Nevin Manimala

Clin Trials. 2026 May 9:17407745261442586. doi: 10.1177/17407745261442586. Online ahead of print.

ABSTRACT

Covariate adjustment uses baseline prognostic variables to improve the precision of treatment effect estimates. Recent Food and Drug Administration guidance and scientific consensus emphasize three principles for its use, namely estimand-focused analyses, assumption-lean robustness, and fit-for-purpose variance estimation. Despite substantial methodological progress, practical guidance for trial practitioners remains fragmented. We review covariate adjustment strategies for continuous, discrete, and time-to-event endpoints in randomized trials that adhere to these three principles. We show how unadjusted estimators, as well as linear and non-linear adjusted estimators, can be viewed as special cases of the general augmented inverse probability weighting framework. For time-to-event endpoints, we describe how covariate adjustment can be applied to Kaplan-Meier estimators, log-rank tests, and estimation of the unconditional hazard ratio without altering the estimand or introducing additional assumptions. We also synthesize recent developments in multi-arm trials, covariate-adaptive randomization, data-adaptive covariate selection, and covariate adjustment in interim analyses, and we provide practical insights for implementation. Covariate-adjusted estimators target the same marginal estimands as unadjusted analyses but typically achieve greater efficiency. Linear adjustment with Analysis of Heterogeneous Covariance guarantees asymptotic efficiency gains under minimal assumptions. Augmented inverse probability weighting generalizes covariate adjustment to flexible modeling frameworks and remains consistent even under model misspecification. For survival analysis, covariate-adjusted versions of the log-rank test and Cox model improve power without altering the estimand or requiring additional assumptions. Properly accounting for covariate-adaptive randomization is essential for valid inference. The reviewed methods are implemented in the RobinCar family of R packages: RobinCar and RobinCar2. Covariate adjustment is a principled and practical approach for improving trial efficiency, aligned with current regulatory guidance. By adhering to the principles of estimand-focus, assumption-lean robustness, and fit-for-purpose variance estimation, practitioners can apply covariate adjustment with confidence across diverse trial settings. Further work on evaluating finite-sample performance and re-analyses of completed trials will deepen understanding of covariate adjustment in practice.

PMID:42104833 | DOI:10.1177/17407745261442586

Acta Med Port. 2026 May 4;39(5):332-339. doi: 10.20344/amp.24246. Epub 2026 May 4.

ABSTRACT

INTRODUCTION: This study aimed to describe and compare patient-reported outcome measures (PROMs) and objective clinical outcome measures (CROMs) in the treatment of age-related macular degeneration (AMD), exploring the concordance between these measures within a value-based healthcare (VBH) framework.

METHODS: This prospective, multicenter, observational, real-world study was conducted at three tertiary referral hospitals specializing in the treatment of neovascular AMD. Clinical outcomes (CROMs) and patient-reported outcomes (PROMs) were analyzed using the National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) questionnaire as a functional assessment tool. Data were collected at baseline and at three, six, and 12 months following initiation of intravitreal anti-vascular endothelial growth factor (anti-VEGF) therapy. Statistical analysis was primarily descriptive. The comparison between baseline and 12 months in the global NEI VFQ-25 score was performed using the Wilcoxon signed-rank test for paired samples. Concordance between CROMs and PROMs was assessed using the intraclass correlation coefficient (ICC).

RESULTS: A total of 235 eyes were included, receiving 2338 intravitreal injections. The mean age of participants was 81 years (SD = 8.57), and 55.8% were female. The mean baseline NEI VFQ-25 score was 67.83 (SD = 10.39). The median best-corrected visual acuity was 63 ETDRS letters (interquartile range [P25 – P75]: 41 – 75) at baseline, increasing to 65 letters at three months and remaining stable through 12 months of follow-up. The comparison between baseline and 12 months revealed a statistically significant difference in visual acuity (Wilcoxon signed-rank test, Z = 4.2; p < 0.001). A reduction in the proportion of patients classified as legally blind was observed, together with an increase in the proportion of patients in the reading-vision and driving-vision categories. At 12 months, 58.7% of patients reported stabilization or improvement in visual function on the NEI VFQ-25 questionnaire. Concordance between the variation in visual acuity and the variation in the global NEI VFQ-25 score showed good agreement between CROMs and PROMs (ICC = 0.76; p < 0.001).

CONCLUSION: The integrated analysis of CROMs and PROMs suggests that anti-VEGF treatment for neovascular AMD is associated with stabilization or improvement in visual acuity and patients’ perceived visual function. The implementation of the VBH-AMD model proved feasible in a real-world clinical setting, reinforcing the importance of integrating patient-centered measures into the evaluation of therapeutic outcomes.

PMID:42104825 | DOI:10.20344/amp.24246

Schizophr Bull. 2026 Apr 10;52(3):sbag067. doi: 10.1093/schbul/sbag067.

ABSTRACT

BACKGROUND: People with severe mental illness (SMI) are at an increased risk of infection mortality compared to the general population. Little is known about how this risk might differ across infection types, and the potential impact of sociodemographic and clinical factors. We investigated associations between SMI and infection mortality in a population-based cohort, examining variation by infection type and potential moderating factors.

STUDY DESIGN: This retrospective matched cohort study (January 1, 2000 to December 31, 2019) used national primary care data from the UK Clinical Practice Research Datalink linked with Office of National Statistics mortality data. Competing risks regression and cause-specific hazard models assessed risk of infection mortality in people with SMI versus non-SMI controls. We examined risk across different infection types and assessed the impact of sociodemographic and clinical factors.

STUDY RESULTS: Our cohort comprised 84 494 people with SMI matched on age, gender, and GP practice with 84 494 non-SMI controls. Fully adjusted models showed that people with SMI were more likely to die from any infection compared to non-SMI controls (adjusted hazards ratio (aHR) = 1.58, 95% CI, 1.44-1.74). Infection-specific analyses revealed increased risk of death from respiratory (aHR = 1.69, 95% CI, 1.51-1.89), gastrointestinal (aHR = 2.01, 95% CI, 1.16-3.48), and renal/urinary (aHR = 1.70, 95% CI, 1.32-2.19) infections in the SMI group.

CONCLUSIONS: People with SMI are at increased risk of infection mortality, especially from respiratory, gastrointestinal, and renal/urinary infections. We recommend prioritizing this group for preventative measures including influenza and pneumococcal vaccines.

PMID:42104801 | DOI:10.1093/schbul/sbag067

Schizophr Bull. 2026 Apr 10;52(3):sbag049. doi: 10.1093/schbul/sbag049.

ABSTRACT

BACKGROUND AND HYPOTHESIS: The early phase of psychosis is critical for interventions to modify long-term outcomes. It is unclear what proportion of individuals’ exhibit early persistent psychosis and the long-term implications.

STUDY DESIGN: An epidemiologic sample of individuals with acute psychosis was recruited at first admission and followed for 25 years. Early persistent psychosis was defined as presence of active psychosis for ≥90% of the days of the 4 years after first hospitalization for psychosis. Multivariable regression analyses were conducted, testing the association between baseline predictors and persistent psychosis, and between persistent psychosis and 25-year outcomes.

STUDY RESULTS: Out of 526 individuals (age = 27.4 ± 9.4 years, males = 56.8%, baseline schizophrenia/schizoaffective disorder = 30.0%), 101 (19.2%) had early persistent psychosis. At baseline, low premorbid cognitive performance (odds ratio (OR) = 2.08, 95% CI, 1.05-4.12), lower Global Assessment of Functioning (OR = 1.59, 95% CI, 1.16-2.13), low role function (OR = 1.49, 95% CI, 1.03-2.16) and worse social function (OR = 1.52, 95% CI, 1.03-2.22) were predictive of persistent psychosis. At 25-year follow-up (n = 307, 58.9%), early persistent psychosis was associated with worse avolition ($beta$=0.25, 95% CI, 0.14-0.35), more severe reality distortion ($beta$=0.19, 95% CI, 0.07-0.31), disorganization ($beta$=0.21, 95% CI, 0.09-0.32), worse social ($beta$=-0.18, 95% CI, -0.06 to -0.30), role ($beta$=-0.22, 95% CI, -0.09 to -0.34), and global function ($beta$=-0.28, 95% CI, -0.17 to -0.38), greater odds of being on public assistance (OR = 2.13, 95% CI, 1.15-3.95), lower odds of living independently (OR = 0.43, 95% CI, 0.23-0.80) or recovery (OR = 0.09, 95% CI, 0.02-0.38).

CONCLUSIONS: One in 5 individuals with first-episode psychosis had early persistent psychosis without clearly modifiable premorbid factors, and with strong associations with adverse long-term outcomes. Individuals experiencing early persistent psychosis require focused long-term interventions.

PMID:42104800 | DOI:10.1093/schbul/sbag049

Schizophr Bull. 2026 Apr 10;52(3):sbag060. doi: 10.1093/schbul/sbag060.

ABSTRACT

STUDY DESIGN: Improving Thinking through Everyday Self-Assessment Training combines 16 weeks of daily mobile task-based training in IA with weekly individual coaching in applying IA to everyday behaviors. Sixty individuals with diagnoses of schizophrenia or schizoaffective disorder participated in an open trial of iTEST with assessments at baseline, 8, 12, and 16 weeks. Primary outcomes included IA on 2 trained tasks (mobile verbal learning and facial emotion recognition tests) and 3 untrained tasks (verbal memory, emotion recognition, and executive functioning).

