Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Feb 24.
Published in final edited form as: Nurs Outlook. 2020 May 8;68(5):548–559. doi: 10.1016/j.outlook.2020.03.007

The potential of digital phenotyping to advance the contributions of mhealth to self-management science

Kavita Radhakrishnan 1, Miyong T Kim 2, Marissa Burgermaster 3, Richard Allen Brown 4, Bo Xie 5, Molly S Bray 6, Catherine A Fournier 7
PMCID: PMC12927669  NIHMSID: NIHMS2135878  PMID: 32402392

Abstract

Digital phenotyping consists of moment-by-moment quantification of behavioral data from individual people, typically collected passively from smartphones and other sensors. Within the evolving context of precision health, digital phenotyping can advance the use of mHealth-based self-management tools and interventions by enabling more accurate prediction for prevention and treatment, facilitating supportive strategies, and informing the development of features to motivate self-management behaviors within real-world conditions. This represents an advancement in self-management science: with digital phenotyping, nurse scientists have opportunities to tailor interventions with increased precision. In this paper, we discuss the emergence of digital phenotyping, the historical background of ecological momentary assessment, and the current state of the science of digital phenotyping, with implications for research design, computational requirements, and ethical considerations in self-management science, as well as limitations.

Introduction

The role of mobile health (mHealth) interventions in self-management research is expanding exponentially. These interventions, which deliver self-management support via the internet, short messaging services, and mobile applications, show promise for improving self-management outcomes (Hamine, Gerth-Guyette, Faulx, Green, & Ginsburg, 2015; Tinschert, Jakob, Barata, Kramer, & Kowatsch, 2017; Whitehead & Seaton, 2016). Examples of the major benefits of mHealth interventions include tracking and monitoring of symptoms for better symptom control (Whitehead & Seaton 2016), on-going education and support for disease management (Heinrich, Schaper, & de Vries (2015), and access to timely communication with healthcare providers (Lancaster et al., 2018; Vorungati, Grunfeld, Makuwaza, & Bender, 2017). However, the rapid development and deployment of mHealth technology has so far resulted in a one-size-fits-all approach. Current mHealth interventions are generally developed for the average person, with insufficient emphasis on differences among individuals or on the impacts of mHealth interventions on health outcomes (Noah et al., 2018). This lack of precision tailoring can reduce a patient’s engagement and even engender a reactive mistrust of intervention technology (López, Green, Tan-McGrory, King, & Betancourt, 2011; O’Leary, Vizer, Eschler, Ralston, & Pratt 2015). Nevertheless, if such problems with engagement can be resolved, mHealth still holds great promise for supporting patients’ self-management.

In this article, we present the concept of digital phenotyping, a technique that can help realize the promise of mHealth interventions for self-management. We provide current definitions of digital phenotyping and suggest potential applications of this technique for self-management in the precision health era. We include a discussion of digital phenotyping’s emergence, the influence of ecological momentary assessment (EMA), and the current state of the science of digital phenotyping, with implications for research design, computational requirements, and ethical considerations in self-management science—as well as limitations.

Current mobile technology, with the increasing availability of affordable smartphones and wearable devices including sensors, has resulted in the profuse collection of physiological or behavioral data. Instead of periodic cross-sectional assessments of self-reports that are prone to recall bias and under-reporting, this technology allows measurement in situ of an individual’s physiological parameters such as heart rate or calories expended or of an individual’s behavioral patterns such as daily activity or sleep. This moment-by-moment quantification of one’s passively collected physiological and behavioral data is termed digital phenotyping (Torrous, Kiang, Lorme, & Onnela, 2016; Onnela 2016). The tailoring of mHealth interventions through personalization can be facilitated by digital phenotyping and can help predict health outcomes and stratify risk for individual patients (Onnela, 2016).

The National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK, 2018) has defined a behavioral or psychological phenotype as “a pattern of behavior or psychological characteristics that are measurable/quantifiable and distinct (explains individual variation).” Onnela and Rauch (2016) have argued that the digital phenotype (Jain, Powers, Hawkins, & Brownstein, 2015), a static digital construct that can define the attributes of an individual’s behavior (e.g., the digital profile of an insomniac could include night-time posts on the social media platform Twitter), is distinct from digital phenotyping, which is a dynamic construct that alludes to passive, ongoing, multidimensional digital measurement of the individual’s behaviors (e.g., clicks on the phone, time of texts, sleep tracking, etc.). Both constructs, however, are useful for the self-management of chronic conditions. Another term that describes digital patterns in behavior and physiology is digital biomarkers, defined as the “objective, quantifiable, physiological, and behavioral data that are collected and measured by means of digital devices, such as embedded environmental sensors or wearables” (Piau, Wild, Mattek, & Kaye 2019, p. 1). Our definition of digital phenotyping is inspired by the NIDDK’s definition; here we are proposing digital phenotyping of patterns of health behaviors and personal characteristics that are measurable, quantifiable, and distinct; digital phenotyping can thus explain individual variation in self-management.

Precision health—targeted, predictive, personalized healthcare—has the potential to improve the efficiency and efficacy of chronic disease management interventions. The digital data for health behaviors and parameters, such as, for example, times in the day when physical activity interventions are most amenable for certain sets of individual characteristics, can inform precision mHealth interventions (Kitsiou, Paré, Jaana, & Gerber, 2017). In developing such interventions, one must gain an understanding of patterns in a subpopulation’s adoption of, engagement with, and response to mHealth interventions. Moreover, to optimize an individual’s self-management responses to precision mHealth interventions, one must also determine the optimal intervention dose in terms of frequency and intensity for the individual. Digital phenotyping can enable more accurate prediction for treatment and prevention, facilitate supportive strategies for adherence and maintenance of change, and inform the development of mHealth intervention features in order to “effectively and efficiently identify who will maximally benefit” (Hickey et al., 2019, p. 463). Together, multidimensional data gathered passively through digital phenotyping, such as patterns in daily activity, calories expended, heart rate, body temperature, sleep, sweat, and mood, can quantify an individual’s physical behaviors, which is often a focus in self-management interventions. Interventions promoting self-management can then be concentrated on those who will benefit, sparing expense and potential unintended adverse consequences for those who will not. As knowledge in the field accumulates, more personalized tailoring methods can be applied on the basis of different digital phenotypes. Efforts are already underway to provide recommendations for developing and evaluating digital behavioral interventions based on varying responses to required durations and dimensions of engagement, pace of interventions, efficiency, effectiveness, and cost (Michie, Yardley, West, Patrick, & Greaves 2017; Voils et al., 2014).

Emergence of Digital Phenotyping

Historically, digital phenotyping derives from methods associated with ecological momentary assessment (EMA), the repeated measurement of an individual’s behaviors and experiences in real time within the context of the individual’s natural environment (Shiffman, Stone, & Hufford, 2008). EMA’s benefits include minimizing recall bias, maximizing ecological validity, and enabling the study of factors or processes that influence behavior in the real world (Shiffman et al., 2008). Early tools for EMA ranged from written diaries and telephone interviews to electronic diaries and physiological sensors (Bertz, Epstein, & Preston, 2018), all of which required active participant input. Since EMA was originally used by behavioral scientists to study mental health issues (e.g., addictive behaviors such as smoking, drinking alcohol, or overeating, as well as mood affects), it is not surprising that digital phenotyping was first applied by behavioral scientists in research on mental health.

