Next Article in Journal
Review of Thermal and Physiological Properties of Human Breast Tissue
Next Article in Special Issue
Inertial-Measurement-Unit-Based Novel Human Activity Recognition Algorithm Using Conformer
Previous Article in Journal
LAG: Layered Objects to Generate Better Anchors for Object Detection in Aerial Images
Previous Article in Special Issue
Smartphone as a Disease Screening Tool: A Systematic Review
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:

Opportunities for Smartphone Sensing in E-Health Research: A Narrative Review

Department of Human Centered Computing, Faculty of IT, Monash University, Clayton, VIC 3168, Australia
Author to whom correspondence should be addressed.
Sensors 2022, 22(10), 3893;
Submission received: 1 May 2022 / Revised: 16 May 2022 / Accepted: 19 May 2022 / Published: 20 May 2022
(This article belongs to the Special Issue Advances in E-health and Mobile Health Monitoring)


Recent years have seen significant advances in the sensing capabilities of smartphones, enabling them to collect rich contextual information such as location, device usage, and human activity at a given point in time. Combined with widespread user adoption and the ability to gather user data remotely, smartphone-based sensing has become an appealing choice for health research. Numerous studies over the years have demonstrated the promise of using smartphone-based sensing to monitor a range of health conditions, particularly mental health conditions. However, as research is progressing to develop the predictive capabilities of smartphones, it becomes even more crucial to fully understand the capabilities and limitations of using this technology, given its potential impact on human health. To this end, this paper presents a narrative review of smartphone-sensing literature from the past 5 years, to highlight the opportunities and challenges of this approach in healthcare. It provides an overview of the type of health conditions studied, the types of data collected, tools used, and the challenges encountered in using smartphones for healthcare studies, which aims to serve as a guide for researchers wishing to embark on similar research in the future. Our findings highlight the predominance of mental health studies, discuss the opportunities of using standardized sensing approaches and machine-learning advancements, and present the trends of smartphone sensing in healthcare over the years.

1. Introduction

Over the past decade, there has been a substantial number of technological advancements, especially in terms of the availability of mobile sensing on smartphones. Mobile-phone technology in particular has witnessed tremendous growth, both in functionality and consumer adoption [1]. Traditional mobile phones of the early 2000s have been superseded by smartphones across the world; these offer a greater variety of features beyond calling and messaging, such as browsing the internet and capturing audio/video content. Importantly for eHealth studies, they also include a suite of sensors, which have the capability to capture human behavior and the ability to transmit data collected.
This has been a result of both hardware and software advances, which have led to improvements in performance and power consumption. Devices today are embedded with a rich set of low-power sensors, which can passively record contextual user data. For example, accelerometers are commonly used in smartphones to record the acceleration forces acting upon the device, whereas light sensors are used to measure the amount of light the device is exposed to. While these sensors facilitate data collection, software frameworks facilitate the transformation of this raw data into meaningful insights relevant to human health and wellbeing. If we take the accelerometer as an example, data extracted from this sensor can be used for activity recognition, such as providing an estimate of the steps taken by the user, which might indicate general activity levels of the user [2,3]. Data from the light sensor (which is generally used for the screen auto-brightness feature of smartphones) could give an indication of the time spent in a dark room, which could potentially indicate a decline in the mood or mental health of a user [4,5]. Most smartphones also have the potential to track additional usage statistics, such as screen time, application usage, and battery usage; as such, they have the capacity to capture a large amount of user information and are becoming increasingly context-aware.
In addition, users spend a considerable amount of time using their phones. According to a survey study conducted in February 2021 [6], 46% of US adults stated that they use their phones for an average of 5–6 h per day. The number of hours users spend using their phones, combined with the rich data and user context that is captured by the phones, can provide significant insights about a person. For example, the Global Positioning System (GPS) data collected can provide a snapshot of the places a person visits in a day [7,8,9]. Combined with human activity recognition (“HAR”) algorithms, one could identify how the person commutes to these places—walking, running or by using a vehicle—as well as indicating their level of physical activity [3,10,11]. By combining the insights gained from all of the sensors, one can glean a range of insights about a person’s daily activities. Thus, smartphones to some extent can act as a proxy data source, providing information about a person that would otherwise require significant human effort to collect (e.g., through paper or electronic diaries).
These capabilities, high user adoption, and the relatively unobtrusive nature of smartphones make them an appealing choice for ‘in-the-wild’ healthcare studies [3,10,12,13]. Several studies have utilized smartphones to remotely collect data from users and gain health insights from these data. Studies have explored how smartphone-based sensing can be used to infer mood [14,15,16,17,18,19,20,21], physical activity [3,11,22,23], and mental health [13,24,25,26,27], and also to track the effects of chronic health conditions [28,29,30,31]. Moreover, software frameworks have also been developed to help researchers use smartphone sensing to conduct their own studies, such as AWARE [32], Beiwe [33], and EARS [34] to name a few. These offer considerable promise for the future of eHealth research and, in turn, improved opportunities and outcomes for patients.
At present, there is a significant amount of multidisciplinary work being carried out in the space of smartphone sensing, with diverse motivations, methods, and approaches being used by health practitioners, HCI researchers and computer scientists. The sheer number of studies and different approaches can make it challenging for researchers to evaluate the appropriateness of smartphone-based sensing for different use cases. In this paper, we analyze the existing literature in this space to address this issue and provide a high-level overview of the various use cases, opportunities, and challenges of smartphone sensing in healthcare. After setting out our literature-selection approach (Section 2), we present a full account (Section 3) of the various health conditions studied, types of data collected, tools used, and barriers identified in the literature, which will benefit future researchers by helping them to make high-level decisions regarding different approaches for future studies using smartphone sensing. We then present the wider discussion points (Section 4), including our specific findings: (1) predominance of mental health studies, (2) opportunities for standardization of sensing approaches, (3) opportunities for using machine-learning advancements in eHealth research, and (4) sensing trends over the years and future scope. The discussion is followed up by limitations and future work (Section 5), with the last section being the conclusion (Section 6).

2. Materials and Methods

The main goal of this study was to understand the potential of smartphone sensing in healthcare, with a view towards providing a full understanding of existing practice. As such, the review aimed to understand the following high-level concerns:
  • What health conditions have been examined using smartphone sensing?
  • What data-collection approaches have been used and what are their trade-offs?
  • What applications and resources have been used?
  • What were the researchers’ motivations for their data-collection and -analysis approaches?
To address these aims, our work utilizes a narrative literature-review methodology. Narrative literature reviews (NLRs) provide a broad overview of a given topic by synthesizing previously published literature. They are valuable, as they condense information from a variety of sources and provide an understanding of the domain without having the reader interrogate the entire literature [35]. There are a few key reasons for choosing this methodology over a systematic literature review, which is the approach typically followed in the clinical literature. Firstly, the field of smartphone sensing is very broad and there are numerous subdomains in this space, such as human-activity recognition, behavior prediction, and self-tracking to name a few. This work aims to present a broad perspective of all such areas to provide insights for new researchers in eHealth around the opportunities and challenges in this specific context.
Secondly, there are numerous systematic reviews that provide a comprehensive insight on smartphone sensing for various specific health conditions [31,36,37,38], smartphone-based activity recognition [39,40], and sensing apps and datasets [41]. Rather than replicating these and providing a quantitative insight into the fields, a narrative review allows for a qualitative approach to highlight overall advancements. This review aims to create new knowledge by synthesizing the insights from all these subdomains, thus providing a synopsis of current practices, to aid in the design and deployment of future smartphone-based health studies.
Whilst systematic reviews have guidelines such as the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) [42] that discuss the methods and reporting the findings, no such formal specifications exist for NRLs. However, a few authors have provided a set of recommendations to follow while writing [35,43] and evaluating [44] NRLs. This review follows these recommendations by systematically discussing the literature search terms, search strategy, and inclusion/exclusion criteria.

2.1. Search Terms

The multidisciplinary research team (with expertise in chronic and mental health conditions, disability, assistive technologies, smartphone and physical sensing, and ubiquitous computing) met on multiple occasions to iteratively brainstorm and decide on a set of search terms to explicitly explore how smartphone sensing has been used in the health space. Four keywords were identified, which were (1) “smartphone sensing”; (2) “self-tracking” AND “smartphone sensing”; (3) “quantified self” AND “smartphone sensing”; (4) “mobile sensing” AND “human health”. As we were interested in exploring smartphone-sensing research across the domain of human health (and not focused on a specific condition), these search terms were selected to encompass the breadth of smartphone-sensing literature, allowing for further assessment of the relevance of articles in the human health space, while restricting articles discussing other aspects of mobile sensing and wearables. The fourth search term was selected because mobile sensing is analogous to smartphone sensing, but the additional keyword was added to make the search more relevant to the area of focus.

2.2. Search Strategy

Two databases were selected for the search, Google Scholar and PubMed. Given the rapid development in smartphone technology and software regulatory changes impacting sensing approaches [45,46], the search was restricted to the past 5 years (2017–2022), to discover the latest developments in the space. The four keywords were then individually used to find literature on the two databases. The titles of the search results were read and a list of potentially relevant peer-reviewed, full-text articles in English was created. After the removal of duplicate articles, the abstracts were read to evaluate whether the article met the inclusion criteria.

2.3. Inclusion Criteria

The goal of this literature review was to understand the opportunities and challenges of smartphone sensing in health studies. As a large population over the globe has access to smartphone technology [1], the aim was to explore the feasibility of smartphones as a standalone device in a ‘real-life’ scenario. As such, only studies that discussed smartphone sensing were considered, and literature using body-worn sensors was excluded. Articles were selected if they met the following inclusion criteria:
  • Empirical research that uses smartphone sensing to monitor a health condition;
  • Empirical research that explores perceptions and challenges of smartphone sensing.
The initial literature search provided a total of 3080 hits: 3034 from Google Scholar and 46 from PubMed. After reading the titles for relevance (and excluding duplicates), a total of 222 articles were present. The abstracts of each of these were read to determine if they met the inclusion criteria. For example, although [47] used a smartphone-sensing app to monitor user behavior, the aim was to understand compliance to COVID-19 policies and was thus excluded. Another exclusion example was a study examining smartphone-based self-tracking adherence for chronic health conditions [48]. Although the study used a smartphone self-tracking app, it also allowed the participants to import data from a wearable device and was thus excluded. After the removal of such examples, a total of 86 articles were present.

3. Results

Out of the 86 articles, 71 were empirical studies using smartphone sensing to monitor health conditions; 9 were explicitly discussing the technical and ethical challenges of smartphone sensing; and 6 papers were discussing user perceptions of smartphone sensing. The challenges, user concerns, and opportunities that were discussed in the 15 papers have been described in various sections. The results section first discusses the various health conditions that have been examined, followed by the data-collection approaches used. We then discuss the frameworks used to facilitate data collection, and lastly, the purpose of the data collection (i.e., what the authors were trying to find out in their studies).

