Digital Health as an Enabler for Patient-Centric, Outcome-Based Care

Article

Considering challenges and opportunities in digital health solutions for mental health care.

sdecoret_AdobeStock

sdecoret_AdobeStock

There has been a recent revolution in the development of digital health solutions for various chronic conditions. However, there are relatively few examples of digital health solutions that have been widely adopted in real-world health care settings and are measurably improving patient outcomes or helping health care systems deliver efficient and patient-centric care. From our vantage, there are some fundamental reasons for this “translational gap,” which, if addressed, can go a long way in helping realize the full potential of digital health technologies.

We use mental health/illness as an exemplar because brain-related illnesses will be at the forefront of digital health revolution, as their course can be monitored, outcomes can be measured via digital phenotyping, and interventions can be delivered digitally and remotely. Furthermore, the challenges and opportunities in digital mental health have been further highlighted by the ongoing COVID-19 pandemic, which has added fuel to an already brewing mental health crisis.

Mental Illnesses: The Need and the Opportunity

Mental illnesses represent some of the most prevalent and disabling conditions. Depressive disorders ranked number 4 in age group 10-24 and number 6 in age group 25-49 in the list of leading global causes of level 3 disability-adjusted life-years (DALYS)1 and are associated with a wide range of negative outcomes, including loss of occupational function,2 reduced quality of life,3 and premature mortality.4 A key barrier to addressing these challenges stems from inadequate metrics to assess disease status, treatment effectiveness, and patient-relevant outcomes.

Mental illnesses do not have a clear outcome measure such as HbA1C in diabetes; instead, efficacy and outcome measures utilized in clinical practice today are complex and subjective, and are only sporadically collected or recorded in databases. As a result, assessment of quality of mental health care often defaults to the more easily measured process and structure metrics, which are distal from true patient outcomes, and, although they are easier to measure, improvement in these metrics does not always translate to improvement in patient-centric outcomes.5 This hampers outcome-relevant assessment of quality, which in turns hampers creation of value-based incentive structures that would lead to true patient-centric care and direct resources based toward the most impactful interventions (“pay for performance”).

The relative value of patient-generated digital data in driving measurement- and outcome-based care will be disproportionately high in mental health since brain-based disorders manifest with changes in behavior, cognition, and mood that are ripe for digital phenotyping. For example, recent and ongoing digital phenotyping efforts in patients with depression are demonstrating correlations between symptom severity and smartphone and wearable-derived features relevant to daily functioning such as physical activity and sleep as measured by actigraphy,6 cognition as inferred from key-strokes,7 home-stay as measured by GPS mobility,8 sociability as measured by blue-tooth handshakes,9 and anxiety or stress as detected in speech features (natural language processing, or NLP) or physiological read-outs such as Galvanic Skin Response (GSR) and Heart Rate/Heart Rate Variability (HR/HRV). Such digital readouts can be more sensitive to real-time measures of symptom severity, functional impairment, and recovery status than episodic clinical assessments. These digital measures may also serve as proxy measures to assess value and efficacy of clinical care and interventions.

Of course, payers and clinicians will require that evidence for this be generated through long-term digital phenotyping studies that relate digital biomarkers with clinical milestones and cost. This will require long-term adherence to digital phenotyping solutions from the ultimate data source: the patient. Adoption and adherence have been challenges10 that can only be addressed if, regardless of the eventual utility for clinicians and payers, digital phenotyping solutions treat the patient as their first and foremost customer.

Using Patient-Generated Data and Digital Tech for Patient-Centric Care

Utilizing digital health technologies and patient-generated data for enabling outcome-based care hinges on developing a “data partnership” with the patient that is built on principles of trust, mutual benefit, and informed data access and use. This will become more urgent as advances in wearable technology and mobile health provide patients greater control over, access to, and use of their personalized health data. In fact, eventually the boundary between health care data primarily housed in electronic health records (EHRs) and so-called patient-generated health data will blur as agency shifts to the patient and we realize that all data—whether in EHRs or generated by wearable devices—is ultimately patient-generated, and thus, is patient owned and controlled.

Matters of privacy, transparent consenting, and, more broadly, trust are clearly “table stakes” for patient participation in a digital data collection regime. Beyond that, in our experience, conducting long-term multimodal digital phenotyping studies in patients with mental illnesses, 3 factors are most important for patient acceptance, engagement, and long-term adherence. First, any sustained digital phenotyping data collection, in the context of research studies or clinical care, must provide immediate “here and now” value and utility back to the patient, over the friction and burden it creates. This equation can be helped by relying on passive data collection and developing bandwidth-frugal applications with minimal impact on smartphone battery life and other factors important to patients.

Furthermore, the end-user interface and visualizations must be engaging and useful to patients.11,12 Designing such a solution requires a systematic data-driven understanding of patients’ attitudes, motivations, and barriers toward data collection.13 It is important to follow the common tech industry practice of co-designing solutions in close partnership with end-customers (patients), with iterative learn and improve cycles. We may find ourselves surprised at how often the end-users’/patients’ views on which features are important to them diverge from experts’ preconceived notions.14

Second, altruism remains a significant motivator; most patients diagnosed with a serious condition are willing—even eager—for their data to be utilized in research that helps their fellow patients, assuming privacy and security safeguards. This propensity to be part of citizen-science efforts should be leveraged to develop “learning engines” wherein pooled, deidentified digital phenotyping data can be used to solve important problems of broad utility and impact.