The current emergence of digital phenotyping is tied to the proliferation of wearable sensors and mHealth applications (apps). Revenue for the sale of wearable devices for the year of 2019 is estimated at more than $3 billion in the U.S., with nearly 40 million users (Statista, 2019). Over 318,000 health apps are available from the top app stores worldwide, with more than 200 being added each day (IQVIA, 2017). Nearly half (48%) of healthcare consumers are using mHealth apps, up from just 16% in 2014 (Francis, 2018). Current innovations in wireless connectivity have enabled mobile phones to seamlessly capture real-time behavior data from behavior-tracking physical devices and sensors by routing the data to mobile phones (Zanella, Bui, Castellani, Vangelista, & Zorzi, 2014). However, mHealth and sensor tools can suffer from limitations in quality, usability, accuracy, or sustained interest (Liew, Zhang, See, & Ong, 2019; Rosenberger, Buman, Haskell, McConnell, & Carstensen, 2016), which suggests that one should proceed cautiously in the application of digital phenotyping in mHealth. For example, the popular augmented reality mHealth game Pokemon Go was purported to promote fun physical activity, but it suffered a rapid drop-off in gameplay rate within a single season. Nevertheless, the increased use of smartphones and wearable sensor devices offers opportunities to track individuals’ self-management routines and behaviors, which can inform the tailoring of self-management interventions through digital phenotyping.

Digital Phenotyping for Precision Tailoring

Nursing science has made significant contributions to self-management science, especially by focusing on populations that suffer from health disparities (Grey, Knafl, & McCorkle, 2006; Grady, 2017; Riegel & Dickson, 2008). Nurse scientists have advanced ways to tailor self-management interventions to address the unique features of target populations. For example, nurse scientists have tailored self-management intervention protocols for cultural sensitivity (Brown, Garcia, Orlander, & Hanis, 2017; H. Kim, Song, Han, Kim, & Kim, 2013; M. T. Kim et al., 2015) and have adapted them for levels of literacy (Rosal et al., 2011), cognitive ability (Stuifbergen et al., 2012), and physical function (Bragina et al., 2017). Influenced by the premises of precision health initiatives, nurse scientists now have a new opportunity to expand the methodology of tailoring (Hickey et al., 2019) through the use of digital phenotyping.

Unfortunately, however, generalizations based on demographic characteristics such as age, ethnicity, or income level have led to stereotypical assumptions about certain subpopulations’ ability to engage with mobile technology-assisted self-management interventions. Such stereotypical assumptions can worsen disparities in access to quality care (O’Leary et al., 2015). Yet the assumption that low English proficiency and older age are barriers to the use of mobile technology interventions has been invalidated by recent studies (Huh et al., 2018; Radhakrishnan et al., 2016). In addition to demographic variables, individual characteristics helpful for personalizing or tailoring mHealth interventions can include physical and cognitive functioning (Bateman et al., 2017; Koo & Vizer, 2019); affective states of fatigue, anxiety, and depression (Moberg, Niles, & Beermann, 2019); sleep (Moon et al., 2017); health literacy (Rosal et al., 2011); digital literacy/eHealth literacy (Kreps, 2017); and attitudes toward the adoption of mHealth technology (Cajita, Hodgson, Lam, Yoo, & Han, 2018). All of these characteristics can inform our understanding of individuals’ responses to mobile technology interventions for the self-management of chronic conditions. Digital phenotyping may help explain why individuals provided with the same self-management technology intervention and clinical diagnosis may respond to the intervention in different ways and have dissimilar outcomes for health and quality of life (QOL). As the design of mHealth tools and interventions becomes adapted and more personalized to meet the unique needs of individuals, “the ‘self’ in self-management will become a focal point for tailored interventions” (O’Leary et al., 2015, p. 992).

Digital Data Collection for Phenotyping

There is no shortage of digital data that can be obtained through both passive and active data collection for the purpose of digital phenotyping. Passive forms of measurement record individuals’ behaviors at a granular level, removing the response burden typically involved in active self-monitoring (Skinner et al., 2017). Smartphones and connected sensors can passively monitor self-management behavior patterns related to physical activity, sleep, medication adherence, eating, smoking, or mobility, through changes in steps, calories, heart rate, hand movements, medication absorption, position of lids of pill boxes, or body temperature. The research platform Beiwe, for example, includes a smartphone app designed specifically for digital phenotyping. This app passively collects metadata minute-by-minute, using GPS, WiFi, accelerometry, text and telephone logs, and times for screen on and off (Onnela & Rauch, 2016). The latest versions of smartwatches can measure activity, sleep, and EKG, with caveats for validity of intense physical activity (Khushhal et al., 2017) or sleep (Lee, Byuon, Keill, Dinkel & Seo, 2018), as well as a lack of regulatory approvals for use of most smartwatches as medical devices (Larson, 2018). To provide actionable insights about addiction or lifestyle behaviors for preventive management, other smartphone apps can collect data related to hand movements associated with smoking or eating (Chun, Bhattacharya, & Thomaz, 2018; Senyurek, Imtiaz, Belsare, Tiffany, & Sazonov, 2019; Skinner, Stone, Doughty, & Munafò, 2019).

Continuous monitoring and improved tracking of an individual’s symptoms, activities, sleep, behaviors, and treatment adherence can allow precise determination of triggers in disease exacerbations and a reliable assessment of a self-management intervention’s effectiveness. Such data can define a digital phenotype for self-management to better understand intrapersonal factors that influence an individual’s self-management behaviors, the effectiveness of specific features in the individual’s self-management intervention, and resulting health outcomes. Despite challenges, mHealth interventions thus hold great potential to facilitate self-management interventions on a large scale, even though the interventions focus precisely on specific individuals.

State of the Science of Digital Phenotyping

In September 2019, our search of the PubMed and CINAHL databases using the terms “digital phenotype” or “digital phenotyping” yielded a total of 82 journal articles, 12 of which reported original studies on the application of digital phenotyping for chronic diseases. All 12 studies employed smartphone apps, with the most commonly collected data being from self-report surveys (10 studies), followed by activity and movement (8); screen time, phone call and text logs (7); geo-location (6); and sleep (3). Across the 12 studies, 9 variables were examined in relation to mental health functioning (e.g., stress levels or depressive symptoms). Although none of the 12 studies included nurses as authors, research centers led by nurse scientists have begun the training and mentoring of nurse researchers to design studies that utilize digital data for informing insights on self-management behaviors (Hickey et al., 2019).

