C.G. Lyketsos1, S.B. Roberts2, E.K. Swift2, A. Quina2, G. Moon2, I. Kremer3, P. Tariot4, H. Fillit5, D.E. Bovenkamp6, P.P. Zandi7, J.G. Haaga8
1. Johns Hopkins Bayview, Johns Hopkins Medicine, Baltimore, MD, USA; 2. The MITRE Corporation, McLean, VA, USA; 3. LEAD Coalition (Leaders Engaged on Alzheimer’s Disease), Washington, DC, USA; 4. Banner Alzheimer’s Institute and University of Arizona College of Medicine, Phoenix, AZ, USA; 5. Alzheimer’s Drug Discovery Foundation NY, NY, USA; 6. BrightFocus Foundation, Clarksburg, MD, USA; 7. Johns Hopkins University School of Medicine, Baltimore, MD, USA;
8. Independent consultant, Bethesda, MD, USA
Corresponding Author: Elaine K. Swift, PhD, The MITRE Corporation, McLean, VA, USA, firstname.lastname@example.org
J Prev Alz Dis 2022;
Published online May 2, 2022, http://dx.doi.org/10.14283/jpad.2022.47
Improving the prevention, detection, and treatment of Alzheimer’s disease and Alzheimer’s disease related dementias (AD/ADRD) across racial, ethnic, and other diverse populations is a national priority. To this end, this paper proposes the development of the Standard Health Record for Dementia (SHRD, pronounced “shared”) for collecting and sharing AD/ADRD real-world data (RWD). SHRD would replace the current unstandardized, fragmented, or missing state of key RWD with an open source, consensus-based, and interoperable common data standard. This paper describes how SHRD could leverage the best practices of the Minimal Common Oncology Data Elements (mCODETM) initiative to advance prevention, detection, and treatment; gain adoption by clinicians and electronic health record (EHR) vendors; and establish sustainable business and governance models. It describes a range of potential use cases to advance equity, including strengthening public health surveillance by facilitating AD/ADRD registry reporting; improving case detection and staging; and diversifying participation in clinical trials.
Key words: Alzheimer’s, electronic health records, equity, real-world data, data standards.
Alzheimer’s disease and Alzheimer’s disease related dementias (AD/ADRD) are devastating conditions resulting in progressive, chronic decline in memory, language, and other cognitive domains as well as functional decline. More than six million people in the United States currently live with AD/ADRD, a number projected to jump to 14 million by 2060. The social and emotional costs are immense, and disproportionately affect minoritized and other disadvantaged populations. Older Black adults are about twice—and older Hispanic Americans about 1.5 times—as likely to have AD/ADRD than older White adults. Genetic factors do not account for the large differences in prevalence or incidence among racial and ethnic groups. Instead, higher prevalence appears to stem from variations in medical conditions, health-related behaviors, and socioeconomic and environmental risk factors. More females than males live with AD/ADRD, but longevity does not fully explain women’s greater risk (1-3). AD/ADRD may be more prevalent among people living in rural than urban areas (4). Roughly half of the persons with Down’s Syndrome will develop AD (5).
The following areas continue to pose major challenges in health equity:
• Prevention/Risk Reduction: Decreasing age-related incidence of AD/ADRD in high income Western countries, including the United States (6), suggests that prevention/risk reduction may represent a comparative success in efforts to address the disease. Decreases appear to be related to improvements in cardiovascular risk factors (e.g., hypertension) and/or lifestyle habits (e.g., smoking, mental and social activity) (7, 8). There is an urgent need to better target prevention strategies in light of widening social and economic inequality and the threats to effective prevention posed by pandemics, environmental crises, and other developments.
• Detection/Diagnosis: Advances in biomarkers could improve clinical management of individual patients and drug trial design. While cerebrospinal fluid and imaging data probably will continue to be used, blood biomarkers will help clinicians better define patient subsets for which these more refined and resource-intensive diagnostics are indispensable. It will be essential to ensure the equitable use of blood biomarkers by populations that often have lacked access to cutting edge diagnostics (9, 10).