Finally, we see physicians’ use of patient-generated data to improve clinical care as crucial motivator for patients to adhere to a digital data regime. In fact, this is perhaps the single most important criterion for patients with serious mental illness who are under the care of a clinician. From the physician’s perspective, the data and visualizations should make the clinic visit more efficient and should integrate readily with their current workflows. Such a system will also improve outcomes by improving patient engagement and enabling more measurement-based care,15 which is known to improve outcomes.16 We believe digital health solutions that enhance the patient-physician interface and generate near-term value for both may require more effort for initial uptake, but are most likely to stick than direct-to-patient approaches often utilized by insurance companies and digital health solution providers.

While physicians’ awareness of patient-derived digital data and overall digital platforms was already on the upswing prior to COVID, the pandemic has provided further tailwinds for such solutions.17 The initial deployment of digital phenotyping in clinical care settings will necessarily rely on validated self-reported scales of symptoms, function, and quality of life such as the disease set recommendations by the International Consortium for Health Outcomes and Measurement (ICHOM)18 to implement simple, interpretable flags and known treatment algorithms to aid clinical decision making.

Eventually, growing datasets will enable validation of digital biomarkers and predictive algorithms that are leading indicators of disease burden, health care utilization, crises, relapses, and other cost-driving events that could provide payers with real-time, leading quality metrics that can duly reward risk reduction and preemption. The Figure summarizes the framework of a “learning engine”19 digital health solution that we think can be deployed in mental health to bring together patient-derived digital data and clinical data from EHRs to deliver value to all key stakeholders: patients, physicians, and payers.

Figure. Digital Health Solution Framework for Filling the Measurement Gap to Enable Outcome-Based Care in Mental Health

Figure. Digital Health Solution Framework for Filling the Measurement Gap to Enable Outcome-Based Care in Mental Health

The previously described digital phenotyping framework could become an extension of existing EHRs and clinical decision-enabling software systems, but given the already noted challenges in integrating patient-generated health data (PGHD) into EMRs,20 we expect earlier evolution of stand-alone digital phenotyping solutions that are interoperable with multiple EMR systems.

Collection of longitudinal digital phenotyping data at scale in the real world would necessarily imply reliance on data from patient-owned consumer devices. This will require a concerted effort to develop platform-independent data standards and APIs that can ensure that digital biomarker readouts across different devices are comparable.

Increasingly, data will be analyzed and appropriate features extracted in-situ on the device, which will ease privacy concerns and reduce bandwidth requirements, both of which can severely limit scalability and deployment. We should also note that digital phenotyping solutions that collect continuous data for asynchronous use during patient-physician interaction are distinct from—and in fact, complementary to and synergistic with—synchronous telemedicine solutions that enhance physician reach and accessibility.

In addition to lack of actionable patient-generated outcomes data, true patient-centric and value-based care in mental health is also hindered due to fissures between mental and physical care (ie, a lack of a “whole person” approach). For instance, there is mounting evidence that patients suffering from mental health conditions have markedly poorer outcomes while incurring significantly higher direct medical costs for their physical comorbidities.21 However, such comorbid depression is often undiagnosed or otherwise poorly treated, if at all. Remote assessments and digital diagnostics now offer new scalable ways to detect depression and anxiety and set individuals on a path to formal diagnosis. Digital health solutions can also provide evidence-based tools for self-monitoring, self-care, and management of depression, anxiety, and stress for patients with a primary physical health diagnosis.

Furthermore, comorbid patients can be provided access to high-quality behavioral health care virtually. The virtual behavioral health team can work in partnership with primary/secondary providers treating their chronic physical comorbidities. In this context, use of physician-supervised or autonomous digital therapeutics can also be considered to further address behavioral health co-morbidities in a scalable and effective manner.

Discussion

The COVID-19 pandemic has illustrated the value of digital technology in health care and has created significant momentum that should be leveraged. Although significant advances have been made, we believe digital solutions can play an even greater role in transforming care. This is particularly true in mental health, where there are inadequate objective measures of disease status or progression. These gaps hinder effective measurement-based patient care as well as an opportunity to better align incentives and scarce resources through more meaningful outcomes-based care. We believe an appropriately designed and safeguarded digital health solution can be used to improve patient care and, at the same time, begin to address critical data gaps needed to enable patient-centric, value-based care.

Dr Narayan is head of strategy and innovation, Davos Alzheimer’s Collaborative. Mr Roy is the founder and CEO of Holmusk Technologies Inc. Dr Insel is chairman of the board at TheSteinberg Institute. Dr Eyre is lead of the Brain Capital Alliance, co-lead of the OECD Neuroscience-inspired Policy Initiative in collaboration with the PRODEO Institute, and senior fellow for Brain Capital with the Meadows Mental Health Policy Institute. Ms Kelley is director of the San Bernardino County Department of Behavioral Health. Dr Manji is visiting professor, Oxford University and Duke University, and previously global therapeutic head of neuroscience for Janssen Research & Development, LLC.