3.1. Health Conditions That Have Been Studied

Smartphone sensing has been used to study a wide range of conditions. Not only have studies used smartphone sensing to evidence factors impacting physical wellbeing, but they have also been used to obtain insights about mental health. The health conditions can be categorized into three categories: general wellbeing, chronic health conditions, and mental health conditions. From the 71 empirical studies, mental health conditions were by far the most common (42 articles), followed by general wellbeing (26 articles) and lastly chronic conditions (3 articles), as shown in Figure 1.
Under each of these categories, several specific subconditions were studied. Figure 2 presents an overview of the subconditions, which are discussed in more detail in the following sections.

3.1.1. General Wellbeing

This category discusses various aspects that have an overall impact on wellbeing. The factors that were studied include physical activity, sociability, sleep, diet, and substance use (alcohol, tobacco, cannabis).
The most common factors examined were physical activity/mobility [2,3,10,11,22,23,49,50,51,52,53] and sociability [3,9,54,55]. Sensor data from the accelerometers, gyroscope, and location data (using GPS, Bluetooth) were used to infer factors such as places visited [50], amount of time spent at home [22], types of activity performed (sitting, standing, walking, running) [11] and regularity of movement [3]. On the other hand, location data [53], frequency and duration of social media app use [55], microphone recordings [3] and phone-call and text-message logs [54] were used to provide information about a person’s physical and virtual sociability.
Six studies collected and examined smartphone data as a proxy indicator of a person’s sleep [19,51,56,57,58,59]. Given that a smartphone cannot directly record a person’s sleep, studies used a combination of passively sensed data and user reports to detect sleeping behaviors. Identifying smartphone-usage behaviors during bedtime, such as putting the phone on charge [58] and lack of screen activity [19] and sound [57] assisted in estimating the user’s sleeping behavior.
A further six studies examined the use of smartphone sensing to detect intake and intoxication due to substances such as alcohol [60,61,62], tobacco [63], and cannabis [12,64]. They utilized location and activity data to understand the contexts of consumption and its impact on mobility. For example, cannabis intake would impact gait and reduce physical activity, which can be monitored through accelerometer [64] and location data [12]. This ability to detect intake instances was seen to provide useful opportunities for intervention using smartphones [63].
Finally, one study investigated smartphone-usage behaviors related to food consumption [65]. They explored whether factors such as sociability, location, and activity could be used to infer food-consumption levels of college students and detect episodes of overeating.

3.1.2. Chronic Health Condition (Parkinson’s)

Chronic health conditions have a long-lasting impact on a person’s health and require ongoing medical support [66]. Parkinson’s disease [28,29,30] was the only condition that was discussed in the articles (another study worked with people with diabetes but focused on symptoms of depression [67]). These studies explored the use of smartphone sensing to monitor gait (manner of walking) [28,29,30], posture [28], and voice [30]. Accelerometer and gyroscope data were commonly used in all of the studies, as they examined movement-related symptoms. Additionally, Ref. [30] explored the impact of medication intake on voice by using the microphone sensor.

3.1.3. Mental Health Conditions

The majority of articles (59%) used smartphone sensing to infer mental health conditions. While six articles examined overall mental health [13,24,25,26,27,68], other studies examined specific factors such as mood [15,16,17,18,19,20,21,69,70,71,72,73,74,75] and stress [76,77,78]. Additionally, studies also examined specific mental health conditions such as depression [7,8,67,79,80,81,82,83,84,85,86,87,88,89], schizophrenia [90,91,92,93], and bipolar disorder [94,95].
Given that physical activities and general wellbeing have an impact on mental health, there was some overlap in the type of data used. Studies in this category used a variety of smartphone data, ranging from sensor data (e.g., accelerometers, light sensors) to software features such as application usage, phone and message logs, etc. Such data, along with self-reports, was used to infer the mental health of the users. For example, both [84,86] used microphone-sensor data combined with self-reports to infer depressive behavior by measuring symptoms of depression such as decreased sociability and disturbed sleep. By combining passively sensed data and different forms of user self-reports (e.g., clinical questionnaires [67,79,86] and ecological momentary assessments [16,70,95]), smartphone sensing was widely used to monitor and predict the mental wellbeing of individuals.

3.2. Data-Collection Approaches

This section discusses the different approaches and types of data collected by the studies. Data collection was conducted using two approaches: actively collected data and passively sensed data. As the name suggests, active data collection involved collecting user input and relied on user compliance. On the other hand, passive data collection used the embedded device sensors to collect user data and relied less heavily on user input and compliance. The majority of the 71 studies (84.5%) used both types of data collection, while the remainder used either active (7%) or passive (8.5%) data. The types of active and passively collected data are discussed in more detail in the following subsections.

3.2.1. Actively Collected Data

Active data collection requires regular user input and relies on user compliance. It requires users to ‘track’ or ‘log’ one or more factors about their health, thus forming a data source, which is then analyzed to obtain insights. The ability to set up schedules and prompt users for input facilitates active data collection. Although these types of data are not sensed and collected directly via user input, they are crucial to consider as they provided useful context to passively sensed data. Studies relied on actively collected data to understand factors that could not be sensed directly (such as mood and mental health) and to label passively collected data as a means to obtain ‘ground truth’ (e.g., users self-labeling their mobility activities to assist in activity-recognition systems). Most papers (91.5%) factored in some level of active data collection and collected different types of data from the users. The types of data collected were as follows:
Demographic information
As a one-off means of data collection, studies collected demographic data from the participants prior to study commencement via smartphone (e.g., [3,55,62]). Information such as age, sex, ethnicity, education, etc., were collected, which assisted in understanding the participants and contextualized findings. The vast majority of studies incorporated demographic-information collection.
Clinical scale/questionnaire responses
Self-collected clinical questionnaires were commonly administered via smartphone to assess factors related to the user’s wellbeing. They include a set of questions which can evaluate health conditions and their severity. Such questionnaires are validated by the medical community and have proven to be reliable indicators of health factors.
Studies used different clinical questionnaires based on the health condition being studied. For example, most studies monitoring depression used the Patient Health Questionnaire (PHQ) [67,79,80,86], which is used to measure depression severity [96]. Another example is that of the UCLA Loneliness Scale [97], which was used by studies monitoring sociability of the participants [22,27]. There are numerous other scales available, which could potentially be used in future studies.
Ecological Momentary Assessments (EMA)/self-reports
The majority of the studies used some forms of Ecological Momentary Assessments to collect self-reported user data. Ecological Momentary Assessments involve repeatedly collecting real-time user information in natural environments [98].
Studies collected information about a variety of metrics to obtain momentary insights about the users. These included factors such as food intake [65], perceived loneliness [9], mood [16,70,95], and stress [22,70], to name a few. Although the questions are not standardized across studies (unlike clinical questionnaires), these too are useful in understanding the user contexts. Moreover, self-reports have been used to add additional context by ‘labeling’ passively sensed data and obtaining ground truth [11,53,57,88], which aids in prediction processes using machine learning.
To summarize, active data collection has been used in most studies and enables the collection of data that may be difficult to interpret from passively sensed data. Table 1 highlights the advantages and disadvantages of active sensing.

3.2.2. Passively Sensed Data

The majority of studies (93%) utilized passively sensed data, by using smartphone sensors for data collection. While there were a few common in-built hardware sensors such as the accelerometer [3,11,53] and GPS/location sensor [9,60,69] used in most studies (62% and 53%, respectively), some also looked at software features such as application usage [61,91]. Studies used a range of sensors and software features to collect various types of contextual information. Table 2 describes the various sensors used, their sensing capabilities, and the type of inferences that have been made from collected data.
Besides these conventional types of sensor and software features, several used less common approaches to infer health conditions. For example, two studies used custom keyboards to collect keystroke/keypress data—one to analyze the sentiment of typed text [78], and the other to infer alcohol intoxication (greater duration between consecutive keypresses, more text deletions and insertions) [62]. One study used the barometer sensor for activity recognition and found that it reduced the misclassification of stair-climbing/descending activity [11]. One study examined the correlation between internet-usage data and PHQ (patient health questionnaire) scores, which provide an indication of depression [87]. Finally, one study explored if social media data (Twitter and Instagram) could provide insight into the mental health of individuals [17].
The potential of smartphone sensing is evident from the wide range of sensor data used in the studies. However, a few important considerations were discussed when deploying passive sensing. These are as follows:
Trade-off between power consumption and data-collection rate
Although the wide range of available sensors facilitate diverse data collection, the types of sensors used and the data-collection rate influence power consumption. A higher sampling rate provides larger amounts of data points, but also has a greater impact on battery life [3,53,101]. It is crucial to ensure that the power consumption of the sensing application is minimal, to ensure user acceptance [15,99,101,102]. To minimize the impact on the battery, studies reduced the rate at which data were collected to ensure that their applications were able to collect data for extended periods of time [3,53,84,86]. Moreover, few studies compared the power consumption to other commonly used applications to ensure that it is within acceptable standards [15,67]. As such, this is an important trade-off to consider when using passive sensing.
Placement of device
The placement of the device largely influences the accuracy of passive data collection. In contrast to active data collection, where the user provides data when the device is with them, this may not always be the case with passive sensing. There may be circumstances where the users might not carry their devices with them or might not carry them as intended [2,3,11,103]. For example, users may not carry their devices at all or keep their devices in handbags, which may affect the accelerometer and gyroscope readings, leading to inaccuracies in human-activity recognition systems [2,11]. Another example (using the light sensor) might be in situations where the device is kept facing downwards, which would give inaccurate readings of users’ surrounding light conditions [19,84]. Such factors must be accounted for while setting up passive sensing.
Data storage and transmission.
Data collected by sensors is naturally required to be stored, processed, and analyzed to obtain insights. Deciding whether the data would be stored locally on the device or transmitted to an external database is another factor to consider [101]. For example, one study that performed real-time analysis stored the data locally on the device [57], whereas other studies temporarily stored data locally and uploaded it to external servers when a Wi-Fi network was available [11,52,53] to reduce data costs. Data storage and transmission costs are important considerations for user acceptance [99].
Device operating system
The availability of different sensors/software features is dependent on the underlying operating system (Android, iOS). Both systems have different policies in terms of data collection and access. The iOS platform has stricter data-protection regulations and prohibits the direct collection of certain types of data [55]. Studies discussed limitations on collecting data from iOS such as application usage [54,55], GPS and Bluetooth [8,9,88], keyboard presses [62], and phone-call/text-message logs [54,55]. Given such restrictions, it was unsurprising that the majority of the articles used Android devices for their study. Figure 3 presents the distribution of operating systems used in the studies. (Note: one of the studies did not specify the platform that was used [68]).
Privacy concerns.
Given the variety and volume of data collected, it is essential to ensure data privacy and confidentiality. Although the topic of privacy in mobile sensing is a whole issue in itself [99,100,104,105,106,107,108], few studies have discussed approaches to minimize privacy invasion of users. Data such as audio recordings from the microphone, keyboard presses, and content of text messages are some examples of particularly sensitive data. For example, studies using the microphone sensor avoided collecting raw, continuous audio data. They collected limited samples to infer higher-level attributes such as amplitude levels [69,84] and presence/absence of voice [3,84,86]. Similarly, the study using keyboard presses only collected every third word and did not collect any passwords/credit card information [78]. Such approaches should be considered to minimize privacy of the users, which is a key factor for user participation and acceptance [99,107].
To summarize, passive sensing has considerable potential for objective data collection while minimizing user burden. However, the aforementioned factors are crucial considerations while collecting passive data.