• Treatment/Care Management: While the drug development pipeline provides reason for cautious optimism on disease-modifying treatment, AD/ADRD care management remains vital. There are multiple drugs under investigation that target underlying AD/ADRD pathophysiology with the intent of disease modification. However, few drugs have been approved (11), and there is ongoing controversy surrounding AduhelmTM (12). Multiple similar therapeutics are in Phase 3 clinical trials, and several companies are pursuing accelerated Food and Drug Administration approval (13, 14). Here, too, ensuring equity will be essential, including greater access to clinical trials and approved drugs as well as inclusion in surveillance studies to identify differences in effectiveness and to detect adverse events across populations.
Proposed Solution: SHRD
To advance progress and equity in AD/ADRD prevention/risk reduction, detection/diagnosis, and treatment/care management, this paper calls for increased attention to the overlooked role of a single, but critical element: standardized real-world data (RWD) in electronic health records (EHRs). RWD, or data on patient health and/or health care collected outside of traditional clinical trials (15), include patient demographics and other characteristics; co-occurring physical and behavioral conditions, symptoms and other disease-related elements; pharmacology and other treatments; outcomes and measures of patient status; imaging studies, lab results, and vital signs; and genomics. RWD are collected and stored in EHRs as part of routine clinical care.
AD/ADRD RWD in EHRs often are unstandardized, incomplete, or missing (16, 17). For example, biomarker data (e.g., results from magnetic resonance imaging and positron emission tomography scans) and genomic information from large-scale gene panels and sequencing lab results (e.g., variants, fusion events, and information on clinical significance) commonly are unstandardized and/or are not available as structured elements. Physicians capture cognitive assessment supportive of diagnosis of dementia or mild cognitive impairment in clinical notes or ad hoc tools, or frequently not at all. Similarly, data critical to tracking the cognitive and functional status of persons living with AD/ADRD, as well as neuropsychiatric signs and symptoms and neurological features, rarely are captured in standardized form. Providers generally make extensive use of progress notes or other unstructured formats when recording their impressions of patient status. The same is true for non-pharmacological interventions such as diet, exercise, smoking cessation, targeted counseling, adult daycare, caregiver support groups, and home care.
Heavy reliance on unstructured text renders valuable information on patient clinical trajectories/outcomes and specific interventions difficult to extract at scale. Elements essential to developing precision medicine approaches to AD/ADRD, by defining subgroups of clinical interest, also typically are neither standardized nor available as structured data.
To retrieve, structure, and analyze unstandardized RWD, researchers increasingly turn to manual extraction, natural language processing, and machine learning. However, data capture through these approaches is generally incomplete, can be cost prohibitive, and is difficult to scale. Moreover, these approaches are usually more effective with data that are structured before they are applied (18).
In addition, a lack of standards means that other key data that one would expect to be reliably available from EHRs, such as race and ethnicity, may be missing or have limited utility. The lack of standardized key data for surveillance and other initiatives contributes to missed opportunities to reduce systemic biases that both reflect and lead to inequities in detection, diagnosis, and treatment (19, 20).
This paper proposes the Standard Health Record for Dementia (SHRD, pronounced “shared”), an open source, consensus-based standard for collecting and sharing AD/ADRD EHR-based RWD. SHRD would help transition the current unstandardized, incomplete, or missing state of EHR-based RWD to a standardized, structured solution. By leveraging the near-universal adoption of EHRs and emerging approaches to interoperability (21, 22), it could support data collection over long time periods from millions of people, including diverse populations.
SHRD would therefore complement and greatly supplement important research datasets such as the National Alzheimer’s Coordinating Center Uniform Data Set (NACC UDS), the Alzheimer’s Disease Neuroimaging Initiative (ADNI), the Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing (AIBL), and the Health and Retirement Study (HRS). Compared to the limited number of volunteers that contribute data to these initiatives, SHRD offers the potential to tap interoperable RWD drawn from the EHRs of millions of patients. SHRD would also make it possible to access data from, and therefore to deepen our understanding of, AD/ADRD in populations underrepresented in research datasets, namely racial, ethnic, and other minority populations as well as those under the age of 65 or over the age of 90 (23-26).
