Susan Allbritton Murphy

Susan Allbritton Murphy

Auto-generated profile, based on publicly available information

Susan Allbritton Murphy

Harvard University

Research Focus & Metrics

Academic Career Length

15Years (2012-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
27 August 2018 in BMC Medicine
Verified
Journal Article
Open access
PDF file
Holger FröhlichHolger Fröhlich
Holger Fröhlich
RBRudi Balling
Rudi Balling
NBNiko Beerenwinkel
Niko Beerenwinkel
OKOliver Kohlbacher
Oliver Kohlbacher
SKSantosh Kumar
Santosh Kumar
BZBlaz Zupan
Blaz Zupan
Personalized, precision, P4, or stratified medicine is understood as a medical approach in which patients are stratified based on their disease subtype, risk, prognosis, or treatment response using specialized diagnostic tests. The key idea is to base medical decisions on individual patient characteristics, including molecular and behavioral biomarkers, rather than on population averages. Personalized medicine is deeply connected to and dependent on data science, specifically machine learning (often named Artificial Intelligence in the mainstream media). While during recent years there has been a lot of enthusiasm about the potential of ‘big data’ and machine learning-based solutions, there exist only few examples that impact current clinical practice. The lack of impact on clinical practice can largely be attributed to insufficient performance of predictive models, difficulties to interpret complex model predictions, and lack of validation via prospective clinical trials that demonstrate a clear benefit compared to the standard of care. In this paper, we review the potential of state-of-the-art data science approaches for personalized medicine, discuss open challenges, and highlight directions that may help to overcome them in the future. There is a need for an interdisciplinary effort, including data scientists, physicians, patient advocates, regulatory agencies, and health insurance organizations. Partially unrealistic expectations and concerns about data science-based solutions need to be better managed. In parallel, computational methods must advance more to provide direct benefit to clinical practice.

29 November 2018 in JMIR mHealth and uHealth
Verified
Journal Article
Open access
NBNiranjan Bidargaddi
Niranjan Bidargaddi
DADaniel Almirall
Daniel Almirall
SMSusan Murphy
Susan Murphy
INInbal Nahum-Shani
Inbal Nahum-Shani
MKMichael Kovalcik
Michael Kovalcik
VSVictor Strecher
Victor Strecher
Journal of Medical Internet Research - International Scientific Journal for Medical Research, Information and Communication on the Internet #Preprint #PeerReviewMe: Warning: This is a unreviewed preprint. Readers are warned that the document has not been peer-reviewed by expert/patient reviewers or an academic editor, may contain misleading claims, and is likely to undergo changes before final publication, if accepted, or may have been rejected/withdrawn. Readers with interest and expertise are encouraged to sign up as peer-reviewer, if the paper is within an open peer-review period. Please cite this preprint only for review purposes or for grant applications and CVs (if you are the author). Background: Mobile health applications (apps) provide an opportunity for easy, just-in-time, access to health promotion and self-management support. The number and breadth of health-related apps in the marketplace has provided a programmatic resource with immense potential for public health. However, poor user engagement with these apps remains a significant unresolved challenge. Objectives: Given the information on when a user might benefit from which type of support, the primary aim of this quality improvement study was to assess the effect of offering this support via a push notification containing a tailored health message versus not sending a push notification on proximal engagement with the app. Secondary aims examined whether this effect varies by number of weeks in the program, day of the week, and time of day. Methods: The quality improvement study analyzes the causal effects of push notifications on proximal engagement in 1,255 users of a commercial workplace wellbeing intervention product over 89 days. The test application employs a micro-randomized trial (MRT) design to send push notifications. At one of six available times per day (8:30AM, 12:30PM, 5:30PM, 6:30PM, 7:30PM or 8:30PM; selected randomly), available users were randomized with equal probability to be sent or not sent a push notification containing a tailored health message. Availability rules were pre-specified and depended on time since the previous sent push notification, time since previous user engagement with the app, and day of the week. Push notifications contained health messages tailored to the user as well as to the user’s current context. The primary outcome was whether the user engaged with the app over the next 24 hours. A generalization of log-linear regression analysis, adapted for use with data arising from a MRT, was used to examine the effect of sending a push vs not sending a push on the probability of engagement over the next 24 hours. Results: Users were 3.9% more likely to engage with the app in the next 24 hours when a tailored health message was sent vs when it was not sent (95%CI, 1.01 to 1.08; p<0.05). The effect of sending the message attenuated over the course of the study, but not significantly (p=0.84). When sent a tailored health message on weekends, users were 8.7% more likely to engage with the app (95% CI, 1.01 to 1.17; p<0.05); whereas on weekdays users were 2.5% more likely to engage with the app (95% CI, 0.98 to 1.07; p=0.27). The effect of sending a tailored health message was greatest at 12:30PM on weekends: 11.8% more likely to engage (90% CI, 1.02 to 1.13; p<0.10). Conclusion: Sending a push notification containing a tailored health message was associated with greater engagement in a mobile health app. Results suggested that users are more likely to engage with the app within 24 hours when push notifications are sent at mid-day on weekends. This has potential implications for other mobile health apps that rely on user-engagement and self-monitoring.

