Medicine X | Nokia
Digital Health
Challenge

A global community of researchers, patients,
and technologists. Five teams challenged to
rethink how research is done.

ei_logo3-300x300
digital health research
that includes everyone

 

We believe patient-generated data is the future of digital health research: therefore patients must have a seat at the table in the design and implementation of digital health research.

Our Everyone Included™ research principles, first developed for President Obama’s Precision Medicine Initiative, have been adapted for digital health research and implemented for our Medicine X | Nokia Digital Health Challenge.

medx_ei  | Powering Digital Health Research

chu_white

“Digital Health is about collaboration, and Nokia has created a research platform that–along with our Everyone Included™ co-creation model– allows our patient- and provider-led research teams to co-develop research in digital health”

 

~Larry Chu, MD, Professor, Stanford Medicine X

the finalists

the top five Everyone Included™ teams selected from the Medicine X community

 

Francois PatouResearch Hypothesis

This research project aims at leveraging the potential of consumer-grade smartwatches and sleep trackers to determine the long-term preventive effect of daily physical activity against the progression of cognitive decline in people living with Mild Cognitive Impairment (MCI). We hypothesize that increased levels and diversity (e.g. walking, cycling, etc.) of physical activity in individuals with MCI can prevent the progression of cognitive decline, including by promoting a better quality of sleep. We finally hypothesize that the objective measurement of physical activity and quality of sleep will reveal stronger associations with cognitive performances than self-reported measures. We intend to test these hypotheses over a two-year longitudinal, controlled, comparative experimental study, possibly together with interested partners within the Nokia research network.

Research Methodology

This project proposal was drafted as natural next step for the follow-up of a current initiative led in Denmark, within the CACHET consortium (www.cachet.dk). The present “Smart-wearables for the monitoring of dementia” project intends to demonstrate the value of consumer-grade wearables to support the pursuit of self-determined, personalized goals for individuals with MCI8,9. This CACHET project of course addresses singular research questions and relies on different assessment methods than proposed for this Medicine X | Nokia challenge. Yet the expertise required for conducting such a study is the same. We, at the Technical University of Denmark and within CACHET, possess the advanced data mining and analytics skills necessary to exploit activity and sleep data, while our clinical research collaborators at Rigshospitalet Glostrup (Glostrup Hospital, Denmark) possess the expertise necessary to leverage our results for clinical inference on MCI and dementia.

Our overall implementation plan involves: – Establishing a research protocol together with interested partners. – Revise the study design depending on available resources and input of interested partners. For instance, a randomization could be envisaged, with one arm following a specific structured physical training program and one arm used for control. The duration, type and frequency of activities carried as part of the training program could be informed by the physiotherapist overseeing the program while a more objective measure of these activities could be recovered using the Activité data. – Recruiting participants through the Nokia Health Research networks and through our current recruitment channel at Rigshospitalet Glostrup (Denmark). – Conduct the study. – Proceed to post-implementation analyses and publication of findings in collaboration with established partners. The modalities of the study itself will be discussed with interested partners, but they will at least require: – The standardized cognitive functional assessment of included participants (e.g. Montreal Cognitive Assessment (MoCA)), as well as assessment tests for possible confounders (e.g. depression, etc.) at baseline, as well as a few cognitive assessment tests throughout the study and at the study end in order to observe trends. – Collecting physical activity data from the study participants via the Withings Health Research platform, using in particular Withings Activité Pop, daily, in particular the intraday activities (i.e. steps, strokes, pool_lap, duration, distance, time of the activity, etc.). – Collecting sleep summary data of each night and associate them with the activities of the previous day. Sleep summary data will include: startdate, enddate, wakeupduration, lightsleepduration, deepsleepduration, remsleepduration, wakeupcount, durationtosleep and du2ationtowakeup. – Processing the data, within our team here at the Technical University of Denmark and in association with members of the CACHET network.