3.3. Applications, Frameworks, and Resources Used in the Studies

Twelve studies discussed the use of existing datasets to test different machine-learning models and algorithms (e.g., StudentLife dataset [27,76,80], mPower dataset [28,29], UCI machine-learning repository [49]). These papers did not collect any primary data during the course of their research.
The majority of researchers, however (41 studies), did collect data and developed their own bespoke applications for the collection and analysis [15,23,55,67,69]. These ranged from apps collecting data from a single category of sensors [23] (e.g., accelerometer, gyroscope for activity recognition) to apps collecting data from both sensor and software features (e.g., app usage [55], battery status [57]). On the other hand, several studies used third-party software and services for data collection, which had similar functionality as the developed applications. Examples of such (paid) services were MovieSens XS [109] and BeHapp [110], used in [93,94], respectively. Ten studies used existing open-source frameworks to develop their applications (see Table 3).

3.4. Motivations for Data-Collection and -Analysis Approaches

The studies reported several motivations for data collection, which governed the analysis approaches that were used. There were five common motivations identified:
  • Exploratory studies:
    Four studies were exploratory, presenting the design of their sensing systems and evaluated the data-collection capabilities of their applications. For example, ref. [58] presents a nonobtrusive sleep-detecting application and evaluates how reliably it could detect sleeping behaviors.
  • Monitoring change in behavioral patterns:
    In seven studies, the emphasis was to monitor human behavior using smartphone-sensed data. For example, Refs. [24,25] monitored changes in mental health and behavior during the COVID-19 pandemic, by examining changes in smartphone-sensed data. They examined how factors such as physical activity, sociability, and mobility of students changed due to the pandemic, which provided an indication of their mental health.
  • Identifying correlation between smartphone-sensed features and wellbeing factors:
    In 24 studies, the emphasis was to examine the statistical significance of features extracted from smartphones with wellbeing behaviors. For example, Ref. [86] collected data from the microphone sensor to evaluate if audio features were correlated to self-reported measures of depression. In another example, Ref. [7] collected location data to determine if there was a correlation between time spent at home and self-reported depressive symptoms.
  • Identifying feature correlations and using machine learning to predict behavior:
    These types of studies (22 studies) not only identified correlation between smartphone-sensed features, but also built machine-learning models to evaluate if these were able to predict user behavior. For example, Ref. [60] found location and activity features that correlated with drinking episodes. They then built a machine-learning framework to classify instances of drinking vs nondrinking and tested the performance of their system.
  • Comparing activity-recognition performance of machine-learning models:
    Such studies aimed to evaluate the activity recognition of different machine-learning models. For example, Ref. [49] the performance of five types of ensemble classifiers to classify six activities (walking, walking upstairs, walking downstairs, sitting, standing, and lying).
Given such motivations, many of the systems inherently used some form of machine-learning pipelines, more specifically human-activity recognition (HAR). Across the systems, there were a wide variety of different sensor combinations used, with individual systems aimed at assessing a different range of conditions. However, there was considerable commonality in the underlying approach, with several works (e.g., [11,49,77]) implementing some version of the standard pipeline for HAR systems: (1) data preprocessing, (2) data segmentation, (3) feature extraction, and (4) model training and classification (as explained in [114]), albeit using different classifiers and feature representations. There was, however, a small minority who employed more modern deep-learning approaches, such as neural networks [21,30,53,65]. Overall, only a few studies used more recent approaches of HAR, a matter which we address in more detail within our discussion.

4. Discussion

The discussion section presents four themes from the findings of the literature review. These are (1) predominance of mental health studies, (2) opportunities for standardization of sensing approaches, (3) opportunities for using machine-learning advancements in eHealth research, and (4) sensing trends over the years and future scope

4.1. Predominance of Mental Health Studies

Over the past 5 years, the literature in this space has shown an overwhelming prominence for mental health as an area of study, with a mix of conditions such as depression [8,79,86,89], bipolar disorder [94,95], and schizophrenia [91,93]. The majority of studies explored how behavioral patterns influenced the mental health of individuals by primarily examining a range of wellbeing factors. The sheer number of works in this space indicate the scope and potential of smartphone sensing in mental health research.
On the other hand, studies focused on physical health and wellbeing, with factors such as physical activity [22,50,53] and sleep [19,57,58] being studied. Although these are not health conditions in themselves, they correlate to other conditions (e.g., reduced sociability influencing depressive symptoms). Of the one physical chronic condition covered (Parkinson’s), the emphasis was on evaluating its impact on motor symptoms such as gait and posture [28,29]. As such, the focus was primarily on physical-activity recognition. Previous works have studied the use of the camera sensor to monitor factors such as heart rate and skin diseases [115]; however, very few studies in this sample collected such data.
The fact that the numbers for general health and chronic conditions were lower than those of mental health conditions could be due to several reasons. Firstly, it is arguable that mental health conditions are a particularly important concern to measure, not only due to their prevalence over the globe [116,117], but also due to limitations that exist around actively collecting data with participants experiencing mental health concerns. For example, the subjectivity of self-reports [2,20,100], reliance on user input and compliance [61,99], and recall bias [70,81] might make passive sensing a more favorable approach to monitor mental health. Secondly, one of the limitations around smartphone sensing that was discussed was the impact of smartphone placement on physical data [2,3,11,103]. Studies monitoring factors such as sleep [19,84] and physical activity [2,11] discussed how different smartphone placements would render data from one or more sensors less reliable, thus impacting the accuracy of results. There is plausibly a higher need for accuracy when measuring physical data longitudinally [11,103], and physical symptoms or their proxies might be challenging to monitor passively through the smartphone. As such, objective monitoring of physical health conditions could potentially be better suited to wearables or smart home devices. There is significant literature looking at wearable and ambient sensing for monitoring physical conditions (e.g., [118,119,120,121,122]), which was beyond the scope of this literature review but should be considered by researchers wishing to explore chronic health conditions in the future.

4.2. Opportunities for Standardization of Sensing Approaches

Perhaps one of the more striking findings was the variety of different approaches employed across the literature in terms of sensing (e.g., active/passive) and the applications that were deployed for data collection. Although a small number of studies used existing frameworks and datasets [9,27,75,76,81], the majority of them developed custom applications from the ground up for their research. As such, significant time and resources would have been expended on the bespoke development of data-collection applications and analysis pipelines, unique to each individual study. Studies collecting similar types of data could benefit from using standardized approaches and existing frameworks, with the effect of conserving resources and accelerating up research.
Researchers across the globe have made significant efforts in recent years to provide open-source frameworks for development and analysis [32,33,34,113], which have already garnered numerous citations in the literature to date. Frameworks such as AWARE [123] and RADAR-base [124] have communities of practice and support available (e.g., Slack channels) and encourage people to contribute to ongoing development and feature development through open GitHub repositories. As such, there is scope for increased collaboration and innovation towards novel approaches, rather than the redevelopment of similar tools from scratch. That said, given their limited use to date, it might be the case that there are barriers to overcome around sharing resources such as these (e.g., lack of broad awareness, skill sets of research teams, or that these frameworks do not precisely fit the design of a specific study). Further exploration focusing on potential barriers towards the use of existing frameworks would be useful to understand how we might overcome these as a research community.
Besides reducing the amount of work involved, standardization could also enable deployment of the systems to wider populations in the future. The literature showed that the majority of studies used Android devices for data collection. This could partly be due to restrictions on collecting certain types of data with iOS [8,9,54,55,88], but also due to the additional overheads required to develop systems for both platforms. Using existing frameworks could potentially ease the development of multiplatform applications, enabling greater inclusion of iOS users in future work. This would ultimately reduce selection bias (i.e., by limiting the participant groups based on the type of device they use) [55] and enable increased access to participant groups.
In short, there is significant promise in this space, and increasing awareness and collaboration amongst practitioners could pave the way for standardization, and thus is an important consideration for future work.

4.3. Opportunities for Using Machine-Learning Advancements in E-Health Research

The effective use of machine learning (ML) is essential for studies exploring human-activity recognition (HAR) with systems using passive sensing for predicting user behavior. Many papers in the sample had some form of HAR pipeline, but there were very few that focused more on recent advances made in HAR over the last few years, such as deep-learning approaches [125] or even more advanced feature representations [126]. Whilst this might simply be due to the skillsets of the research teams, or the focus of the research questions/aims, it does show a gap in the literature that could explore the scope for implementing more modern approaches in machine learning for smartphone sensing, which could potentially increase the performance of HAR. Future work is required to investigate how we can best bring together the knowledge of multidisciplinary groups, to make use of combined expertise.

4.4. Sensing Trends over the Years and Future Scope

A substantial amount of research has been carried out in the space of smartphone sensing for healthcare, which has led to several systematic reviews providing an overview of the advancements at different points in time. Most previous reviews have focused on specific aspects of smartphone sensing. For example, prior works reviewed the literature with a focus on specific conditions or set of conditions (e.g., Parkinson’s [31], drug use [38], mental health [127,128], and physical activity [129]), whilst others narrowed their purview by some studying specific forms of data collection (e.g., EMAs [130], multimedia sensors [116]), and analysis approaches [40]. In contrast, our review presents a broad overview of all such factors and does not focus on specific conditions and sensing approaches.
There was, however, another complementary systematic review conducted in 2019 that presented the broader perspective of smartphone sensing [131]. They reviewed passive-sensing literature from 2014–2019 and found that the majority of the studies were examining physical activity and mental health conditions. This highlights a continuing trend of the prevalence of mental health studies. This is further showcased by the number of systematic reviews examining mental health conditions over the past decade [37,127,128,132,133,134]. While their review focuses solely on passive-sensing aspects, we have also discussed the several forms of active data collection used in the literature, which were widely used in the studies. Additionally, our review also highlights the applications and resources used in the literature, which can serve as a guide for future research.
All this said, it is crucial to consider how smartphone sensing may be impacted in the future due to changes in privacy laws and operating system restrictions on sensor data collection. Notably, there has been a significant push to protect user privacy and provide users with greater control over their data. For example, Google has made several changes to Android user permissions over the years that provide greater control over the data collected by smartphone applications. These privacy changes have impacted several factors such as application access to certain sensors, the granularity of data collected and applications collecting data in the background [46]. Additionally, changes to privacy laws will also impact data collection, storage, and sharing practices. The impact of such changes is not predictable; for example, the GDPR provides expansive opportunities to conduct research, provided certain data-protection obligations are met (e.g., data security, see [135] for a summary of relevant general obligations), yet the approach adopted by platform maintainers (e.g., Google in respect of Android, Apple for iOS) is a lot more conservative. This implies that the requirements of the privacy laws such as GDPR cannot be used to determine the level of restrictions that would be applied in the future. Such changes will have a significant impact on the potential of smartphone sensing and researchers may have greater restrictions on collecting user data in future studies.

