Generalizability Of Clinical Trial Results Research Articles

Characterize demographics, comorbidities and co‐medications in newly diagnosed United States (US) Alzheimer’s Disease patients using Medicare claims

AbstractBackgroundAccess to Medicare claims provides a unique opportunity to broadly characterize diseases affecting US adults 65+ years of age. This can provide insights into real world populations and a basis of comparison to those in clinical trial populations. To this end, this study characterizes demographics, comorbidities and co‐medications in patients newly diagnosed with Alzheimer’s Disease (AD).MethodsThis study utilized Medicare Part A and B claims (Fee‐For‐Service and Medicare Advantage) limited to patients with at least one 2019 medical encounter. The study’s source population was limited to those 65+ years of age with Medicare Part D coverage and 1‐year pre‐encounter continuous enrollment. From this source population, three newly diagnosed AD cohorts were compiled; cohort 1 utilized the 1‐year Bynum‐Standard Alzheimer’s disease and related dementia (ADRD) claims algorithm, cohort 2 adapted the ADRD algorithm to include only specific AD diagnoses, and cohort 3 was selected based on new use of cholinesterase inhibitors (irrespective of ADRD diagnosis). Demographics, comorbidities, and co‐medications present in the 1‐year pre‐cohort qualification are described.ResultsFrom 53,890,958 patients with at least one 2019 medical encounter, 33,315,038 qualified for the source population. Of these patients, 861,727 (2.6%) were newly diagnosed with ADRD (cohort 1), 373,798 patients (1.1%) were newly diagnosed with AD (cohort 2), and 395,319 patients (1.2%) were new users of cholinesterase inhibitors (cohort 3). The cohorts were not mutually exclusive. Overall, the three cohorts were similar in terms of demographics (table 1) with 75.1‐85.3% of patients aged 75+ and 61.5‐66.4% female. Common comorbidities (table 2) across the 3 cohorts include cardiovascular (CV) comorbidities such as myocardial infarction (34.1‐41.9%) and heart failure (39.8‐51.4%), diabetes (38.5‐41.6%), arthritis (44.0‐46.6%), stroke (28.9‐34.4%) and anxiety/depression (48.6‐54.5%). Co‐medications received (table 3) such as anti‐hypertensives, lipid modifying drugs, anti‐diabetics, systemic corticosteroids, and anti‐depressants reflect these conditions.ConclusionExtensive understanding of a drug’s real‐world target population can inform generalizability of clinical trial results and highlight unaddressed risk. This study stresses the importance of studying new AD drugs in populations with a history of CV risk, diabetes and stroke as well as those that may be immunocompromised due to treatments with systemic corticosteroids.

Alzheimer’s disease blood tests of amyloid‐β 42/40, p‐tau217, p‐tau181 and p‐tau205 ratios in a diverse community‐based population: Preliminary results from SEABIR

AbstractBackgroundAlzheimer’s disease (AD) blood tests that can identify and quantify amyloid plaques, tau tangles, and cognitive and clinical decline are needed in clinical practice, including primary care. However, these tests require validation in diverse groups and real‐world settings. The goals of the Study to Evaluate Amyloid in Blood and Imaging Related to Dementia (SEABIRD) were to enroll 1120 participants to determine the accuracy and validity of blood amyloid‐β (Aβ) and phosphorylated tau (p‐tau) tests in a diverse, community‐based sample of older adults compared with amyloid PET, and to explore the impact of key factors (age, race, years of education, cognition, APOE genotype, and medical conditions) on validity estimates.MethodSEABIRD measured blood plasma Aβ42/40 ratio and ratios of phosphorylated to non‐phosphorylated p‐tau217, p‐tau181, and p‐tau205 (%p‐tau) in a diverse population of older adults. In addition, cognitive screening with the AD8® dementia screen and Montreal Cognitive Assessment (MoCA) were obtained. Approximately 25% of participants were selected for amyloid PET and CDR as a reference standard.ResultOf the 859 participants enrolled in this ongoing study, 20.6% self‐identified as Black or African American and the percentage of APOE ε4 carriers was 30.2%, as expected for a population‐based sample (Table 1). For a subset of 76 participants with completed amyloid PET and all blood measures, plasma Aβ42/40 had ROC‐AUC with amyloid PET status of 0.85, %p‐tau217 of 0.97, %p‐tau181 of 0.83, and %p‐tau205 of 0.81 (Figure 1). Spearman correlations with amyloid PET SUVR were ‐0.59 for plasma Aβ42/40, 0.69 for p‐tau217 ratio, 0.43 for p‐tau181 ratio and 0.48 for p‐tau205 ratio.ConclusionThese preliminary results demonstrate that AD blood tests have similar accuracies in a diverse population‐based sample to those reported in AD research cohorts. These findings suggest that rapid screening and diagnosis of amyloid pathology with AD blood tests is possible in a broader population with higher comorbid disease burden and greater diversity than typical AD research cohorts. Screening with AD blood tests could accelerate clinical trial recruitment and increase the representativeness of clinical trial populations, which may improve the generalizability of clinical trial results.