SHRD would be modeled after best practices and processes used by the Minimal Common Oncology Data Elements (mCODE™), an open-source standard health record for oncology (21, 22). mCODE is based on Fast Healthcare Interoperability Resources® (FHIR), an open-source standards framework created by Health Level Seven International® (HL7) that provides a common language and process for all health information technology. FHIR promotes health information exchange and data interoperability by making it easier for different systems to communicate and share information (27).
mCODE was launched in 2018 by a core group that included the American Society of Clinical Oncology (ASCO), which represents oncology professionals; MITRE, which operates a federally funded research and development center for the U.S. Department of Health and Human Services (HHS); and the Alliance for Clinical Trials in Oncology, a National Cancer Institute-sponsored clinical trial research consortium. mCODE is now being implemented by numerous leading health systems and EHR vendors (21, 22).
The SHRD Value Proposition: Impact, Feasibility, and Sustainability
While the need for standardizing EHR-based RWD is clear, a successful initiative must aim to develop more than just uniform data requirements. It must also aim for standards that are:
• Impactful, by advancing equity-oriented prevention, detection, and treatment;
• Feasible, by gaining adoption by both clinicians and EHR vendors; and
• Sustainable, by developing sound governance and business and governance models.
Lessons learned from mCODE point to strategies that SHRD could adapt to meet these criteria:
Impact: Like mCODE, SHRD should rely on a strong clinical and research equity-based value proposition through the development of standards that is driven by use cases and pilots and informed by guiding principles. mCODE use cases involve major scientific and clinical challenges that motivate researchers and/or clinicians to participate. Use cases also indicate key domains and elements that must be standardized to achieve success. Pilots involve multiple users to help refine standards and show practical impact. Both use cases and pilots serve aims of equity in cancer surveillance, treatment, and outcomes. In addition, mCODE is informed by guiding principles, including consensus-based decision making to enhance rigor and increase uptake, and expectations of a minimal set of elements to limit clinician burden and increase uptake (21, 22).
Section 4.0 describes some potential use cases to advance prevention/risk reduction, detection/diagnosis, and treatment/care management. Some, like clinical trial recruitment, are similar to those in mCODE. Others, including evaluation of biomarker performance across diverse populations, would also likely engage interest and support from researchers and clinicians and elicit wide participation.
Feasibility: mCODE offers SHRD several strategies to drive adoption by both clinicians and EHR vendors. In addition to enhancing impact, use cases help to motivate uptake by focusing on major problems that clinicians wish to solve, while pilots build collegial relationships as clinicians participate in communities of practice to address implementation challenges. In addition to reducing burden through standardizing a parsimonious set of elements, SHRD could further increase efficiency by using—to the extent possible— existing data standards for use case elements, such as those developed for the NACC UDS.
To advance AD/ADRD prevention, detection, and treatment in ways that serve diverse populations, it will be critical for clinicians across care settings to use SHRD both in—and beyond—academic health centers. As has been the case with mCODE, this can be accomplished by piloting use cases involving problems that affect patients across health care settings, such as recruiting clinical trial participants from community and primary care organizations. Over time, uptake will also spread as SHRD elements are adopted by EHRs serving a variety of health care settings.
EHR vendor uptake will be critical for SHRD, which can leverage the factors that have advanced vendor adoption of mCODE. In part, vendor support has been prompted by important rules enacted by the HHS Office of the National Coordinator for Health Information Technology (ONC) and the Centers for Medicare & Medicaid Services (CMS) requiring them to enable patient access to data by implementing FHIR-based application programming interfaces (28). Federal requirements for FHIR implementation encourage vendors to integrate FHIR-based standards such as mCODE into their source systems. Adoption has also been driven by EHRs responding to customer requests by health systems and other pilot participants to incorporate mCODE standards. In addition, specialty societies have been influential by advising vendors on product development based on mCODE (21, 22).
Sustainability: SHRD will need both a shorter- and longer-term business model to launch and grow. Here, too, mCODE provides guidance. Its initial business model was developed by ASCO, MITRE, and other collaborators. They launched mCODE with limited resources contributed by each member organization. Over time, the maintenance of the standard has been established through the HL7 FHIR AcceleratorTM program (29). CodeX is the HL7 FHIR Accelerator that administers mCODE, and that enables funding from a number of resources, including public and private organizations sponsoring or participating in pilots (21, 22).