5. Limitations and Future Work

In common with all literature reviews, there are criteria used for paper selection. This implies that relevant works in the space may have been excluded. Specifically, articles not focusing on smartphones may have been overlooked, even if that work in part uses them within a wider sensing infrastructure (e.g., smartphones used along with wearables and external sensors). The focus was solely on smartphones due to the additional concerns surrounding wearable sensors/sensing infrastructure. Factors such as cost, intrusiveness, ease of use, adoption by various population groups [136,137,138], etc., have an impact on the uptake of wearables. Such factors must be explored in depth when studying wearables. Considering these factors and using wider terminology would have surfaced a larger number of papers, albeit at the risk of providing too many results to be practically reviewed. This all said, a wider review might have also complicated our analysis and occluded the clear issues that we identified in this account.
Our findings also highlight several important directions for future work. First, there is a pressing need to investigate the barriers that may be impacting standardization of sensing approaches across research teams. Identifying these could pave the way for collaboration and lead to greater focus on innovation. Second, there is scope for leveraging machine-learning advancements such as deep learning in smartphone-sensing applications. Future work can explore the feasibility of using such approaches and evaluate if these can enhance the capabilities of smartphone sensing. Lastly, future studies could further explore the integration of smartphone sensing and other sensing modalities (e.g., [139]). There is also scope for such systems to contribute to user wellbeing, rather than studying it; for example, one could see such systems integrated into persuasive technology [140]. As such, there are a wide range of opportunities for future work that arise from our findings.

6. Conclusions

In this paper, we have presented a narrative review of smartphone sensing for health, covering papers published over the last five years. This narrative review has covered a diverse body of work that has used a great variety of different approaches and tools to monitor a broad range of health conditions. In addition to providing a map of the current state of the art, this paper has also presented a clear agenda forward for capitalizing upon this work, which highlights the need to move towards standardization and investigate potential barriers in the process. It highlights the scope for collaboration of clinical and human-activity research (HAR) communities together, which could enhance the potential of smartphone-based activity-recognition systems. We provide an overview of the advancements in smartphone sensing which could assist future researchers to quickly decide between different approaches and assist in making high-level decisions for using smartphone sensing.

Author Contributions

Conceptualization, P.K., R.K. and R.M.; methodology, P.K., R.K. and R.M.; formal analysis, P.K.; investigation, P.K.; data curation, P.K.; writing—original draft preparation, P.K. and R.K.; writing—review and editing, R.K. and R.M.; visualization, P.K.; supervision, R.K. and R.M.; All authors have read and agreed to the published version of the manuscript.


This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.


GPSGlobal Positioning System
HARHuman Activity Recognition
NLRNarrative Literature Review
PRISMAPreferred Reporting Items for Systematic Reviews and Meta-Analyses
PHQPatient Health Questionnaire
UCLAUniversity of California, Los Angeles
EMAEcological Momentary Assessments
Wi-FiWireless Fidelity
MLMachine Learning