17 March 2022 in American Psychologist
Verified
Journal Article
Closed access
INInbal Nahum-Shani
Inbal Nahum-Shani
SSSteven D. Shaw
Steven D. Shaw
SCStephanie M. Carpenter
Stephanie M. Carpenter
SMSusan A. Murphy
Susan A. Murphy
CYCarolyn Yoon
Carolyn Yoon
The notion of “engagement,” which plays an important role in various domains of psychology, is gaining increased currency as a concept that is critical to the success of digital interventions. However, engagement remains an ill-defined construct, with different fields generating their own domain-specific definitions. Moreover, given that digital interactions in real-world settings are characterized by multiple demands and choice alternatives competing for an individual’s effort and attention, they involve fast and often impulsive decision making. Prior research seeking to uncover the mechanisms underlying engagement has nonetheless focused mainly on psychological factors and social influences and neglected to account for the role of neural mechanisms that shape individual choices. This paper aims to integrate theories and empirical evidence across multiple domains to define engagement and discuss opportunities and challenges to promoting effective engagement in digital interventions. We also propose the AIM-ACT framework, which is based on a neurophysiological account of engagement, to shed new light on how in-the-moment engagement unfolds in response to a digital stimulus. Building on this framework, we provide recommendations for designing strategies to promote engagement in digital interventions and highlight directions for future research.

13 January 2022 in Psychological Methods
Verified
Journal Article
Closed access
TQTianchen Qian
Tianchen Qian
AWAshley E. Walton
Ashley E. Walton
LCLinda M. Collins
Linda M. Collins
PKPredrag Klasnja
Predrag Klasnja
SLStephanie T. Lanza
Stephanie T. Lanza
SMSusan A. Murphy
Susan A. Murphy
Just-in-time adaptive interventions (JITAIs) are time-varying adaptive interventions that use frequent opportunities for the intervention to be adapted—weekly, daily, or even many times a day. The micro-randomized trial (MRT) has emerged for use in informing the construction of JITAIs. MRTs can be used to address research questions about whether and under what circumstances JITAI components are effective, with the ultimate objective of developing effective and efficient JITAI.