To achieve sustainability, SHRD will also require a governance model. As with mCODE, affiliation with HL7 offers a potential avenue, with comprehensive and internationally recognized governance practices covering such key areas as standards development process, standards review, balloting, and conflict of interest.
Working closely with HL7 would also lay the basis for international application of SHRD and improved across-country interoperability as called for by the Real world Outcomes across the Alzheimer’s Disease spectrum for better care: Multi-modal data Access Platform (ROADMAP) project that aimed to further the application of real-world data across European countries, health systems, and data sources to AD/ADRD (29).
How Can We Harness SHRD to Advance Equity?
The following illustrate some general areas and possible use cases that SHRD could address to drive equity-based improvements in prevention/risk reduction, detection/diagnosis, and treatment/care management:
• Strengthening public health data for prevention/risk detection by using SHRD to facilitate AD/ADRD registry reporting. Reporting AD/ADRD patient data to registries is currently a forms-based process that can take months to provide the data to intended recipients. Easier registry reporting using SHRD would enable low burden, standardized, automated, protected reporting of AD/ADRD data from AD/ADRD research centers and other providers to registries, which would provide key information across diverse populations for research and public health purposes. The plan would be to leverage the SHRD data standard with additional strategically selected, registry-specific data element extensions. This solution could meet the needs of a wide variety of registries, including those focused on standards of care, public health, clinical trials, post-market surveillance, quality reporting, and comparative effectiveness.
• Improving case detection/diagnosis by structuring data from standardized cognitive, functional, and neuropsychiatric assessments, biomarkers, and other key inputs. Early identification of persons living with AD/ADRD is key for timely referral to specialists, optimal selection of pharmacologic and non-pharmacologic interventions, and disparity reduction. Currently, clinicians often capture cognitive diagnoses such as dementia or mild cognitive impairment in different ways, including in unstructured clinical notes or possibly billing codes, or perhaps not at all, even when they go on to prescribe specific medications relevant to dementia. Improved case detection via SHRD would allow multiple kinds of studies based on EHR-based RWD, anchored in standardized diagnoses. Moreover, SHRD would allow more accurate staging of AD/ADRD by facilitating the use of standard, validated, widely endorsed cognitive, functional, and neuropsychiatric assessments.
• Ensuring that cutting edge clinical trial research benefits diverse populations by using SHRD to make EHR-based RWD more integral to clinical trial-related issues such as referral, safety monitoring, and real-world efficacy. Currently, AD/ADRD clinical trials commonly use separate and expensive manual processes to collect data. One potential use case for SHRD would be to demonstrate that EHRs can provide high quality data less expensively and more efficiently. This use case would make it easier for clinical trial researchers to analyze standardized RWD from diverse populations to broaden their understanding of the safety, effectiveness, and value of new treatments.
• Diversifying participants in AD/ADRD clinical trials by developing SHRD-based open data standards and open application programming interfaces as is now being piloted through mCODE for oncology patients. These interfaces enable interoperable, scalable, and accessible “coarse” clinical trial matching services that are integrated into existing clinical workflow, by using the data standards to identify automatically clinical trials for which a given patient may qualify. This makes it easier for all patients—as well as care teams in settings serving economically disadvantaged, rural, and other underserved populations—to identify relevant clinical trials offering promising treatments in accessible locations. This in turn supports efforts to diversify clinical trial participants.
• Improving treatment quality and care management for diverse populations by making it easier for clinicians to seamlessly share complex, nonpharmacologic intervention summary information (e.g., support groups for caregivers, home care), including end-of-treatment and in-progress summary data, across health systems and specialty vendor systems. Clinicians could use these data to make informed decisions related to a patient’s health. Researchers could use them to analyze comparative effectiveness research, improve patient safety, and develop effective interventions.