  1. Smartphone Ownership Is Growing Rapidly around the World, but Not Always Equally. Pew Research Center. Available online: (accessed on 6 April 2022).
  2. Miškinytė, A.; Dėdelė, A. Objective Assessment of Physical Activity Patterns Based on Accelerometer and GPS Data in Adults. Travel Behav. Soc. 2021, 25, 112–119. [Google Scholar] [CrossRef]
  3. Harari, G.M.; Gosling, S.D.; Wang, R.; Chen, F.; Chen, Z.; Campbell, A.T. Patterns of Behavior Change in Students over an Academic Term: A Preliminary Study of Activity and Sociability Behaviors Using Smartphone Sensing Methods. Comput. Hum. Behav. 2017, 67, 129–138. [Google Scholar] [CrossRef]
  4. Grimaldi, S.; Partonen, T.; Saarni, S.I.; Aromaa, A.; Lönnqvist, J. Indoors Illumination and Seasonal Changes in Mood and Behavior Are Associated with the Health-Related Quality of Life. Health Qual. Life Outcomes 2008, 6, 56. [Google Scholar] [CrossRef] [PubMed]
  5. Espiritu, R.C.; Kripke, D.F.; Ancoli-Israel, S.; Mowen, M.A.; Mason, W.J.; Fell, R.L.; Klauber, M.R.; Kaplan, O.J. Low Illumination Experienced by San Diego Adults: Association with Atypical Depressive Symptoms. Biol. Psychiatry 1994, 35, 403–407. [Google Scholar] [CrossRef]
  6. Time Spent on Average on a Smartphone in the U.S. 2021. Statista. Available online: (accessed on 6 April 2022).
  7. Laiou, P.; Kaliukhovich, D.A.; Folarin, A.A.; Ranjan, Y.; Rashid, Z.; Conde, P.; Stewart, C.; Sun, S.; Zhang, Y.; Matcham, F.; et al. The Association Between Home Stay and Symptom Severity in Major Depressive Disorder: Preliminary Findings From a Multicenter Observational Study Using Geolocation Data From Smartphones. JMIR Mhealth Uhealth 2022, 10, e28095. [Google Scholar] [CrossRef] [PubMed]
  8. Yue, C.; Ware, S.; Morillo, R.; Lu, J.; Shang, C.; Bi, J.; Kamath, J.; Russell, A.; Bamis, A.; Wang, B. Fusing Location Data for Depression Prediction. IEEE Trans. Big Data 2018, 7, 355–370. [Google Scholar] [CrossRef]
  9. Wu, C.; Barczyk, A.N.; Craddock, R.C.; Harari, G.M.; Thomaz, E.; Shumake, J.D.; Beevers, C.G.; Gosling, S.D.; Schnyer, D.M. Improving Prediction of Real-Time Loneliness and Companionship Type Using Geosocial Features of Personal Smartphone Data. Smart Health 2021, 20, 100180. [Google Scholar] [CrossRef]
  10. Ehatisham-ul-Haq, M.; Azam, M.A.; Asim, Y.; Amin, Y.; Naeem, U.; Khalid, A. Using Smartphone Accelerometer for Human Physical Activity and Context Recognition In-the-Wild. Procedia Comput. Sci. 2020, 177, 24–31. [Google Scholar] [CrossRef]
  11. O’Brien, M.K.; Shawen, N.; Mummidisetty, C.K.; Kaur, S.; Bo, X.; Poellabauer, C.; Kording, K.; Jayaraman, A. Activity Recognition for Persons With Stroke Using Mobile Phone Technology: Toward Improved Performance in a Home Setting. J. Med. Internet Res. 2017, 19, e184. [Google Scholar] [CrossRef]
  12. Bae, S.W.; Chung, T.; Islam, R.; Suffoletto, B.; Du, J.; Jang, S.; Nishiyama, Y.; Mulukutla, R.; Dey, A. Mobile Phone Sensor-Based Detection of Subjective Cannabis Intoxication in Young Adults: A Feasibility Study in Real-World Settings. Drug Alcohol Depend. 2021, 228, 108972. [Google Scholar] [CrossRef]
  13. Wang, R.; Chen, F.; Chen, Z.; Li, T.; Harari, G.; Tignor, S.; Zhou, X.; Ben-Zeev, D.; Campbell, A.T. StudentLife: Using Smartphones to Assess Mental Health and Academic Performance of College Students. In Mobile Health; Rehg, J., Murphy, S., Kumar, S., Eds.; Springer: Cham, Switzerland, 2017. [Google Scholar] [CrossRef]
  14. Zhang, H.; Hanci, E.; Gashi, S.; Matthews, O.; Kimm, H. Moodbook: An Application for Continuous Monitoring of Social Media Usage and Mood. In Proceedings of the 2018 ACM International Joint Conference on Pervasive and Ubiquitous Computing and 2018 ACM International Symposium on Wearable Computers, Singapore, 8–12 October 2018; pp. 1150–1155. [Google Scholar] [CrossRef]
  15. Zhang, X.; Li, W.; Chen, X.; Lu, S. MoodExplorer: Towards Compound Emotion Detection via Smartphone Sensing. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2018, 1, 1–30. [Google Scholar] [CrossRef]
  16. Torkamaan, H.; Ziegler, J. Mobile Mood Tracking. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2020, 4, 1–30. [Google Scholar] [CrossRef]
  17. Dogrucu, A.; Perucic, A.; Isaro, A.; Ball, D.; Toto, E.; Rundensteiner, E.A.; Agu, E.; Davis-Martin, R.; Boudreaux, E. Moodable: On Feasibility of Instantaneous Depression Assessment Using Machine Learning on Voice Samples with Retrospectively Harvested Smartphone and Social Media Data. Smart Health 2020, 17, 100118. [Google Scholar] [CrossRef]
  18. Rebolledo, M.; Eiben, A.E.; Bartz-Beielstein, T. Bayesian Networks for Mood Prediction Using Unobtrusive Ecological Momentary Assessments. In Applications of Evolutionary Computation; Castillo, P.A., Jiménez Laredo, J.L., Eds.; Lecture Notes in Computer Science; Springer: Cham, Switzerland, 2021; Volume 12694, pp. 373–387. [Google Scholar] [CrossRef]
  19. Niemeijer, K.; Mestdagh, M.; Kuppens, P. Tracking Subjective Sleep Quality and Mood With Mobile Sensing: Multiverse Study. J. Med. Internet Res. 2022, 24, e25643. [Google Scholar] [CrossRef] [PubMed]
  20. Zhang, X.; Zhuang, F.; Li, W.; Ying, H.; Xiong, H.; Lu, S. Inferring Mood Instability via Smartphone Sensing: A Multi-View Learning Approach. In Proceedings of the 27th ACM International Conference on Multimedia, Nice, France, 21–25 October 2019; pp. 1401–1409. [Google Scholar] [CrossRef]
  21. Suhara, Y.; Xu, Y.; Pentland, A.S. DeepMood: Forecasting Depressed Mood Based on Self-Reported Histories via Recurrent Neural Networks. In Proceedings of the 26th International World Wide Web Conference, WWW 2017, Perth, Australia, 3–7 April 2017; pp. 715–724. [Google Scholar] [CrossRef]
  22. Müller, S.R.; Peters, H.; Matz, S.C.; Wang, W.; Harari, G.M. Investigating the Relationships between Mobility Behaviours and Indicators of Subjective Well–Being Using Smartphone–Based Experience Sampling and GPS Tracking. Eur. J. Personal. 2020, 34, 714–732. [Google Scholar] [CrossRef]
  23. Fahim, M.; Baker, T.; Khattak, A.M.; Alfandi, O. Alert Me: Enhancing Active Lifestyle via Observing Sedentary Behavior Using Mobile Sensing Systems. In Proceedings of the 2017 IEEE 19th International Conference on e-Health Networking, Applications and Services, Healthcom 2017, Dalian, China, 12–15 October 2017; pp. 1–4. [Google Scholar] [CrossRef]
  24. Huckins, J.F.; da Silva, A.W.; Wang, W.; Hedlund, E.; Rogers, C.; Nepal, S.K.; Wu, J.; Obuchi, M.; Murphy, E.I.; Meyer, M.L.; et al. Mental Health and Behavior of College Students During the Early Phases of the COVID-19 Pandemic: Longitudinal Smartphone and Ecological Momentary Assessment Study. J. Med. Internet Res. 2020, 22, e20185. [Google Scholar] [CrossRef] [PubMed]
  25. Mack, D.L.; DaSilva, A.W.; Rogers, C.; Hedlund, E.; Murphy, E.I.; Vojdanovski, V.; Plomp, J.; Wang, W.; Nepal, S.K.; Holtzheimer, P.E.; et al. Mental Health and Behavior of College Students During the COVID-19 Pandemic: Longitudinal Mobile Smartphone and Ecological Momentary Assessment Study, Part II. J. Med. Internet Res. 2021, 23, e28892. [Google Scholar] [CrossRef]
  26. MacLeod, L.; Suruliraj, B.; Gall, D.; Bessenyei, K.; Hamm, S.; Romkey, I.; Bagnell, A.; Mattheisen, M.; Muthukumaraswamy, V.; Orji, R.; et al. A Mobile Sensing App to Monitor Youth Mental Health: Observational Pilot Study. JMIR Mhealth Uhealth 2021, 9, e20638. [Google Scholar] [CrossRef]
  27. Thakur, S.S.; Roy, R.B. Predicting Mental Health Using Smart-Phone Usage and Sensor Data. J. Ambient. Intell. Humaniz. Comput. 2021, 12, 9145–9161. [Google Scholar] [CrossRef]
  28. Abujrida, H.; Agu, E.; Pahlavan, K. Machine Learning-Based Motor Assessment of Parkinson’s Disease Using Postural Sway, Gait and Lifestyle Features on Crowdsourced Smartphone Data. Biomed. Phys. Eng. Express 2020, 6, 035005. [Google Scholar] [CrossRef]
  29. Abujrida, H.; Agu, E.; Pahlavan, K. Smartphone-Based Gait Assessment to Infer Parkinson’s Disease Severity Using Crowdsourced Data. In Proceedings of the 2017 IEEE Healthcare Innovations and Point of Care Technologies, HI-POCT 2017, Bethesda, MD, USA, 6–8 November 2017; pp. 208–211. [Google Scholar] [CrossRef]
  30. Zhang, H.; Guo, G.; Song, C.; Xu, C.; Cheung, K.; Alexis, J.; Li, H.; Li, D.; Wang, K.; Xu, W. PDLens: Smartphone Knows Drug Effectiveness among Parkinson’s via Daily-Life Activity Fusion. In Proceedings of the Annual International Conference on Mobile Computing and Networking, MOBICOM, London, UK, 21–25 September 2020; Volume 20, pp. 150–163. [Google Scholar] [CrossRef]
  31. Postolache, G.; Postolache, O. Smartphone Sensing Technologies for Tailored Parkinson’s Disease Diagnosis and Monitoring. In Mobile Solutions and Their Usefulness in Everyday Life; EAI/Springer Innovations in Communication and Computing; Springer: Cham, Switzerland, 2019; pp. 251–273. [Google Scholar] [CrossRef]
  32. Ferreira, D.; Kostakos, V.; Dey, A.K. AWARE: Mobile Context Instrumentation Framework. Front. ICT 2015, 2, 6. [Google Scholar] [CrossRef] [Green Version]
  33. Torous, J.; Kiang, M.V.; Lorme, J.; Onnela, J.P. New Tools for New Research in Psychiatry: A Scalable and Customizable Platform to Empower Data Driven Smartphone Research. JMIR Ment. Health 2016, 3, e5165. [Google Scholar] [CrossRef] [PubMed]
  34. Lind, M.N.; Byrne, M.L.; Wicks, G.; Smidt, A.M.; Allen, N.B. The Effortless Assessment of Risk States (EARS) Tool: An Interpersonal Approach to Mobile Sensing. JMIR Ment. Health 2018, 5, e10334. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  35. Green, B.N.; Johnson, C.D.; Adams, A. Writing Narrative Literature Reviews for Peer-Reviewed Journals: Secrets of the Trade. J. Chiropr. Med. 2006, 5, 101–117. [Google Scholar] [CrossRef] [Green Version]
  36. Torous, J.; Larsen, M.E.; Depp, C.; Cosco, T.D.; Barnett, I.; Nock, M.K.; Firth, J. Smartphones, Sensors, and Machine Learning to Advance Real-Time Prediction and Interventions for Suicide Prevention: A Review of Current Progress and Next Steps. Curr. Psychiatry Rep. 2018, 20, 51. [Google Scholar] [CrossRef] [PubMed]
  37. Kerst, A.; Zielasek, J.; Gaebel, W. Smartphone Applications for Depression: A Systematic Literature Review and a Survey of Health Care Professionals’ Attitudes towards Their Use in Clinical Practice. Eur. Arch. Psychiatry Clin. Neurosci. 2020, 270, 139–152. [Google Scholar] [CrossRef] [PubMed]
  38. Nguyen, N.; Nguyen, C.; Thrul, J. Digital Health for Assessment and Intervention Targeting Tobacco and Cannabis Co-Use. Curr. Addict. Rep. 2020, 7, 268–279. [Google Scholar] [CrossRef]
  39. Konsolakis, K.; Hermens, H.; Villalonga, C.; Vollenbroek-Hutten, M.; Banos, O. Human Behaviour Analysis through Smartphones. Proceedings 2018, 2, 1243. [Google Scholar] [CrossRef] [Green Version]
  40. Straczkiewicz, M.; James, P.; Onnela, J.P. A Systematic Review of Smartphone-Based Human Activity Recognition Methods for Health Research. Npj Digit. Med. 2021, 4, 148. [Google Scholar] [CrossRef]
  41. Mendes, J.P.M.; Moura, I.R.; van de Ven, P.; Viana, D.; Silva, F.J.S.; Coutinho, L.R.; Teixeira, S.; Rodrigues, J.J.P.C.; Teles, A.S. Sensing Apps and Public Data Sets for Digital Phenotyping of Mental Health: Systematic Review. J. Med. Internet Res. 2022, 24, e28735. [Google Scholar] [CrossRef]