Digital phenotyping has primarily been applied to classify or stratify individuals by levels of daily stress, depressive symptoms, or suicidal thoughts (Kleiman et al., 2018; Saeb, Lattie, Schueller, Kording, & Mohr, 2016; Sarda, Munuswamy, Sarda, & Subramanian, 2019; Wahle, Kowatsch, Fleisch, Rufer, & Weidt, 2016). Algorithms based on dense digital data were used to predict onset of mental health crises such as relapses in schizophrenia or increased severity of mood disturbances (Barnett, Torous, Staples, Sandoval, et al., 2018; Choet al., 2019; Zulueta et al., 2018). These algorithms were also used to explore the validity of objective digital data collected from mobile or wearable devices in comparison with self-reported mental health symptoms or status (Rohani, Faurholt-Jepsen, Kessing, & Bardram, 2018; Sano et al., 2018) and to obtain insights for providing timely support for mental illness crises (Wahle et al., 2016). Three of the 12 studies reported on the application of digital phenotyping for physiological diseases. These applications included stratifying treatment compliance levels of children with asthma (Jaimini et al., 2018), quantifying the progression of amyotrophic lateral sclerosis (Berry et al., 2019), and quantifying mobility and QOL in people with spine diseases (Cote, Barnett, Onnela, & Smith, 2019). In other studies, the term digital phenotyping was not specifically used, but the concept has been applied in forecasting glucose levels on the basis of typical self-monitoring practices in individuals with diabetes (Albers et al., 2017).

Digital Phenotyping Considerations for Evaluating and Promoting Self-Management Behaviors

Research Study Designs

Evidence from digital phenotyping studies includes numerous repeated measures over time on individuals’ health behaviors obtained with digital tools (Vieira, McDonald, Araújo-Soares, Sniehotta, & Henderson, 2017). This requires research designs that test hypotheses and interventions at the individual level and that accommodate individualized sequences and strategies involved in adaptive intervention delivery (e.g., to reduce frequencies of alerts, switch incentives, increase the intensity of messages; Murphy, Lynch, Oslin, McKay, & TenHave, 2007). Study designs for just-in-time adaptive interventions (JITAIs) require a “sequence of decision rules that specify whether, how, when (timing), and based on which measures, to alter the dosage (duration, frequency or amount), type, or delivery of treatment(s) at decision stages in the course of care” (Almirall, Nahum-Shani, Sherwood, & Murphy, 2014, p. 262).

Conventional parallel-group randomized controlled trials (RCTs) can assess what is best on average for a given population, but they are limited with respect to the individual participant. Study designs such as n-of-1 trials or micro-randomized trials (MRTs) assess what is best for individual patients. Such designs are thus particularly well suited to precision health and to the emerging field of digital phenotyping (Lillie et al., 2011; Mirza, Punja, Vohra, & Guyatt, 2017). For JITAI trials, which adjust to the individual’s evolving status, n-of-1 trials and MRT designs (Klasnja et al., 2015; Klasnja et al., 2019) provide the rigor and processes necessary to elicit evidence for tailoring interventions at the individual level. Moreover, such designs incorporate key elements of RCTs that reduce bias, such as randomization (of intervention sequence) and blinding (of patients, care providers, outcome assessors, and data analysts). Interventions in which adaptations are delivered in a momentary way, usually by relying on data from smartphones or other mobile devices with sensors, are embedded within an MRT or n-of-1 trial design for testing and evaluation. Momentary interventions allow testing of the tailoring variables and the intervention components in the same trial, and they allow clinicians to develop the best decision rules based on research rather than a priori decisions (Murphy et al., 2007). Whereas RCTs evaluate interventions that have already been developed, individuals in MRTs are randomized hundreds or thousands of times over the course of the study, so that the trial can construct multicomponent interventions by assessing the effects of intervention components and their variance over time, as well as each individual’s context (Klasnja et al., 2015). In n-of-1 trials, on the other hand, interventions and the order of interventions are randomized multiple times within a person (Methodology Center, n.d.-a). In response to concerns about the generalizability of insights from n-of-1 trials, aggregated analysis of several n-of-1 trials is also a possibility (Vieira et al., 2017).

Such study designs, however, do have the downside of increased cost and complex logistics in assigning individuals to several interventions at multiple timepoints. Multiphase optimization strategy (MOST) is a research design framework in which many interventions are tested in phases before the most optimized interventions are carried out in an RCT (Collins, Murphy, & Stretcher, 2007; Methodology Center, n.d.-b). Such research design frameworks are most suited for evaluating multicomponent behavioral interventions, and can provide a balance between cost, complexity, and rigor. The sequential multiple-assignment randomized trial (SMART) is a special tool of MOST, in which a dynamic treatment regimen can be efficiently optimized by “identifying the best set of decision points, decision rules, treatment options, and tailoring variables” (Collins, Nahum-Shani, & Almirall, 2014, p.429). Although a full comparison of the different adaptive study designs is beyond the scope of this article, the JITAI resource page at The Pennsylvania State University and its cited articles provide good resources for designing such studies (Methodology Center, n.d.-a). Future digital phenotyping studies can provide further evidence on the feasibility and value of different study designs for evaluating longitudinal interventions with long follow-up periods and many repeated measures that can be adapted to the individual’s real-world conditions.

Analysis and Computation

Mobile and sensor-based tools have the potential to provide longitudinal, high-frequency, rich data from individuals’ everyday self-management behaviors—substantial quantities of data. Up to 1,000,000 data points per day have been obtained in studies that collect raw sensor data from smartphone-based tracking of individuals’ activities or behaviors (Barnett, Torous, Staples, Keshavan, & Onnela, 2018). Converting this vast amount of information into actionable insights for self-management behaviors necessitates the extraction of meaningful behavioral features from digital data as well as appropriate statistical analysis of associations of behaviors with relevant health outcomes (Barnett, Torous, Staples, Keshavan, & Onnela, 2018). Since digital phenotyping generates big data, machine learning algorithms and related statistical methods can be useful for inferences. Barnett, Torous, Staples, Keshavan, and Onnela (2018) have demonstrated the application of three statistical methods—mixed models, multiple comparisons of correlated tests, and dimension reduction for correlated behavioral covariates—for drawing valid conclusions in digital phenotyping studies. Table 1 presents a comparison of the three statistical methods with respect to use case scenarios, decisions about preferred options, and considerations for usage.

Table 1:

Approaches for Statistical Analysis of Digital Phenotyping Data

Analysis Method Use Case Scenario Preferred Approach When: Considerations during analysis
1. Mixed models
1.1. Generalized estimating equations (GEE)


1.2. Generalized linear mixed models (GLMM)		 Correlations between daily behaviors and symptoms 1.1. Model misspecification is high or outcome distribution is unknown.
1.2.Model misspecification is not severe and large numbers of observations are obtained for each subject; has higher accuracy.		 a. Random effect structures and computational considerations based on sample size
b. Missing data: Data imputation prior to analysis of complete data, or multiple imputation approaches (Barnett & Onnela, 2016)
2. Correction for multiple statistical tests to avoid the possibility of abundant false positive results
2.1. Family-wise error rate (FWER)


2.2. False discovery rate (FDR)		 Correlations between each self-reported outcome and a behavioral marker, requiring the need for many simultaneous statistical tests 2.1. More conservative approach is needed, but may lead to high rate of false-negatives
2.2. More lenient approach is suited. May be more suited to the exploratory nature of current digital phenotyping studies		 Many outcomes and behavioral features can be highly correlated, so there is a need for correction of the assumption of independence of tests, to optimally reduce the rate of false negatives.
3. Dimensionality reduction, to reduce the number of redundant predictors
3.1. Principal component analysis (PCA) or factor analysis
3.2. Least squares PCA		 Correlation blocks obtained by clustering of digital phenotyping predictors around a specific behavior type 3.1. There are no missing values in the data