• Monitoring the safety, efficacy, and value of approved therapies and care interventions across diverse populations in real time by using SHRD to make it easier to disseminate high-quality, reliable EHR-based RWD and conduct comprehensive post-market surveillance. Larger and more diverse populations than the populations typically participating in clinical trials will access approved AD/ADRD therapies, such as Aduhelm, and care interventions. These populations may experience different kinds, or increased numbers, of adverse events or clinical benefits. SHRD standards in EHRs would speed detection of post-market issues and provide stronger evidence for changes in labeling or additional required studies, along with adaptations in care interventions. Clinical trials using SHRD could help to support more effective post-market surveillance by providing data structured for ready comparison to post-market outcomes. Taken cumulatively, these post-market data would provide a basis for more discreet and precise value assessments through lenses including health equity, direct or indirect impact on co-occurring conditions and associated health care utilization, and health status of care partners.
RWD represent important and underutilized evidence for accelerating AD/ADRD research and health equity and improving clinical care and community-based supports for diverse populations. Enhancing EHR-based RWD via the SHRD project would represent a wise investment for AD/ADRD research. In addition to enhancing the value of the nation’s considerable investment in EHRs and interoperability, SHRD could unlock invaluable equity-centered evidence for advancing AD/ADRD. It could also improve assessment of the effectiveness, safety, and value of new AD/ADRD treatments and care interventions. Such evidence is crucial in international efforts to combat AD/ADRD and in informing questions on complex payment policy that will continue to arise with continued progress in preventing, detecting, and treating this disease across the many populations it affects deeply.
Funding: Constantine G. Lyketsos MD received support from the Richman Family Precision Medicine Center of Excellence in Alzheimer’s Disease at Johns Hopkins. The work of other authors was not supported by any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Acknowledgments: The authors thank Heather M. Snyder PhD of the Alzheimer’s Association and Steve Bratt PhD of the MITRE Corporation for their comments.
Conflict of interest: The authors declare no conflicts.
1. Alzheimer’s Association. Alzheimer’s Disease Facts and Figures. 2021. alzheimers-facts-and-figures.pdf. Accessed 22 Dec 2021.
2. Olivari BS, French ME, McGuire LC. The Public Health Road Map to Respond to the Growing Dementia Crisis. Innovation in Aging. 2020: v4(1). igz043. https://doi.org/10.1093/geroni/igz043.
3. National Institute on Aging. Alzheimer’s Disease and Related Dementias. Available at Alzheimer’s Disease and Related Dementias | National Institute on Aging (nih.gov). Accessed 21 Feb 2022.
4. Cross SH, Warraich HJ. Rural-Urban Disparities in Mortality from Alzheimer’s and Related Dementias in the United States, 199-2018. JAGS 2020; 69(4): 1095-1096. https://doi.org/10.1111/jgs.16996.
5. National Institute on Aging. Alzheimer’s Disease in People with Down’s Syndrome. Available at Alzheimer’s Disease in People With Down Syndrome | National Institute on Aging (nih.gov). Accessed 4 March 2022.
6. Serrano-Pozo A, Growdon JH. Is Alzheimer’s Disease Risk Modifiable? J Alzheimers Dis. 2019; 67(3): 795–819. Doi: 10.3233/JAD181028.
7. Livingston G, Huntley J, Sommerlad A et al. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet 2020; 396 (10248):413-446. doi: 10.1016/S0140-6736(20)30367-6.
8. Crous-Bou M, Minguillón C, Gramunt N, Molinuevo JN. Alzheimer’s disease prevention: from risk factors to early intervention. Alz Res Therapy 2017; 9: Article 71. https://doi.org/10.1186/s13195-017-0297-z.
9. Mattsson-Carlgren N, Palmqvist S, Blennow K, Hansson O. Increasing the reproducibility of fluid biomarker studies in neurodegenerative studies. Nature communications 2020; 11(1): 6252. https://doi.org/10.1038/s41467-020-19957-6.
10. Meldolesi J. News about the Role of Fluid and Imaging Biomarkers in Neurodegenerative Diseases. Biomedicines. 2021;9(3):252. https://doi.org/10.3390/biomedicines9030252.
11. Cummings J, Lee G, Zhong K, et al. Alzheimer’s disease drug development pipeline: 2021. Alzheimer’s Dement. 2021; 7:e12179. https://doi.org/10.1002/trc2.12179.
12. Doody R. Consequences of the FDA Decision on Aducanumab for Patient Care and Research. 2021. Consequences of the FDA Decision on Aducanumab for Patient Care and Research. J Prev Alzheimers Dis 8, 393–395 (2021). https://doi.org/10.14283/jpad.2021.43.