  42. Moher, D.; Liberati, A.; Tetzlaff, J.; Altman, D.G. Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. Ann. Intern. Med. 2009, 151, 264–269. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  43. Ferrari, R. Writing Narrative Style Literature Reviews. Eur. Med. Writ. Assoc. 2015, 24, 230–235. [Google Scholar] [CrossRef]
  44. Baethge, C.; Goldbeck-Wood, S.; Mertens, S. SANRA—A Scale for the Quality Assessment of Narrative Review Articles. Res. Integr. Peer Rev. 2019, 4, 5. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  45. Privacy in Android 11. Android Developers. Available online: (accessed on 29 April 2022).
  46. Privacy Changes in Android 10. Android Developers. Available online: (accessed on 29 April 2022).
  47. Ibrahim, A.; Zhang, H.; Clinch, S.; Poliakoff, E.; Parsia, B.; Harper, S. Digital Phenotypes for Understanding Individuals’ Compliance With COVID-19 Policies and Personalized Nudges: Longitudinal Observational Study. JMIR Form. Res. 2021, 5, e23461. [Google Scholar] [CrossRef] [PubMed]
  48. Jones, S.L.; Hue, W.; Kelly, R.M.; Barnett, R.; Henderson, V.; Sengupta, R. Determinants of Longitudinal Adherence in Smartphone-Based Self-Tracking for Chronic Health Conditions. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2021, 5, 16. [Google Scholar] [CrossRef]
  49. Rahim, K.N.K.A.; Elamvazuthi, I.; Izhar, L.I.; Capi, G. Classification of Human Daily Activities Using Ensemble Methods Based on Smartphone Inertial Sensors. Sensors 2018, 18, 4132. [Google Scholar] [CrossRef] [Green Version]
  50. Chen, A.T.; Russell, S.; Wu, C.; Fritz, H.; Miller, M.; Craddock, C.; Kinney, K.; Castelli, D.; Schnyer, D. Exploring Post COVID-19 Outbreak Intradaily Mobility Pattern Change in College Students: A GPS-Focused Smartphone Sensing Study. Front. Digit. Health 2021, 3, 765972. [Google Scholar] [CrossRef]
  51. Vaghela, M.; Sasidhar, K. Analyzing the Human Behavior Using Pervasive Sensing System. In Proceedings of the 21st International Conference on Distributed Computing and Networking, Kolkata, India, 4–7 January 2020. [Google Scholar]
  52. Kelly, D.; Curran, K.; Caulfield, B. Automatic Prediction of Health Status Using Smartphone-Derived Behavior Profiles. IEEE J. Biomed. Health Inform. 2017, 21, 1750–1760. [Google Scholar] [CrossRef]
  53. di Mitri, D.; Asyraaf Mat Sanusi, K.; Trebing, K.; Bromuri, S. MOBIUS: Smart Mobility Tracking with Smartphone Sensors. In Science and Technologies for Smart Cities; Lecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST; Springer: Cham, Switzerland, 2021; Volume 372. [Google Scholar] [CrossRef]
  54. Harari, G.M.; Müller, S.R.; Stachl, C.; Wang, R.; Wang, W.; Bühner, M.; Rentfrow, P.J.; Campbell, A.T.; Gosling, S.D. Sensing Sociability: Individual Differences in Young Adults’ Conversation, Calling, Texting, and App Use Behaviors in Daily Life. J. Personal. Soc. Psychol. 2020, 119, 204–228. [Google Scholar] [CrossRef]
  55. Wetzel, B.; Pryss, R.; Baumeister, H.; Edler, J.S.; Gonçalves, A.S.O.; Cohrdes, C. “How Come You Don’t Call Me?” Smartphone Communication App Usage as an Indicator of Loneliness and Social Well-Being across the Adult Lifespan during the COVID-19 Pandemic. Int. J. Environ. Res. Public Health 2021, 18, 6212. [Google Scholar] [CrossRef]
  56. Schoedel, R.; Pargent, F.; Au, Q.; Völkel, S.T.; Schuwerk, T.; Bühner, M.; Stachl, C. To Challenge the Morning Lark and the Night Owl: Using Smartphone Sensing Data to Investigate Day–Night Behaviour Patterns. Eur. J. Personal. 2020, 34, 733–752. [Google Scholar] [CrossRef]
  57. Montanini, L.; Sabino, N.; Spinsante, S.; Gambi, E. Smartphone as Unobtrusive Sensor for Real-Time Sleep Recognition. In Proceedings of the 2018 IEEE International Conference on Consumer Electronics, Las Vegas, NV, USA, 12–14 January 2018; pp. 1–4. [Google Scholar] [CrossRef]
  58. ElBolock, A.; Amr, R.; Abdennadher, S. Non-Obtrusive Sleep Detection for Character Computing Profiling. Adv. Intell. Syst. Comput. 2018, 722, 249–254. [Google Scholar] [CrossRef]
  59. Lim, J.; Jeong, C.Y.; Lim, J.M.; Chung, S.; Kim, G.; Noh, K.J.; Jeong, H. Assessing Sleep Quality Using Mobile EMAs: Opportunities, Practical Consideration, and Challenges. IEEE Access 2022, 10, 2063–2076. [Google Scholar] [CrossRef]
  60. Phan, T.T.; Labhart, F.; Muralidhar, S.; Gatica-Perez, D. Understanding Heavy Drinking at Night through Smartphone Sensing and Active Human Engagement. In Proceedings of the PervasiveHealth: Pervasive Computing Technologies for Healthcare, Atlanta, GA, USA, 18–20 May 2020; Volume 12, pp. 211–222. [Google Scholar] [CrossRef]
  61. Meegahapola, L.; Labhart, F.; Phan, T.T.; Gatica-Perez, D. Examining the Social Context of Alcohol Drinking in Young Adults with Smartphone Sensing. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2021, 5, 26. [Google Scholar] [CrossRef]
  62. Bae, S.; Chung, T.; Ferreira, D.; Dey, A.K.; Suffoletto, B. Mobile Phone Sensors and Supervised Machine Learning to Identify Alcohol Use Events in Young Adults: Implications for Just-in-Time Adaptive Interventions. Addict. Behav. 2018, 83, 42–47. [Google Scholar] [CrossRef] [PubMed]
  63. Naughton, F.; Brown, C.; High, J.; Notley, C.; Mascolo, C.; Coleman, T.; Barton, G.; Shepstone, L.; Sutton, S.; Prevost, A.T.; et al. Randomised Controlled Trial of a Just-in-Time Adaptive Intervention (JITAI) Smoking Cessation Smartphone App: The Quit Sense Feasibility Trial Protocol. BMJ Open 2021, 11, e048204. [Google Scholar] [CrossRef]
  64. Li, R.; Agu, E.; Balakrishnan, G.; Herman, D.; Abrantes, A.; Stein, M.; Metrik, J. WeedGait: Unobtrusive Smartphone Sensing of Marijuana-Induced Gait Impairment by Fusing Gait Cycle Segmentation and Neural Networks. In Proceedings of the 2019 IEEE Healthcare Innovations and Point of Care Technologies, HI-POCT 2019, Bethesda, MD, USA, 20–22 November 2019; pp. 91–94. [Google Scholar] [CrossRef]
  65. Meegahapola, L.; Ruiz-Correa, S.; Robledo-Valero, V.D.C.; Hernandez-Huerfano, E.E.; Alvarez-Rivera, L.; Chenu-Abente, R.; Gatica-Perez, D. One More Bite? Inferring Food Consumption Level of College Students Using Smartphone Sensing and Self-Reports. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2021, 5, 26. [Google Scholar] [CrossRef]
  66. About Chronic Diseases. CDC. Available online: (accessed on 6 April 2022).
  67. Sarda, A.; Munuswamy, S.; Sarda, S.; Subramanian, V. Using Passive Smartphone Sensing for Improved Risk Stratification of Patients with Depression and Diabetes: Cross-Sectional Observational Study. JMIR Mhealth Uhealth 2019, 7, e11041. [Google Scholar] [CrossRef] [Green Version]
  68. Singh, V.K.; Long, T. Automatic Assessment of Mental Health Using Phone Metadata. Proc. Assoc. Inf. Sci. Technol. 2018, 55, 450–459. [Google Scholar] [CrossRef]
  69. Servia-Rodríguez, S.; Rachuri, K.K.; Mascolo, C.; Rentfrow, P.J.; Lathia, N.; Sandstrom, G.M. Mobile Sensing at the Service of Mental Well-Being: A Large-Scale Longitudinal Study. In Proceedings of the 26th International World Wide Web Conference, WWW 2017, Perth, Australia, 3–7 April 2017; pp. 103–112. [Google Scholar] [CrossRef] [Green Version]
  70. Messner, E.M.; Sariyska, R.; Mayer, B.; Montag, C.; Kannen, C.; Schwerdtfeger, A.; Baumeister, H. Insights—Future Implications of Passive Smartphone Sensing in the Therapeutic Context. Verhaltenstherapie 2019, 1–10. [Google Scholar] [CrossRef]
  71. Kalogirou, I.P.; Kallipolitis, A.; Maglogiannis, I. Passive Emotion Recognition Using Smartphone Sensing Data. Stud. Comput. Intell. 2020, 891, 17–28. [Google Scholar] [CrossRef]
  72. Buda, T.S.; Khwaja, M.; Matic, A. Outliers in Smartphone Sensor Data Reveal Outliers in Daily Happiness. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2021, 5, 19. [Google Scholar] [CrossRef]
  73. Piskioulis, O.; Tzafilkou, K.; Economides, A.A. Emotion Detection through Smartphone’s Accelerometer and Gyroscope Sensors. In Proceedings of the 29th ACM Conference on User Modeling, Adaptation and Personalization, Utrecht, The Netherlands, 21–25 June 2021; Volume 8, pp. 130–137. [Google Scholar] [CrossRef]
  74. Ruensuk, M.; Cheon, E.; Hong, H.; Oakley, I. How Do You Feel Online. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2020, 4, 1–32. [Google Scholar] [CrossRef]
  75. Wen, H.; Sobolev, M.; Vitale, R.; Kizer, J.; Pollak, J.P.; Muench, F.; Estrin, D. MPulse Mobile Sensing Model for Passive Detection of Impulsive Behavior: Exploratory Prediction Study. JMIR Ment. Health 2021, 8, e25019. [Google Scholar] [CrossRef]
  76. Wang, F.; Wang, Y.; Wang, J.; Xiong, H.; Zhao, J.; Zhang, D. Assessing Mental Stress Based on Smartphone Sensing Data: An Empirical Study. In Proceedings of the 2019 IEEE SmartWorld, Ubiquitous Intelligence and Computing, Advanced and Trusted Computing, Scalable Computing and Communications, Internet of People and Smart City Innovation (SmartWorld/UIC/ATC/SCALCOM/IOP/SCI), Leicester, UK, 19–23 August 2019; pp. 1031–1038. [Google Scholar] [CrossRef]
  77. Etienne, N.; Agu, E. Investigating Transfer Learning of Smartphone-Sensed Stress in University Populations. In Proceedings of the 2020 IEEE International Conference on Big Data (Big Data), Atlanta, GA, USA, 10–13 December 2020; pp. 4850–4858. [Google Scholar] [CrossRef]
  78. Byrne, M.L.; Lind, M.N.; Horn, S.R.; Mills, K.L.; Nelson, B.W.; Barnes, M.L.; Slavich, G.M.; Allen, N.B. Using Mobile Sensing Data to Assess Stress: Associations with Perceived and Lifetime Stress, Mental Health, Sleep, and Inflammation. Digit. Health 2021, 7. [Google Scholar] [CrossRef]
  79. Asare, K.O.; Visuri, A.; Ferreira, D.S.T. Towards Early Detection of Depression through Smartphone Sens-Ing. In Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and 2019 ACM International Symposium on Wearable Computers, London, UK, 9–13 September 2019; pp. 1158–1161. [Google Scholar] [CrossRef] [Green Version]
  80. Nguyen, B.; Kolappan, S.; Bhat, V.; Krishnan, S. Clustering and Feature Analysis of Smartphone Data for Depression Monitoring. In Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS, Mexico, 1–5 November 2021; pp. 113–116. [Google Scholar] [CrossRef]
  81. Asare, K.O.; Terhorst, Y.; Vega, J.; Peltonen, E.; Lagerspetz, E.; Ferreira, D. Predicting Depression From Smartphone Behavioral Markers Using Machine Learning Methods, Hyperparameter Optimization, and Feature Importance Analysis: Exploratory Study. JMIR Mhealth Uhealth 2021, 9, e26540. [Google Scholar] [CrossRef]
  82. Müller, S.R.; Chen, X.; Peters, H.; Chaintreau, A.; Matz, S.C. Depression Predictions from GPS-Based Mobility Do Not Generalize Well to Large Demographically Heterogeneous Samples. Sci. Rep. 2021, 11, 14007. [Google Scholar] [CrossRef]
  83. Huckins, J.F.; DaSilva, A.W.; Hedlund, E.L.; Murphy, E.I.; Rogers, C.; Wang, W.; Obuchi, M.; Holtzheimer, P.E.; Wagner, D.D.; Campbell, A.T. Causal Factors of Anxiety and Depression in College Students: Longitudinal Ecological Momentary Assessment and Causal Analysis Using Peter and Clark Momentary Conditional Independence. JMIR Ment. Health 2020, 7, e16684. [Google Scholar] [CrossRef]
