Use of Mobile Health Applications in Low-Income Populations
See Editorial by Merid et al
Goals and Vision of the Program
Applications accessible on digital platforms are increasingly popular in healthcare settings and have the potential to help patients engage with their health through monitoring of health behaviors (eg, physical activity, dietary habits) and physiological parameters (eg, heart rate, blood pressure, blood glucose). Some mobile health (mHealth) applications provide patients access to their medical data through patient portals and the ability to participate in real-world, pragmatic research using their smart devices.1 For low-income populations, however, unique barriers to mHealth utilization exist, amplifying the impact of social determinants of health. These barriers include fluency with mobile applications, limited health literacy, lack of empowerment, and historical mistrust of healthcare systems.2 As mHealth platforms play a larger role in healthcare delivery, the digital divide could serve to worsen health disparities.3,4
mHealth applications that are sensitive to the user needs of vulnerable populations have the potential to gain uptake in more diverse communities. Consideration to the visual and linguistic design of mHealth applications, along with how mHealth applications are introduced to patients, may increase adoption and acceptability. One promising intervention is the use of community health workers (CHWs), whose role is to support patients in their health-related needs and address social determinants of health, to promote greater usage and understanding of health applications, although this has not been studied.
Accordingly, to better understand the real-world challenges and potential catalysts, including the use of CHWs, in increasing mHealth technology uptake in vulnerable populations, we evaluated the onboarding, perspectives, and experiences of underserved patients using a novel mHealth application. We aimed to learn how a digital health platform could be customized to better serve low-income patients and the supports necessary to facilitate mHealth adoption with sustained engagement.
Design of the Initiative
We conducted a community-based participatory research project in partnership with a community health organization, a primary care center, and a digital health company. Project Access-New Haven is a nonprofit organization with the mission of improving access to care for the underserved. CHWs at Project Access-New Haven are bilingual in English and Spanish, often come from the same communities as patients, and are knowledgeable about resources within the community to address social determinants of health, including digital connectivity. CHWs function in similar ways to a patient navigator—building trusting and supportive relationships and helping to advance health and well-being—but they are not integrated in the medical system to navigate care. The Yale Primary Care Center is a hospital-based clinic serving primarily patients with Medicaid. Hugo Health is a digital health company that produced its eponymic application, enables people to connect to their electronic health data, sync wearables, and participate in research. With patients’ permission, patients can opt to share portions of their data with research teams and answer surveys from researchers using their email or smart devices.
Our study’s first phase, preimplementation, involved designing a seamless participant experience at initial enrollment and throughout the study, and developing and β-testing survey questions in English and Spanish to achieve readability and understanding (Figure). Project Access-New Haven CHWs provided specific feedback on terminology and syntax that may vary between speakers from different Spanish-speaking countries. This was achieved via biweekly phone meetings that included all partners. The second phase of patient enrollment and study participation took place in 2 sequential phases to allow time for the development of a Hugo Spanish interface: English language-preferred (ELP) patients were enrolled from April 25, 2018 to May 31, 2018, and Spanish language-preferred (SLP) patients were enrolled from October 11, 2018 to December 20, 2018. We obtained approval from the Yale University Institutional Review Board (protocol ID: 2000022503), and all patients provided informed consent.
Figure. Components of successful adoption of a multifaceted digital health application (app) in a low-income primary care population. When the app was introduced with these supports, participants were interested in using the health app to access health records, keep track of their own health information, get appointment reminders, sync wearable devices, and participate in research. CHW indicates community health worker.
Implementation of the Initiative
Patients attending primary care visits were eligible if they had ongoing healthcare needs for which the Hugo application would be helpful, demonstrated basic literacy in English or Spanish, and had access to a smart device. A CHW stationed at the healthcare centers assessed eligibility, gauged interest in study participation, and consented participants for enrollment (Figure). If patients did not want to participate, the CHW logged the reason(s) for nonparticipation. For all enrolled patients, the CHW downloaded the application on their mobile devices, connected their health portal(s), and taught them how to use application features that were most applicable to the patient (eg, following up on test results, sharing health data with clinicians/family members; tracking symptoms and physiological data from connected devices). The CHW practiced using these different functionalities with the patient. Patients were informed that they would receive 6 biweekly follow-up surveys (totaling 3 months) through Hugo. The CHW texted with and called patients at 1 and 3 months to check in and remind them to take the surveys. Incentives were provided for the time and effort involved with study participation, with $25 provided at 2 time points—upon completing the baseline survey and following completion of the final survey. Finally, the CHW kept a qualitative log of barriers encountered and strategies utilized to address those barriers during the enrollment process.
Results were shared with participants at the end of the study (Figure).
