The Contribution of “Individual Participant Data” Meta-Analyses of Psychotherapies for Depression to the Development of Personalized Treatments: A Systematic Review

While randomized trials typically lack sufficient statistical power to identify predictors and moderators of outcome, “individual participant data” (IPD) meta-analyses, which combine primary data of multiple randomized trials, can increase the statistical power to identify predictors and moderators of outcome. We conducted a systematic review of IPD meta-analyses on psychological treatments of depression to provide an overview of predictors and moderators identified. We included 10 (eight pairwise and two network) IPD meta-analyses. Six meta-analyses showed that higher baseline depression severity was associated with better outcomes, and two found that older age was associated with better outcomes. Because power was high in most IPD meta-analyses, non-significant findings are also of interest because they indicate that these variables are probably not relevant as predictors and moderators. We did not find in any IPD meta-analysis that gender, education level, or relationship status were significant predictors or moderators. This review shows that IPD meta-analyses on psychological treatments can identify predictors and moderators of treatment effects and thereby contribute considerably to the development of personalized treatments of depression.

Hydroxychloroquine/Chloroquine for the Treatment of Hospitalized Patients with COVID-19: An Individual Participant Data Meta-Analysis

Importance: Results from observational studies and randomized clinical trials (RCTs) have led to the consensus that hydroxychloroquine (HCQ) and chloroquine (CQ) are not effective for COVID-19 prevention or treatment. Pooling individual participant data (IPD), including unanalyzed data from trials terminated early, enables further investigation of the efficacy and safety of HCQ/CQ. Objective: To assess efficacy of HCQ/CQ in patients hospitalized with COVID-19, both overall and in prespecified subgroups. Data Sources: ClinicalTrials.gov was searched multiple times in May-June 2020. Principal investigators of US-based RCTs evaluating HCQ/CQ in hospitalized COVID-19 patients were invited to collaborate in this IPD meta-analysis. Study Selection: RCTs in which: (1) HCQ/CQ was a treatment arm; (2) patient informed consent and/or individual study IRB approval allowed for data sharing; (3) principal investigators/their institutions signed a data use agreement for the present study; and (4) the outcomes defined in this study were recorded or could be extrapolated. Data Extraction and Synthesis: Wherever possible, harmonized de-identified data were collected via a common template spreadsheet sent to each principal investigator, then shared via a secure online data sharing platform to create a pooled data set. When this was not possible, individual study data were harmonized and merged manually. Data were analyzed by fitting a prespecified Bayesian ordinal regression model and standardizing the resulting predictions. Main Outcome(s) and Measure(s): 7-point ordinal scale, measured between day 28 and 35 post-enrollment. Results: Eight of 19 trials met eligibility criteria and agreed to participate. Patient-level data were available from 770 participants (412 HCQ/CQ vs 358 control). Baseline characteristics were similar between groups. We found no evidence of a difference in ordinal scores between days 28 and 35 post-enrollment in the pooled patient population (odds ratio, 0.97; 95% credible interval, 0.76-1.24; higher favors HCQ/CQ), and no convincing evidence of meaningful treatment effect heterogeneity among prespecified subgroups. Adverse event and serious adverse event rates were numerically higher with HCQ/CQ vs control (0.39 vs 0.29 and 0.13 vs 0.09 per patient, respectively). Conclusions and Relevance: The findings of this IPD meta-analysis reinforce those of individual RCTs that HCQ/CQ is not efficacious for treatment of COVID-19 in hospitalized patients.

Concurrent chemoradiotherapy with cisplatin + S-1 versus cisplatin + other third-generation agents for locally advanced non-small-cell lung cancer: a meta-analysis of individual participant data