3.2. Data include some missing values.		 Block correlation structures can allay the potential of poor interpretability of transformed predictor variables.
Open in a new tab

Implications of Digital Phenotyping for Self-Management Science

Smartphones and wearable devices capture both an individual’s behavior and environment (e.g., sleep, light, temperature). Therefore, digital phenotyping data can facilitate an understanding of how individuals perform behaviors (e.g., pace of activity or napping) in real time and in response to specific features of mHealth interventions (e.g., timing of alerts or notifications) under real-world conditions (Marsch, 2018). Digital phenotyping can also enable an understanding of the reactive effects of self-monitoring actions on behavior. Digital phenotyping data can inform guidelines for optimal dosage and for frequency and intensity of self-management behaviors that are realistic and relevant to an individual’s level of motivation, intrapersonal factors, and environment (Marsch, 2018). These behavior guidelines can inform the design of effective JITAIs that provide the right amount of self-management support to individuals precisely when it is most effective, in order to gradually advance the individual’s behaviors in a desired direction (Marsch, 2018; Nahum-Shani, Hekler, & Spruijt-Metz, 2015; Skinner et al., 2017). Such precise prescriptions for behaviors that individuals can actually attain are in line with precision health initiatives and can contribute to improved health outcomes.

Digital phenotyping also provides a novel approach to identify and predict as well as incentivize an individual’s self-management behaviors (Jain et al., 2015). For example, data from digital phenotyping can be used to address personal or environmental attributes that predict an individual’s need for human facilitators or technology brokers, such as community health workers or family members for mHealth-based self-management interventions. Digital phenotyping could also capture aspects of social support such as reactions to competition, intellectual stimulation, peer support, family support, and accountability to clinicians, or promote the perception of being accountable to a digital monitoring device that can motivate the specific individual. Digital phenotyping can inform structures and timing of incentives within a mobile digital device based on individuals’ interactions with the digital device (e.g., via voice, texts, alert frequency and timing, reminders; Torous, Onnela, & Keshavan, 2017). Finally, digital phenotyping can inform feedback messages from the device that support an individual’s engagement in self-management behaviors as well as in mHealth-based self-management interventions.

Insights from digital phenotyping studies can also inform the adaptation of mHealth technology interventions based on individuals’ responses to interventions in the presence of varying levels of physical or cognitive functioning (e.g., presence of physical disability or gait variances; varying levels of cognitive impairment). Subcultural influences (e.g., suburban vs. rural, caregivers, military) on responses to messages or preferences for incentives are still unclear. Digital phenotyping has the potential to provide insights on when individuals may be most receptive to messages providing information or support through mHealth self-management interventions (Marsch, 2018). Digital phenotyping research can develop algorithms that match specific intervention elements with individuals’ attributes and engagement in self-management behaviors, enabling the delivery of precise interventions to motivate behavior changes.

Ethical Implications

Traditionally, the term digital divide alludes to lower access to technology among those who face disparities due to lower socio-economic status, older age, or rural residency with little high-speed internet (Anderson & Perrin, 2019). This divide can be a significant barrier to scaling up digital self-management interventions and realizing digital phenotyping’s promise. However, the current global increase of mobile phone ownership offers low-cost possibilities for introducing digital self-management and thus empowering individuals (Huh et al., 2018). Seventy-seven percent of Americans now own smartphones, up from just 35% in 2011 (Pew Research Center, 2018). In the groups with a yearly income of less than $30,000, smartphone ownership drops to 67%, but it is over 80% in groups with incomes above that level (Pew Research Center, 2018). Although disparities in smartphone ownership remain, sweeping generalizations about lack of access to technology or lack of motivation among end users as core barriers to digital intervention adoption in underserved populations are no longer accurate (Bennet et al., 2014). Digital phenotyping studies can provide valuable insights on how to personalize mHealth interventions to support the self-management needs of diverse socio-economic groups and cultures (Huh et al., 2018).

The promise of digital phenotyping to achieve the aims of precision health initiatives for self-management can be realized only if individuals are willing to wear and engage with sensors or other devices that mHealth technology-assisted self-management interventions require (Skinner et al., 2017). Therefore, research on digital phenotyping can benefit greatly from user-centered design (Norman & Draper, 1986) and from usability studies that address the contexts, needs, and preferences of each user. Moreover, to reduce inequities traditionally associated with mHealth interventions, usability considerations are especially pertinent for tailoring digital phenotyping interventions for individuals with lower literacy or those who are differently abled (Veinot, Mitchell, & Ancker, 2018). Finally, the advancement of self-management science through digital phenotyping must not exclude the consideration of the human touch (K. B. Kim et al., 2016; Radhakrishnan, Jacelon, & Roche, 2012). Individuals with chronic diseases in nursing studies using technology interventions such as telehealth have articulated their acceptance for digital interventions as long as it does not completely replace human touch. The value-added benefit of digital phenotyping should be to accentuate the human touch, as opposed to diminishing it, by freeing up clinicians to spend more quality time with their patients. Insights provided by digital phenotyping might be most beneficial by helping to sort out contextual clues that support self-management, ultimately enhancing human touch.

Major ethical implications of digital phenotyping especially relevant today are implicated in issues of data ownership, security, and privacy. With the use of data acquisition tools in digital phenotyping studies come formidable challenges to protecting the privacy and security of individuals and their information. Geolocation data obtained through GPS and integrated with behavior data can have negative consequences if the data fall into wrong hands (e.g., of stalkers or criminals). Other unintended consequences might consist of decisions on health insurance and employment eligibility based on digital health behavior data, which might be sold to third parties by digital device companies. To demonstrate the value of digital phenotyping for self-management interventions, one must be able to ensure trust. It is imperative to provide individuals with ownership of their digital data, the right to grant access to view their data and any related application of their data, and the right to consent regarding which data they might share with researchers and clinicians as well as their family caregivers.

Several policy initiatives have been proposed and endorsed to protect individuals’ rights over their personal digital health data. The European Union has passed the General Data Protection Regulation (European Commission, 2018), which allows consumers to demand erasure of their personal information or prohibit companies from selling their digital data to third parties. Along the same lines, U.S. Senators Klobuchar and Murkowski have introduced the Protecting Personal Health Data Act (2019; see also Ackerman, Kraus, Ponder, & Tonsager, 2019), which proposes greater consumer control over how personal health data are utilized, as well as the right to delete and amend health data and to access a copy of one’s personal health data free of cost, in an accessible format. In addition, among other measures, the bill proposes a National Task Force on Health Data Protection to explore encryption standards, transfer protocols, and de-identification techniques. As of 2020, the California Consumer Privacy Act (Privacy, 2018) will require all internet of things devices that are made or sold in California (e.g., activity trackers, refrigerators, thermostats) to be equipped with unique passwords or to require users to set their own unique passwords. Forcing password measures on internet of things devices could prevent cybersecurity attacks that might be carried out by hijacking such devices and infecting laptop/desktop computers in the home (Information Privacy, 2019). In 2018, a similar bill at the federal level called the Internet of Things Cybersecurity Improvement Act was proposed by Senators Warner and Gardner, but it has yet to gain traction (Hawkins, 2018; Warner, 2017). The Food and Drug Administration’s software precertification program is an example of a regulatory mechanism to ensure that commercialization of digital health products is “high-quality, safe and effective while providing appropriate patient safeguards” such as protecting individuals’ rights over their digital data and their data’s applications (U.S. Food and Drug Administration, 2019, p. 1). It is imperative for researchers and clinicians to be aware of these policy initiatives to protect digital health data and ensure necessary data protections in their research and practice in order to realize the value of digital phenotyping ethically.