18 July 2018 in JMIR Research Protocols
Verified
Journal Article
Open access
MRMashfiqui Rabbi
Mashfiqui Rabbi
MKMeredith Philyaw Kotov
Meredith Philyaw Kotov
RCRebecca Cunningham
Rebecca Cunningham
EBErin E Bonar
Erin E Bonar
INInbal Nahum-Shani
Inbal Nahum-Shani
SMSusan Murphy
Susan Murphy
Background: Substance use is an alarming public health issue associated with significant morbidity and mortality. Adolescents and emerging adults are at particularly high risk because substance use typically initiates and peaks during this developmental period. Mobile health apps are a promising data collection and intervention delivery tool for substance-using youth as most teens and young adults own a mobile phone. However, engagement with data collection for most mobile health applications is low, and often, large fractions of users stop providing data after a week of use. Objective: Substance Abuse Research Assistant (SARA) is a mobile application to increase or sustain engagement of substance data collection overtime. SARA provides a variety of engagement strategies to incentivize data collection: a virtual aquarium in the app grows with fish and aquatic resources; occasionally, funny or inspirational contents (eg, memes or text messages) are provided to generate positive emotions. We plan to assess the efficacy of SARA’s engagement strategies over time by conducting a micro-randomized trial, where the engagement strategies will be sequentially manipulated. Methods: We aim to recruit participants (aged 14-24 years), who report any binge drinking or marijuana use in the past month. Participants are instructed to use SARA for 1 month. During this period, participants are asked to complete one survey and two active tasks every day between 6 pm and midnight. Through the survey, we assess participants’ daily mood, stress levels, loneliness, and hopefulness, while through the active tasks, we measure reaction time and spatial memory. To incentivize and support the data collection, a variety of engagement strategies are used. First, predata collection strategies include the following: (1) at 4 pm, a push notification may be issued with an inspirational message from a contemporary celebrity; or (2) at 6 pm, a push notification may be issued reminding about data collection and incentives. Second, postdata collection strategies include various rewards such as points which can be used to grow a virtual aquarium with fishes and other treasures and modest monetary rewards (up to US $12; US $1 for each 3-day streak); also, participants may receive funny or inspirational content as memes or gifs or visualizations of prior data. During the study, the participants will be randomized every day to receive different engagement strategies. In the primary analysis, we will assess whether issuing 4 pm push-notifications or memes or gifs, respectively, increases self-reporting on the current or the following day. Results: The microrandomized trial started on August 21, 2017 and the trial ended on February 28, 2018. Seventy-three participants were recruited. Data analysis is currently underway. Conclusions: To the best of our knowledge, SARA is the first mobile phone app that systematically manipulates engagement strategies in order to identify the best sequence of strategies that keep participants engaged in data collection. Once the optimal strategies to collect data are identified, future versions of SARA will use this data to provide just-in-time adaptive interventions to reduce substance use among youth. Trial Registration: ClinicalTrials.gov NCT03255317; https://clinicaltrials.gov/show/NCT03255317 (Archived by WebCite at http://www.webcitation.org/70raGWV0e) Registered Report Identifier: RR1-10.2196/9850

22 July 2022 in Algorithms
Verified
Journal Article
Open access
PDF file
Anna Li TrellaAnna Li Trella
Anna Li Trella
KZKelly W. Zhang
Kelly W. Zhang
INInbal Nahum-Shani
Inbal Nahum-Shani
VSVivek Shetty
Vivek Shetty
FDFinale Doshi-Velez
Finale Doshi-Velez
SMSusan A. Murphy
Susan A. Murphy
Online reinforcement learning (RL) algorithms are increasingly used to personalize digital interventions in the fields of mobile health and online education. Common challenges in designing and testing an RL algorithm in these settings include ensuring the RL algorithm can learn and run stably under real-time constraints, and accounting for the complexity of the environment, e.g., a lack of accurate mechanistic models for the user dynamics. To guide how one can tackle these challenges, we extend the PCS (predictability, computability, stability) framework, a data science framework that incorporates best practices from machine learning and statistics in supervised learning to the design of RL algorithms for the digital interventions setting. Furthermore, we provide guidelines on how to design simulation environments, a crucial tool for evaluating RL candidate algorithms using the PCS framework. We show how we used the PCS framework to design an RL algorithm for Oralytics, a mobile health study aiming to improve users’ tooth-brushing behaviors through the personalized delivery of intervention messages. Oralytics will go into the field in late 2022.