Survival Disparities By Frontline Clinical Trial Enrollment Status for Treatment of Pediatric Acute Myeloid Leukemia

Introduction: Studies have suggested superior outcomes for patients enrolled on clinical trials for their cancer treatment. While sociodemographic disparities in trial enrollment are well established in adults, the literature in children is mixed and limited. A deeper understanding of disparities in pediatric clinical trial enrollment and potential survival differences is critical to ensure the generalizability of trial results and equitable outcomes. In a cohort of children with AML, we sought to examine (1) differences in sociodemographic and clinical characteristics by trial enrollment status, (2) the impact of trial enrollment on survival, and (3) the potential drivers of survival differences. Methods: A retrospective cohort including all pediatric (age <19 years) patients with AML treated at 10 institutions in the United States from 2011-2018 was established and is being actively expanded to include diagnoses through 2023. Chart abstraction by trained study personnel collected detailed demographic and clinical information. Resource utilization data from the Pediatric Health Information System was merged to ascertain patient acuity. Higher acuity was defined as requiring ICU-level resources for 2 or more organ systems within 72 hours of diagnosis. Only patients who had a frontline therapeutic trial option available at their treating institution were included. This was operationalized as a center-level restriction to patients initiating therapy while any frontline trial was enrolling at the center. We additionally included only patients with complete data for all covariates in this analysis. Statistical Analysis: Demographic and clinical characteristics by trial enrollment status were compared using t-test or Fisher's exact test as appropriate. Cox proportional hazards regression estimated hazard ratios (HR) with a doubly robust estimator (adjusted models only) of 4-year OS and EFS by trial enrollment status. A stepwise propensity score (PS) approach assessed the impact of potential drivers of the associations between trial enrollment and overall survival (OS) and event-free survival (EFS). PS were computed by logistic regression and generated stabilized inverse probability weights used for adjustment. First, a Base PS model including a predefined set of potential confounders (age, sex, weight category, preferred language, risk classification, mean annual hospital AML volume) was created. Additional variables of interest were then sequentially added to the PS. The change in adjusted HR relative to the unadjusted comparison was used to quantify the proportion of the association explained at each step. Results: 342 pediatric patients diagnosed with AML with representation from three distinct frontline clinical trials (AAML1031, AML08, and AML16) were included. Patients who were Black (11% vs. 24%), Hispanic (24% vs. 28%), publicly insured (44% vs. 55%), higher acuity (3% vs. 8%), and high-risk classification (29% vs. 34%) were underrepresented in the trial enrollment group compared to non-trial enrolled (Table 1). In unadjusted analyses, trial enrollment was significantly associated with superior OS (HR=0.69, 95% CI 0.47-0.99, p=0.04); there was no significant difference in EFS (HR=0.95, 95% CI 0.69-1.31, p=0.75). Adjustment for the Base PS accounted for 23% of the observed association between trial enrollment and OS (Figure 1). Acuity explained an additional 7% of the association. Inclusion of race/ethnicity and insurance did not result in further adjustment. In fully adjusted analysis, the association was substantially attenuated (HR=0.93, 95% CI 0.59-1.46, p=0.76). Conclusions: Significant sociodemographic and clinical differences exist in frontline trial enrollment for pediatric AML, which may contribute to worse overall survival outcomes for patients not enrolled on trials. Acuity at diagnosis may be a potential modifiable driver of survival differences. These findings support the need to evaluate enrollment patterns across pediatric clinical trials, including how enrollment disparities impact generalizability of clinical trial results, as well as a critical need for interventions designed to improve equity in healthcare access.

Distance to Care Predicts Clinical Trial Enrollment on COG ALL Clinical Trials: A Report from the Children's Oncology Group