Limitations and Future Directions

Although advances in technology have vastly improved the capability of smartphones and sensors to collect behavioral data, there are still limitations to mHealth technologies’ ability to passively collect accurate information for certain critical behaviors, such as diet. For the foreseeable future, a need remains for digital phenotyping to be combined with traditional types of data collection. A meaningful area for research would be the development of methods to combine diverse sources and types of data including genomics or other omics data in such a way as to provide an interpretable picture of an individual’s self-management behaviors.

In self-management science, one must also acknowledge the merit of actively collecting data such as an individual’s self-reflection on his/her behavior in juxtaposition with digital phenotyping. Self-reflection on one’s own behavior is emerging as a major component of behavioral change, aided today by mHealth-based self-monitoring activities. The intensity of such active data collection should be commensurate with an individual’s optimal response burden.

Conclusion

Digital phenotyping holds the potential to enable researchers to realize the promise of mHealth-based self-management interventions within precision health initiatives. Digital phenotyping is a novel approach to identify and predict as well as incentivize individuals’ self-management behaviors within the context of their social, cognitive, and physical functioning under real-world conditions. However, the advancement of digital phenotyping in self-management must comply with ethical considerations as well as the demands of computation and research design. The data from digital phenotyping will also provide answers to missing gaps regarding intermediary ingredients in mHealth-based interventions that must be addressed to enable the interventions’ success in motivating self-management behaviors in individuals within real-world conditions.

Acknowledgment

Editorial support with manuscript development was provided by the Cain Center for Nursing Research and the Center for Transdisciplinary Collaborative Research in Self-Management Science (P30, NR015335) at The University of Texas at Austin School of Nursing.

Funding Sources

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Contributor Information

Kavita Radhakrishnan, Assistant Professor, School of Nursing, The University of Texas – Austin, 1710 Red River Street, Austin, TX 78701-1499, UNITED STATES.

Miyong T. Kim, La Quinta Centennial Endowed Professor, School of Nursing, The University of Texas – Austin, 1710 Red River Street, Austin, TX 78701-1499, UNITED STATES.

Marissa Burgermaster, Assistant Professor, Department of Population Health, Department of Nutritional Sciences, The University of Texas – Austin, 103 W 24TH ST, Austin, TX 78712, UNITED STATES.

Richard Allen Brown, Research Professor, School of Nursing, The University of Texas – Austin, 1710 Red River Street, Austin, TX 78701-1499, UNITED STATES.

Bo Xie, Professor, School of Nursing, School of Information, The University of Texas – Austin, 1710 Red River Street, Austin, TX 78701-1499, UNITED STATES.

Molly S Bray, Julian C. Barton Endowed Professor, School of Nutrition, Professor, Department of Pediatrics, The University of Texas – Austin, 103 W 24TH St, Austin, TX 78712, UNITED STATES.

Catherine A. Fournier, Robert Woods Johnson Fellow, Research Assistant, School of Nursing, The University of Texas – Austin, 1710 Red River Street, Austin, TX 78701-1499, UNITED STATES.

