jpad journal

AND option

OR option

MAIL AND TELEPHONE OUTREACH FROM ELECTRONIC HEALTH RECORDS FOR RESEARCH PARTICIPATION ON COGNITIVE HEALTH AND AGING

 

K. Pun1, C.W. Zhu1,2, M.T. Kinsella1, M. Sewell1, H. Grossman1,2, J. Neugroschl1, C. Li1, A. Ardolino1, N. Velasco1, M. Sano1,2

 

1. Icahn School of Medicine at Mount Sinai, New York, NY, USA; 2. James J. Peters VA Medical Center, Bronx, NY, USA

Corresponding Authors: Carolyn W. Zhu, PhD, Department of Geriatrics & Palliative Medicine, Icahn School of Medicine at Mount Sinai and JJP VA Medical Center, 130 West Kingsbridge Road, Bronx NY 10468, USA. Email: carolyn.zhu@mssm.edu, Telephone: 718-584-9000 ext. 6146, Fax: 718-741-4211.

J Prev Alz Dis 2021;
Published online April 24, 2021, http://dx.doi.org/10.14283/jpad.2021.18

 


Abstract

Objectives: This report describes the efficacy and utility of recruiting older individuals by mail to participate in research on cognitive health and aging using Electronic Health Records (EHR).
Methods: Individuals age 65 or older identified by EHR in the Mount Sinai Health System as likely to have Mild Cognitive Impairment (MCI) were sent a general recruitment letter (N=12,951). A comparison group of individuals with comparable age and matched for gender also received the letter (N=3,001).
Results: Of the 15,952 individuals who received the mailing, 953 (6.0%) responded. 215 (1.3%) declined further contact. Overall rate of expression of interest was 4.6%. Of the 738 individuals who responded positively to further contact, 321 indicated preference for further contact by telephone. Follow-up of these individuals yielded 30 enrollments (0.2% of 15,952). No differences in response rate were noted between MCI and comparison groups, but the comparison group yielded higher enrollment. 6 individuals who were not the intended recipients of mailing but nevertheless contacted our study were also enrolled.
Conclusions: Mailings to individuals identified through a trusted source, such as a medical center from which they have received clinical care, may be a viable means of reaching individuals within this age group as this effort yielded a low rejection rate. However, EHR information did not enhance study enrollment. Implications for improving recruitment are discussed.

Key words: Recruitment methods, electronic health records, cognitive health, mild cognitive impairment.


 

Introduction

It is increasingly recognized that the pathophysiological process of Alzheimer’s disease (AD) begins years and maybe decades prior to the onset of clinical symptoms (1-7). Over the past several decades, both pharmacological and non-pharmacological lifestyle interventions have been studied for the prevention of cognitive decline or dementia in older adults with or without risk factors for AD. While important innovations in ongoing trials include identification of novel targets, development of multidomain interventions, identification and validation of biomarker or genetic targets, and improving outcome measures, the biggest challenge remains the recruitment of participants, espcially for long studies of non or minimally affected individuals (8-11).
One of the critical components for the success of these studies is identifying and recruiting individuals at high risk of developing dementia, both for observational studies investigating the natural history of prodromal and early disease stages and for interventional studies aimed at disease prevention or modification. Individuals with Mild Cognitive Impairment (MCI) have an increased risk of developing dementia compared with their cognitively normal peers (12). However, outreach to older adults for studies in these areas is often difficult because disease may be undetected in its mildest forms and awareness of future problems may be low. Studies have used diagnostic codes in administrative data or medical records to identify cases, however accuracy of diagnostic codes for cognitive impairment is limited (13-16). Identification of non-demented individuals with MCI from electronic health records (EHR) can be challenging, often depending on unstructured text for detection but several algorithms have been reported for such case ascertainment (17-19). Machine learning algorithms that have been developed to increase precision of identification have yet to be used in outreach efforts (20).
Campaigns to improve outreach and recruitment have often used mass mailings, defined as letters without a specific addressee, as they permit an inexpensive way to reach large numbers of potential participants and do not require technologies that may be less available to older cohorts. In general, rates of recruitment by this method can be low but the large numbers reached can allow studies to achieve needed sample sizes. For example, the Lifestyle Interventions and Independence for Elders (LIFE) Study reported that directly mailing a study brochure to households with age-eligible residents obtained from commercial databases and voter registration lists yielded 59.5% of randomized cases (21). However, the number of contacts needed to achieve this recruitment is quite high.
We made the assumption that trusted sources such as a medical center from which individuals have received clinical care may increase their interest in research participation and improve response to outreach efforts. The Alzheimer’s Disease Research Center (ADRC) at the Icahn School of Medicine at Mount Sinai has been continuously funded by the National Institutes of Health and has more than 30 years of experience that is highly recognized in the community. The ADRC offers ongoing opportunities to participate in a variety of clinical studies ranging from observational studies with and without imaging to pharmacological and non-pharmacological intervention studies. Building on that reputation, we undertook a focused mailing outreach effort using EHR, to engage individuals in research on cognitive health and aging. This report describes our experience and evaluates the efficacy of implementing a mass mailing within a trusted healthcare system to a cohort likely to have MCI based on their EHR.

 

Methods

Using EHR to Identify MCI and Comparison Groups

The EHR was used to create a group of individuals likely to have MCI. The MCI group included those age 55 to 90 who received care from the Mount Sinai Health System between spring 2013 and 2018. Inclusion criteria were: presence of an International Classification of Diseases, Ninth Revision (ICD-9) diagnosis code consistent with memory loss (MCI, memory loss, or dementia), or a similar Tenth Revision (ICD-10) diagnosis code of MCI or other amnesia. Exclusion criteria included presence of Alzheimer’s disease, Parkinson’s disease, schizophrenia, Huntington’s Chorea, epilepsy, multiple sclerosis, substance use disorders, tobacco use disorders, bipolar and major depressive disorders, and use of anti-dementia medications. Full inclusion/exclusion criteria are available upon request.
To evaluate whether the MCI group provided a recruitment benefit over that from an unselected comparison group, a comparison group was selected from individuals age 65 years or older, seen in the same health care system over the same time interval. They were further matched by sex to the MCI group. Those with a dementia diagnosis were excluded. The sample size of the comparison group was approximately 20% of the MCI group.
The ADRC staff were blinded to group status when contacting interested individuals, but group status was available for analysis at the end of recruitment efforts.

Mailing Process

A letter printed on institutional letterhead without specific salutation was sent to all individuals in the MCI and comparison groups. The letter 1) acknowledged the individual’s connection to the health system; 2) invited the individual to learn more about research in cognitive health and aging; 3) provided options to contact the ADRC via telephone, mail, or email; 4) informed the individual that they could “opt out” of further contact; and 5) informed the individual that the ADRC may reach out to them in the future if interest in further contact was expressed.
The letter included a returnable mailing slip on which individuals could confirm or deny interest in further contact by the ADRC and specify their preferred method of contact. A prepaid envelope to return the mailing slip was also included. Mailing materials are shown in Supplemental Materials Figure 1.

Response to Mailing and Expressions of Interest

Individuals who received the mailing were able to actively express interest in learning more about research programs at the ADRC either by 1) contacting the ADRC directly by phone; 2) contacting the ADRC directly by email; or 3) returning the included mailing slip, denoting interest for further contact, and specifying interest in further contact by telephone, email, or mail. A small number of individuals contacted us directly by email, all of whom also contacted us by telephone and expressed interest in further contact by telephone. Because the majority of responses across all contact methods preferred telephone, the current analysis describes our outreach efforts to those individuals regardless of initial response method. These included individuals who called the ADRC directly as well as those who emailed us directly or returned the mailing slip and expressed interest in further contact by telephone.

Recruitment Efforts

ADRC staff proceeded to call all individuals who expressed interest in further contact by telephone. Interested individuals were offered the opportunity to participate in observational and interventional studies available at the ADRC at the time of contact. ADRC staff explained study details and then offered to schedule initial screening appointments for all individuals who remained interested.

Estimating Staff Time and Effort

ADRC staff attempted to contact interested individuals by telephone using the following protocol: When individuals answered calls or responded to voice messages left by ADRC staff, calls continued until a decision regarding research participation was reached. If there was no response to a voice message within two weeks or it was not possible to leave a message, staff made up to two additional follow-up calls. A total of three call attempts were made to minimally or non-responsive individuals, with attempts made to vary the day and time of the call. Staff time and effort were estimated as follows: scheduling visits: 20 minutes, handling requests for more information or time: 15 minutes, determining ineligibility: 15 minutes, determining not interested: 10 minutes, and leaving a voice message: 5 minutes.

 

Results

Response to Mailing and Expression of Interest

Of the 15,952 individuals who were sent the mailing, 114 contacted the ADRC by telephone directly and expressed interest, 839 sent returnable mailing slips, and the remaining 14,999 did not respond (Figure 1). Of the 839 who returned mailing slips, 624 (3.9% of all individuals who were sent the mailing) expressed interest in further contact, while 215 (1.3%) declined further contact. Among the 624 mailing slip respondents who expressed interest in further contact, 207 (1.3% of all individuals who were sent the mailing) indicated preference for further contact by telephone, 240 (1.5%) by email, and 177 (1.1%) by mail. Taken together, overall rate of expression of interest from the mailing was 4.6%.

Figure 1. Response to Mailing and Expression of Interest

 

This analysis focused on the 321 individuals who expressed interest in further contact by telephone (i.e., 2% of the entire mailing). Therefore, the 417 individuals who expressed interest in further contact through mail and email are not described in this report, and follow up on this cohort was left for future efforts.

Recruitment Outcomes

Of these 321 individuals who expressed interest for further contact from the ADRC by telephone, 30 (9.3%) enrolled in a study, 57 (17.8%) were not eligible due to medical comorbidities or study contraindications, 82 (25.6%) were not interested in research participation after speaking with ADRC staff, 137 (42.7%) did not definitively respond after at least three contact attempts, and 15 (4.7%) were not eligible for currently enrolling studies but remained interested in future participation (Table 1).

Table 1. Recruitment Outcomes for Individuals who Expressed Interest in Further Contact by Telephone (N=321)

Study Participation by MCI and Comparison Group

Among the 321 individuals who expressed interest in further contact by telephone, 227 (70.7%) individuals were from the MCI group (Table 2). This represents 1.8% of the 12,951 individuals from the MCI group. 53 of the 321 individuals (16.5%) were from the comparison group, representing 1.8% of the 3,001 individuals in the comparison group. An additional 41 (12.8%) individuals were incidental contacts who were not the intended recipients of the mailing but nevertheless contacted the ADRC. Group status for these individuals is by definition unknown.

Table 2. MCI and Comparison Group Status of Individuals who Expressed Interest and Individuals who Enrolled

 

Of the individuals who enrolled in studies, 15 of 30 were from the MCI group, representing a 0.1% enrollment rate from the total MCI group, while 9 of 30 were from the comparison group, representing a 0.3% enrollment rate from the total comparison group. A two-sample test of proportions shows that the difference in enrollment rate between MCI and comparison groups is statistically significant (p=0.019). An additional 6 of the 30 enrolled individuals were incidental contacts whose group status was unknown.

Enrollment by Study Type

Of the 30 individuals who enrolled, 7 (23.3%) enrolled in observational studies which did not include imaging, 21 (70.0%) enrolled in observational studies which included imaging, and 2 (6.7%) enrolled in an intervention trial (Table 3). Of the 7 individuals enrolled in observational studies without imaging, 5 (71.4%) were from the MCI group and 2 (28.6%) were from the comparison group. Among the 21 individuals who enrolled in observational studies with imaging, 9 (42.9%) were from the MCI group, 6 (28.6%) were from the comparison group, and 6 (28.6%) were from the incidental contact group. Finally, of the 2 individuals who enrolled in interventional trials, 1 (50.0%) was from the MCI group and 1 (50.0%) was from the comparison group. Detailed descriptions of these studies are included in Supplemental Materials Table 1.

Table 3. Enrollment by Study Type

Staff Time and Effort

ADRC staff logged 463 calls and spent an estimated total of 87 hours communicating with the 321 individuals who expressed interest in further contact with the ADRC by telephone (Table 4). Distribution of call logs and estimated time spent by recruitment outcome are also reported. On average, three hours (177 minutes) of staff time were required to enroll one participant. These estimates are limited to communication with individuals by telephone and offer a conservative summary of the total time and effort dedicated to this recruitment effort.

Table 4. Staff Time and Effort Required to Contact Individuals who Expressed Interest for Further Contact by Telephone

 

Discussion

In this report we described our experience with a mailing outreach effort to engage individuals in research on cognitive health and aging. Individuals who had recent contact with the medical center were identified through the Mount Sinai Health System’s Electronic Health Record. Despite recent contact with the medical center and proximity to the site (over 95% of mailing addresses were within New York, New Jersey, and Connecticut), overall rate of expression of interest was approximately 5%, which is somewhat higher than comparable efforts of targeted mailing in similar age groups. For example, the Medical, Epidemiologic and Social Aspects of Aging (MESA) Study used a commercial mailing list to recruit women age 55-80 for a trial of behavioral techniques for the prevention of urinary incontinence that reached over 48,000 individuals and reported a 3.3% initial positive response rate (22). Efforts to recruit for dementia prevention trials using Medicare beneficiary lists and voter registration polls reported response rates between 0.4 to 2.0%.23 In the AD Anti-inflammatory Prevention Trial (ADAPT), over 3.5 million mailings were sent to Medicare beneficiaries 70 years or older in specified zip codes. The trial enrolled 2,518 volunteers at 6 sites over 44 months. Across trial sites, enrollment ranged from 0.4 to 1.9 participants per 1,000 mailings (24, 25). Unlike our study, this report included calls made to all individuals who received the mailing and did not opt out. The Ginkgo Evaluation of Memory (GEM) study to evaluate ginkgo biloba to prevent dementia mailed brochures to approximately 243,000 individuals age 75 and over who were assumed to be dementia free from purchased lists, voter registration lists, and university lists. Using telephone follow up to those who received the mailing, the study team attempted to reach 14,603 mail recipients, reached 12,186 and enrolled 1.3% of those who received the initial mailing (26).

Higher response rates have been reported in the literature when mailings came directly from primary care providers and when research activity was home-based and required no travel. For example, the Screening, Recruitment, and Baseline Characteristics for the Strategies to Reduce Injuries and Develop Confidence in Elders (STRIDE) Study for fall prevention found that approximately 12% of mailings resulted in expressions of interest (27). Higher response rates have been noted outside the US. Andersen et al mailed a questionnaire examining self-reported cognitive function to more than 11,807 community residents in Norway. 3,767 (31%) responded to the mailing. Of these, 438 met criteria for cognitive impairment and 292 were willing to undergo clinical evaluation (28). Notably, the cohort was invited to join a study for symptomatic individuals which may be of greater interest than studies of disease prevention.
In our study, about one third of the individuals who responded to the mailing contacted us by telephone. Of note, while 1.5% (240) used the returnable mailing slip and indicated preference for future contact by email, those who contacted the ADRC directly by email also contacted us by telephone, which may indicate less confidence in initiating email connections in this age group. Those who used the returnable mailing slip to provide telephone contact were nearly twice as likely to be lost to follow up relative to individuals who expressed interest by telephone directly. Future work might attempt to determine if requiring telephone follow-up would identify a more specifically interested group. Optimizing approaches to identifying sufficient numbers of interested individuals is critical to efficient and cost effective outreach.
Contrary to our expectations, we found greater interest and higher enrollment among those in the comparison group than in the MCI group. Of note, among those who enrolled, the MCI group was more likely to participate in observational studies and less likely to enroll in imaging or interventional studies. There are several possible reasons for this difference. Cognitive impairment and dementia in health records may be associated with other serious health problems (29). These health problems may be more prominent than cognitive impairment for these individuals and may reduce their interest in research on this topic. Meanwhile, the comparison group may be less likely to suffer from comorbid conditions and may be motivated by interest in prevention or protection against cognitive decline. Only 2 participants enrolled in clinical trials, highlighting challenges in recruitment to prodromal AD drug trials. Given these low frequencies, results from this study should be replicated. Understanding these themes will be important topics for future research.
While this mailing did not specifically invite non-recipients to join, the recruitment effort was prepared to accept them. We identified several individuals who described the mailing but were unsure of the source from which it came. This incidental interest is obviously a positive outcome for improving recruitment. Some mailing efforts encourage anyone to reply and this approach may be worthy of consideration. In particular, these contacts may be a source of “high interest” individuals, and efforts to identify features to improve outreach to them could provide an important contribution to recruitment science.
Several studies report the use of an opt-out approach to recruitment. In general those who opt-out prior to contact are low but often not reported. In our study, very few mailing recipients refused further contact (<2% of total number of mailings). Others have reported similarly low rates (23). The opt-out option can be executed in different ways including the “pre-mailing” requirement or a delay in outreach before initiating contact to those who received mail (23). These options require significant expense and time delay. In our study we reached out only to those who contacted us. Without an opt-out option, when unsolicited telephone calls followed the initial mailing, rate of no-interest is high, with reports of over 50% in several studies (24-26). The discordance between the low opt-out rate and the high no-interest rate reported suggest that opt-out resources may be effective in allowing studies to call more people. However, the resources and expenses needed to call those who do not demonstrate any initial interest are high. Future work may focus on identifying criteria for more efficient recruitment of individuals who do not opt out.
There are several limitations to fully appreciate this report in the context of recruitment to studies of cognitive health and aging. While there is growing use of algorithms to identify undetected dementia using EHR, little work has been done on identifying MCI (30, 31). Accuracy of diagnostic codes for dementia and cognitive impairment in medical records is limited (13-15). The study aimed to target a wide group of potential participants, however, it is possible that suitable participants were excluded if they were prescribed cholinesterase inhibitors for their MCI but did not have a formal diagnosis of MCI recorded in their medical records. Furthermore, our algorithm was based on EHR data collected across a five year period, and key patient information may not be the most up to date. We also did not have important variables such as time since diagnosis to examine their usefulness as potential predictors of response. We have little data on the accuracy of our mailing as “return to sender” information was not available to us. Recruitment efforts served multiple studies which may not have been open throughout the entire outreach period. However, the center had a variety of observational and interventional studies open at any time, and the option to be contacted for future studies was always provided. Ongoing efforts include study-specific mailing outreach to individuals who did not decline further contact, including those who expressed their preference for future contact by mail or email. Finally, while response to this mailing effort was slightly higher than expected, our experience is limited to a single site.

 

Conclusions

Focused mailing outreach efforts continue to represent a valuable means of engaging older adults in research on cognitive health and aging. However, using EHR to identify individuals who likely have cognitive impairment did not appear to increase response or participation rates. As demonstrated among the incidental contacts, a written letter has the potential to spark interest beyond its intended recipients. Mailings from known healthcare systems build upon an established relationship and may foster the development of a preliminarily defined, trial-ready cohort.

 

Acknowledgments and funding: Funding for this initiative was provided by the Alzheimer’s Disease Research Center (ADRC) at the Icahn School of Medicine at Mount Sinai (P50AG005138), the Alzheimer’s Therapeutic Research Institute (R01AG047992), and the National Center for Advancing Translation Sciences (UL1TR001433). Drs. Sano, Grossman, and Zhu also are supported by the Department of Veterans Affairs, Veterans Health Administration. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. The sponsors had no role in the design and conduct of the study; in the collection, analysis, and interpretation of data; in the preparation of the manuscript; or in the review or approval of the manuscript.

Conflict of Interest: No conflict of interest has been declared by the authors.

Ethics standards: The study was approved by the Icahn School of Medicine Institutional Review Board (IRB). All participants gave informed consent before participating.

 

SUPPLEMENTARY MATERIAL

 

References

1. Schneider LS. Recruitment methods for United States Alzheimer disease prevention trials. J Nutr Health Aging. 2012;16(4):331-335.
2. Sperling R, Mormino E, Johnson K. The evolution of preclinical Alzheimer’s disease: implications for prevention trials. Neuron. 2014;84(3):608-622.
3. Vellas B. Update on Prevention Trials in Alzheimer’s Disease. J Prev Alzheimers Dis. 2014;1(3):168-175.
4. Weninger S, Carrillo MC, Dunn B, et al. Collaboration for Alzheimer’s Prevention: Principles to guide data and sample sharing in preclinical Alzheimer’s disease trials. Alzheimers Dement. 2016;12(5):631-632.
5. Alber J, Lee AKW, Menard W, Monast D, Salloway SP. Recruitment of At-Risk Participants for Clinical Trials: A Major Paradigm Shift for Alzheimer’s Disease Prevention. J Prev Alzheimers Dis. 2017;4(4):213-214.
6. Brookmeyer R, Abdalla N. Design and sample size considerations for Alzheimer’s disease prevention trials using multistate models. Clin Trials. 2019;16(2):111-119.
7. Rafii MS, Aisen PS. Alzheimer’s Disease Clinical Trials: Moving Toward Successful Prevention. CNS Drugs. 2019;33(2):99-106.
8. Galvin JE. Prevention of Alzheimer’s Disease: Lessons Learned and Applied. J Am Geriatr Soc. 2017;65(10):2128-2133.
9. 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.
10. Wang J, Tan L, Yu JT. Prevention Trials in Alzheimer’s Disease: Current Status and Future Perspectives. J Alzheimers Dis. 2016;50(4):927-945.
11. Vellas B, Bain LJ, Touchon J, Aisen PS. Advancing Alzheimer’s Disease Treatment: Lessons from CTAD 2018. J Prev Alzheimers Dis. 2019;6(3):198-203.
12. Petersen RC, Morris JC. Mild cognitive impairment as a clinical entity and treatment target. Arch Neurol. 2005;62(7):1160-1163; discussion 1167.
13. Albrecht JS, Hanna M, Kim D, Perfetto EM. Predicting Diagnosis of Alzheimer’s Disease and Related Dementias Using Administrative Claims. J Manag Care Spec Pharm. 2018;24(11):1138-1145.
14. Gianattasio KZ, Prather C, Glymour MM, Ciarleglio A, Power MC. Racial disparities and temporal trends in dementia misdiagnosis risk in the United States. Alzheimers Dement (N Y). 2019;5:891-898.
15. Zhu CW, Ornstein KA, Cosentino S, Gu Y, Andrews H, Stern Y. Misidentification of Dementia in Medicare Claims and Related Costs. J Am Geriatr Soc. 2019;67(2):269-276.
16. Dell’Agnello G, Desai U, Kirson NY, et al. Reliability of coded data to identify earliest indications of cognitive decline, cognitive evaluation and Alzheimer’s disease diagnosis: a pilot study in England. BMJ Open. 2018;8(3):e019684.
17. Gilmore-Bykovskyi AL, Block LM, Walljasper L, Hill N, Gleason C, Shah MN. Unstructured clinical documentation reflecting cognitive and behavioral dysfunction: toward an EHR-based phenotype for cognitive impairment. J Am Med Inform Assoc. 2018;25(9):1206-1212.
18. Graham SA, Lee EE, Jeste DV, et al. Artificial intelligence approaches to predicting and detecting cognitive decline in older adults: A conceptual review. Psychiatry Res. 2020;284:112732.
19. Amra S, O’Horo JC, Singh TD, et al. Derivation and validation of the automated search algorithms to identify cognitive impairment and dementia in electronic health records. J Crit Care. 2017;37:202-205.
20. Uspenskaya-Cadoz O, Alamuri C, Wang L, et al. Machine Learning Algorithm Helps Identify Non-Diagnosed Prodromal Alzheimer’s Disease Patients in the General Population. J Prev Alzheimers Dis. 2019;6(3):185-191.
21. Marsh AP, Lovato LC, Glynn NW, et al. Lifestyle interventions and independence for elders study: recruitment and baseline characteristics. J Gerontol A Biol Sci Med Sci. 2013;68(12):1549-1558.
22. Messer KL, Herzog AR, Seng JS, et al. Evaluation of a mass mailing recruitment strategy to obtain a community sample of women for a clinical trial of an incontinence prevention intervention. Int Urol Nephrol. 2006;38(2):255-261.
23. Sano M, Jacobs D, Andrews H, et al. A multi-center, randomized, double blind placebo-controlled trial of estrogens to prevent Alzheimer’s disease and loss of memory in women: design and baseline characteristics. Clin Trials. 2008;5(5):523-533.
24. Adapt Research Group, Meinert CL, McCaffrey LD, Breitner JC. Alzheimer’s Disease Anti-inflammatory Prevention Trial: design, methods, and baseline results. Alzheimers Dement. 2009;5(2):93-104.
25. Meinert CL, Breitner JC. Chronic disease long-term drug prevention trials: lessons from the Alzheimer’s Disease Anti-inflammatory Prevention Trial (ADAPT). Alzheimers Dement. 2008;4(1 Suppl 1):S7-S14.
26. Fitzpatrick AL, Fried LP, Williamson J, et al. Recruitment of the elderly into a pharmacologic prevention trial: the Ginkgo Evaluation of Memory Study experience. Contemp Clin Trials. 2006;27(6):541-553.
27. Gill TM, McGloin JM, Latham NK, et al. Screening, Recruitment, and Baseline Characteristics for the Strategies to Reduce Injuries and Develop Confidence in Elders (STRIDE) Study. J Gerontol A Biol Sci Med Sci. 2018;73(11):1495-1501.
28. Andersen F, Engstad TA, Straume B, et al. Recruitment methods in Alzheimer’s disease research: general practice versus population based screening by mail. BMC Medical Research Methodology. 2010;10(1):35.
29. Wilkinson T, Ly A, Schnier C, et al. Identifying dementia cases with routinely collected health data: A systematic review. Alzheimers Dement. 2018.
30. 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;68(1):103-111.
31. Reuben DB, Hackbarth AS, Wenger NS, Tan ZS, Jennings LA. An Automated Approach to Identifying Patients with Dementia Using Electronic Medical Records. J Am Geriatr Soc. 2017;65(3):658-659.