Success of the Initiative
The overwhelming majority of patients were interested in engaging in an mHealth study. Among 108 ELP and 52 SLP patients approached for the study, 81 and 50 were enrolled, respectively (Table I in the Data Supplement). Reasons for declining participation (n=29) were as follows: did not own smartphone (n=4), did not have email (n=6), and not interested (n=17). Survey response rates varied but were particularly high for the surveys sent at 1 month (CHW outreach) and 3 months (CHW outreach, second monetary incentive; Table II in the Data Supplement).
Participants were interested in having access to their health records and were open to participating in research, with some differences by language preference. Specifically, most ELP (41 of 52) but only half of SLP (16 of 32) respondents reported feeling comfortable sharing data gathered on connected devices, such as FitBits and blood pressure cuffs, with a trusted research team (Table III in the Data Supplement). The vast majority of ELP and SLP respondents were somewhat interested in being the owner of their health records (56 of 67; 43 of 45, respectively), having their health records in one place (61 of 66; 40 of 40, respectively) and being able to view their records on their phones or tablets (68 of 68; 43 of 44, respectively; Figure I in the Data Supplement). Furthermore, 61 of 69 ELP and 42 of 46 SLP participants indicated that they felt somewhat comfortable with participating in research using their mobile phone or tablets (Figure I in the Data Supplement). Most preferred to receive research surveys via text message (24 of 43; 22 of 29, respectively) rather than email (15 of 43; 3 of 29, respectively; Table III in the Data Supplement).
Still, participants expressed concerns about privacy and data security. The majority wanted to learn more about privacy policies explaining how data are collected and used (38 of 46; 25 of 27, respectively), feeling somewhat concerned about sharing health information with research teams (32 of 48; 14 of 27, respectively), and that information collected through applications could be sold or shared without their permission (39 of 48; 19 of 27, respectively; Figures II and III in the Data Supplement).
There were several requests throughout the study to speak with a CHW about different application features or privacy policies. The median percentage of patients requesting CHW assistance (across survey points) was 51.6% for SLP participants and 28.1% for ELP respondents (Figure IV in the Data Supplement).
Compared with before the study, ELP and SLP participants reported feeling more comfortable using health applications (29 of 43; 23 of 28, respectively) and more likely to participate in research using applications (32 of 42; 21 of 29, respectively; Figure V in the Data Supplement). Moreover, 25 of 70 ELP and 8 of 45 SLP respondents stated that they used other mHealth applications during the study (Table III in the Data Supplement). A greater proportion of SLP (41 of 47) than ELP (41 of 70) respondents reported being somewhat or likely to share health data with a trusted research team (Figure I in the Data Supplement). Participants randomly selected to be surveyed at study conclusion reported that they trusted the CHW and Hugo team, and they were excited to use many of the applications’ features (Text I in the Data Supplement).
Participants responded that they were interested in additional functionalities, including using applications that bring together people who share similar health conditions (28 of 43 ELP respondents; 20 of 27 SLP respondents), learning about other people’s experiences (20 of 28; 6 of 19, respectively), and learning about research studies and clinical trials (16 of 28; 2 of 19, respectively; Table III in the Data Supplement).
Local Challenges in Implementation
There were several logistical challenges during enrollment, including technological and economic concerns (eg, owning a smartphone/tablet, limited storage space, texting plans, and consistent access to Wi-Fi or data plans to download the application and send/receive surveys). Many patients did not remember their email address or password or did not know how to use email to download the application. In these cases, the CHW often set up new email accounts and made wallet-sized cards with participants’ login information. Based on this experience, the platform was adapted to accommodate email and text messaging options to receive surveys.
Survey data also revealed several data-sharing concerns, including mistrust related to where health data are stored, who has access to that data, and concerns about data security. To better understand and address these contextual factors, the CHW served as a resource for any questions or concerns regarding application privacy. By maintaining a close relationship with the digital health company and research team through regular meetings, the CHW was able to communicate these questions and concerns and provide participants with specific information, when indicated. These communications prompted updates to workflows and features of the platform in real time that could benefit all participants and future studies.
Translation to Other Settings
Most participants had access to smartphones, which is more than previously reported by the Pew Foundation finding that 53% of adults with annual incomes below $30 000 owned a smartphone.5 Still, there is an opportunity to bolster mHealth application usage among low-income populations by partnering with community organizations, hiring CHWs to facilitate engagement, and increasing hot spots to free high-speed Wi-Fi. Given widespread usage of smartphones and the efficiencies of virtual check-ins by CHWs, we believe that translation and scalability are feasible to provide increased digital connectivity in low-income populations with the goal of improving health and health outcomes.