References

  1. Ackerman W, Kraus AD, Ponder J, & Tonsager l. (2019, June 25). Senators introduce legislation to regulate privacy and security of wearable health devices and genetic testing kits. Covington Digital Health. Retrieved from https://www.covingtondigitalhealth.com/2019/06/senators-introduce-legislation-to-regulate-privacy-and-security-of-wearable-health-devices-and-genetic-testing-kits/
  2. Albers DJ, Levine M, Gluckman B, Ginsberg H, Hripcsak G, & Mamykina L (2017). Personalized glucose forecasting for type 2 diabetes using data assimilation. PLOS Computational Biology, 13(4), Article e1005232. 10.1371/journal.pcbi.1005232 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Almirall D, Nahum-Shani I, Sherwood NE, & Murphy SA (2014). Introduction to SMART designs for the development of adaptive interventions: With application to weight loss research. Translational Behavioral Medicine, 4(3), 260–274. 10.1007/s13142-014-0265-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Anderson M, & Perrin A (2019). 10% of American’s don’t use the internet. Who are they? Retrieved from https://www.pewresearch.org/fact-tank/2019/04/22/some-americans-dont-use-the-internet-who-are-they/
  5. Barnett I, & Onnela J-P (2016). Inferring mobility measures from GPS traces with missing data. Retrieved from https://arxiv.org/abs/1606.06328 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Barnett I, Torous J, Staples P, Keshavan M, & Onnela J-P (2018). Beyond smartphones and sensors: Choosing appropriate statistical methods for the analysis of longitudinal data. Journal of the American Medical Informatics Association, 25(12), 1669–1674. 10.1093/jamia/ocy121 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Barnett I, Torous J, Staples P, Sandoval L, Keshavan M, & Onnela J-P (2018). Relapse prediction in schizophrenia through digital phenotyping: A pilot study. Neuropsychopharmacology, 43(8), 1660–1666. 10.1038/s41386-018-0030-z [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Bateman DR, Srinivas B, Emmett TW, Schleyer TK, Holden RJ, Hendrie HC, & Callahan CM (2017). Categorizing health outcomes and efficacy of mHealth apps for persons with cognitive impairment: A systematic review. Journal of Medical Internet Research, 19(8), Article e301. 10.2196/jmir.7814 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bennett GG, Steinberg DM, Stoute C, Lanpher M, Lane I, Askew S, ... & Baskin ML (2014). Electronic health (eHealth) interventions for weight management among racial/minority adults: A systematic review. Obesity Reviews, 15(S4), 146–158. 10.1111/obr.12218 [DOI] [PubMed] [Google Scholar]
  10. Berry JD, Paganoni S, Carlson K, Burke K, Weber H, Staples P, … Onnela JP (2019). Design and results of a smartphone-based digital phenotyping study to quantify ALS progression. Annals of Clinical and Translational Neurology, 6(5), 873–881. 10.1002/acn3.770 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bertz JW, Epstein DH, & Preston KL (2018). Combining ecological momentary assessment with objective, ambulatory measures of behavior and physiology in substance-use research. Addictive Behaviors, 83, 5–17. 10.1016/j.addbeh.2017.11.027 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Bragina I, Lippke S, Pischke C, Meyer J, Muellmann S, Rost E, ... Voelcker-Rehage C (2017). PROMOTE: tailoring physical activity interventions to promote health ageing. Innovation in Aging, 1(Suppl. 1), 221. 10.1093/geroni/igx004.828 [DOI] [Google Scholar]
  13. Brown SA, García AA, Orlander PR, & Hanis CL (2017). A randomized clinical trial of diabetes self-management for Mexican Americans: Are there serendipitous health benefits for supporters of study participants? SAGE Open Medicine, 5, Article 2050312116682125. 10.1177/2050312116682125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Cajita MI, Hodgson NA, Lam KW, Yoo S, Han HR (2018). Facilitators of and barriers to mHealth adoption in older adults with heart failure. Computers, Informatics and Nursing, 36(8), 376–382. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Cho CH, Lee T, Kim MG, In HP, Kim L, & Lee HJ Mood prediction of patients with mood disorders by machine learning using passive digital phenotypes based on the circadian rhythm: Prospective observational cohort study. (2019). Journal of Medical Internet Research, 21(4), e11029. 10.2196/11029 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Chun KS, Bhattacharya S, & Thomaz E (2018). Detecting eating episodes by tracking jawbone movements with a non-contact wearable sensor. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies, 2(1), Article 4. 10.1145/3191736 [DOI] [Google Scholar]
  17. Collins LM, Murphy SA, & Strecher V (2007). The multiphase optimization strategy (MOST) and the sequential multiple assignment randomized trial (SMART): New methods for more potent eHealth interventions. American Journal of Preventive Medicine, 32(5, Suppl.), S112–S118. 10.1016/j.amepre.2007.01.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Collins LM, Nahum-Shani I, & Almirall D (2014). Optimization of behavioral dynamic treatment regimens based on the sequential, multiple assignment, randomized trial (SMART). Clinical Trials, 11(4), 426–434. 10.1177/1740774514536795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Cote DJ, Barnett I, Onnela J-P, & Smith TR (2019). Digital phenotyping in patients with spine disease: A novel approach to quantifying mobility and quality of life. World Neurosurgery, 126, e241–e249. 10.1016/j.wneu.2019.01.297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. European Commission. (2018). EU data protection rules. Retrieved September 1, 2019 from https://ec.europa.eu/commission/priorities/justice-and-fundamental-rights/data-protection/2018-reform-eu-data-protection-rules_en#abouttheregulationanddataprotection [Google Scholar]
  21. Francis J (2018, March 6). Accenture study finds growing demand for digital health services revolutionizing delivery models: Patients, doctors + machines [Accenture news release]. Retrieved from https://newsroom.accenture.com/news/accenture-study-finds-growing-demand-for-digital-health-services-revolutionizing-delivery-models-patients-doctors-machines.htm [Google Scholar]
  22. Grady PA (2017). Investigating the determinants of health: The role of nursing science. Nursing Outlook, 65(5), 489–491. 10.1016/j.outlook.2017.04.010 [DOI] [PubMed] [Google Scholar]
  23. Grey M, Knafl K, & McCorkle R (2006). A framework for the study of self- and family management of chronic conditions. Nursing Outlook, 54(5), 278–286. 10.1016/j.outlook.2006.06.004 [DOI] [PubMed] [Google Scholar]
  24. Hamine S, Gerth-Guyette E, Faulx D, Green BB, & Ginsburg AS (2015). Impact of mHealth chronic disease management on treatment adherence and patient outcomes: A systematic review. Journal of Medical Internet Research, 17(2), Article e52. 10.2196/jmir.3951 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hawkins D (2018, October 8). The Cybersecurity 202: California’s new Internet of Things law only protects against a small portion of cyberthreats. Washington Post. https://www.washingtonpost.com/news/powerpost/paloma/the-cybersecurity-202/2018/10/08/the-cybersecurity-202-california-s-new-internet-of-things-law-only-protects-against-a-small-portion-of-cyberthreats/5bba75781b326b7c8a8d1885/?noredirect=on
  26. Heinrich E, Schaper NC, & de Vries NK (2015). Self-management interventions for type 2 diabetes: A systematic review. European Diabetes Nursing, 7(2), 71–76. 10.1002/edn.160 [DOI] [Google Scholar]
  27. Hickey KT, Bakken S, Byrne MW, Bailey DE, Demiris G, Docherty SL, . . . Grady PA (2019). Precision health: Advancing symptom and self-management science. Nursing Outlook, 67(4), 462–475. 10.1016/j.outlook.2019.01.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Huh J, Koola J, Contreras A, Castillo AKP, Ruiz M, Tedone KG, . . . Schiaffino MK (2018). Consumer health informatics adoption among underserved populations: Thinking beyond the digital divide. Yearbook of Medical Informatics, 27(1), 146–155. 10.1055/s-0038-1641217 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Information privacy: Digital health feedback systems, California Assembly Bill AB-384. (2019). https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201920200AB384 [Google Scholar]
  30. IQVIA Institute for Human Data Science. (2017, November 7). The growing value of digital health: Evidence and impact on human health and the healthcare system [Institute report]. Retrieved from https://www.iqvia.com/institute/reports/the-growing-value-of-digital-health