Introduction: Equitable participation in clinical trials for all populations is essential to ensure the generalizability of clinical trial results. Historical data suggest that the majority of pediatric patients are enrolled on a trial when treated at a cooperative group institution and minoritized patients are proportionally represented. We proposed to evaluate clinical trial enrollment on Children's Oncology Group (COG) acute lymphoblastic leukemia (ALL) studies to determine if among those who participate in registry studies, there are disparities in participation in therapeutic clinical trials. We hypothesized that minoritized children, those living in disadvantaged neighborhoods, and those living distant from trial sites are less likely to enroll on therapeutic ALL studies. Methods: This study leverages the Childhood Cancer Research Network (CCRN), a childhood cancer case recruitment protocol that enrolled patients between 2008 and 2015. Within the CCRN registry cohort, we evaluated predictors of therapeutic and non-therapeutic clinical trial enrollment for patients with ALL. Patients >21 years at diagnosis and those without race or ethnicity reported were excluded. Enrollment on an upfront therapeutic trial (that overlapped with the timeframe during which the CCRN protocol was active) was defined as participation on one of the following 10 ALL trials: AALL0232, AALL0331, AALL0433, AALL0434, AALL0631, AALL0932, AALL1131, AALL1231, AALL1331, AALL1621. Enrollment on a non-therapeutic trial was defined as participation in any biology repository (including AALL03B1, AALL05B1, AALL08B1), supportive care, or late effects COG study. Potential predictors of interest included a composite of self-reported race and ethnicity, age at diagnosis, neighborhood socioeconomic status (SES) as measured by the Yost Index at the census tract, and distance to care (calculated as distance from home address to COG hospital). Poisson regression with log link and robust error variance was used to estimate adjusted risk ratios (aRR) and corresponding 95% confidence intervals (CI) for enrollment on a COG therapeutic clinical trial. Results: Of the 13,392 patients with ALL enrolled in the CCRN registry, 65% (n=8650) enrolled on a therapeutic COG clinical trial, 9% (1199) enrolled on a non-therapeutic trial without enrolling on a therapeutic trial, and 26% (3543) enrolled in the registry, but no other COG clinical trial. The majority of patient were between 1-9 years old and 58% were non-Hispanic (NH) White, 26% were Hispanic, and 7.2% were NH Black. Almost 11% of patients lived more than 100 miles from the COG center where they received treatment. Neither race nor ethnicity was associated with participation on a clinical trial (as shown in Table 1). Similarly, neighborhood SES was not associated with clinical trial participation. In multivariable models, age and distance to care were both meaningful predictors of therapeutic COG clinical trial enrollment with infants (aRR 0.64, 95%CI 0.57, 0.72) and adolescents (aRR 0.76, 95%CI 0.73, 0.80) less likely to enroll. Those that lived 100-250 miles from their COG center were 8% less likely to enroll (aRR 0.92, 95%CI 0.87, 0.97) and those that lived >250 miles from their COG center were 26% less likely to enroll (aRR 0.74, 95%CI 0.67, 0.81). Age was also a predictor of enrollment on non-therapeutic trials, but distance to care was not associated with enrollment on non-therapeutic trials. Conclusion: Among those who enroll on COG registry trials (which previous data suggests accounts for 55% of patients with ALL), race, ethnicity, and neighborhood SES are not associated with participation in therapeutic or non-therapeutic clinical trials. While age impacts risk stratification in ALL and thus may have impacted eligibility for various therapeutic trials, greater distance to care emerged as a relevant barrier to enrollment on therapeutic clinical trials. This novel finding in pediatric oncology suggests that interventions focused on ensuring access to clinical trials among those that live more than 100 miles from a COG site should be implemented. Approaches to address this disparity could include allowing study labs to be drawn locally and study visits to be conducted by telehealth when feasible.

Use of modernized eligibility criteria in pancreatic cancer clinical trials.

6545 Background: In 2017, ASCO + Friends recommended modernized eligibility criteria to improve generalizability of clinical trial results. Traditional criteria contribute to disparities in trial participation, yet modernized criteria can equalize eligibility rates. We aimed to assess use of modernized eligibility criteria in pancreatic ductal adenocarcinoma (PDAC) clinical trials since the updated guidelines. Methods: Modernized eligibility criteria were obtained from ASCO + Friends guidelines. Clinicaltrials.gov was queried for “pancreas adenocarcinoma” trials from Jan. 1, 2018 onward and eligibility criteria were assessed for compliance with modernized criteria. Kruskal-Wallis and Fisher’s exact tests were used to determine differences in compliance based on disease stage(s), starting year, phase, and funder. Results: In total, 108 clinical trials were identified - 31.5% for resectable, 56.5% for unresectable, 12.0% for both resectable/unresectable disease. Most trials were phase 1 (24.1%), phase 1|2 (15.7%) or phase 2 (45.4%). Funding sources included industry (38.0%), NIH (22.2%) and other sources (39.8%). No trial was compliant with all modernized criteria. Only 34.7% of trials allowed patients with ECOG performance status of 2. HIV+ patients were not allowed in 56.5% of trials. Only 7 trials (6.5%) were compliant with assessing renal function whereas 64.8% appropriately used upper limit of normal cutoffs for labs. Among trials that excluded patients for CHF, 59.6% used a classification system for exclusion. Specific cardiac abnormalities were excluded in 60.2% of trials. Prior malignancy was allowed in 71.3% of trials if treated > 2 years prior and concurrent malignancy, if stable, was allowed in 21.3% of trials. Prior PDAC treatment was allowed for 38.0% and required for 24.1% of trials. If prior treatment was allowed, 94% justified drug exclusions, 40.3% did not require a washout period, and 35.8% required recovery from prior adverse events. Stable/treated brain metastases were allowed in 33.8% of trials for unresectable cancer. Compliance with prior malignancy criteria improved from 2018-2019 to 2020-2022 (56.0% vs 84.5%, respectively; p = 0.001). Requirement of prior PDAC treatment was associated with disease stage (8.8% for resectable, 32.8% for unresectable, and 23.1% for both; p = 0.024), as was exclusion based on prior therapy (58.8% for resectable disease, 14.8% for unresectable, and 30.8% for both; p < 0.0001). Earlier trial phase was associated with prior treatment allowance (p = 0.002), whereas later phase trials were less likely to allow prior treatment (p = 0.021). NIH-funded trials required prior treatment (37.5%) more than industry (29.3%) or other (11.6%) funded trials (p = 0.034). Conclusions: Modernized eligibility criteria are inadequately implemented in PDAC clinical trials. Efforts to improve their use are warranted to produce more generalizable results and advance inclusivity.

Predicting Treatment Response with Sensory Phenotyping in Post-Traumatic Neuropathic Pain.