13. Biogen. Alzheimer’s disease. 2021. https://www.biogen.com/en_us/alzheimers-disease.html. Accessed 19 Dec 2021.
14. Eli Lilly to seek FDA approval for potential Alzheimer’s drug. CNBC. https://www.cnbc.com/2021/06/24/eli-lilly-to-seek-fda-approval-for-potential-alzheimers-drug-.html. Published June 24, 2021. Accessed 19 Dec 2021.
15. Office of the Commissioner. Real-world evidence. U.S. Food and Drug Administration. https://www.fda.gov/science-research/science-and-research-special-topics/real-world-evidence. Published July 16, 2021. Accessed 18 Dec 2021.
16. Maserejian N, Krzywy H, Eaton S, Galvin JE. Cognitive measures lacking in EHR prior to dementia or Alzheimer’s disease diagnosis. Alzheimers Dement. 2021 Jul;17(7):1231-1243. doi: 10.1002/alz.12280.
17. Barnes DE, Zhou J, Walker RL et al. “Development and Validation of eRADAR: A Tool using EHR Data to Detect Unrecognized Dementia.” J Am Geriatr Soc. 2020 Jan; 68(1):103–111. https://doi.org/10.1111/jgs.16182.
18. Henry J, Pylypchuk Y, Searcy T, Patel V. Adoption of Electronic Health Record Systems among U.S. Non-Federal Acute Care Hospitals: 2008-2015. ONC Data Brief 2016:35. https://dashboard.healthit.gov/evaluations/data-briefs/non-federal-acute-care-hospital-ehr-adoption-2008-2015.php. Accessed 20 Dec 2021.
19. Polubriaginof FCG, Ryan P, Salmasian H et al. Challenges with quality of race and ethnicity data in observational databases. Journal of the American Medical Informatics Association August/September 2019; 26 (8-9):730 736. https://doi.org/10.1093/jamia/ocz113.
20. Agency for Healthcare Research and Quality. Race, Ethnicity, and Language Data: Standardization for Healthcare Quality Improvement. Available at 5. Improving Data Collection Across the Health Care System. Agency for Healthcare Research and Quality (ahrq.gov). Accessed 21 Feb 2022.
21. Osterman TJ, Terry M, Miller RS. Improving Cancer Data Interoperability: The Promise of the Minimal Common Oncology Data Elements (mCODE) Initiative. JCO Clin Cancer Inform. 2020;4:993-1001. https://doi.org/10.1200/CCI.20.00059.
22. mCODETM. Available at mCODE – CodeX – Confluence (hl7.org). Accessed 21 Feb 2022.
23. National Alzheimer’s Coordinating Center Uniform Data Set. Available at Uniform Data Set version 3 | National Alzheimer’s Coordinating Center (naccdata.org). Accessed 01 March 2022.
24. Alzheimer’s Disease Neuroimaging Initiative. About. Available at ADNI | About (usc.edu). Accessed 01 March 2022.
25. Australian Imaging, Biomarker & Lifestyle Flagship Study of Ageing (AIBL). About. Available at About – AIBLAIBL (csiro.au). Accessed 04 March 2022.
26. Health and Retirement Study (HRS). About. Available at About | Health and Retirement Study (umich.edu). Accessed 04 March 2022.
27. FHIR overview – v4.0.1. HL7 FHIR. 2019. https://www.hl7.org/fhir/overview.html. Accessed 16 Dec 2021.
28. Brooks-LaSure, Chiquita. Interoperability and the Connected Health System. CMS.gov Blog, Dec. 08, 2021. Available at Interoperability and the Connected Health Care System | CMS. Accessed 21 Feb 2022.
29. HL7 FHIR Accelerator Program. See HL7 FHIR Accelerator™ Program | HL7 International. Accessed 4 March 2022.
30. Real world Outcomes across the Alzheimer’s Disease spectrum for better care: Multi-modal data Access Platform (ROADMAP) project. See ROADMAP | ROADMAP Alzheimer (roadmap-alzheimer.org). Accessed 22 Dec 2021.