  31. Jaimini U, Thirunarayan K, Kalra M, Venkataraman R, Kadariya D, & Sheth A (2018). “How is my child’s asthma?” Digital phenotype and actionable insights for pediatric asthma. JMIR Pediatrics and Parenting, 1(2), Article e11988. 10.2196/11988 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Jain SH, Powers BW, Hawkins JB, & Brownstein JS (2015). The digital phenotype. Nature Biotechnology, 33(5), 462–463. 10.1038/nbt.3223 [DOI] [PubMed] [Google Scholar]
  33. Khushhal A, Nichols S, Evans W, Gleadall-Siddall DO, Page R, O’Doherty AF, … Abt G (2017). Validity and reliability of the Apple watch for measuring heart rate during exercise. Sports Medicine International Open, 1(6), E206–E211. 10.1055/s-0043-120195 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Kim H, Song H-J, Han H-R, Kim KB, & Kim MT (2013). Translation and validation of the Dietary Approaches to Stop Hypertension for Koreans intervention: Culturally tailored dietary guidelines for Korean Americans with high blood pressure. The Journal of Cardiovascular Nursing, 28(6), 514–523. 10.1097/JCN.0b013e318262c0c1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Kim KB, Kim MT, Lee HB, Nguyen T, Bone LR, & Levine D (2016). Community health workers versus nurses as counselors or case managers in a self-help diabetes management program. American Journal of Public Health, 106(6), 1052–1058. 10.2105/AJPH.2016.303054 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kim MT, Kim KB, Huh B, Nguyen T, Han H-R, Bone LR, & Levine D (2015). The effect of a community-based self-help intervention: Korean Americans with type 2 diabetes. American Journal of Preventive Medicine, 49(5), 726–737. 10.1016/j.amepre.2015.04.033 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Kitsiou S, Paré G, Jaana M & Gerber B (2017). Effectiveness of mHealth interventions for patients with diabetes: An overview of systematic reviews. PLoS ONE, 12(3), Article e0173160. 10.1371/journal.pone.0173160 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Klasnja P, Hekler EB, Shiffman S, Boruvka A, Almirall D, Tewari A, & Murphy SA (2015). Microrandomized trials: An experimental design for developing just-in-time adaptive interventions. Health Psychology, 34(Suppl.), 1220–1228. 10.1037/hea0000305 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Klasnja P, Smith S, Seewald NJ, Lee A, Hall K, Luers B, . . . Murphy SA (2019). Efficacy of contextually tailored suggestions for physical activity: A micro-randomized optimization trial of HeartSteps. Annals of Behavioral Medicine, 53(6), 573–582. 10.1093/abm/kay067 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Kleiman EM, Turner BJ, Fedor S, Beale EE, Picard RW, Huffman JC, & Nock MK (2018). Digital phenotyping of suicidal thoughts. Depression and Anxiety, 35(7), 601–608. 10.1002/da.22730 [DOI] [PubMed] [Google Scholar]
  41. Koo BM, & Vizer LM (2019). Mobile technology for cognitive assessment of older adults: A scoping review. Innovation in Aging, 3(1), Article igy038. 10.1093/geroni/igy038 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Kreps GL (2017). The relevance of health literacy to mHealth. Information Services & Use, 37, 123–130. 10.3233/ISU-170828 [DOI] [PubMed] [Google Scholar]
  43. Lancaster K, Abuzour A, Khaira M, Mathers A, Chan A, Bui V, ... Dolovich L (2018). The use and effects of electronic health tools for patient self-monitoring and reporting of outcomes following medication use: Systematic review. Journal of Medical Internet Research, 20(12), Article e294. 10.2196/jmir.9284 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Larson RS (2018). A path to better-quality mHealth apps. JMIR mHealth and uHealth, 6(7), Article e10414. 10.2196/10414 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Lee J-M, Byun W, Keill A, Dinkel D, & Seo Y (2018). Comparison of wearable trackers’ ability to estimate sleep. International Journal of Environmental Research and Public Health, 15(6), Article 1265. 10.3390/ijerph15061265 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Liew MS, Zhang J, See J, & Ong YL (2019). Usability challenges for health and wellness mobile apps: Mixed-methods study among mHealth experts and consumers. JMIR mHealth and uHealth, 7(1), Article e12160. 10.2196/12160 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Lillie EO, Patay B, Diamant J, Issell B, Topol EJ, & Schork NJ (2011). The n-of-1 clinical trial: The ultimate strategy for individualizing medicine? Personalized Medicine, 8(2), 161–173. 10.2217/pme.11.7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. López L, Green AR, Tan-McGrory A, King R, & Betancourt JR (2011). Bridging the digital divide in health care: The role of health information technology in addressing racial and ethnic disparities. The Joint Commission Journal on Quality and Patient Safety, 37(10), 437–445. 10.1016/S1553-7250(11)37055-9 [DOI] [PubMed] [Google Scholar]
  49. Marsch LA (2018). Opportunities and needs in digital phenotyping. Neuropsychopharmacology, 43(8), 1637–1638. 10.1038/s41386-018-0051-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. The Methodology Center, Pennsylvania State University College of Health and Human Development (n.d.-a). Just-in-time adaptive interventions and adaptive interventions. Retrieved July 19, 2019 from https://www.methodology.psu.edu/ra/adap-inter/
  51. The Methodology Center, Pennsylvania State University College of Health and Human Development (n.d.-b). Overview of MOST. Retrieved July 19, 2019 from https://www.methodology.psu.edu/ra/most/research/
  52. Michie S, Yardley L, West R, Patrick K, & Greaves F (2017). Developing and evaluating digital interventions to promote behavior change in health and health care: Recommendations resulting from an international workshop. Journal of Medical Internet Research, 19(6), Article e232. 10.2196/jmir.7126 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Mirza RD, Punja S, Vohra S, & Guyatt G (2017). The history and development of N-of-1 trials. Journal of the Royal Society of Medicine, 110(8), 330–340. 10.1177/0141076817721131 [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Moberg C, Niles A, & Beermann D (2019). Guided self-help works: Randomized waitlist controlled trial of Pacifica, a mobile app integrating cognitive behavioral therapy and mindfulness for stress, anxiety, and depression. Journal of Medical Internet Research, 21(6), Article e12556. 10.2196/12556 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Moon RY, Hauck FR, Colson ER, Kellams AL, Geller NL, Heeren T, ... Corwin MJ (2017). The effect of nursing quality improvement and mobile health interventions on infant sleep practices: A randomized clinical trial. JAMA, 318(4), 351–359. 10.1001/jama.2017.8982 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Murphy SA, Lynch KG, Oslin D, McKay JR, & TenHave T (2007). Developing adaptive treatment strategies in substance abuse research. Drug and Alcohol Dependence, 88(Suppl. 2), S24–S30. 10.1016/j.drugalcdep.2006.09.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Nahum-Shani I, Hekler EB, & Spruijt-Metz D (2015). Building health behavior models to guide the development of just-in-time adaptive interventions: A pragmatic framework. Health Psychology, 34(Suppl.), 1209–1219. 10.1037/hea0000306 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. National Institute of Diabetes and Digestive and Kidney Diseases. (2018). Ancillary studies to identify behavioral and/or psychological phenotypes contributing to obesity (Program Announcement PAR-18–105). Retrieved June 18, 2019 from https://grants.nih.gov/grants/guide/pa-files/par-18-105.html [Google Scholar]
  59. Noah B, Keller MS, Mosadeghi S, Stein L, Johl S, Delshad S, ... Spiegel BMR (2018). Impact of remote patient monitoring on clinical outcomes: An updated meta-analysis of randomized controlled trials. npj Digital Medicine, 1, Article 20172. https://www.nature.com/articles/s41746-017-0002-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Norman DA, & Draper SW (Eds.). (1986). User centered system design: New perspectives on human-computer interaction. Hillsdale, NJ: Lawrence Erlbaum Associates; 1986. [Google Scholar]
  61. O’Leary K, Vizer L, Eschler J, Ralston J, & Pratt W (2015). Understanding patients’ health and technology attitudes for tailoring self-management interventions. AMIA Annual Symposium Proceedings, 2015, 991–1000. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4765611/ [PMC free article] [PubMed] [Google Scholar]