Currently available treatments for neuropathic pain are only modestly efficacious when assessed in randomized clinical trials and work for only some patients in the clinic. Induced-pain or gain-of-function phenotypes have been shown to predict response to analgesics (vs placebos) in patients with neuropathic pain. However, the predictive value of these phenotypes has never been studied in post-traumatic neuropathic pain. Mixed-effects models for repeated measures were used to evaluate the efficacy of pregabalin vs placebo in subgroups with induced-pain phenotypes (i.e., hyperalgesia or allodynia) in data from a recent, multinational randomized clinical trial (N = 539) that identified phenotypic subgroups through the use of a structured clinical exam. The difference in mean pain score between the active and placebo groups (i.e., delta) after 15 weeks of treatment for the subgroup with hyperalgesia was -0.76 (P = 0.001), compared with 0.19 (P = 0.47) for the subgroup that did not have hyperalgesia. The treatment-by-phenotype interaction, which tests whether subgroups have statistically different treatment responses, was significant (P = 0.0067). The delta for the subgroup with allodynia was -0.31 (P = 0.22), compared with -0.30 (P = 0.22) for the subgroup that did not have allodynia (treatment-by-phenotype interaction P = 0.98). These data suggest that hyperalgesia, but not allodynia, predicts response to pregabalin in patients with chronic post-traumatic neuropathic pain. This study extends the growing data supporting the utility of induced-pain phenotypes to predict response to analgesics in post-traumatic neuropathic pain. Sensory phenotyping in large, multisite trials through the use of a structured clinical exam has the potential to accelerate the development of new analgesics and improve the generalizability of clinical trial results.

Comparison of Racial, Ethnic, and Geographic Location Diversity of Participants Enrolled in Clinic-Based vs 2 Remote COVID-19 Clinical Trials

Racial and ethnic diversity among study participants is associated with improved generalizability of clinical trial results and may address inequities in evidence that informs public health strategies. Novel strategies are needed for equitable access and recruitment of diverse clinical trial populations. To investigate demographic and geographical location data for participants in 2 remote COVID-19 clinical trials with online recruitment and compare with those of a contemporaneous clinic-based COVID-19 study. This cohort study was conducted using data from 3 completed, prospective randomized clinical trials conducted at the same time: 2 remotely conducted studies (the Early Treatment Study and Hydroxychloroquine COVID-19 Postexposure Prophylaxis [PEP] Study) and 1 clinic-based study of convalescent plasma (the Expanded Access to Convalescent Plasma for the Treatment of Patients With COVID-19 study). Data were collected from March to August 2020 with 1 to 28 days of participant follow-up. All studies had clinical sites in Seattle, Washington; the 2 remote trials also had collaborating sites in New York, New York; Syracuse, New York; Baltimore, Maryland; Boston, Massachusetts; Chicago, Illinois; New Orleans, Louisiana; and Los Angeles, California. Two remote trials with inclusive social media strategies enrolled 929 participants with recent SARS-CoV-2 exposure (Hydroxychloroquine COVID-19 PEP Trial) and 231 participants with COVID-19 infection (Early Treatment Study); the clinic-based Expanded Access to Convalescent Plasma for the Treatment of Patients With COVID-19 study enrolled 250 participants with recent COVID-19 infection. Data were analyzed from April to August 2021. Remote trials used inclusive social media strategies and clinician referral for recruitment and telehealth, courier deliveries, and self-collected nasal swabs for remotely conducted study visits. For the clinic-based study, participants were recruited via clinician referral and attended in-person visits. Google Analytics data were used to measure online participant engagement and recruitment. Participant demographics and geographical location data from remote trials were pooled and compared with those of the clinic-based study. Statistical comparison of demographic data was limited to participants with COVID infections (ie, those in the remotely conducted Early Treatment Study vs those in the clinic-based study) to improve accuracy of comparison given that the Hydroxychloroquine COVID-19 PEP Trial enrolled participants with COVID-19 exposures and thus had different enrollment criteria. A total of 1410 participants were included. Among 1160 participants in remote trials and 250 participants in the clinic-based trial, the mean (range) age of participants was 39 (18-80) years vs 50 (19-79) years and 676 individuals (58.3%) vs 131 individuals (52.4%) reported female sex. The Early Treatment Study with inclusive social media strategies enrolled 231 participants in 41 US states with increased rates of racial, ethnic, and geographic diversity compared with participants in the clinic-based study. Among 228 participants in the remotely conducted Early Treatment Study with race data vs participants in the clinic-based study, 39 individuals (17.1%) vs 1 individual (0.4%) identified as Alaska Native or American Indian, 11 individuals (4.8%) vs 22 individuals (8.8%) identified as Asian, 26 individuals (11.4%) vs 4 individuals (1.6%) identified as Black, 3 individuals (1.3%) vs 1 individual identified as Native Hawaiian or Pacific Islander, 117 individuals (51.3%) vs 214 individuals (85.6%) identified as White, and 32 individuals (14.0%) vs 8 individuals (3.2%) identified as other race (P < .001). Among 230 individuals in the Early Treatment Study vs 236 individuals in the clinic-based trial with ethnicity data, 71 individuals (30.9%) vs 11 individuals (4.7%) identified as Hispanic or Latinx (P<.001). There were 29 individuals in the Early Treatment Study with nonurban residences (ie, rural, small town, or peri-urban; 12.6%) vs 6 of 248 individuals in the clinic-based trial with residence data (2.4%) (P < .001). In remote trial online recruitment, the highest engagement was with advertisements on social media platforms; among 125 147 unique users with age demographics who clicked on online recruitment advertisements, 84 188 individuals (67.3%) engaged via Facebook. These findings suggest that remote clinical trials with online advertising may be considered as a strategy to improve diversity among clinical trial participants.