  84. di Matteo, D.; Fotinos, K.; Lokuge, S.; Mason, G.; Sternat, T.; Katzman, M.A.; Rose, J. Automated Screening for Social Anxiety, Generalized Anxiety, and Depression From Objective Smartphone-Collected Data: Cross-Sectional Study. J. Med. Internet Res. 2021, 23, e28918. [Google Scholar] [CrossRef]
  85. McIntyre, R.S.; Lee, Y.; Rong, C.; Rosenblat, J.D.; Brietzke, E.; Pan, Z.; Park, C.; Subramaniapillai, M.; Ragguett, R.M.; Mansur, R.B.; et al. Ecological Momentary Assessment of Depressive Symptoms Using the Mind.Me Application: Convergence with the Patient Health Questionnaire-9 (PHQ-9). J. Psychiatr. Res. 2021, 135, 311–317. [Google Scholar] [CrossRef]
  86. di Matteo, D.; Fotinos, K.; Lokuge, S.; Yu, J.; Sternat, T.; Katzman, M.A.; Rose, J. The Relationship Between Smartphone-Recorded Environmental Audio and Symptomatology of Anxiety and Depression: Exploratory Study. JMIR Form. Res. 2020, 4, e18751. [Google Scholar] [CrossRef] [PubMed]
  87. Yue, C.; Ware, S.; Morillo, R.; Lu, J.; Shang, C.; Bi, J.; Kamath, J.; Russell, A.; Bamis, A.; Wang, B. Automatic Depression Prediction Using Internet Traffic Characteristics on Smartphones. Smart Health 2020, 18, 100137. [Google Scholar] [CrossRef] [PubMed]
  88. Ware, S.; Yue, C.; Morillo, R.; Lu, J.; Shang, C.; Bi, J.; Kamath, J.; Russell, A.; Bamis, A.; Wang, B. Predicting Depressive Symptoms Using Smartphone Data. Smart Health 2020, 15, 100093. [Google Scholar] [CrossRef]
  89. Chow, P.I.; Fua, K.; Huang, Y.; Bonelli, W.; Xiong, H.; Barnes, L.E.; Teachman, B.A. Using Mobile Sensing to Test Clinical Models of Depression, Social Anxiety, State Affect, and Social Isolation Among College Students. J. Med. Internet Res. 2017, 19, e6820. [Google Scholar] [CrossRef]
  90. Fulford, D.; Mote, J.; Gonzalez, R.; Abplanalp, S.; Zhang, Y.; Luckenbaugh, J.; Onnela, J.P.; Busso, C.; Gard, D.E. Smartphone Sensing of Social Interactions in People with and without Schizophrenia. J. Psychiatr. Res. 2021, 137, 613–620. [Google Scholar] [CrossRef]
  91. Wang, R.; Wang, W.; Aung, M.S.H.; Ben-Zeev, D.; Brian, R.; Campbell, A.T.; Choudhury, T.; Hauser, M.; Kane, J.; Scherer, E.A.; et al. Predicting Symptom Trajectories of Schizophrenia Using Mobile Sensing. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 2017, 1, 1–24. [Google Scholar] [CrossRef] [Green Version]
  92. Wang, W.; Mirjafari, S.; Harari, G.; Ben-Zeev, D.; Brian, R.; Choudhury, T.; Hauser, M.; Kane, J.; Masaba, K.; Nepal, S.; et al. Social Sensing: Assessing Social Functioning of Patients Living with Schizophrenia Using Mobile Phone Sensing. In Proceedings of the Conference on Human Factors in Computing Systems, Honolulu, HI, USA, 25–30 April 2020. [Google Scholar] [CrossRef]
  93. Jongs, N.; Jagesar, R.; van Haren, N.E.M.; Penninx, B.W.J.H.; Reus, L.; Visser, P.J.; van der Wee, N.J.A.; Koning, I.M.; Arango, C.; Sommer, I.E.C.; et al. A Framework for Assessing Neuropsychiatric Phenotypes by Using Smartphone-Based Location Data. Transl. Psychiatry 2020, 10, 211. [Google Scholar] [CrossRef]
  94. Ebner-Priemer, U.W.; Mühlbauer, E.; Neubauer, A.B.; Hill, H.; Beier, F.; Santangelo, P.S.; Ritter, P.; Kleindienst, N.; Bauer, M.; Schmiedek, F.; et al. Digital Phenotyping: Towards Replicable Findings with Comprehensive Assessments and Integrative Models in Bipolar Disorders. Int. J. Bipolar Disord. 2020, 8, 35. [Google Scholar] [CrossRef]
  95. Busk, J.; Faurholt-Jepsen, M.; Frost, M.; Bardram, J.E.; Kessing, L.V.; Winther, O. Forecasting Mood in Bipolar Disorder From Smartphone Self-Assessments: Hierarchical Bayesian Approach. JMIR Mhealth Uhealth 2020, 8, e15028. [Google Scholar] [CrossRef] [Green Version]
  96. Kroenke, K.; Spitzer, R.L.; Williams, J.B.W. The PHQ-9: Validity of a Brief Depression Severity Measure. J. Gen. Intern. Med. 2001, 16, 606. [Google Scholar] [CrossRef]
  97. Russell, D.W. UCLA Loneliness Scale (Version 3): Reliability, Validity, and Factor Structure. J. Personal. Assess. 2010, 66, 20–40. [Google Scholar] [CrossRef] [PubMed]
  98. Shiffman, S.; Stone, A.A.; Hufford, M.R. Ecological Momentary Assessment. Annu. Rev. Clin. Psychol. 2008, 4, 1–32. [Google Scholar] [CrossRef] [PubMed]
  99. Rooksby, J.; Morrison, A.; Murray-Rust, D. Student Perspectives on Digital Phenotyping the Acceptability of Using Smartphone Data to Assess Mental Health. In Proceedings of the Conference on Human Factors in Computing Systems, Glasgow, UK, 4–9 May 2019; p. 14. [Google Scholar] [CrossRef] [Green Version]
  100. di Matteo, D.; Fine, A.; Fotinos, K.; Rose, J.; Katzman, M. Patient Willingness to Consent to Mobile Phone Data Collection for Mental Health Apps: Structured Questionnaire. JMIR Ment. Health 2018, 5, e9539. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  101. González-Pérez Ignacio Miralles, A.; Granell Sven Casteleyn, C.; González-Pérez, A.; Miralles, I.; Granell, C. Technical Challenges to Deliver Sensor-Based Psy-Chological Interventions Using Smartphones. In Proceedings of the 2019 ACM International Joint Conference on Pervasive and Ubiquitous Computing and 2019 ACM International Symposium on Wearable Computers, London, UK, 9–13 September 2019. [Google Scholar] [CrossRef]
  102. Tonti, S.; Marzolini, B.; Bulgheroni, M. Smartphone-Based Passive Sensing for Behavioral and Physical Monitoring in Free-Life Conditions: Technical Usability Study. JMIR Biomed. Eng. 2021, 6, e15417. [Google Scholar] [CrossRef]
  103. Keusch, F.; Wenz, A.; Conrad, F. Do You Have Your Smartphone with You? Behavioral Barriers for Measuring Everyday Activities with Smartphone Sensors. Comput. Hum. Behav. 2022, 127, 107054. [Google Scholar] [CrossRef]
  104. Harari, G.M. A Process-Oriented Approach to Respecting Privacy in the Context of Mobile Phone Tracking. Curr. Opin. Psychol. 2020, 31, 141–147. [Google Scholar] [CrossRef]
  105. Breslin, S.; Shareck, M.; Fuller, D. Research Ethics for Mobile Sensing Device Use by Vulnerable Populations. Soc. Sci. Med. 2019, 232, 50–57. [Google Scholar] [CrossRef]
  106. Delgado-Santos, P.; Stragapede, G.; Tolosana, R.; Guest, R.; Deravi, F.; Vera-Rodriguez, R. A Survey of Privacy Vulnerabilities of Mobile Device Sensors. ACM Comput. Surv. 2021. [Google Scholar] [CrossRef]
  107. Beierle, F.; Tran, V.T.; Allemand, M.; Neff, P.; Schlee, W.; Probst, T.; Zimmermann, J.; Pryss, R. What Data Are Smartphone Users Willing to Share with Researchers?: Designing and Evaluating a Privacy Model for Mobile Data Collection Apps. J. Ambient. Intell. Humaniz. Comput. 2020, 11, 2277–2289. [Google Scholar] [CrossRef]
  108. Mulvenna, M.D.; Bond, R.; Delaney, J.; Dawoodbhoy, F.M.; Boger, J.; Potts, C.; Turkington, R. Ethical Issues in Democratizing Digital Phenotypes and Machine Learning in the Next Generation of Digital Health Technologies. Philos. Technol. 2021, 34, 1945–1960. [Google Scholar] [CrossRef]
  109. Experience Sampling—MovisensXS—Movisens GmbH. Available online: (accessed on 27 April 2022).
  110. Behapp—Digital Phenotyping. Available online: (accessed on 27 April 2022).
  111. EARS—Ksana Health. Available online: (accessed on 27 April 2022).
  112. Lathia, N.; Rachuri, K.K.; Mascolo, C.; Roussos, G. Open Source Smartphone Libraries for Computational Social Science. In Proceedings of the 2013 ACM Conference on Ubiquitous Computing, Zurich, Switzerland, 8–12 September 2013; pp. 911–919. [Google Scholar] [CrossRef] [Green Version]
  113. Ranjan, Y.; Rashid, Z.; Stewart, C.; Conde, P.; Begale, M.; Verbeeck, D.; Boettcher, S.; Dobson, R.; Folarin, A. RADAR-Base: Open Source Mobile Health Platform for Collecting, Monitoring, and Analyzing Data Using Sensors, Wearables, and Mobile Devices. JMIR Mhealth Uhealth 2019, 7, e11734. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  114. Bulling, A.; Blanke, U.; Schiele, B. A Tutorial on Human Activity Recognition Using Body-Worn Inertial Sensors. ACM Comput. Surv. 2014, 46, 1–33. [Google Scholar] [CrossRef]
  115. Parra, L.; Sendra, S.; Jiménez, J.M.; Lloret, J. Multimedia Sensors Embedded in Smartphones for Ambient Assisted Living and E-Health. Multimed. Tools Appl. 2016, 75, 13271–13297. [Google Scholar] [CrossRef] [Green Version]
  116. Mental Health. Available online: (accessed on 28 April 2022).
  117. Mental Health—Our World in Data. Available online: (accessed on 28 April 2022).
  118. Schwenk, M.; Mohler, J.; Wendel, C.; D’Huyvetter, K.; Fain, M.; Taylor-Piliae, R.; Najafi, B. Wearable Sensor-Based In-Home Assessment of Gait, Balance, and Physical Activity for Discrimination of Frailty Status: Baseline Results of the Arizona Frailty Cohort Study. Gerontology 2015, 61, 258–267. [Google Scholar] [CrossRef] [Green Version]
  119. Wang, J.B.; Cadmus-Bertram, L.A.; Natarajan, L.; White, M.M.; Madanat, H.; Nichols, J.F.; Ayala, G.X.; Pierce, J.P. Wearable Sensor/Device (Fitbit One) and SMS Text-Messaging Prompts to Increase Physical Activity in Overweight and Obese Adults: A Randomized Controlled Trial. Telemed. E-Health 2015, 21, 782–792. [Google Scholar] [CrossRef] [Green Version]
  120. Papagiannaki, A.; Zacharaki, E.I.; Kalouris, G.; Kalogiannis, S.; Deltouzos, K.; Ellul, J.; Megalooikonomou, V. Recognizing Physical Activity of Older People from Wearable Sensors and Inconsistent Data. Sensors 2019, 19, 880. [Google Scholar] [CrossRef] [Green Version]
  121. Rao, A.K. Wearable Sensor Technology to Measure Physical Activity (PA) in the Elderly. Curr. Geriatr. Rep. 2019, 8, 55–66. [Google Scholar] [CrossRef]
  122. Luo, H.; Lee, P.A.; Clay, I.; Jaggi, M.; de Luca, V. Assessment of Fatigue Using Wearable Sensors: A Pilot Study. Digit. Biomark. 2020, 4, 59–72. [Google Scholar] [CrossRef]
  123. AWARE—Open-Source Context Instrumentation Framework for Everyone. Available online: (accessed on 30 April 2022).
  124. Home—RADAR-Base. Available online: (accessed on 30 April 2022).
  125. Ordóñez, F.J.; Roggen, D.; Liu, Y.; Xiao, W.; Chao, H.-C.; Chu, P. Deep Convolutional and LSTM Recurrent Neural Networks for Multimodal Wearable Activity Recognition. Sensors 2016, 16, 115. [Google Scholar] [CrossRef] [Green Version]
  126. Hammerla, N.Y.; Kirkham, R.; Andras, P.; Plötz, T. On Preserving Statistical Characteristics of Accelerometry Data Using Their Empirical Cumulative Distribution. In Proceedings of the 17th Annual International Symposium on International Symposium on Wearable Computers, Zurich, Switzerland, 8–12 September 2013. [Google Scholar] [CrossRef]
  127. Ben-Zeev, D.; Scherer, E.A.; Wang, R.; Xie, H. Next-Generation Psychiatric Assessment: Using Smartphone Sensors to Monitor Behavior and Mental Health. Psychiatr. Rehabil. J. 2015, 38, 218–226. [Google Scholar] [CrossRef] [Green Version]
  128. Harari, G.M.; Lane, N.D.; Wang, R.; Crosier, B.S.; Campbell, A.T.; Gosling, S.D. Using Smartphones to Collect Behavioral Data in Psychological Science: Opportunities, Practical Considerations, and Challenges. Perspect. Psychol. Sci. 2016, 11, 838–854. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  129. Kim, H.N.; Seo, K. Smartphone-Based Health Program for Improving Physical Activity and Tackling Obesity for Young Adults: A Systematic Review and Meta-Analysis. Int. J. Environ. Res. Public Health 2019, 17, 15. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  130. de Vries, L.P.; Baselmans, B.M.L.; Bartels, M. Smartphone-Based Ecological Momentary Assessment of Well-Being: A Systematic Review and Recommendations for Future Studies. J. Happiness Stud. 2021, 22, 2361–2408. [Google Scholar] [CrossRef] [PubMed]