Background For decades, concurrent chemo-radiotherapy with cisplatin-based regimen has been a standard therapy for locally advanced stage III non-small-cell lung cancer (NSCLC). We conducted individual-participant-data (IPD) meta-analyses to compare S-1/cisplatin versus other third-generation anti-cancer medications plus cisplatin regimens with the goal of determining whether or not S-1/cisplatin was the ideal choice for treatment accompanied by radiotherapy (RT). Methods A thorough search was performed using multiple electronic databases. We integrated the IPD of each trial and analyzed the resulting meta-database. The primary endpoint was the overall survival (OS), and the secondary endpoints included the progression-free survival (PFS), objective response rate (ORR), toxicities, and treatment delivery. Subgroup analyses were conducted based on baseline characteristics. Statistical analyses were stratified by trials. Results Three randomized control trials (WJOG5008L study, SPECTRA study, and TORG1018 study) were found. Of the 316 patients enrolled in those studies, 159 received S-1/cisplatin (SP), and 157 were assigned to other combination chemotherapy. The median OS for the SP arm was 48.2 months, and that of the non-SP arm was 42.4 months. The combined hazard ratio (HR) for the OS was 0.895 (95% confidence interval [CI] 0.638–1.256), and no heterogeneity was noted among the trials (test for heterogeneity, p = 0.87; I2 = 0). The median PFS for the SP and non-SP arms was 12.8 and 14.0 months, respectively. The corresponding HR for the PFS was 1.022 (95% CI 0.776–1.347), and there was evidence of moderate heterogeneity among the trials (test for heterogeneity, p = 0.16; I2 = 0.46). The ORRs were 69.7% (95% CI 62.1–76.7%) and 70.9% (95% CI 63.7–78.1%) in the SP and non-SP arms, respectively. The toxicity profile showed that SP caused significantly fewer instances of grade 3–4 leukopenia and neutropenia than non-SP regimens. Conclusion No marked differences were detected in the OS, PFS, or ORR between the SP and non-SP arms. SP had significantly less myelosuppression and better treatment compliance as a chemotherapy regimen for concurrent chemoradiation in locally advanced NSCLC than non-SP regimens.

Racial/Ethnic Disparities of Depressive Symptoms in the United States A Systematic Review and Meta-Analysis of Individual Participant Data

Although a growing body of research has documented racial/ethnic disparities in depressive symptoms in the United States, the precise magnitude of these differences is not known. We conducted a systematic review and meta-analysis of individual participant data to (1) estimate the average difference of depressive symptoms between Whites and racial/ethnic minorities, as well as differences between (i.e., Asian American, African American, Latinxs, Multiracial, Native American, other race) and within (i.e., Latinx: Central American, Cuban American, Mexican American, Puerto Rican, other Latinx) minority groups, and (2) determine if moderators account for these differences. We screened 2,425 nationally-representative studies from the Inter-university Consortium for Political and Social Research (ICPSR), and identified 127 datasets of studies conducted from 1971 to 2018. We included 73 datasets from 26 nationally-representative studies (N = 2,116,853). The average absolute difference was d = 0.09, 95% CI [0.07, 0.12] between White and minority participants; was d = 0.07, 95% CI [0.06, 0.09] between minority participants; and d = 0.10, 95% CI [0.06, 0.15] within minority Latinx participants. Increases in socioeconomic status exacerbated these disparities. Psychometric analyses showed that measure reliability was related to larger differences. We discuss the implications of these findings.

On the Benefits of Speech-Language Therapy for Individuals Born With Cleft Palate: A Systematic Review and Meta-Analysis of Individual Participant Data

Purpose: Cleft lip and/or palate (CLP) is a common birth defect, and after reconstructive surgery, about 50% of children at 5 years of age have speech deviations and are referred to speech-language therapy (SLT). The peer-reviewed evidence for the benefit of SLT has been uncertain. Our objective was to systematically review and meta-analytically summarize the benefit of SLT for individuals born with CLP. Method: A systematic search was conducted (last search on February 19, 2021) on studies evaluating SLT with pre and post measures on speech production, language ability, intelligibility, and/or patient-reported outcomes. We sought individual participant data (IPD) and evaluated on an individual level if the outcome measure had improved to a clinically relevant degree during SLT and if the outcome measure was on a level with peers or not after SLT. Meta-analyses and meta-regressions were applied to synthesize IPD across studies. Results: Thirty-four eligible studies were found. Nineteen studies provided IPD ( n = 343) for the main analysis on speech production. The synthesized information suggests that, during SLT, speech production improved to a clinically relevant degree for many individuals (95% CI [61%, 87%]) and that speech production was on a level with peers for some individuals after SLT (95% CI [10%, 34%]). Conclusions: The main strength of this meta-analysis is that we evaluated on an individual level pre- and post-intervention data based on considerations of clinical relevance. This approach allowed us to conclude that many individuals benefit from SLT and that further work on evaluating SLT in this patient group is meaningful. Supplemental Material https://doi.org/10.23641/asha.17700992