Estimating attrition in mild‐to‐moderate Alzheimer’s disease and mild cognitive impairment clinical trials

AbstractBackgroundParticipant retention directly affects the validity and generalizability of clinical trial results. Investigators need information about expected attrition when planning trials of varying durations.MethodUsing the Alzforum and clinicaltrials.gov databases we reviewed 351 phase 2 and 3 placebo‐controlled trials conducted since 1998. Eighty‐six mild‐to‐moderate Alzheimer’s disease (AD) and eleven Mild Cognitive Impairment (MCI) trials met criteria for inclusion in this study. We first assessed the frequency with which trial arms differed in overall retention. When trials included more than one active arm, we combined those arms into a single group. To model associations between trial duration and retention, we used binomial regression.ResultFigure 1 illustrates differences in the observed proportion of participants completing in each trial between active and placebo arms for mild‐to‐moderate AD (Figure 1A) and MCI trials (Figure 1B). For each trial, a 95% confidence interval for the true difference between arms is provided. Since there was no consistent difference between arms, subsequent analyses combined active and placebo arms. Figure 2 illustrates the completion rate by trial duration for mild‐to‐moderate AD (Figure 2A) and MCI trials (Figure 2B). A local smoother along with uncertainty regions is added for visual aid. In general, trials with longer study duration had lower retention rates. Using binomial regression with robust variance estimates to account for within‐trial correlation, we estimated that a 6‐month increase in trial duration is associated with a 34% decrease in the odds of trial completion (OR=0.66; 95% CI: 0.58, 0.75; p&lt;0.001) among mild‐to‐moderate AD trials and a 20% decrease (OR=0.80; 95% CI: 0.68, 0.93; p=0.004) among MCI trials. From the binomial regression model, the proportion of participants completing 6, 12, and 18‐month trials were estimated to be 82.9%, 76.3%, and 68.0% for mild‐to‐moderate AD trials and 82.6%, 79.1%, and 75.2% for MCI trials, respectively.ConclusionWhile decisions related to expected trial attrition are guided by several factors, these estimates may assist investigators designing trials and suggest that trial attrition is frequent and related to trial duration.

Novel classifications for systemic sclerosis: challenging historical subsets to unlock new doors.

Systemic sclerosis (SSc) is a severe rheumatic disease characterized by a considerable heterogeneity in clinical presentations and pathophysiological mechanisms. This variability has a substantial impact on morbidity and mortality and limits the generalizability of clinical trial results. This review aims to highlight recent studies that have proposed new innovative approaches to decipher this heterogeneity, in particular, by attempting to optimize disease classification. The historical dichotomy limited/diffuse subsets based on cutaneous involvement has been challenged by studies highlighting an underestimated heterogeneity between these two subtypes and showing that presence of organ damage and autoantibody profiles markedly influenced survival beyond skin extension. Advanced computational methods using unsupervised machine learning analyses of clinical variables and/or high-throughput omics technologies, clinical variables trajectories modelling overtime or radiomics have provided significant insights on key pathogenic processes that could help defining new subgroups beyond the diffuse/limited subsets. We can anticipate that a future classification of SSc patients will integrate innovative approaches encompassing clinical phenotypes, variables trajectories, serological features and innovative omics molecular signatures. It nevertheless seems crucial to also pursue the implementation and standardization of readily available and easy to use tools that can be used in clinical practice.

Evaluation of the representativeness and generalizability of Japanese clinical trials for localized rectal/colon cancer: Comparing participants in the Japan Clinical Oncology Group study with patients in Japanese registries

IntroductionIt is unclear if clinical trial results including patients who meet trial eligibility criteria, are applicable to actual patients in daily practice (generalizability). Moreover, the extent to which are trial participants different from patients seen in daily practice (representativeness) is also unclear. The aim of this study was to evaluate the representativeness of the patients registered in randomized clinical trials to patients in daily practice and examine the generalizability of trial results to daily practice. MethodsWe compared the results of surgical trials conducted by the Japan Clinical Oncology Group with data from two Japanese cancer registries, representing patients seen in daily practice. We compared overall survival (OS) between trial participants and registry patients to evaluate representativeness of trial participants. We then compared the OS of registry patients who received open surgery (OP) and laparoscopic surgery (LAP) to evaluate the generalizability of trial results. ResultsWe analyzed 3051 patients (701 in JCOG0212, 2350 registry patients) with rectal cancer and 3116 patients (1057 in JCOG0404, 2059 registry patients) with colon cancer. Trial participants tended to possess lower clinical stages. Multivariable analyses revealed registry patients with significantly worse survival compared with trial participants. The hazard ratio of LAP to OP among registry patients was 0.305 (95% CI; 0.048–2.188), which did not meet the prespecified generalizability criteria of 0.9. ConclusionsOur results failed to ensure either the representativeness or generalizability of clinical trial results, compared to daily practice. Careful considerations are required when applying trial results to patients in daily practice.

Abstract P1-19-20: Safety of palbociclib in African American women with hormone receptor positive HER2 negative advanced breast cancer and benign ethnic neutropenia: PALINA study

Abstract Background: There are significant racial disparities in breast cancer outcomes between African American women (AAW) and Caucasian women (CW) in the US. It has been reported that AAW have more chemotherapy dose reductions than CW which results in worse clinical outcomes (Schneider BP et al. JCO Precis Oncol 2017). Given the high incidence of benign ethnic neutropenia (BEN) in AAW, we sought to evaluate the hematological safety of palbociclib in this group of patients. A polymorphism (SNP rs2814778 at chromosome 1q23.2) in the Duffy Antigen Receptor Chemokine (DARC) gene is implicated in the pathophysiology of BEN. AAW have been historically underrepresented in clinical trials including studies that led to the FDA approval of palbociclib. This diversity gap can compromise the generalizability of clinical trial results and further exacerbate minority health disparities. Methods: PALINA is a phase II study that evaluated the hematological safety of palbociclib with endocrine therapy (ET) in 35 AAW with HR positive HER2 negative advanced breast cancer (ABC) and ANC ≥1,000 cells/mm3. Eligible patients were self-identified as African, African-American or Black, PS 0-2 and had not received a prior CDK4/6 inhibitor (CDK4/6i). Treatment included palbociclib 125mg daily for 21 days followed by 7 days off and either an aromatase inhibitor (AI) or fulvestrant per investigator discretion. Maximum time allowed on study was 12 months. Responding patients could continue palbociclib off trial. Presence of the Duffy null polymorphism as a predictive marker for neutrophil count was assessed at baseline. Metabolite and exosomal signature (proteins and RNA) of drug resistance were also evaluated at different time points and will be reported separately. Primary endpoint was the proportion of patients who completed planned oncologic therapy without the development of a hematological event defined as episodes of febrile neutropenia (FN) or treatment discontinuation due to neutropenia. A two-stage design was used to test if the completion rate of planned oncologic therapy without a hematological event was at least 80% versus if it was below 60%, with 80% power at a significance level of 5%. Results: 35 women were enrolled from 5 different institutions. Mean age was 64 years (30-90). 9% (3/35) of patients had PS of 2. Visceral involvement was present in 51% (18/35) of patients; 60% (21/35) received an AI and 40% (14/35) received fulvestrant. Baseline median ANC was 3,100 (1,300-11,100) cells/mm3. None of the patients had FN or required discontinuation of therapy due to neutropenia. Grade 3 and 4 neutropenia were experienced by 46% (16/35) and 3% (1/35) of patients respectively. Dose delays occurred in 17 patients, and 13 patients required dose reduction (5 to 100mg and 8 to 75 mg). Clinical benefit rate (CR+PR+SD persisting for ≥6 months) was 70%. Of the 29 patients who have completed study participation, the median time on treatment was 280 days (14, 385) and 15 continued on commercial palbociclib. Last visit on study estimated for 10/2019. The Duffy polymorphism was evaluated in 94% (33/35) of patients. The Duffy null phenotype was present in 58% (19/33) of the patients. Lower baseline ANC (2,400 vs 4,300 cells/mm3, p0.006), grade 3 neutropenia (63.2 vs 21.4%, p0.003) and dose reductions (52.6 vs 7.1%, p0.009) were more common in patients with the Duffy null polymorphism. Conclusion: This is the first trial specifically designed to evaluate the hematologic toxicity of a CDK4/6i in AAW. Although patients with the Duffy null phenotype had more dose reductions and treatment delays due to neutropenia, this did not result in episodes of FN or treatment discontinuation due to neutropenia. This analysis suggests that palbociclib with ET is safe in AAW including those with the Duffy null phenotype. Citation Format: Filipa Lynce, Rebecca Zhuo, Matthew Blackburn, Christopher Gallagher, Olwen Hahn, Maysa Abu-Khalaf, Mahsa Mohebtash, Tianmin Wu, Paula Pohlmann, Asma Dilawari, Shruti Tiwari, Ami Chitalia, Robert Warren, Ming Tan, Ayesha Shajahan-Haq, Claudine Isaacs. Safety of palbociclib in African American women with hormone receptor positive HER2 negative advanced breast cancer and benign ethnic neutropenia: PALINA study [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P1-19-20.

Generalizability of Clinical Trial Results for Adolescent Major Depressive Disorder.

Although there have been a number of clinical trials evaluating treatments for adolescents with major depressive disorder (MDD), the generalizability of those trials to samples of depressed adolescents who present for routine clinical care is unknown. Examining the generalizability of clinical trials of pharmacological and psychotherapy interventions for adolescent depression can help administrators and frontline practitioners determine the relevance of these studies for their patients and may also guide eligibility criteria for future clinical trials in this clinical population. Data on nationally representative adolescents were derived from the National Comorbidity Survey: Adolescent Supplement. To assess the generalizability of adolescent clinical trials for MDD, we applied a standard set of eligibility criteria representative of clinical trials to all adolescents in the National Comorbidity Survey: Adolescent Supplement with a Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition diagnosis of MDD (N = 592). From the overall MDD sample, 61.9% would have been excluded from a typical pharmacological trial, whereas 42.2% would have been excluded from a psychotherapy trial. Among those who sought treatment (n = 412), the corresponding exclusion rates were 72.7% for a pharmacological trial and 52.2% for a psychotherapy trial. The criterion leading to the largest number of exclusions was "significant risk of suicide" in both pharmacological and psychotherapy trials. Pharmacological and, to a lesser extent, psychotherapy clinical trials likely exclude most adolescents with MDD. Careful consideration should be given to balancing eligibility criteria and internal validity with applicability in routine clinical care while ensuring patient safety.