Summary of the Experience, Future Directions, and Challenges
We found marked interest from ELP and SLP participants in using a digital health platform to monitor health and participate in research. A CHW was able to facilitate the efficient enrollment and teaching/utilization of the platform by addressing basic technical issues during enrollment and by elevating more complex questions and challenges to the digital health company (Hugo) and the research team.
Our study had some limitations. First, we only enrolled patients who had access to a smart device, and therefore, we do not know about the interest of using mHealth applications among those who do not own smart devices. Second, approximately one-fifth of participants indicated that they did not use the application by the end of the study; therefore, study findings may not be applicable to those participants or the broader population.
Several lessons around implementation emerged that can be applied to future efforts to increase the use of a digital health platform. First, we designed a preimplementation plan to optimize workflow and roles and to minimize disruption to clinician workflow and burden to patients. This process resulted in the integration of the CHW into morning clinician huddles and established trust and promoted confidence in the research team. Second, enrollment in the application via email was a significant barrier. Alternative login methods could be considered (eg, utilizing social media profiles, face ID, or fingerprint recognition), recognizing that these may also be associated with privacy concerns. Third, surveys routed through text and communication via text were the mainstay of communication, though some still preferred email, suggesting that mHealth application efforts should be attuned to participant preferences. Fourth, the consistent reliance on CHWs for help with the application and questions about research participation, especially by the SLP participants, suggests that CHWs played a vital role. Future efforts to expand mHealth to low-income populations should consider using CHWs, which may better ensure that vulnerable populations are not left behind as mHealth technologies advance.
Acknowledgments
We wish to thank the Hugo team for providing mobile health application and support, as well as the Aetna Foundation for funding the study.
Sources of Funding
This study was supported by grant 17-3735 from the Aetna Foundation—a national foundation based in Hartford, Connecticut, that supports projects to promote wellness, health, and access to high-quality health care for everyone. The views presented here are those of the authors and not necessarily those of the Aetna Foundation, its directors, officers, or staff.
Disclosures
Dr Spatz receives support from the Centers for Medicare and Medicaid Services to develop and maintain performance measures used in public reporting programs and from the Food and Drug Administration to support projects within the Yale-Mayo Clinic Center of Excellence in Regulatory Science and Innovation. She also receives support from the National Institute on Minority Health and Health Disparities (U54MD010711-01) to study precision-based approaches to diagnosing and preventing hypertension and from the National Institute of Biomedical Imaging and Bioengineering (R01 EB028106-01) to study a cuff-less blood pressure device. She is a board member of Project Access-New Haven. The other authors report no conflicts.
Footnotes
References
- 1.
Marquis-Gravel G, Roe MT, Robertson HR, Harrington RA, Pencina MJ, Berdan LG, Hammill BG, Faulkner M, Muñoz D, Fonarow GC, Nallamothu BK, Fintel DJ, Ford DE, Zhou L, Daugherty SE, Nauman E, Kraschnewski J, Ahmad FS, Benziger CP, Haynes K, Merritt JG, Metkus T, Kripalani S, Gupta K, Shah RC, McClay JC, Re RN, Geary C, Lampert BC, Bradley SM, Jain SK, Seifein H, Whittle J, Roger VL, Effron MB, Alvarado G, Goldberg YH, VanWormer JL, Girotra S, Farrehi P, McTigue KM, Rothman R, Hernandez AF, Jones WS . Rationale and design of the aspirin dosing – a patient-centric trial assessing benefits and long-term effectiveness (ADAPTABLE) trial.JAMA Cardiol. 2020; 5:598–607. doi: 10.1001/jamacardio.2020.0116CrossrefMedlineGoogle Scholar - 2.
Levy H, Janke A . Health literacy and access to care.J Health Commun. 2016; 21(suppl 1)43–50. doi: 10.1080/10810730.2015.1131776MedlineGoogle Scholar - 3.
Jhamb M, Cavanaugh KL, Bian A, Chen G, Ikizler TA, Unruh ML, Abdel-Kader K . Disparities in electronic health record patient portal use in nephrology clinics.Clin J Am Soc Nephrol. 2015; 10:2013–2022. doi: 10.2215/CJN.01640215MedlineGoogle Scholar - 4.
Roblin DW, Houston TK, Allison JJ, Joski PJ, Becker ER . Disparities in use of a personal health record in a managed care organization.J Am Med Inform Assoc. 2009; 16:683–689. doi: 10.1197/jamia.M3169MedlineGoogle Scholar - 5.
Duggan M . Cell phone activities 2013. Pew Research Center Internet & Technology; 2013. https://www.pewresearch.org/internet/2013/09/19/cell-phone-activities-2013/. Accessed April 30, 2020Google Scholar