in schizophrenia using LAMP and to demonstrate that the use of the app is safe. In the first phase of the SHARP study, revisions and updates will be made to LAMP to enhance user-experience and meet the specified needs of clinicians, patients, and their family members at sites in Boston, US and Bengaluru and Bhopal, India. In the second phase, study participants will use LAMP to report mood symptoms and behaviors, wear a smartwatch, and utilize cognitive games and assessments available on the app. These active data will be supplemented by passive data additionally captured through the app that highlights the participants' physical and digital behaviors. This will help the study team draw conclusions with regards to the acceptability, feasibility, and clinical impact of LAMP in preventing relapse for persons with schizophrenia.
| Critique|| |
From a technical and research perspective, although information from smartphone appears ideal for understanding behavior of an individual, sensors generate thousands of data points in a day which can make analysis difficult. Further, data from sensors can have errors due to malfunction or damage. Sensor-based data arises from multiple sources, and this heterogeneity is challenging for data mining. Our understanding of the association of these signals with mental health conditions is still nascent. Wearable sensing devices such as glasses and smartwatches, despite its immense potential applicability, have been found to be intrusive and unacceptable to participants. Text content analysis is hampered by the use of slangs, abbreviations, emoticons, or misspellings. Content generated may also be fake which makes data analysis prone to erroneous conclusions. Data from social media platform and from electronic health care are also huge (hundreds of exabytes) and will continue to rise. From a user perspective, it is important to understand that most apps do not have peer-reviewed research to support their claim, and it is unlikely that every mental health app will go through a randomized, controlled research trial to test for effectiveness. More importantly, users, including patients, are concerned about issues related to privacy. User confidence can be improved by measures to increase data security through encryption and password protection. Asymmetric data encryption, wherein the phone cannot read its own data once encrypted is also helpful. Users receiving feedback on their passive data on social media use may develop anxiety or paranoia, i.e., symptoms may be produced by providing feedback.
To conclude, smartphone-based digital phenotyping can offer significant insight into human behavior. Implications for such insights can be thought of as akin to what telescopes did to astronomy and what microscopes did to understanding infective pathology. The growing popularity of such devices combined with evolving analytic techniques such as machine learning and NLP can provide alternate understanding and classification of behavior pathology. Judicious use of such information can provide avenues for early identification, treatment monitoring, and relapse prevention for those with mental illness. As alluded to in the beginning, the deliberations during the 26th NCIASP, Bhubaneshwar will hopefully set the scene as to how “social psychiatry” can embrace the concept of “digital technology” and utilize it judiciously and productively for the future in order to ensure that its identity remains maintained by conducting cutting edge research, ensuring quality service delivery and cost-cum-quality effective resource development.
| References|| |
Onnela JP, Rauch SL. Harnessing smartphone-based digital phenotyping to enhance behavioral and mental health. Neuropsychopharmacology 2016;41:1691-6.
Huckvale K, Venkatesh S, Christensen H. Toward clinical digital phenotyping: A timely opportunity to consider purpose, quality, and safety. NPJ Digit Med 2019;2:88.
Torous J, Kiang MV, Lorme J, Onnela JP. New tools for new research in psychiatry: A Scalable and customizable platform to empower data driven smartphone research. JMIR Ment Health 2016;3:e16.
Liang Y, Zheng X, Zeng DD. A survey on big data-driven digital phenotyping of mental health. Inf Fusion 2019:52;290-307.
Shiffman S, Stone AA, Hufford MR. Ecological momentary assessment. Annu Rev Clin Psychol 2008;4:1-32.
Sano A, Taylor S, McHill AW, Phillips AJ, Barger LK, Klerman E, et al.
Identifying objective physiological markers and modifiable behaviors for self-reported stress and mental health status using wearable sensors and mobile phones: Observational study. J Med Internet Res 2018;20:e210.
Place S, Blanch-Hartigan D, Rubin C, Gorrostieta C, Mead C, Kane J, et al.
Behavioral indicators on a mobile sensing platform predict clinically validated psychiatric symptoms of mood and anxiety disorders. J Med Internet Res 2017;19:e75.
O'Dea B, Larsen ME, Batterham PJ, Calear AL, Christensen H. A linguistic analysis of suicide-related twitter posts. Crisis 2017;38:319-29.
Santani D, Trinh-Minh-Tri Do, Labhart F, Landolt S, Kuntsche E, Gatica-Perez D. DrinkSense: Characterizing youth drinking behavior using smartphones. IEEE Trans Mob Comput 2018;17:2279-92.
Nandakumar R, Gollakota S, Sunshine JE. Opioid overdose detection using smartphones. Sci Transl Med 2019;11. pii: eaau8914.
Barnett I, Torous J, Staples P, Sandoval L, Keshavan M, Onnela JP. Relapse prediction in schizophrenia through digital phenotyping: A pilot study. Neuropsychopharmacology 2018;43:1660-6.
Fang R, Pouyanfar S, Yang Y, Chen C, Iyengar SS. Computational health informatics in the big data age: a survey. ACM Comput Surv 2016;49:36.