Computer-aided assessment of the generalizability of clinical trial results

BackgroundThe effects of an intervention on patients from populations other than that included in a trial may vary as a result of differences in population features, treatment administration, or general setting. Determining the generalizability of a trial to a target population is important in clinical decision making at both the individual practitioner and policy-making levels. However, awareness to the challenges associated with the assessment of generalizability of trials is low and tools to facilitate such assessment are lacking. MethodsWe review the main factors affecting the generalizability of a clinical trial results beyond the trial population. We then propose a framework for a standardized evaluation of parameters relevant to determining the external validity of clinical trials to produce a “generalizability score”. We then apply this framework to populations of patients with heart failure included in trials, cohorts and registries to demonstrate the use of the generalizability score and its graphic representation along three dimensions: participants’ demographics, their clinical profile and intervention setting. We use the generalizability score to compare a single trial to multiple “target” clinical scenarios. Additionally, we present the generalizability score of several studies with regard to a single “target” population. ResultsSimilarity indices vary considerably between trials and target population, but inconsistent reporting of participant characteristics limit head-to-head comparisons. ConclusionWe discuss the challenges involved in performing automatic assessment of trial generalizability at scale and propose the adoption of a standard format for reporting the characteristics of trial participants to enable better interpretation of their results.

In and Out, Good and Bad News, of Generalizability of SWOG Treatment Trial Results

The issue of generalizability of clinical trial results has challenged researchers and practitioners of adult cancer patients for years. It is one reason why the adult cancer cooperative group enterprise in the United States is undergoing yet another transformation (1). Figure 1 may capture the essence of the situation. Its foreground shows the most recent data available to this commentator on the estimated proportion of cancer patients in the United States entered onto cooperative group treatment trials as a function of individual year of patient age. The background shows the rest of the United States. With only 2% to 4% of adults aged 20 to 65 years entered onto clinical trials, how can we rely on such a small sample size to be applicable to the general population? In a comparison of survival outcomes among cancer patients treated in and out of clinical trials published in this issue of the Journal, Unger and colleagues analyzed 21 SWOG treatment trials enrolled during the period from 1986 to 2007 with comparable patients in the US Surveillance, Epidemiology and End Results (SEER) database (2). Their primary subset analysis was good- vs poor-prognosis trials defined by a 2-year average study-specific survival rate of greater than 50% or less than or equal to 50%, respectively. The former had 11 trials that included melanoma, myeloma, and carcinoma of the breast, bladder, uterine cervix, and prostate. The latter had 10 trials that included acute myelogenous leukemia, brain tumors, lung cancer, and pancreatic carcinoma. The results may be interpreted as a series of good- vs bad-news scenarios (Table 1), with several advantages as a source of disadvantage and vice versa. Among the best news is that patients in SWOG trials had a survival benefit, albeit only in poor-prognosis trials. The bad news for them was that it “endured for only 1 year” (2). However, according to Supplementary Figure 2 in the Unger et al. report (2), which shows the SWOG and SEER survival curves for each of 10 poor-prognosis trials, the survival benefit is better than the authors concluded, lasting at least 5 years in two trials, 3 years in three trials, and 2 years in two trials. In only two of the 10 trials was it less than 1 year. Whether or not the apparent benefit of the SWOG trial in patients with a poor prognosis is more durable than concluded, the striking difference between the broad categories of good and poor prognosis as defined has major implications for the interpretation of cancer clinical trial outcomes and application to the general population:

Generalizability of Clinical Trial Results for Bipolar Disorder to Community Samples

Research on the generalizability of clinical trial results for bipolar disorder is limited. The present post hoc study sought to quantify the generalizability of clinical trial results in individuals with DSM-IV bipolar disorder to a large representative community sample. Data were derived from the 2001-2002 National Epidemiologic Survey on Alcohol and Related Conditions (NESARC), a large, nationally representative sample of 43,093 adults from the United States population. We applied a standard set of eligibility criteria representative of clinical trials to all adults with DSM-IV bipolar depression (n = 785) or mania (n = 724) in the past 12 months and then to a subsample of participants seeking treatment for bipolar depression (n = 276). Our aim was to determine the proportion of participants with bipolar depression or acute mania who would have been excluded from a clinical trial by typical eligibility criteria. We found that more than 5 of 10 participants with bipolar depression (58.17%) or mania (55.75%) would have been excluded by at least 1 eligibility criterion. In the subgroup of participants with bipolar depression who sought treatment, the exclusion rate by at least 1 criterion was higher (63.87%). Having a significant risk of suicide was the criterion excluding the highest percentage of participants in the bipolar depression samples, while having a current DSM-IV diagnosis of alcohol abuse or dependence was the one leading to the greatest exclusion rate in clinical trials for participants with acute mania. Exclusion rates were higher for participants with bipolar I depression compared with those with bipolar II depression. Traditional clinical trials tend to exclude a majority of individuals with bipolar disorder. Clinical trials should carefully consider the impact of eligibility criteria on the generalizability of their results and explain the rationale for their use. Future trials should weigh the trade-offs between internal validity and the representativeness of the study.