  131. Trifan, A.; Oliveira, M.; Oliveira, J.L. Passive Sensing of Health Outcomes Through Smartphones: Systematic Review of Current Solutions and Possible Limitations. JMIR Mhealth Uhealth 2019, 7, e12649. [Google Scholar] [CrossRef]
  132. Bardram, J.E.; Matic, A. A Decade of Ubiquitous Computing Research in Mental Health. IEEE Pervasive Comput. 2020, 19, 62–72. [Google Scholar] [CrossRef]
  133. Antosik-Wójcińska, A.Z.; Dominiak, M.; Chojnacka, M.; Kaczmarek-Majer, K.; Opara, K.R.; Radziszewska, W.; Olwert, A.; Święcicki, Ł. Smartphone as a Monitoring Tool for Bipolar Disorder: A Systematic Review Including Data Analysis, Machine Learning Algorithms and Predictive Modelling. Int. J. Med. Inform. 2020, 138, 104131. [Google Scholar] [CrossRef]
  134. Bauer, M.; Glenn, T.; Geddes, J.; Gitlin, M.; Grof, P.; Kessing, L.V.; Monteith, S.; Faurholt-Jepsen, M.; Severus, E.; Whybrow, P.C. Smartphones in Mental Health: A Critical Review of Background Issues, Current Status and Future Concerns. Int. J. Bipolar Disord. 2020, 8, 2. [Google Scholar] [CrossRef]
  135. Binns, R.; Kirkham, R. How Could Equality and Data Protection Law Shape AI Fairness for People with Disabilities? ACM Trans. Access. Comput. 2021, 14, 1–32. [Google Scholar] [CrossRef]
  136. Ferguson, C.; Hickman, L.D.; Turkmani, S.; Breen, P.; Gargiulo, G.; Inglis, S.C. “Wearables Only Work on Patients That Wear Them”: Barriers and Facilitators to the Adoption of Wearable Cardiac Monitoring Technologies. Cardiovasc. Digit. Health J. 2021, 2, 137–147. [Google Scholar] [CrossRef]
  137. Sergueeva, K.; Shaw, N.; Lee, S.H. Understanding the Barriers and Factors Associated with Consumer Adoption of Wearable Technology Devices in Managing Personal Health. Can. J. Adm. Sci. Rev. Can. Sci. L’adm. 2020, 37, 45–60. [Google Scholar] [CrossRef]
  138. Loncar-Turukalo, T.; Zdravevski, E.; da Silva, J.M.; Chouvarda, I.; Trajkovik, V. Literature on Wearable Technology for Connected Health: Scoping Review of Research Trends, Advances, and Barriers. J. Med. Internet Res. 2019, 21, e14017. [Google Scholar] [CrossRef] [PubMed]
  139. Nolazco-Flores, J.A.; Faundez-Zanuy, M.; Velázquez-Flores, O.A.; Del-Valle-Soto, C.; Cordasco, G.; Esposito, A. Mood State Detection in Handwritten Tasks Using PCA–MFCBF and Automated Machine Learning. Sensors 2022, 22, 1686. [Google Scholar] [CrossRef] [PubMed]
  140. Baumeister, H.; Kraft, R.; Baumel, A.; Pryss, R.; Messner, E.-M. Persuasive E-Health Design for Behavior Change. In Digital Phenotyping and Mobile Sensing; Springer: Cham, Switzerland, 2019; pp. 261–276. [Google Scholar] [CrossRef]
Figure 1. Types of conditions studied using smartphone sensing.
Figure 1. Types of conditions studied using smartphone sensing.
Sensors 22 03893 g001
Figure 2. Specific subconditions studied using smartphone sensing.
Figure 2. Specific subconditions studied using smartphone sensing.
Sensors 22 03893 g002
Figure 3. Types of operating systems used in the studies. ‘Both’ refers to the study being conducted on iOS and Android.
Figure 3. Types of operating systems used in the studies. ‘Both’ refers to the study being conducted on iOS and Android.
Sensors 22 03893 g003
Table 1. Advantages and disadvantages of active sensing.
Table 1. Advantages and disadvantages of active sensing.
Highly customizable and can collect as much or as little data required [83]Requires regular user input, places burden on the user. This may impact user acceptance, compliance, and retention [61,99].
Ability to collect data about conditions that cannot be sensed directly, such as mental health [16,70,95]Self-reported data can be subjective and susceptible to bias [2,20,100]
Ability to provide additional context or complementary data to passively sensed data [11,53,57,88]Reliance on user memory and recall, which may not always be accurate [70,81].
Table 2. Smartphone-sensing capabilities and application scenarios.
Table 2. Smartphone-sensing capabilities and application scenarios.
Sensor (S)/
Software Feature (SF)
What Does It Collect?What Has It Been Used for?Key Advantages (+)/Disadvantages (−)
Acceleration forces along x, y, and z axes of the deviceIt has been used to detect physical activity (such as standing, walking, running, etc.) and sedentary behavior [11,23,53].
Physical activity has also been used to infer mental wellbeing of individuals [15,67,92] (e.g., decline in physical activity impacting mental health)
+ Relatively privacy-sensitive.
+ Low power
− Accuracy impacted by sampling rate.
− Performance negatively impacted by device placement.
Ambient Light
Amount of light the device is exposed toIt has been used alongside other sensors to understand the user surroundings. Studies used the data to infer when the user was asleep [13,57,58] and infer the amount of spent in the dark, which could provide an indication of mood/mental health [15,26,80]− Only able to make very limited inferences by itself, used in conjunction with other sensors
− Potentially impacted by device placement
Application usage
Information about the applications used on the deviceIt has been used to infer the communication behavior of users. Information such as application use time and genres of applications (e.g., social media) used provided an insight into the user’s sociability and wellbeing [55,70,92].+ Can be used to infer a wide range of user interactions
− Privacy concerns depending on what information is captured.
Battery status
Indicates the phone charging status (on/off)It was used as a proxy measure to infer phone-usage behavior. For example, studies monitoring sleep used it as an indicator of the person sleeping, assuming they charge their phone overnight [19,57].+ Privacy-sensitive
− Only able to make limited inferences by itself, used in conjunction with other sensors
Information about nearby Bluetooth-enabled devicesIt has been used to infer the sociability of the user. By collecting information such as count of nearby Bluetooth devices, number of recurring devices etc., studies were able to infer the social context of users [9,61,76].− Not all nearby devices may have Bluetooth turned on
Capture images and videosIt has been used to infer the user’s emotions by capturing facial images [71]. Another study used the camera to capture eye-movement data and checked if such features could provide an indication of the user emotions [74]. + Ability to visually monitor user behavior
− Higher impact on battery life
− Relatively serious privacy concerns, due to video recording.
Global Positioning System (GPS)
Latitudinal and longitudinal coordinates indicating physical locationIt has been used to infer the mobility of a user (number of places visited, time spent outdoors, time spent at home) which has an impact on wellbeing [26,27,84] (e.g., too much time spent at home indicating a decline in sociability and in turn mental health [7])+ Can use location to make a wide range of inferences about behavior and wellbeing.
− Higher impact on battery life compared to other modes of sensing.
− Privacy concerns, especially when used with a high degree of granularity.
Rotational forces along the x, y, and z axes of the deviceIt has been used in conjunction with the accelerometer for activity recognition. Assisted in detecting activities such as walking, standing, laying etc. [11,30,49]+ Can increase recognition accuracy compared to an accelerometer alone, due to the provision of additional rotational information.
+ Low power
− Impacted by device placement
Collect audio recordings from the surroundingsIt has been used to infer surrounding sound, which can provide information about the user’s context. Some studies used it to detect if the user was alone (i.e., sociability) by listening for conversation [3,54,84]. Some used it to detect if the user was sleeping if the surroundings were quiet (along with other sensor data such as light) [57,58].+ Has utility in respect of social sensing.
− Impacted by device placement
− Relatively serious privacy concerns due to audio recording.
Phone lock/unlock status
Indicates whether the phone is locked or unlockedIt was used to infer phone usage behavior. By calculating the time between the unlock and lock states, studies estimated the phone usage time [24,25,91]. Additionally, this was also used as one of the factors to infer sleep (i.e., phone in locked state for long time during bedtime hours) [57,58,91]+ Privacy-sensitive.
− Unreliable by itself, used in conjunction with other sensors
Phone-call and text-message logs
Logs/records of text messages and phone callsIt has been used to infer the communication patterns of users, which correlate to social wellbeing. For example, decreased frequency of such communication features could indicate decreased sociability of individuals [55,69,85]− Privacy concerns depending on what information is captured.
Screen status
Indicates screen on/off statusSimilar to phone lock/unlock status, it was used to infer phone-user behavior. Screen on/off indicated when the device was being used, which could further indicate distracted/anxious behavior [84], or infer sleep [19,91]− Unreliable by itself, used in conjunction with other sensors
− Can be impacted by phone notifications (resulting in screen on state)
Indicates nearby Wi-Fi connectivityThese types of data were used as a complimentary source to infer location and indicated indoor mobility [8,51,60,88]+ Can increase accuracy of location determination
Table 3. Smartphone-sensing capabilities and application scenarios (Data as of April 2022).
Table 3. Smartphone-sensing capabilities and application scenarios (Data as of April 2022).
[Original Ref]
Platforms SupportedCodebaseLast Updated (Year)Cited by
AWARE [32]Android, iOSAndroid: (accessed on 1 May 2022)
iOS: (accessed on 1 May 2022)
Android: 2020
iOS: 2021
Beiwe (Both open-source and Software-as-a-Service (SaaS) framework for data collection and analysis) [33]Android, iOS (accessed on 1 May 2022)Android: 2021
iOS: 2022
EARS (Initially open-source, now available as SaaS for data collection and analysis [34,111]) [34]Android, iOS (accessed on 1 May 2022)Android: 2020
iOS: 2020
Emotion Sense [112]Android (accessed on 1 May 2022)2017
Project is no longer maintained
RADAR—base [113]Android, iOS (accessed on 1 May 2022)Android: 2022
iOS: 2021
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Share and Cite

MDPI and ACS Style

Kulkarni, P.; Kirkham, R.; McNaney, R. Opportunities for Smartphone Sensing in E-Health Research: A Narrative Review. Sensors 2022, 22, 3893.

AMA Style

Kulkarni P, Kirkham R, McNaney R. Opportunities for Smartphone Sensing in E-Health Research: A Narrative Review. Sensors. 2022; 22(10):3893.

Chicago/Turabian Style

Kulkarni, Pranav, Reuben Kirkham, and Roisin McNaney. 2022. "Opportunities for Smartphone Sensing in E-Health Research: A Narrative Review" Sensors 22, no. 10: 3893.

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop