Navigating the Ocean of Apps

Transcription

1 Navigating the Ocean of Apps P. Pearl O Rourke, MD Partners HealthCare Boston, MA What we have. What we want. How it feels.

2 Agenda What are mhealth apps Why the popularity? Use in clinical care and research What are the concerns What should the IRB ask?

3 Definitions App is a shortening of the term "software application 2010 Word of the Year A mobile app is a software program designed to run on a computer or mobile device such as a phone/tablet or watch. mhealth app is a mobile app focused on health related issues

4 mhealth Apps Everywhere

5 /news/mental healt there s an app=for t

6 10 mental health apps/slideshow/2608/

7 kit

8 The explosion of Apps September 2017: Total number of apps on Google Play 3.3 million (doubled over 1 year) The number of mhealth apps may have also doubled, from 165,000 to 330,000. Of the mhealth apps: Roughly 76% relate to fitness, lifestyle, diet, etc.. The balance, about 80,000, closer to the definition of medical applications Apps for specific diseases: ~ 35% Apps for Women s healthcare and pregnancy: ~ 30% Medication reminders and information: ~ 25% Apps for insurance and healthcare providers: ~10% medical application approval fda/

9 Examples of new FDA cleared Apps YuGo Tracks movement and creates record for PT Contact lens that tracks glaucoma 24 hour monitor MoMe Kardia tracks ECG, respiration and motion Wearable for pain relief to release natural opiods Smart Pump a pump tracks pumping sessions, feedings, manages breastmilk inventory Trak system: app connected home device for male fertility smartphone app to tell sperm count low, med Qbtech clinical decision making for dx and rx of ADHD Bluetooth enabled oximeter

10 Why the popularity? Feeds off the iphone junkie mentality 80% of population has a cell phone Phones have been medicalized ECG, blood pressure, respiratory function, blood glucose Immediate feedback Have developed ways to keep people engaged: Gamification Congratulatory messaging

11 Why the popularity in clinical care? Access to more real time real life data Expanded feed May be able to look at behaviors and social interactions Remote access Patient/subject involvement Opportunity for feedback May improve compliance

12 For research: mhealth promises better information Can reach more people Gather more information Physiological parameters continuously and in real time Measures of behavior and related health states Environmental exposure Measures of social interaction GPS geographic space issues Improve retention and compliance Ease of engagement

13 kit

14 kit

15 kit

16 kit

17 General App Concerns App overload Do app developers understand the needs? Premature commercialization Accuracy Participant/patient interface Sustaining interest Privacy and Security

18 App Overload 330,000 mhealth apps Roughly 76% relate to fitness, lifestyle, diet and other non specific categories. The balance, about 80,000, closer to the definition of medical applications. medical application approval fda/

19 Who is making the apps? Health apps are increasingly being developed by purely digital entities

20 healththere s an app=for that but few have been thoroughly tested

21 Mental Health Apps bare fact is that most apps haven t been tested at all. A 2013 review identified more than 1,500 depression related apps in commercial app stores but just 32 published research papers on the subject...australian researchers... assessed how commercially available apps affected mentalhealth symptoms or disorder. They found eight papers on five different apps. health there s an app for that

22 Is the data accurate?

23 one app claiming to give a smartphone the ability to measure blood pressure was assessed by an emergency physician and found to give inaccurate readings. [5] The doctor contacted the app developer, who told him the app was only in the "betatesting stage" (ie, not ready for prime time) and that its use was for "entertainment purposes only." None of these caveats were included in the app description at the online store from which it could be downloaded for $3.99.

24 On the 108 cases used for evaluation the algorithm scored 80% sensitivity and 78% specificity in detecting (pre)malignant conditions.

25 heart rate 5/5/1717 May 05, 2017 About a fifth of measurements recorded by heart rate monitoring mhealth apps for consumers can be inaccurate by upwards of 20 beats per minute, compared to clinical monitoring

26 apps for cardiac caresanctioned 5/5/17

27 Who uses apps and can their interest be sustained?

28 Limitations of Apps Nature Biotechnology, DOI: /nbt.3826 Apple Asthma app Selection Bias Participants are younger, wealthier, more educated, male, sicker Low participant retention 40.6k downloaded: 7593 enrolled: 131 (1.7%) completed week of surveys and 6 month report Results in incomplete data sets

29 Potential challenges with this technology include selection bias, low retention rates, reporting bias, and data security. These issues require attention to realize the full potential of mobile platforms in research and patient care.

30 nce.com/mhealth appmarket gettingcrowded mhealth apps new/

31 Issues: Unappreciated amount of information being collected Secondary uses of that info Uncertain Privacy Protections

32 App functions Global positioning When people are aware many refuse to allow activation of specific functions (e.g., GPS) Camera Microphone Movement, accelerometer, gyroscope Recordings can capture people around you

33 Apps as Trojan Horses Permissions, which users must accept to download an app: authorized collection and modification of sensitive information, including: tracking location (17.5%), activating the camera (11.4%), activating the microphone (3.8%), and modifying or deleting information (64.0%) Privacy Policies of Android Diabetes Apps and Sharing of Health Information JAMA. 2016;315(10): doi: /jama

34 Third party Apps on Facebook: Privacy and the Illusion of Control ACM CHIMIT 2011 WSJ: numerous third party app listed on Facebook extract identifiable user information and share with advertising companies Situations in which 3 rd party app overrides users privacy settings Happy Calendar User wants to block birthday disclosure Then adds Happy Calendar agreed to app disclosures and voila a calendar with her birthday available to all When adding an app no control to limit/restrict the app s access only after adding the app can users adjust privacy settings Facebook can enable 3 rd party apps to override users privacy settings