Conflicting interests: Dr Insel is an advisor to Alto Neuroscience, Cerebral, Compass Pathways, Koa Health, and Valera Health. Drs Narayan and Manji are employees of Jansen Research & Development, LLC, and may hold company stock.

References

1. GBD 2019 Diseases and Injuries Collaborators. Global burden of 369 diseases and injuries in 204 countries and territories, 1990-2019: a systematic analysis for the Global Burden of Disease Study 2019 [published correction appears in Lancet. 2020 Nov 14;396(10262):1562]. Lancet. 2020;396(10258):1204-1222.

2. Cambridge OR, Knight MJ, Mills N, Baune BT. The clinical relationship between cognitive impairment and psychosocial functioning in major depressive disorder: a systematic reviewPsychiatry Res. 2018;269:157-171.

3. Saragoussi D, Christensen MC, Hammer-Helmich L, et al. Long-term follow-up on health-related quality of life in major depressive disorder: a 2-year European cohort studyNeuropsychiatr Dis Treat. 2018;14:1339-1350.

4. Miloyan B, Fried E. A reassessment of the relationship between depression and all-cause mortality in 3,604,005 participants from 293 studiesWorld Psychiatry. 2017;16(2):219-220.

5. Kilbourne AM, Beck K, Spaeth-Rublee B, et al. Measuring and improving the quality of mental health care: a global perspectiveWorld Psychiatry. 2018;17(1):30-38.

6. Zhang Y, Folarin AA, Sun S, et al. Relationship between major depression symptom severity and sleep collected using a wristband wearable device: multicenter longitudinal observational studyJMIR Mhealth Uhealth. 2021;9(4):e24604.

7. Vesel C, Rashidisabet H, Zulueta J, et al. Effects of mood and aging on keystroke dynamics metadata and their diurnal patterns in a large open-science sample: a BiAffect iOS studyJ Am Med Inform Assoc. 2020;27(7):1007-1018.

8. Laiou P, Kaliukhovich DA, Folarin AA, 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 smartphonesJMIR Mhealth Uhealth. 2022;10(1):e28095.

9. Zhang Y, Folarin AA, Sun S, et al. Predicting depressive symptom severity through individuals' nearby Bluetooth device count data collected by mobile phones: preliminary longitudinal studyJMIR Mhealth Uhealth. 2021;9(7):e29840.

10. 25% of users abandon apps after one use. Upland Software, Inc. Accessed July 20, 2022. https://uplandsoftware.com/localytics/resources/blog/21-percent-of-users-abandon-apps-after-one-use/

11. Simblett S, Matcham F, Curtis H, et al. Patients' measurement priorities for remote measurement technologies to aid chronic health conditions: qualitative analysisJMIR Mhealth Uhealth. 2020;8(6):e15086.

12. Richardson S, Sinha A, Vahia I, et al. Brain health living labsAm J Geriatr Psychiatry. 2021;29(7):698-703.

13. Simblett S, Greer B, Matcham F, et al. Barriers to and facilitators of engagement with remote measurement technology for managing health: systematic review and content analysis of findingsJ Med Internet Res. 2018;20(7):e10480.

14. Hickie IB, Davenport TA, Burns JM, et al. Project Synergy: co-designing technology-enabled solutions for Australian mental health services reformMed J Aust. 2019;211 Suppl 7:S3-S39.

15. Hong RH, Murphy JK, Michalak EE, et al. Implementing measurement-based care for depression: practical solutions for psychiatrists and primary care physiciansNeuropsychiatr Dis Treat. 2021;17:79-90.

16. Chang D, Carlo AD, Khor S, et al. Transforming population-based depression care: a quality improvement initiative using remote, centralized care managementJ Gen Intern Med. 2021;36(2):333-340.

17. Andrews JA, Craven MP, Jamnadas-Khoda J, et al. Health care professionals' views on using remote measurement technology in managing central nervous system disorders: qualitative interview studyJ Med Internet Res. 2020;22(7):e17414.

18. Depression & anxiety. ICHOM Connect. Accessed July 20, 2022. https://connect.ichom.org/patient-centered-outcome-measures/depression-anxiety/

19. Narayan VA, Manji HK. Moving from 'diagnose and treat' to 'predict and pre-empt' in neuropsychiatric disordersNat Rev Drug Discov. 2016;15(2):71-72.

20. Cresswell K, McKinstry B, Wolters M, et al. Five key strategic priorities of integrating patient generated health data into United Kingdom electronic health recordsJ Innov Health Inform. 2019;25(4):254-259.

21. Davenport S, Gray TJ, Melek S. How do individuals with behavioral health conditions contribute to physical and total healthcare spending? Milliman, Inc. August 13, 2020. Accessed July 20, 2022. https://www.milliman.com/-/media/milliman/pdfs/articles/milliman-high-cost-patient-study-2020.ashx

Related Videos
© 2024 MJH Life Sciences

All rights reserved.