Background

Mild cognitive impairment (MCI) is a syndrome characterized by a cognitive decline greater than expected for an individual of a given age and education level, but that does not interfere notably with activities of daily life [1]. The estimated prevalence of MCI in the general population averages 18.7% [2]. This number significantly increases as the sample population’s mean-age increases, especially beyond 65. In about 50% of the cases, MCI develops into dementia within 5 years. Today, about 36 million individuals live with dementia worldwide [2]. Aside from the psychological and social distress it often causes, dementia weighs heavy on the global economy with a burden expected to account to one trillion US dollars in 2018 [3]. Existing pharmacological treatments generally fail to counteract the progression of dementia, and a body of evidence suggests that physical, cognitive and social activities, especially in earlier life, may prevent the onset or progression of MCI [2]. In particular, physical activity has been correlated with preserving cognitive functioning in older adults, both with and without cognitive impairment. Larger and longer trials are yet still needed in order to provide more conclusive evidence on the preventive effect of physical activity against cognitive decline.

A common limitation of available studies is that they often include multimodal physical exercise which does not differentiate which type of activity (i.e. aerobic, resistance, or stretching) is most influential on the outcome or whether confounding factors such as the socialization aspects associated with physical activity could be at the origin of the observed benefits. Furthermore, these studies often rely on less rigorous assessments of physical activity (e.g. questionnaire or simple pedometer) over relatively short periods of time, possibly weakening the association of unevenly distributed physical activity levels among participants with their respective cognitive scores. Smartwatches, such as the Nokia Activité Pop, could be valuable in objectively-assessing activity levels (e.g. swimming, cycling, etc.) and thus better at determining the activity’s influence for people with MCI on the progression of cognitive impairment.

The seamless integration of the Nokia products in everyday life is paramount, as well as the capabilities they offer for stimulating user engagement and social interactions (e.g. role of family members). Just as day time sedentary behaviour in individuals with MCI is associated with increased cognitive decline, night time insomnia, sleep quality, duration and sleep-disordered breathing have also been associated with neuropsychiatric features, daily functioning, quality of life, disability, carer stress and the need for institutionalization in individuals at risk of developing advanced dementia [4,5]. Although these associations are widely acknowledged, the root cause of the observed patterns is yet to be determined. Opinions seem to have recently shifted regarding the direction of the causality relationship between sleep and cognition. An increasing belief is that sleep disturbances promote cognitive decline [4]. In their review, Yaffe et al. denounce the heterogeneity of sleep assessment methods and the short duration of the reviewed studies as being potentially responsible for the contradictory results associating sleep and cognition. The imprecisions and biases of self-reported sleep quality assessment methods can today be avoided using sleep trackers such as the Aura system, which provides a seamlessly integrated solution for the monitoring of sleep initiation, duration, fragmentation, REM sleep, etc.

Although physical exercise is known to positively affect quality of sleep in older adults [6], only one short prospective study (7 days) recently suggested that physical activity could mediate improved cognitive function in women through better quality of sleep [7]. The study we propose here, aims principally at more thoroughly evaluating the associations and causal pathways between and among physical activity, sleep quality, and changes in cognition. Secondary to these clinical considerations, we would like to leverage the proposed study to investigate the influence of the participants’ perception of the objective measures of their physical activity/sleep quality and their engagement in physical activity. More generally speaking, we would like to study the potential for consumer-grade wearables to promote behavioural change in the MCI patient population and research the implications of our observations for the design of next generation pervasive, assistive technologies.

  1. Gauthier S, Reisberg B, Zaudig M, et al. Mild cognitive impairment. Lancet. 2006;367(9518):2006. doi:10.1016/S0140-6736(06)68542-5.
  2. Andrieu S, Coley N, Lovestone S, Aisen PS, Vellas B. Prevention of sporadic Alzheimer’s disease: Lessons learned from clinical trials and future directions. Lancet Neurol. 2015;14(9):926-944. doi:10.1016/S1474-4422(15)00153-2.
  3. Prince M, Wimo A, Guerchet M, Gemma-Claire A, Wu Y-T, Prina M. World Alzheimer Report 2015: The Global Impact of Dementia – An analysis of prevalence, incidence, cost And trends. 2015:84. doi:10.1111/j.0963-7214.2004.00293.x.
  4. Yaffe K, Falvey CM, Hoang T. Connections between sleep and cognition in older adults. Lancet Neurol. 2014;13(10):1017-1028. doi:10.1016/S1474-4422(14)70172-3.
  5. Naismith SL, Rogers NL, Hickie IB, Mackenzie J, Norrie LM, Lewis SJG. Sleep Well, Think Well: Sleep-Wake Disturbance in Mild Cognitive Impairment. J Geriatr Psychiatry Neurol. 2010;23(2):123-130. doi:10.1177/0891988710363710.
  6. Institute WH. How Does Physical Activity Impact Sleep?; 2016.
  7. Lambiase MJ, Gabriel KP, Kuller LH, Matthews K a. Sleep and Executive Function in Older Women: The Moderating Effect of Physical Activity. J Gerontol A Biol Sci Med Sci. 2014;69(9):1170-1176. doi:10.1093/gerona/glu038.
  8. Crilly N, Maier A, Clarkson PJ. Representing artefacts as media: Modelling the relationship between designer intent and consumer experience. Int J Des. 2008;2(3):15-27. doi:http://dx.doi.org/10.1108/17506200710779521.
  9. Thorpe JR, Rønn-andersen KVH, Bien P, Özkil AG, Forchhammer HB, Maier AM. Pervasive assistive technology for people with dementia : a user-centred design case. Healthc Technol Lett. 2016. doi:10.1049/htl.2016.0057.