35 What can be done with GPS data!

36

37 What about Privacy Policies? 125 iphone apps for dementia 33 had available privacy policies 271 diabetes apps/ 211 used in the sample 81% had no privacy policy Those with policies majority collected user data and approximately 50% shared the data 31 of 41 without privacy policy shared user information 19/24 with privacy policies shared user data How Weak Mobile Health App Privacy, Security Affects Patients weak mobile health app privacysecurity affects patients

38 What about Privacy Policies? A separate analysis found that 35 of the mobile health apps originally listed.transmitted identifying information such as e mail addresses, names and birthdates about users over the internet, and two thirds of these did not encrypt the data. health there s an app for that

39 What about Privacy Policies? 271 diabetes apps (211 at time of review) Only 19% had privacy policies and not all actually protected privacy (eg, 80.5% collected user data and 48.8% shared data) Only 4 policies said they would ask users for permission to share data. Privacy Policies of Android Diabetes Apps and Sharing of Health Information JAMA. 2016;315(10): doi: /jama

40 What about Privacy Policies? Health information (eg, insulin and blood glucose) routinely collected and shared with third parties 56 of 65 apps (86.2%) placed tracking cookies; User information sharing: 76% of apps with no privacy policy 79% of apps with a privacy policy The 19 apps with privacy policies that shared data with third parties, 11 apps disclosed sharing 8 apps did not. Privacy Policies of Android Diabetes Apps and Sharing of Health Information JAMA. 2016;315(10): doi: /jama

41 Splinter News 8/29/16 Facebook recommended that this psychiatrist's patients friend each other Based on single cell phone noted Facebook coordinated with others and put her patients in contact with each other

42 Privacy Policy Comprehension 64 popular apps: reading level of or equivalent to 1 st year in college Privacy Policies for Apps Targeted toward Youth: Descriptive analysis of Readability

43 The IRB What to do when a protocol includes an app? Or several apps? How to assess risk benefit when an app is involved?

44 What should an IRB ask about: the App? How many apps and what is their status? Marketed app? New app? Is it an approved app? By whom? Accuracy of the data? What is the security of the app? What does the app platform keep/have access to? Will the data be shared? If so with whom? When? Is there a EULA (End User Licensing Agreement)? Is it comprehensible?

45 What should the IRB ask about: Research Design? What is the user interface? Who is the user? And how to monitor? Fit bit on the dog? Passive data flow Active input Parkinson s screen tapping Diary think school student Response to prompts Is data about close contacts captured?

46 What should the IRB ask about: The Participants? Who are the participants? Vulnerable and/or small group? Sensitive disease/condition What is their Experience with devices and apps? If experienced may be a biased population! Will they use their own device/app or will the study provide one If provided, what happens if lost? Stolen? End of study?

47 What should the IRB ask about: Subject Engagement and Retention? Differences between the routine app user and the app naïve user Duration and intensity of the intervention Are there incentives to maintain engagement? Does the study design and statistical plan account for possible retention problems

48 What should the IRB ask about: Privacy and Security? Privacy Policy of the app/s Specifically ask: Who has access to the data? (App platform, 3 rd parties) Will data from other sources be merged with the study data? E.g., if 23andMe, how does this change identifiability? Risk? Secondary uses of the data by the app owner Specifically any commercial uses? JMIR Mhealth Uhealth :e126

49 Limitations of Apps Nature Biotechnology, DOI: /nbt.3826 Suggestion Narrow research question Short duration study Passive data collection Statistical Analysis Plan to deal with low retention and missing data

50 What should the IRB ask about: App approval or clearance? Uncertainty re: review and approval of various apps. Lines are fuzzy at best. In the US: The FDA is taking a tailored, risk based approach that focuses on the small subset of mobile apps In general, an app that claims to prevent, diagnose or treat a specific disease is likely to be considered a medical device and to attract regulatory scrutiny, whereas one that promises to 'boost mood' or provide 'coaching' might not. medicalapplication approval fda/

51 Examples of Apps that will NOT require FDA application/registration Help patients/users self manage their disease or condition without providing specific treatment suggestions; Provide patients with simple tools to organize and track their health information; Provide easy access to information related to health conditions or treatments; Help patients document, show or communicate potential medical conditions to health care providers; ations/default.htm

52 Is the App worth it? 201 diabetes apps reviewed none met all of the functional criteria desired for a diabetes app Majority included education or recommendations Many were simply food recipes given the high level of heterogeneity in the apps we found, we believe there is a need for standard mhealth app certification criteria developed and regulated by a credible organization Diabetes mhealth Apps: designing for greater uptake IOS doi: /

53 What should the IRB do about: Informed Consent? Will consent be remote? Or in person? Comprehension? Opportunity for asking questions? Consent should include: Details about the app Privacy and security issues in an understandable presentation EULA details in an understandable presentation Where there information will go and how it may be used. Any costs: For the device, the app, increased phone use time. Responsibility if lost etc. Ability to withdraw from the study Any special limitations with apps?

54 What should the IRB do about: Ancillary Reviews? Privacy and Security review Most IRBs do not have adequate expertise consider having specific privacy and security review Legal review EULAs can be very complex, difficult to understand and include liability language, holding the user responsible

55 Common Responses from Investigators These apps are being used for clinical care why is the IRB being so picky? These people already have their own Smart phone with this app why is the IRB being so picky?

56 Summary Apps and mhealth apps are everywhere And not going away Challenges in research User interface Over promise Privacy and security Informed consent

57 Navigating the Ocean of Apps What we have. What we want. How it feels.