BACKGROUND:Soluble species of multimeric amyloid-beta including globular amyloid-beta oligomers (AβOs) and linear amyloid-beta protofibrils are toxic to neurons. Sabirnetug (ACU193) is a humanized monoclonal antibody, raised against globular species of soluble AβO, that has over 650-fold greater binding affinity for AβOs over monomers and appears to have relatively little binding to amyloid plaque. OBJECTIVES:To assess safety, pharmacokinetics, and exploratory measures including target engagement, biomarker effects, and clinical efficacy of sabirnetug in participants with early symptomatic Alzheimer's disease (AD; defined as mild cognitive impairment and mild dementia due to AD). DESIGN:Randomized, double-blind, placebo-controlled, ascending dose first-in-human phase 1 study. SETTING:Fifteen study centers in the United States. PARTICIPANTS:Sixty-five participants with early symptomatic AD. INTERVENTION:Participants received one infusion of sabirnetug 2 mg/kg, 10 mg/kg, 25 mg/kg, 60 mg/kg, or placebo (Part A) or three infusions of sabirnetug 10 mg/kg, 25 mg/kg, 60 mg/kg, or placebo (Part B). MEASUREMENTS:Safety, tolerability, serum pharmacokinetics, and central target engagement of single and multiple doses of sabirnetug, cerebrospinal fluid (CSF) concentrations of sabirnetug, and amyloid plaque load, as determined by positron emission tomography. RESULTS:Sabirnetug was generally well tolerated. A larger percentage of participants receiving sabirnetug (56.3%) versus placebo (42.9%) had at least one treatment emergent adverse event, with approximately 29% in each group considered related to study drug. Most events were mild-to-moderate in severity. Of 48 participants given sabirnetug, five developed amyloid related imaging abnormalities - edema/effusion, including one instance that was mildly symptomatic in a participant who had received one dose sabirnetug 60 mg/kg. Notably, none of the six apolipoprotein E Ɛ4 homozygotes who received sabirnetug developed amyloid related imaging abnormalities - edema/effusion or - hemorrhage/hemosiderin deposition. Infusion reactions, such as rash, pain, or erythema, were not frequent (6.3% for sabirnetug versus 0.0% for placebo). Sabirnetug exposure was dose proportional in both serum and CSF. Target engagement, defined as drug bound to AβOs in CSF, was shown to be dose and exposure dependent. Over three months, approximately 25% and 20% reduction in amyloid plaques, respectively, were observed in participants receiving three infusions of sabirnetug 60 mg/kg every four weeks and 25 mg/kg every two weeks. CONCLUSIONS:The Phase 1 INTERCEPT-AD study provided safety, tolerability, dosing, and target engagement data that supported the design of the ongoing ALTITUDE-AD study (NCT06335173).

Qualitative patient interviews are increasingly conducted alongside clinical trials; interviews administered early in drug development can yield insight into the patient journey. Examination among patient subgroups may identify factors that influence the patient experience, including trial participation. As part of the INTERCEPT-AD phase 1 study evaluating the safety and tolerability of the Aβ oligomer-targeting monoclonal antibody ACU193, we conducted semi-structured qualitative interviews among a subset of participants with mild cognitive impairment (MCI) or mild dementia due to Alzheimer’s disease (AD) and their study partners to obtain feedback on their trial experience and the decision-making process preceding trial enrollment; results were compared between study participant genders. A subset of study participants and study partners completed qualitative interviews following the final study visit. Topics included referral source, motivations for participating, enrollment decision-making, concerns regarding study medication and procedures, and positive and negative aspects of participation. Study participant/study partner dyads were defined as female or male according to the gender of the study participant as it was captured in the main study protocol; study partner gender was not obtained. Transcripts were coded and analyzed using principles of qualitative thematic analysis and grounded theory. Differences in concepts mentioned between genders were reported when >20%, and when sample size was sufficient (≥5 dyads reporting a concept with a difference of >2 mentions between genders). Twenty-eight participants (64.2% female) and their study partners were interviewed, representing 43% of the total randomized trial population (n = 65; 53.8% female). Female dyads accounted for 62% and 70% of those who first heard about INTERCEPT-AD from social media or physicians, respectively. More female dyads desired additional information about AD (86% of all dyads reporting this concept), to benefit others (75%), and to benefit self (65%) by participating. Regarding the decision to enroll, those who decided independently were exclusively female dyads; female dyads accounted for 71% among those who had pre-trial concerns about side effects. More female dyads expressed dislike for cognitive tests (83%) and blood draws (78%). Certain factors of trial participation differed by participant gender, including motivations for participation and perceived burden of some study procedures.

Objective: Sabirnetug (ACU193) is a humanized monoclonal antibody selective for soluble amyloid beta oligomers (A/3Os), synaptotoxins that are early and persistent triggers of Alzheimer's disease (AD). Sabirnetug pharmacodynamics were examined in the INTERCEPT-AD phase 1 study in mild cognitive impairment and mild dementia to AD (NCT04931459) using biofluid biomarkers associated with A /3 and tau pathology, synaptic dysfunction, neuroinflammation, and neurodegeneration. Methods: INTERCEPT-AD was a randomized, first-in-human study of sabirnetug versus placebo administered a single (SAD; 2, 10, 25, 60 mg/kg) or multiple (MAD; three doses of 10 or 60 mg/kg every 4 weeks [Q4W] 25 mg/kg Q2W) ascending doses. Biomarkers were measured pre-/post-dose in CSF and EDTA-plasma. tions of biomarker changes versus dose, exposure duration, and target engagement were determined. Results: In MAD cohorts, CSF pTau181 decreased significantly (60 mg/kg Q4W, p = 0.049). VAMP2 decreased significantly at all doses ( p <= 0.041); neurogranin decreased significantly at 60 mg/kg Q4W ( p = A /31-42 /A /31-40 trended upward with sabirnetug dose. A/31-42/A/31-40 and neurogranin changes correlated sabirnetug-A/3O target engagement ( p <= 0.01). Decreases in tTau, VAMP2, and neurogranin correlated posure duration ( p <= 0.007). Plasma pTau181, pTau217, GFAP, and NfL trended lower. Discussion: Following three sabirnetug doses, changes in CSF and plasma biomarkers were observed. biomarker response increased with increasing dose and exposure duration, consistent with previous reports sabirnetug reaches the central compartment and engages its A/3O target. The ongoing phase 2 ALTITUDE study (NCT06335173) will test whether sabirnetug's pharmacodynamic effects can be substantiated with sample size and longer treatment duration.

Incorporation of the patient voice into drug development has been recognized as fundamental; recent legislation has addressed the importance of collecting patient experience data. Qualitative patient interviews conducted in tandem with clinical trials, often in later phases, are among the most common means of eliciting these data. To assess aspects of patient experience earlier in development, we conducted semi-structured qualitative interviews following participation in the phase 1 ACU-001 (INTERCEPT-AD) trial, a study evaluating the safety and tolerability of the Aβ oligomer-targeting monoclonal antibody ACU193, among a subset of participants with mild cognitive impairment (MCI) or mild Alzheimer’s disease (AD) and their study partners. Interview topics included disease experience and expectations for treatment. Interviews followed the end-of-study visit and included both participants and study partners. Participants were asked to report their most problematic symptoms, to describe what they would hope to gain from a new AD treatment, and what changes would be meaningful. The interview guide was developed in accordance with FDA Patient Focused Drug Development guidance. Principles of qualitative thematic analysis guided the coding and analysis of transcripts, with additional features drawn from grounded theory, conforming to best practices. Twenty-eight participants and their study partners were interviewed. Participants reported a broad array of problems consistent with AD. The most frequent included difficulty with memory or cognitive functioning (96.4%), getting lost (71.4%), mood (60.7%), and difficulty communicating (57.1%). Within each of these groupings, specific problems varied among participants; higher numbers of participants reported forgetting appointments, names, and losing things (81.4%) and getting lost while driving (90.0%), while fewer reported feeling isolated (5.9%) and experiencing confusion (11.1%). Nearly every participant desired treatment that would keep the disease from getting worse or slow progression (89.3%), which was equated with maintenance of functional abilities. Maintaining the ability to recognize loved ones and communicate were also stated as important benefits of a new treatment. Problems reported by participants were typical to AD, although there was heterogeneity in overall AD experience. Changes in mood were frequently reported. These findings highlight the need to contextualize clinical trial results by incorporating the patient voice.

PET tracers enable the in-vivo measurement of amyloid in the brain and have been incorporated in eligibility criteria for clinical trials. Automated quantitative image analysis enables the calculation of a standardised uptake value ratio (SUVr) which can enable a rapid amyloid classification without requiring a visual read. However, quantitative analysis may have an equivocal range where amyloid classification does require a visual read. The range of SUVrs that is equivocal is not well understood. Participants in the INTERCEPT-AD trial of ACU193 (Siemers et al., J Prev Alzheimers Dis. 2023;10(1):19-24) underwent an amyloid [18 F ] Florbetapir PET scan and T1-weighted MRI scans. MRIs were segmented using the LEAP algorithm (Wolz et al., NeuroImage. 2010;49(2):1316-1325), registered to the PET image, and the global cortical average (GCA) SUVr based on brain regions typically with high amyloid load in Alzheimer’s disease (Klunk et al., Alzheimers Dement. 2015;11(1):1-15) was calculated, using the whole cerebellum as the reference region. SUVrs greater than 1.2 were classified as amyloid positive, SUVrs less than 1.0 were classified as amyloid negative, and SUVrs in the equivocal range were assessed by visual read by trained central radiologists. Of 194 participants who underwent screening, 188 had an SUVr available from automated analysis. Of these, 64 were within the equivocal range (SUVr 1.0 – 1.2) and had a visual read performed. For PET scans in the equivocal range, 53 (83%) were classified as amyloid negative and 11 (17%) were classified as amyloid positive by visual read. Within the equivocal range, there was no difference in SUVrs between the images classified by visual read as amyloid negative or positive. SUVrs may be used as an objective quantitative amyloid classification method at values > 1.2 for trial inclusion; however, in the equivocal range, the GCA SUVr did not correspond with amyloid classification based on a visual read. Further study is required to investigate the equivocal range and the use of combined quantitative measures and visual reads.

Alzheimer’s disease (AD) is a neurodegenerative disorder affecting approximately 40 million people worldwide. There are few treatments available that affect the underlying degenerative process, making AD one of the largest unmet medical needs. ACU193 is a humanized, IgG2 subclass monoclonal antibody that targets soluble amyloid beta oligomers (sAβOs). Evidence supports the view that sAβOs are one of the primary instigators of AD neurodegeneration. Thus, blocking the toxicity of sAβOs with ACU193 may be a promising approach for the treatment of AD. INTERCEPT-AD is a phase 1 randomized, placebo-controlled, single- and multiple-dose study investigating the safety, tolerability, and pharmacokinetics of intravenous ACU193 in mild cognitive impairment or mild dementia due to AD. The SAD portion includes four cohorts with 2 mg/kg, 10 mg/kg, 25 mg/kg, and 60 mg/kg ACU193 or placebo. The MAD portion includes three cohorts receiving three doses of ACU193 or placebo: 10 mg/kg once every four weeks (Q4W), 60 mg/kg Q4W and 25 mg/kg once every 2 weeks (Q2W). Key inclusion criteria are 55-90 years of age, PET scan positive for brain amyloid, MMSE 18-30 and CDR-Global score 0.5 or 1.0. The INTERCEPT-AD trial is fully enrolled. At 17 sites in the U.S., 260 participants were screened, 65 were randomized, and 62 received at least one dose of ACU193 or placebo, including 33 females (53%) and 29 males (47%). Negative PET (61%) was the most common reason for screen failure. Baseline age was 72.1 ± 7.8 years (mean ± SD), CDR Sum of Boxes score was 3.5 ± 1.8, ADAS-Cog-13 was 24.2 ± 9.0 and screening MMSE was 24.2 ± 3.6. ApoE e4 homozygotes accounted for 13% of the study population and ApoE e4 heterozygotes for 46.7%. Traditionally underrepresented ethnoracial populations comprised 23% of the study (5% Black/African American, 16% Hispanic/Latino, 2% Other). Additional baseline measures including demographics, clinical characteristics, AD medication, and imaging data will be presented. We describe key baseline demographic, clinical, and cognitive characteristics of participants in the INTERCEPT-AD trial. Data generated from this trial will guide further development of ACU193 for the treatment of early AD.

Clinical trials are essential for testing safety and efficacy of potential treatments for Alzheimer’s disease (AD), but recruitment is often difficult due to a variety of factors. Recruitment tactics in AD clinical trials may include media advertising, physician referrals, patient registries, and referral campaigns. Newer strategies such as social media and community outreach have also been successful in increasing recruitment. However, not all recruitment tactics are successful for every study site and a multi-faceted approach tailored to each site may be more successful. INTERCEPT-AD is a phase 1 randomized, double-blind study of ACU193 in mild cognitive impairment or mild dementia due to AD. Seventeen study sites in the U.S. screened potential participants identified through multiple recruitment tactics which were grouped post hoc into seven categories: site database, external referral, physician referral, site campaign, and three sponsor-initiated campaigns. Recruitment efforts were tailored to each site’s patient population, community, and staff availability. INTERCEPT-AD screened 260 participants and 70 were eligible for study participation. Recruitment was completed in January 2023. The sponsor worked directly with individual sites to develop a tailored recruitment plan that considered each site’s target population, historically successful recruitment tactics, and current ability to screen patients from outside sources. Comparisons of the seven types of recruitment sources showed site databases were used most frequently (12/17 sites, 71%) and accounted for the most screenings and eligible participants (107 [41%] and 32 [46%], respectively). Additional data from all recruitment sources and success rates will be presented. Study sites found success using a combination of tactics specifically tailored to their enrollment goals. Sites with robust internal patient databases had higher rates of screening and enrollment. A combination of site-led and sponsor-driven recruitment initiatives likely maximized enrollment rates. These findings suggest a strategy using diverse recruitment tactics customized to a study site’s capabilities and patient population may be more successful than a one-size-fits-all approach.

The increasingly global scope of clinical trials in Alzheimer’s disease (AD) has led to the recruitment of subjects from various geo-cultural regions. Recent studies suggest that commonly used objective measures of cognition such as the MMSE as well as subjective, clinician-rated global measures such as the CDR might be impacted by cultural and/or linguistic differences. This raises the question regarding the impact of cultural differences on the convergent validity of the CDR and MMSE. In this analysis, we examined the extent to which the correspondence between the CDR and MMSE varied across seven geo-cultural regions in multinational AD clinical trials. This study included electronic CDR (eCDR) and electronic MMSE (eMMSE) data from four multinational AD clinical trials of subjects with Prodromal to Mild AD. Data were collected at screening from seven geo-cultural regions: East Asia (n = 1916), Eastern Europe (n = 1882), South America (n = 1601), Middle East/Africa (n = 188), North America (n = 9311), Oceania (n = 363), and Western Europe (n = 5174). Pearson correlations between eCDR Sum of Boxes (SOB) and eMMSE across the 7 geo-cultural regions ranged from r = -.31 (East Asia, Western Europe) to r = -.48 (Mideast/Africa). A linear regression model with geo-cultural regions entered as covariates further demonstrated that the relationship between eMMSE and eCDR SoB scores differed by region (Adjusted R 2  = 0.14, p<0.001). Findings indicate the strength of the correlation between the CDR SOB and MMSE total score varied by geo-cultural region. These results build on prior findings indicating that CDR SOB and MMSE scores vary across regions independent of each other. A growing body of literature suggests these differences might be due to cultural and/or linguistic factors.

Abstract Background While schizophrenia is observed in different parts of the world across countries, ethnicities, and races, research indicates cultural factors play significant roles in the phenomenology of this illness. Cultural norms and values affect manifestations of this pathology; more specifically, they affect how symptoms are expressed, experienced, and interpreted. Given that culture affects manifestations of schizophrenia, cultural factors should be considered in the assessment of its symptoms in clinical trials. This study explores the differences and patterns in the Positive and Negative Syndrome Scale (PANSS) item ratings across different geocultural regions. Identifying such patterns can give insights into culturally sensitive assessment practices and aid in developing more effective rater training and data surveillance that consider unique cultural factors. Methods Data were obtained from an international group of raters from 37 different countries, representing 6 geocultural regions across 13 different studies. As part of the rater training and qualification process for each of these studies, raters viewed and scored the 30-item PANSS based on a video-recorded PANSS interview that was administered using the Structured Clinical Interview-Positive and Negative Syndrome Scale (SCI-PANSS). Raters were deemed qualified if their scores fell within the defined acceptable item score ranges. Given the cultural diversity of the raters, the acceptable passing score ranges for each country were determined by a combination of expert opinion scores, group modal scores, and clinical analyses. Only the scores from raters who achieved qualification on their first scoring attempt were analyzed. The number of raters per geocultural regions included: Asia Pacific, n = 397; Eastern Europe, n = 412; Latin America, n = 88; Middle East/Africa, n = 29; North America, n = 339; and Western Europe, n = 129. Results A Shapiro-Wilk test for normality was conducted on the scores for each PANSS item and found all significantly different from a normal distribution (all ps < .0001). A Kruskal-Wallis test for rank-ordered differences was conducted for each item for the influence of region on item score. Most items showed a significant influence of region on score after a Bonferroni correction was applied; most ps < .0001 with the following exceptions: N1, p < .001; G15, p < .05; P2, P7, N6, N7, G1, G6, G10, G12, and G13 were not significant. The most significant cross-regional differences were found with P1, P3 and P6, and these items were analyzed further with a post hoc Dunn test to understand cross-regional patterns. On P1, Asia Pacific and Eastern Europe were significantly lower than Latin America, North America, and Western Europe (all ps < .0008) but not Mideast/Africa. On P3, Western Europe were significantly lower than all other regions (all ps < .0005); Asia Pacific were significantly lower than Eastern Europe, North America, and Western Europe (all ps < .006). On P6, Asia Pacific and Eastern Europe were significantly lower than all other regions (all ps < .03). Discussion The present study suggests that ratings of schizophrenia symptoms are influenced by cultural factors. Cultural beliefs and behaviors seem to influence interpretations of schizophrenia pathology. Given that the PANSS is not standardized for cross-cultural contexts, it is important to consider cultural factors when using this scale in clinical studies. In addition, when developing rater training and data surveillance programs, adjustment of acceptable item score ranges for key PANSS items highlighted above for different geocultural region is recommended. Future studies should explore country-level patterning of ratings of the PANSS.

In the last 20+ years, the number of Alzheimer's disease (AD) clinical trials has increased significantly. Despite the uptick in AD drug development, these clinical trials have continued to fail. Proper administration and scoring of screening measures are critical to only enroll appropriate subjects and to properly stage/stratify them when indicated. The Mini-Mental State Examination (MMSE) is commonly used as part of inclusion criteria in AD clinical trials, and is often used for staging of the disease process. Despite its near universal use, MMSE raters continue to struggle with standard scoring and administration conventions. Past research indicates data quality surveillance programs and the advancement of electronic clinical outcome assessments (eCOA) have contributed to better data quality. In this analysis, we evaluated the outcome of data quality surveillance programs on subjects’ MMSE score across three large multi-national AD clinical trials. The impact of scoring errors were reviewed, including their potential impact on both inclusion criteria, and stratification. Calibrated clinicians performed a data quality review in local language of 100% of electronic MMSEs at Screening and Baseline visits for three large multi-national AD clinical trials. Reviews consisted of listening to audio recordings of scale assessments to confirm that standard scoring and administrative conventions were followed. Clinicians initiated a data query when administration and/or scoring errors were identified. Across three trials, a total of 613 data queries were identified as scoring errors. Of the scoring data queries raised, 7% suggested a score change that would result in a change of subject inclusion or exclusion status. For one study that included stratification, 4.7% of suggested score changes would result in a change of randomization group. Virtually all (98.5%) data queries were within ±2 points in total score changes with the maximum change at 8 points. Robust data quality surveillance/monitoring programs should continue to play a crucial role in AD clinical trials. The results of this analysis suggest identification of errors on the MMSE could have significant impact in a study by mitigating concerns related to inclusion of appropriate subjects, randomization of ineligible subjects, or incorrectly stratifying subjects.

Data quality programs that reduce rater error are important to clinical trials, especially those in Alzheimer's disease (AD). Given the high failure rate in AD drug development over the past 10+ years, ensuring optimal data quality is paramount. Prior research has shown that employing enhanced electronic clinical outcome assessment (eCOA) scales in AD clinical trials significantly reduces rater error rates when compared to both paper and pencil and non-enhanced eCOA scales. Additionally, the implementation of a customized in-study data quality program can further improve study data by identifying instances where raters deviate from proper administration and/or scoring conventions. When deviations occur, raters are remediated with the goal of decreasing future error rates. In this analysis we examine the impact of rater remediation, on the Alzheimer's Disease Assessment Scale – Cognitive subtest (ADAS-Cog), on rater performance over time. 100% of ADAS-Cog ratings were evaluated from two multi-national AD clinical trials using an enhanced eCOA version of the scale. Raters were trained and certified on proper scale administration and scoring. All submissions of ADAS-Cog scale data were reviewed and all raters’ ratings were assessed by a calibrated clinician. Raters who committed scoring related errors were remediated. A total of 630 raters and 7,796 assessments were included in the analysis. A logistic regression on the chances of a visit assessment remediation as a function of the length of time a rater has been in a study, controlling for study, was fitted for the ADAS-Cog. The results showed a significant impact of time (b = −0.001, 95% CI [−0.0017, -0.0004], z = −3.01, p = .003). For each day beyond the rater's first ADAS-Cog administration, the odds of being remediated decreased by a factor of 1.00. Prior research has shown enhanced eCOA scales significantly improve overall data quality in clinical trials. This analysis indicates a significant reduction in error rates over time when raters are monitored and receive feedback about discrepancies in scoring. This offers support for the value of data quality programs, used in tandem with enhanced eCOA scales, in AD clinical trials.

Alzheimer's disease (AD) is an irreversible and progressive brain disorder, slowly affecting memory and thinking aptitude and eventually the individual's ability to function. The past several decades have seen significant growth in Alzheimer's disease (AD) clinical research, however, success in developing treatments has been limited. In concert with the growth in AD clinical trials, there has been an increased focus on the importance of collecting reliable and valid data as a means of assessing therapeutic efficacy. The evolution of technology and has allowed for the development of novel ways to help ensure data quality in AD clinical trials including advanced electronic clinician reported outcomes (eCLINRo), the secondary review of key assessments, the incorporation and review of assessment audio, and the utilization of analytics and statistical modeling to provide insight into rater, site, regional performance as well as variability and trends in outcome data. Finally, the intersection of technology with the application of data science-based techniques such as machine learning have the potential to significantly increase outcome validity. Relevant data and publications detailing the innovations and evolution of programs utilized to enhance clinical trial data quality in Alzheimer's disease during the past several years were reviewed as well as current practices and future concepts. The evolution of technology including advanced eCOA and analytic techniques in AD clinical research have led to reduced error rates, decreased variability and greater data validity. Ensuring data quality through evolving technology and data science has the potential to lower trial costs and lead to improved trial outcomes.

The use of Electronic Clinical Outcome Assessments (eCOA) in Alzheimer's disease (AD) clinical trials continues to expand. By using internal scale logic and automatic scoring algorithms the enhanced eCOA versions of the ADAS-Cog and MMSE have been shown to significantly reduce error rates when compared to both paper and non-enhanced eCOA versions of the scales while maintaining clinical validity 1, 2, 3. Further, the addition of audio review of scale assessments has been shown to reveal additional errors that cannot be identified through review of the scale data alone 4. While clinical trials stand to benefit from the overall improvement in data quality, it remains important to understand the areas where raters continue to make errors despite utilizing enhanced eCOA scales. In this analysis, we evaluated both the domain and specific type of errors on the enhanced eCOA version of the MMSE and ADAS-Cog. ADAS-Cog and MMSE ratings were evaluated from several completed multi-national clinical trials in Dementia which utilized enhanced eCOA scales. Raters were trained and certified on the proper scale administration and scoring. Error rates were calculated for each scale, domain and type (scoring vs. administrative). A total of 10,635 visits were reviewed across programs including 4,707 individual MMSE's and 5,928 ADAS-Cog. One sixth (18%) (N = 856) of MMSE reviewed were flagged for an administrative or scoring error as were nearly one quarter of ADAS-Cog's (23%) (N = 1,374). The type of error and corresponding domains are shown in table 1. The increased utilization of technology will continue to improve the quality of data in clinical trials. However, understanding the problem areas that raters have on key outcome measures, despite the use of enhanced eCOA scales, will help to promote more nuanced training, ensure more reliable and valid data and inform future technological advances to scale versions.

In an attempt to successfully recruit Alzheimer's disease (AD) subjects, trials are becoming increasingly global. Given this geographic expansion, identifying regions and countries that are able to provide high quality data continues to be paramount. The aim of this analysis was to determine whether there were any geographic differences in rater administration and/or scoring error rates on the ADAS-Cog. We reviewed data from four multi-national AD clinical trials involving 43 countries. All studies utilized and enhanced electronic ADAS-Cog (eADAS-Cog) which included employing internal logic, automated scoring, standardized instructions and scoring conventions. ADAS-Cog ratings were evaluated from several multi-national AD clinical trials which utilized an enhanced electronic clinical outcome Assessment (eCOA) version of the ADAS-Cog-13. All study raters were trained and certified on the ADAS-Cog and all studies included an active in-study data quality program. Error rates were calculated for the ADAS-Cog by type (scoring versus administrative) for the Screening and Baseline visits only. Countries were included in the analysis if its raters performed a minimum of 30 assessments between Screening and Baseline visits. Twenty-five countries met the analysis criteria for scoring errors. A total of 8,868 eADAS-Cogs were reviewed across programs with an overall error rate of 8.54%. The lowest and highest error rates were found in Croatia (Hrv) (1.54%, N=65) and Slovakia (50.00%, N=96), respectively. Of the 15 countries and 3,578 visits reviewed for administration errors, the overall error rate was 13.19% with the lowest error rate found in Serbia (0.00%, N=70) and the highest found in Slovakia (53.12%, N=96). Prior research has shown an enhanced eADAS-Cog combined with an in-study surveillance program significantly improves overall data quality in clinical trials. Our analysis indicates there can be substantial differences in eADAS-Cog scale performance across counties potentially leading to increased data variability. Continuing to identify raters and sites that provide high data quality remains critical to the success of AD clinical trials.

The Alzheimer's Disease Assessment Scale (ADAS) was developed to provide a measure of change in cognitive and behavioral functions known to be impaired by Alzheimer's disease. Over the last 20 years, the cognitive subscale, (ADAS-Cog), has become the de facto gold-standard for assessing the efficacy of putative anti-dementia treatments, serving as the primary or co-primary outcome for numerous phase 2 and phase 3 drug development trials. Given its importance, there has been an increasing interest in providing greater standardization, automation, and administration consistency to the scale. Recently, electronic versions of the ADAS-Cog (eADAS-Cog) have been utilized in clinical trials and demonstrated significant reductions in frequency of rater error as compared to paper. 1, 2 While the eADAS-Cog has been purported to be equivalent to paper in terms of validity, to date, there has not been a prospective trial comparing a version of the eADAS-Cog with the paper ADAS-Cog in a clinical population. Twenty subjects completed a single-center, randomized, counterbalanced, prospective trial comparing a version of the eADAS-Cog to the paper version of ADAS-Cog (13 item) for men and women ages 50-90 (inclusive). All subjects met NIA-AA criteria for mild to moderate Alzheimer's disease. Subjects who met inclusion criteria (MMSE 12-26) were randomized to one of two study conditions associated with the alternating order in which they received study measures (ADAS-Cog vs eADAS-Cog). Subjects returned for two subsequent visits at 4-week intervals at which the study assessments were repeated. A two-way repeated measures ANOVA was conducted to examine the effect of order and mode of administration on ADAS-Cog scores. The model was not significant F(2, 48) = 3.08, p = .06. Kappa analyses also yielded high agreement between eADAS-Cog and paper version across total and its sub-test scores (κ range: 0.88 - 0.99). While there are various versions of the eADAS-Cog currently being utilized in clinical trials, this pilot study is the first time a version has been validated in a clinical population thus helping to establish concurrent validity between electronic and paper modalities.

The Alzheimer's Disease Assessment Scale – Cognitive subscale (ADAS-Cog) and Mini-Mental State Examination (MMSE) are two of the most commonly used scales in Alzheimer's disease (AD) clinical trials. Previous research has demonstrated error rates among clinical trial raters to be upwards of 32% for the ADAS-Cog and 23% for the MMSE on initial administration when using a paper-pencil version of the scale. The original electronic Clinical Outcome Assessments (eCOA) platform reduced ADAS-Cog and MMSE error rates in the WN28745 Marguerite RoAD trial to 11.3% and 10.8% respectively. The platform and scales were recently enhanced and improved and are currently utilized in the BN29552 CREAD Study. We evaluated their impact on overall data quality in comparison to the initial versions of the scales. The enhancements to the eADAS-Cog and eMMSE included employing internal logic, automated scoring, standardized instructions and scoring conventions taken directly from the scale manual and the study specific training curriculum. These enhancements were intended to further decrease rater error and increase standardization and scoring. After being trained and certified to rate the eADAS-Cog and eMMSE, all raters’ performance regarding accuracy in scale administration and scoring was assessed by a calibrated clinician at the initial Screening visit in both trials. Of the 475 new eADAS-Cogs reviewed at Screening, 34 (7.2%), required a contact with the rater due to an error in scoring. This represents a significant reduction (p < 0.05) in error rates compared to the previously reported 11.3% with the original electronic version of the scale. Similarly, 558 new versions of the eMMSE were reviewed at screening of which 35 (6.3%) required a contact with the rater. Compared to the 10.8% error rate previously reported utilizing the original eMMSE, a significant reduction (p < 0.01) in error rates was observed. Compared to the original version, using an enhanced eCOA platform, coupled with an in-study ratings surveillance program, significantly improves data quality by reducing rater error, promoting standardized scale administration/scoring and minimizing rater drift.

BACKGROUND:Electronic administration of clinician-reported outcomes (eClinROs) has advantages over paper-based methods, but the mode of administration change has the potential to affect the validity of the scale. The literature on migration of patient-reported outcomes (PROs) suggests that there are different levels of modification, which necessitate different approaches to demonstrating mode equivalence. However, little has been written on the migration of ClinROs to electronic administration.METHODS:We propose a method of comparing paper and electronic versions of scales that includes a comparison based on content and a comparison based on format. The determination of whether the eClinRO has undergone minor, moderate, or substantial modification will drive the necessary studies required for validation.RESULTS:The unique characteristics of ClinROs suggest 2 additional types of modifications, including functionality adaptation and adaptation of instructions.CONCLUSIONS:In many respects, the migration of a ClinRO to electronic administration is similar to that of a PRO. This article has explored the ways in which there might be special considerations for ClinROs that have not been elaborated for PROs.