 

Research Hypothesis

We aim to assess the hypotheses that home self-monitoring of vital signs data could affect the practice of physicians providing care to the patients post myocardial infarction in terms of the antihypertensive medications and dosage they prescribe for this patient group.

Research Methodology

Aim: This study intends to assess whether access to data on home self-monitoring of vital signs (Blood Pressure (BP), Heart Rate (HR), Weight), collected using Corrie smartphone application and Bluetooth enabled blood pressure cuff and weighing scale, affects prescription of antihypertensive medications by the provider.

Background: Several early lines of evidence point to the success of digital home-based and/or mobile health (mHealth) strategies and emphasize their high potential. In an invited commentary by Polk and O’Gara [1], it was concluded that “the path forward to improve utilization involves novel approaches that center on the patient. We have seen only glimpses of what can be accomplished with digital and e-health strategies. Wide-scale change will require patients, clinicians, insurers, and health systems to adopt and catch up with what is already digitally achievable.” In particular, mHealth provides an innovative and effective approach to promote prevention and management of chronic conditions including cardiovascular disease (CVD); however, the magnitude of these effects is unclear. There have been some recent studies which look into the opportunities, possibilities and potential approaches to exploit the potential of mHealth with respect to cardiovascular diseases [2, 3].

Research Design: The study will comprise a 1-month, dual institution (Johns Hopkins Hospital and Johns Hopkins Bayview Medical Center) observational study. Patients admitted with ST-elevation Myocardial Infarction or Non-ST elevation Myocardial Infarction will be approached for study participation if they are 18 years or older. A requirement of participation will include personal use of a smartphone capable of downloading an application (“app”) from the Internet, and usage of Apple Watch, Bluetooth enabled blood pressure cuff and weighing scale.

Patient Selection: To recruit 200 patients, we plan to enroll approximately 100 from Johns Hopkins Hospital and 100 from Johns Hopkins Bayview who meet the eligibility criteria above. Exclusion criteria will include: (1) patients who are non-English speaking; (2) the presence of visual, hearing, or motor impairment which precludes use of the intervention (smartphone application and/or the Bluetooth enabled devices); (3) patients admitted for primary non-cardiac diagnosis who developed ACS as a secondary condition (e.g., perioperative MI); (4) planned discharge to nursing home or skilled nursing facility.

Methods: If the patient passes screening questions for study inclusion and is interested in participating, a member of the study team will assist the patient in downloading the application on their phone. Patient will be consented. Next, the clinician (investigator on the study) will guide the participant through setting up their device including inputting medications and completing the profile page and instructions on using the Apple Watch and wireless blood pressure monitoring device and scale. Patients will start recording their vital signs (HR, BP and Weight) post discharge at home through Corrie app. Then, patients will follow up with their Cardiologist. The study will comprise two phases. The first phase of study is evaluation of patient at Cardiology clinic 7 days post discharge. The second phase of the study would be the evaluation of de-identified chart data of the same patient along with data on vital signs through Corrie App, viewed by the same cardiologist. In the first phase, the patient will be evaluated at Cardiology clinic. Regular care will be provided and changes on predefined antihypertensive medications will be recommended and noted at the end of visit. In the second phase, the charts of the patients seen by the cardiologist will be de-identified and presented to the cardiologist together with the data from Corrie app on vital signs (BP, HR, Weight). After evaluating the chart and Corrie data, cardiologist will give appropriate recommendation for adjustment of predefined antihypertensive medications and these recommendations will be recorded, to be compared with the recommendations obtained from phase one. There will be six classes of antihypertensive medications that will be recorded for comparison, which include (1) Angiotensin converting enzyme inhibitors/Angiotensin receptor blockers (2) Beta-blockers (3) Calcium channel blockers (4) Diuretics (5) Potassium sparing diuretic (5) Arterial vasodilators and (6) Nitrates. The two phases will be compared in terms of changes in aforementioned antihypertensive medications. Drug changes will be compared in terms of defined daily dosing [4]. Any increase or decrease in the dose of a drug as well as addition or removal of a drug will be reflected as change in defined daily dosing. Finally, two phases will be compared to assess if there is a statistical significance in the difference defined daily dosing of the prescribed antihypertensive medications in view of data on home self-monitoring of vital signs (BP, HR, Weight), through Corrie app, Apple Watch and Bluetooth enabled blood pressure cuff and weighing scale. The comparative analysis will be performed using Wilcoxon rank-sum test.

  1. Doll, Jacob A., et al. “Participation in cardiac rehabilitation programs among older patients after acute myocardial infarction.” JAMA internal medicine 175.10 (2015): 1700-1702.
  2. Rumsfeld, John S., Karen E. Joynt, and Thomas M. Maddox. “Big data analytics to improve cardiovascular care: promise and challenges.” Nature Reviews Cardiology (2016).
  3. Eapen, Zubin J., Mintu P. Turakhia, Michael V. McConnell, Garth Graham, Patrick Dunn, Colby Tiner, Carlo Rich, Robert A. Harrington, Eric D. Peterson, and Patrick Wayte. “Defining a Mobile Health Roadmap for Cardiovascular Health and Disease.” Journal of the American Heart Association 5, no. 7 (2016): e003119.
  4. World Health Organization. Defined daily dose definition and general considerations: WHO Collaborating Centre for Drug Statistics Methodology; 2013. http://www.whocc.no/ddd /definition_and_general_considera. Website: corriehealth.com

 

brownsteinResearch Hypothesis

Schizophrenia is a devastating mental disorder that affects 1% of the world population and often leads a deteriorating course and premature mortality. Sleep disturbances are commonly seen in schizophrenia patients, with 80% of hospitalized individuals with schizophrenia having some form of sleep disorder (PMC5022006), most commonly obstructive sleep apnea (OSA). OSA is a risk factor in sudden cardiac death in schizophrenia (PMC5022006).

Boston Children’s Hospital has amassed a large cohort of children with juvenile onset psychosis (onset before age 13). As this rare but devastating condition is related to adult-onset schizophrenia, we hypothesize that sleep disorders are also occurring in this population, and that diagnosis and treatment would improve sleep quality. Severe patients often miss appointments due to inability to travel, so employing patient-reported data is ideal for those with this condition.

Research Methodology
Our cohort of patients with juvenile psychosis (N=60) will be provided with Nokia Pulse 0x monitors. Patient data will be available to both parents and Boston Children’s Hospital investigators. Aim 1 goal will be to compare sleep/wake cycles between children with juvenile onset psychosis to typical children’s sleep patterns. Aim 2 analysis will investigate changes in sleep patterns due to parental awareness of problems highlighted by the Nokia Pulse 0x. Patient data will be analyzed in conjunction with other phenotypic and genotypic data. If a sleep disorder is determined, patients will referred to the BCH sleep clinic and treated for their disorder. Patients will continue to wear their Nokia Pulse 0x. 3 months after the initiation of treatment, sleep/wake cycles will be analyzed and compared to a) pre-treatment and b) typical sleep patterns. Correcting childhood sleep disorders in this vulnerable population may lead to improved outcomes later in life.

 


Research Hypothesis

Utilization of Nokia platform for predicting and preventing postpartum depression. One in 8 new mothers experiences postpartum depression (PPD), a common condition that affects the wellbeing of the whole family. Despite its high occurrence, little is known about the development of PPD and even less is done to prevent PPD. I hypothesize that 1) by monitoring daily personal metrics (sleep, activity, heart rate, body composition) one can predict the women at risk of developing postpartum depression and 2) behavioral interventions based on personalized data would prevent or alleviate the symptoms of PPD. I propose to study the potential of early detection and timely intervention through the utilization of Nokia platform for the prevention, diagnosis and treatment of PPD.