  62. Onnela J-P, & Rauch SL (2016). Harnessing smartphone-based digital phenotyping to enhance behavioral and mental health. Neuropsychopharmacology, 41(7), 1691–1696. 10.1038/npp.2016.7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Pew Research Center. (2018, February 5). Mobile fact sheet. Retrieved from https://www.pewinternet.org/fact-sheet/mobile/
  64. Piau A, Wild K, Mattek N, & Kaye J (2019). Current state of digital biomarker technologies for real-life, home-based monitoring of cognitive function for mild cognitive impairment to mild Alzheimer disease and implications for clinical care: Systematic review. Journal of Medical Internet Research, 21(8), e12785. 10.2196/12785 [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Privacy: Personal Information: Businesses, California Assembly Bill AB-375. (2018). https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201720180AB375 [Google Scholar]
  66. Protecting Personal Health Data Act, S.1842, 116th Cong. (2019). https://www.congress.gov/bill/116th-congress/senatebill/1842/
  67. Radhakrishnan K, Jacelon C, & Roche J (2012) Perceptions on the use of telehealth by homecare nurses and patients for heart failure: A mixed method study. Home Healthcare Management & Practice, 24(4), 175–181. 10.1177/1084822311428335 [DOI] [Google Scholar]
  68. Radhakrishnan K, Toprac P, O’Hair M, Bias R, Kim MT, Bradley P, & Mackert M (2016). Interactive digital e-Health game for heart failure self-management: A feasibility study. Games for Health Journal, 5(6), 366–374. 10.1089/g4h.2016.0038 [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Riegel B, & Dickson VV (2008). A situation-specific theory of heart failure self-care. Journal of Cardiovascular Nursing, 23(3), 190–196. 10.1097/01.JCN.0000305091.35259.85 [DOI] [PubMed] [Google Scholar]
  70. Rohani DA, Faurholt-Jepsen M, Kessing LV, & Bardram JE (2018). Correlations between objective behavioral features collected from mobile and wearable devices and depressive mood symptoms in patients with affective disorders: Systematic review. JMIR mHealth and uHealth, 6(8), e165. 10.2196/mhealth.9691 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Rosal MC, Ockene IS, Restrepo A, White MJ, Borg A, Olendzki B, ... Reed G (2011). Randomized trial of a literacy-sensitive, culturally tailored diabetes self-management intervention for low-income Latinos: Latinos en control. Diabetes Care, 34(4), 838–844. 10.2337/dc10-1981 [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Rosenberger ME, Buman MP, Haskell WL, McConnell MV, & Carstensen LL (2016). Twenty-four hours of sleep, sedentary behavior, and physical activity with nine wearable devices. Medine & Science in Sports & Exercise, 48(3), 457–465. 10.1249/MSS.0000000000000778 [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Saeb S, Lattie EG, Schueller SM, Kording KP, & Mohr DC (2016). The relationship between mobile phone location sensor data and depressive symptom severity. PeerJ, 4, Article e2537. 10.7717/peerj.2537 [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Sano A, Taylor S, McHill AW, Phillips AJK, Barger LK, Klerman E, & Picard R (2018). Identifying objective physiological markers and modifiable behaviors for self-reported stress and mental health status using wearable sensors and mobile phones: Observational study. Journal of Medical Internet Research, 20(6), Article e210. 10.2196/jmir.9410 [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. Sarda A, Munuswamy S, Sarda S, & Subramanian V (2019). Using passive smartphone sensing for improved risk stratification of patients with depression and diabetes: Cross-sectional observational study. JMIR mHealth and uHealth, 7(1), e11041. 10.2196/11041 [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Senyurek V, Imtiaz M, Belsare P, Tiffany S, & Sazonov E (2019). Cigarette smoking detection with an inertial sensor and a smart lighter. Sensors, 19(3), Article 570. 10.3390/s19030570 [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Shiffman S, Stone AA, & Hufford MR (2008). Ecological momentary assessment. Annual Review of Clinical Psychology, 4, 1–32. 10.1146/annurev.clinpsy.3.022806.091415 [DOI] [PubMed] [Google Scholar]
  78. Skinner AL, Attwood AS, Baddeley R, Evans-Reeves K, Bauld L, & Munafò MR (2017). Digital phenotyping and the development and delivery of health guidelines and behaviour change interventions. Addiction, 112(7), 1281–1285. 10.1111/add.13746 [DOI] [PubMed] [Google Scholar]
  79. Skinner AL, Stone CJ, Doughty H, & Munafò MR (2019). StopWatch: The preliminary evaluation of a smartwatch-based system for passive detection of cigarette smoking. Nicotine & Tobacco Research, 21(2), 257–261. 10.1093/ntr/nty008 [DOI] [PMC free article] [PubMed] [Google Scholar]
  80. Statista. (2019). Expert tools: Wearables, United States. Retrieved from https://www.statista.com/outlook/319/109/wearables/united-states#market-revenue
  81. Stuifbergen A, Becker H, Perez F, Mor[r]ison J, Kullberg V, & Todd A (2012). A randomized controlled trial of a cognitive rehabilitation intervention for persons with multiple sclerosis. Clinical Rehabilitation, 26(10), 882–893 10.1177/0269215511434997 [DOI] [PMC free article] [PubMed] [Google Scholar]
  82. Tinschert P, Jakob R, Barata F, Kramer J-N, & Kowatsch T (2017). The potential of mobile apps for improving asthma self-management: A review of publicly available and well-adopted asthma apps. JMIR mHealth and uHealth, 5(8), Article e113. 10.2196/mhealth.7177 [DOI] [PMC free article] [PubMed] [Google Scholar]
  83. Torous J, Kiang MV, 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), Article e16. 10.2196/mental.5165 [DOI] [PMC free article] [PubMed] [Google Scholar]
  84. Torous J, Onnela J-P, & Keshavan M (2017). New dimensions and new tools to realize the potential of RDoC: Digital phenotyping via smartphones and connected devices. Translational Psychiatry, 7(3), Article e1053. 10.1038/tp.2017.25 [DOI] [PMC free article] [PubMed] [Google Scholar]
  85. U. S. Food and Drug Administration, Center for Devices and Radiological Health. (2019). Digital health innovation action plan. https://www.fda.gov/media/106331/download
  86. Veinot TC, Mitchell H, & Ancker JS (2018). Good intentions are not enough: How informatics interventions can worsen inequality. Journal of the American Medical Informatics Association, 25(8), 1080–1088. 10.1093/jamia/ocy052 [DOI] [PMC free article] [PubMed] [Google Scholar]
  87. Vieira R, McDonald S, Araújo-Soares V, Sniehotta FF, & Henderson R (2017). Dynamic modelling of n-of-1 data: Powerful and flexible data analytics applied to individualised studies. Health Psychology Review, 11(3), 222–234. 10.1080/17437199.2017.1343680 [DOI] [PubMed] [Google Scholar]
  88. Voils CI, King HA, Maciejewski ML, Allen KD, Yancy WS Jr., & Shaffer JA (2014). Approaches for informing optimal dose of behavioral interventions. Annals of Behavioral Medicine, 48(3), 392–401. 10.1007/s12160-014-9618-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  89. Voruganti T, Grunfeld E, Makuwaza T, & Bender JL (2017). Web-based tools for text-based patient-provider communication in chronic conditions: Scoping review. Journal of Medical Internet Research, 19(10), Article e366. 10.2196/jmir.7987 [DOI] [PMC free article] [PubMed] [Google Scholar]
  90. Wahle F, Kowatsch T, Fleisch E, Rufer M, & Weidt S (2016). Mobile sensing and support for people with depression: A pilot trial in the wild. JMIR mHealth and uHealth, 4(3), e111. 10.2196/mhealth.5960 [DOI] [PMC free article] [PubMed] [Google Scholar]
  91. Warner MR (2017, August 1). Senators introduce bipartisan legislation to improve cybersecurity of “Internet of Things” (IoT) devices [Press release]. https://www.warner.senate.gov/public/index.cfm/2017/8/senators-introduce-bipartisan-legislation-to-improve-cybersecurity-of-internet-of-things-iot-devices
  92. Whitehead L, & Seaton P (2016). The effectiveness of self-management mobile phone and tablet apps in long-term condition management: A systematic review. Journal of Medical Internet Research, 18(5), Article e97. 10.2196/jmir.4883 [DOI] [PMC free article] [PubMed] [Google Scholar]
  93. Zanella A, Bui N, Castellani A, Vangelista L, & Zorzi M (2014). Internet of things for smart cities. IEEE Internet of Things Journal, 1(1), 22–32. 10.1109/JIOT.2014.2306328 [DOI] [Google Scholar]
  94. Zulueta J, Piscitello A, Rasic M, Easter R, Babu P, Langenecker SA, ... Leow A (2018). Predicting mood disturbance severity with mobile phone keystroke metadata: A biaffect digital phenotyping study. Journal of Medical Internet Research, 20(7), Article e241. 10.2196/jmir.9775 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES