The pre-screening visit for an anti-amyloid therapy trial is one of the most operationally expensive moments in Alzheimer’s research. A patient travels to the clinic, a coordinator blocks two to three hours, a neuropsychologist runs a full battery, and then the eligibility picture that emerges sends the case to a multidisciplinary conference. If the cognitive stage turns out to be outside the protocol window, the site absorbs every dollar of that visit as a screen failure. No reimbursement. No enrollment credit. Just cost. A study published September 15, 2026 in medRxiv out of the Mayo Clinic Rochester Alzheimer’s Disease Treatment Clinic describes what happens when you move part of that assessment out of the clinic and into the patient’s home before the clock starts, and the completion numbers are worth paying attention to.
Of 209 patients evaluated through the clinic for anti-amyloid therapy eligibility, 197 initiated the Mayo Test Drive (MTD) remote self-administered digital cognitive assessment. Of those who started, 97.5% completed a session. Seventy percent of completions happened remotely, through a patient-facing healthcare portal, with no staff present. Completion rates held across diagnostic groups: 100% in cognitively unimpaired patients, 99.1% in mild cognitive impairment, 97.0% in mild dementia, and 89.5% in moderate dementia. The difference across stages was not statistically significant (p =.12). That last number is the one that matters operationally. The conventional assumption in trial design is that digital self-administered tools work for early-stage patients and fall apart as impairment increases. These data challenge that directly.
What the Completion Rate Actually Tells Sites
Sites running anti-amyloid trials under protocols like those supporting lecanemab or donanemab already know how compressed the eligibility window is. The Clarity AD trial, which established much of the evidentiary base for current lecanemab use, reported a discontinuation rate of 17.2%. Every coordinator who has worked a study with that kind of attrition understands that the patients who are most burdensome to screen are often the ones least likely to complete. When a remote assessment tool holds 89.5% completion even in moderate dementia, that is not a convenience feature. It reframes which patients a site can realistically characterize before committing to a full in-person workup.
The operational value runs in two directions. For sites, a completed remote MTD session before a scheduled eligibility visit gives the clinical team a cognitive snapshot that can sharpen the visit agenda, surface potential protocol mismatches earlier, and reduce the proportion of full neuropsychological evaluations that terminate in a screen failure. For sponsors, a site that can pre-characterize patients remotely compresses the time from referral to eligibility determination without adding staff hours or scheduling pressure on already-stretched neuropsychology resources. The study found that administration setting (remote versus in-clinic) did not differ significantly by clinical stage, which means the cognitive data from home sessions and clinic sessions are drawing from the same population, not a healthier, more tech-savvy subset.
Assessor variability is the quieter benefit that does not always make it into feasibility papers. Published inter-rater reliability data for digitally administered cognitive tools like the MoCA show ICC values between 0.91 and 0.96 for total scores under structured conditions, but that reliability depends on assessor training and consistency across sites. Research on inter-rater reliability for digital cognitive assessments confirms that even with structured digital formats, scoring variability remains a documented concern. A self-administered platform with fixed item presentation and automated scoring removes that variability at the source. For a multi-site AD trial where neuropsychological assessments are endpoint-adjacent, that consistency has direct implications for data quality, not just operational convenience.
The Regulatory Gap Sites Cannot Ignore
Before a sponsor builds a remote cognitive assessment into a protocol as anything other than a prescreening tool, there is a regulatory boundary that needs to be stated plainly. Mayo Test Drive has not been cleared or approved by the FDA, according to Mayo Clinic Laboratories’ own classification policies. That does not disqualify it from site use in eligibility workflows, where the primary purpose is clinical characterization rather than protocol-defined endpoint collection. But it does define the lane. In the current study, MTD sat alongside, not in place of, the full neuropsychological evaluation conducted at the clinic. That is the right architecture for now.
In December 2023, the FDA finalized guidance titled “Digital Health Technologies for Remote Data Acquisition in Clinical Investigations,” which outlines what sponsors need to demonstrate when using digital tools to collect trial data remotely. The 2023 DHT guidance asks for pre-specified fit-for-purpose validation, evidence of user acceptability across the target population, and a data integrity framework for remote transmission. MTD data collected during a routine clinical eligibility workflow does not automatically satisfy those requirements for use as a clinical trial endpoint. Sites and sponsors who read this feasibility study and want to expand the application need to sequence that regulatory validation work before the protocol is written.
Remote digital cognitive assessment also raises documentation considerations that affect how sites record its use. If MTD results inform an eligibility decision that is captured in the source record, the site’s SOP needs to address how that remote session output is stored, who reviews it, and how it connects to the contemporaneous clinical assessment. Among patients who did not initiate MTD in the Mayo study, reported reasons included not receiving or opening the portal message, as well as other or unknown factors. That failure mode lives entirely on the site-operations side, patient outreach, portal enrollment, and follow-up workflow, not the technology. Any site implementing a similar pathway needs a documented outreach protocol with defined follow-up touchpoints before treating non-initiation as a patient factor.
Making This Work Monday Morning
Sites working with anti-amyloid therapy trials should treat this data as a workflow design prompt, not a validation study. The question it answers is specific: can patients at the cognitive stages targeted by these protocols complete a remote digital assessment reliably enough to use it as a pre-visit characterization tool? At 97.5% overall completion and 97.9% remote completion among those who initiated, the answer from 209 real patients moving through a live clinical eligibility process is yes. That is not a controlled research environment. It is a working clinic.
For sites, the practical step is a pre-screening workflow audit. Map the current path from referral to eligibility determination: how many patients complete a full in-person neuropsychological evaluation and then fall outside the protocol window? If that screen failure rate is substantial, a remote cognitive characterization step added before the visit is worth piloting. Sponsors reviewing site feasibility questionnaires for upcoming AD trials should add a direct question about remote cognitive assessment capacity and portal infrastructure. Sites that already have this pathway built are operationally ahead; those that do not will struggle to pre-characterize patients at the volume these protocols require.
The UK Dementia Research Institute published comparable feasibility data showing ICC values above 0.90 after 10 remote sessions in an autosomal dominant AD cohort, with a remote digital composite slightly outperforming a traditional composite at separating mutation carriers from non-carriers. The Mayo data now extends that signal into a treatment-eligibility clinical workflow, in a broader and more impaired population. The next site that builds this into its standard pre-screening SOP will not just reduce screen failure costs, it will compress time-to-eligibility-decision for a patient population where weeks matter.
References
- medRxiv, “Feasibility of remote self-administered cognitive assessment in an Alzheimer’s disease treatment clinic”
- BMT Advisors, FDA Final Guidance: “Digital Health Technologies for Remote Data Acquisition in Clinical Investigations” (December 2023)
- Mayo Clinic Laboratories, Test Classification and Policies: Mayo Test Drive regulatory status
- 2 Minute Medicine, Clarity AD trial discontinuation rate (17.2%) and completion data
- UK Dementia Research Institute, Remote digital cognitive assessment in trial-ready Alzheimer’s disease cohort: ICC data and composite performance
- PubMed, Diederiks et al.: Inter-rater reliability of digital MoCA administration (ICC 0.91–0.96)