Research Methodology

Methods and Patient Community
Nokia existing data structure and platform will be utilized for the collection of activity, sleep, weight, and heart rate. In addition, subjective data regarding mood and feelings will be collected by brief questions prompted through the app. Below is one example:

Which of the following describes your feelings for today:
I feel as happy as usual.
I feel less happy than usual.
I feel happier than usual.
The retrospective part of the study will use existing data and look for the general trends between these metrics and the mood disturbances. The prospective part of the study will focus on pre- and postpartum period for female (pregnant and/or within 1 year of childbirth) users of 18-50 years of age. For the intervention group, simple suggestions will be provided to guide and support new mothers in their daily activities. If need arises, referrals for professional care will be made to prevent and/or to facilitate timely diagnosis and fast recovery.

Research plan
Part 1: In order to understand the average trends in the wellness data, a retroactive study (spanning Jan-Dec 2016) will be performed to see 1) how each variable fluctuates over time, 2) whether these fluctuations correlate with an underlying mood disturbance and/or disorder, based on self-reports and/or medical diagnosis. For the second part, members with existing data will be prompted to provide some information on their overall health and wellness level. Below is an example:

Have you been diagnosed with a health problem within the last year?
If yes, which of the following categories describes it the best?
Cardiovascular problems
Mood related problems
Orthopedic problems
Gastrointestinal problems

This phase of the study will provide a general background information on the data and be used to refine the parameters/surveys/analytical tools for the second part of the study.

Part 2: Pregnant women (28th weeks and later) will be recruited and assigned into two groups. Extra care will be provided to ensure groups match by age, pregnancy-complications, and socio-economic status. These women will be provided with Nokia devices and will be prompted to answer short-surveys during the study (up to 1 year postpartum). One of the groups will interact with the existing platform. The other, the intervention group, will receive simple behavioral suggestions based on their wellness levels. Below is an example: ‘
You slept less than 5 hours yesterday, try to get a nap today when the baby sleeps.

Data for the last quarter of the pregnancy will be used as a baseline. The first level of analysis will be to extract whether postpartum wellness data can predict PPD. Second level of analysis will include a comparison between the two groups to determine whether interventions were successful in preventing and/or ameliorating PPD symptoms.

Overall, this study will provide crucial information on the possibility of predicting and preventing and/or alleviating PPD using personalized daily wellness data. Results will have a direct impact on not only women’s health but also the healthy development of children and unity of the family structure.

 

bottrosResearch Hypothesis
The United States is unfortunately in an “opioid epidemic” where increased numbers of opioids are being prescribed to patients in pain. This has led to increased addiction rates and deaths related to overdoses. An often overlooked, but significant contributor to this epidemic is the prescribing pattern and use of postoperative opioids. Patients are often being discharged with large quantities of opioids on a “just-in-case” basis because their physician follow-up is in 4 or more weeks, all in the name to increase patient satisfaction. This leads to a large amount of medication left unprotected at home serving as a catalyst for abuse by patients, family members, or friends. Once discharged from the hospital, we often overlook important aspects of post-discharge care, such as opioid consumption and activity level. We hypothesize that more granular data collected post-discharge from activity trackers and apps will help improve opioid prescription patterns and reduce unused opioids.

Research Methodology
We will focus on two surgeries that have grown exponentially in the last few years, namely hip and knee replacements. One hundred patients will be randomized to traditional post-discharge care (TPD) group or personalized post-discharge care (PPD). In the TPD group, patients will be discharged from the hospital with the traditional amounts of opioids and routine post-operative follow-up with their surgeon at the usual 4-6 weeks. In the PPD group, patients will be discharged with a Nokia Pulse Ox as well as Blood Pressure monitor. We will monitor patients’ heart rate and blood pressure and correlate with opioid medication usage using the Health Mate app along with tailored questions related to pain level and medication use. We will also track sleep quality and pulse oximetry as it relates to pain levels, medication use, and the risk of respiratory depression related to opioids. We will finally compare actual medication usage within the groups in an effort to learn more appropriate opioid prescribing patterns for these patients.

Great Talks from the Stanford Medicine X Main Stage