Clinical trial variability: Quality control in a randomized clinical trial

IntroductionA major issue in clinical trials in manual medicine is treatment variability. The challenge is to insure that the bounded treatment options are both representative of field practitioner behavior and consistent among research clinicians. This investigation assesses the treatment comparability of field practitioners and research clinicians, for a flexion-distraction treatment procedure, as quality control for a randomized clinical trial. MethodsUsing a series of vignettes, we studied the level of agreement of treatment protocols between field clinicians, research clinicians and a reference clinician regarding treatment location, range of motion during treatment, and number of repetitions used within the flexion-distraction protocol. ResultsResults indicated that reliability around decision making for anticipated location of spinal treatment was highest regardless of clinician group. For the research clinicians this level of agreement was ICC=0.88. Decision-making for treatment direction was second highest, at kappa=0.64 for the research clinicians. Reliability around the number of repetitions is poor ranging from ICC=0.18 to 0.34 depending on clinician type. DiscussionUnderstanding the disparity in treatment protocols is of value in the construction and maintenance of quality control in an actual randomized clinical trial setting. More work was recommended in the preparation of clinical trials and the understanding of clinical decision-making because these disparate factors may dramatically impact the generalizability of clinical trial results.

Generalizability of Clinical Trial Results for Major Depression to Community Samples

Although emerging data indicate that sample composition may influence the effectiveness of mental health interventions, the extent to which subjects in clinical trials represent affected community samples remains unknown. The goal of this study was to assess the proportion of community-dwelling adults with major depressive episode (MDE) who would meet eligibility criteria for a traditional efficacy trial in patients with MDE. We applied a standard set of exclusion criteria used in clinical trials in patients with MDE to the 2001-2002 National Epidemiologic Survey for Alcohol and Related Conditions (NESARC), the largest psychiatric epidemiologic study in the United States to date (N = 43,093). Because individuals who seek treatment for a disorder may systematically differ from those who do not, we applied the criteria first to all individuals with a current diagnosis of MDE (N = 3119) (diagnosed according to DSM-IV) and then to the subsample of individuals who sought treatment (N = 1359). Among the full sample of individuals with MDE, 75.8% were excluded by one or more study eligibility criteria. Approximately two thirds (66.9%) of the subsample of those who sought treatment were excluded. The percentage of subjects excluded by individual study criteria ranged from 2.4% to 47.4% in the overall sample and 0% to 38.4% in the treatment-seeking sample. For both groups, the presence of comorbid, nondepressive, non-substance use Axis I disorders and the duration of the depressive episode excluded the largest percentage of individuals. The design of traditional clinical trials tends to exclude a majority of individuals with MDE. Selection of exclusion criteria may have a powerful influence on the generalizability of study results. Clinical trials should explain the rationale for their exclusion criteria and estimate the impact of eligibility criteria on the generalizability of trial results.

Generalizability of cancer clinical trial results

BACKGROUND The generalizability of clinical trial results is questionable, because fewer than 5% of cancer patients participate. The authors examined the comparability of clinical trial participants and nonparticipants and the potential impact of differences. METHODS A retrospective cohort of 19,340 cancer patients who were diagnosed between January 1990 and December 1997 was characterized by trial participation. The distributions of prognostically important factors among trial participants were compared with the distributions among nonparticipants and the population of patients diagnosed during the same period in the Surveillance, Epidemiology, and End Results population. The impact of these factors on survival was examined by using a Cox proportional hazards analysis. RESULTS Trial participants were younger and had better performance status and fewer comorbid conditions compared with nonparticipants. However, participants were more likely to have locally advanced disease, positive lymph node status, poorly differentiated tumors, liver metastases, and multiple metastatic sites. The former factors were associated with significantly longer survival, whereas the later factors were associated with significantly shorter survival. CONCLUSIONS The lack of comparability between trial participants and nonparticipants called into question the generalizability of clinical trial results. Although selective recruitment for clinical trials is justified, the authors encourage the use of population-based trials of effectiveness in “all comers.” Cancer 2006. © 2006 American Cancer Society.

Generalizability of cancer clinical trial results

The generalizability of clinical trial results is questionable, because fewer than 5% of cancer patients participate. The authors examined the comparability of clinical trial participants and nonparticipants and the potential impact of differences. A retrospective cohort of 19,340 cancer patients who were diagnosed between January 1990 and December 1997 was characterized by trial participation. The distributions of prognostically important factors among trial participants were compared with the distributions among nonparticipants and the population of patients diagnosed during the same period in the Surveillance, Epidemiology, and End Results population. The impact of these factors on survival was examined by using a Cox proportional hazards analysis. Trial participants were younger and had better performance status and fewer comorbid conditions compared with nonparticipants. However, participants were more likely to have locally advanced disease, positive lymph node status, poorly differentiated tumors, liver metastases, and multiple metastatic sites. The former factors were associated with significantly longer survival, whereas the later factors were associated with significantly shorter survival. The lack of comparability between trial participants and nonparticipants called into question the generalizability of clinical trial results. Although selective recruitment for clinical trials is justified, the authors encourage the use of population-based trials of effectiveness in "all comers."