STUDY RESULTS: Improving Thinking through Everyday Self-Assessment Training showed strong feasibility, retaining 86.7% of participants, and strong adherence with an average daily mobile-training completion rate of 87%. In linear-mixed models with intent-to-treat data, statistically significant IA improvements were observed over time in both trained tasks and in 2 of the 3 untrained tasks (Cohen’s d’s = 0.5-1.28). Significant improvements were also observed in secondary outcomes of real-world function, positive symptoms, and depression.

CONCLUSIONS: This project provides the first data, to our knowledge, to demonstrate that IA in schizophrenia can be improved. Improving Thinking through Everyday Self-Assessment Training also represents one of just a few blended digital health interventions, including remote cognitive training, and may therefore serve as a blueprint for future intervention development.

PMID:42104799 | DOI:10.1093/schbul/sbag060

Schizophr Bull. 2026 Apr 10;52(3):sbag019. doi: 10.1093/schbul/sbag019.

ABSTRACT

BACKGROUND AND HYPOTHESIS: Schizophrenia treatment in young people involves complex pharmacological decisions, yet sex-specific prescribing patterns and pandemic impacts remain poorly understood. We hypothesized that prescribing trends would differ systematically by sex and show pandemic-related disruptions.

STUDY DESIGN: This population-based cohort study analyzed 8092 individuals aged 15-29 years with schizophrenia-spectrum disorders in Hong Kong (2011-2023) using electronic health records from the Hospital Authority system. We examined temporal trends in 11 medication subclasses using generalized least squares models with autoregressive correlation structures, sex differences using interaction terms, and COVID-19 impacts using interrupted time series (ITS) analysis with adjustment for age and comorbidity.

STUDY RESULTS: After covariate adjustment, all medication subclasses increased over time (0.05-3.71 percentage points annually), indicating universal treatment intensification. Males showed steeper increases than females in 5 subclasses after adjustment, with 18 of 21 specific agents increasing significantly more in males. Period-level pandemic comparisons showed minimal effects, but ITS analysis revealed substantial COVID disruptions in 5 medication subclasses namely oral first-generation antipsychotics, injectable second-generation antipsychotics, serotonin and norepinephrine reuptake inhibitor, Z-hypnotics, and benzodiazepines.

CONCLUSIONS: Young people with schizophrenia experienced universal treatment intensification with males receiving more intensive pharmacotherapy after controlling for confounders. The pandemic produced complex sex-specific disruptions masked by aggregate analyses. Whether these prescribing patterns represent appropriate individualization or systematic care variation remains unknown, highlighting the critical need for studies linking prescribing patterns to functional outcomes and quality of life.

PMID:42104796 | DOI:10.1093/schbul/sbag019

Schizophr Bull. 2026 Apr 10;52(3):sbag066. doi: 10.1093/schbul/sbag066.

ABSTRACT

BACKGROUND AND HYPOTHESIS: Cognitive models of schizophrenia-spectrum disorders (SSD) posit dysfunctional beliefs and cognitive biases as maintenance mechanisms of positive and negative symptoms. Although cognitive behavioral therapy (CBT) targets these processes, its effects on mechanism-level outcomes remain unclear. This review examined whether CBT modifies dysfunctional beliefs and cognitive biases in SSD using rigorous randomized evidence.

STUDY DESIGN: PRISMA 2020-compliant systematic review and meta-analysis (PROSPERO registered). Primary analyses were restricted to intention-to-treat (ITT) randomized controlled trials (RCTs) in SSD samples, using random-effects models and between-group post-treatment estimates. Pre-post and nonrandomized studies were analyzed separately as secondary evidence. Subgroup and meta-regression analyses were conducted.

STUDY RESULTS: Thirty-three studies met inclusion criteria. Fourteen ITT RCTs contributed to the primary pooled analysis of dysfunctional beliefs, yielding a small but statistically significant effect favoring CBT (g = 0.154, 95% CI, 0.049-0.259). Effects were strongest for delusional conviction (g = 0.450) and self-related schemas (positive-self g = 0.278; negative-self g = 0.298). Voice-related beliefs did not reach statistical significance. Too few RCTs assessed cognitive biases to support primary pooled analyses; exploratory findings suggested small effects for belief inflexibility and no reliable effect for jumping-to-conclusions. Greater reductions in dysfunctional beliefs were associated with greater improvements in positive symptoms across trials.

CONCLUSIONS: CBT produces small but reliable improvements in dysfunctional beliefs in SSD, although effects vary by domains particularly for delusional conviction and self-schemas, supporting their role as modifiable therapeutic targets and plausible mechanisms of change. Effects on cognitive biases remain limited and understudied.

PMID:42104795 | DOI:10.1093/schbul/sbag066

Schizophr Bull. 2026 Apr 10;52(3):sbag071. doi: 10.1093/schbul/sbag071.

ABSTRACT

BACKGROUND AND HYPOTHESIS: The requirement for hospital admission to initiate clozapine presents a health-systems-related barrier to clozapine prescription and contributes to its underutilization in treatment-resistant schizophrenia (TRS). This study aimed to examine the clinicodemographic characteristics associated with treatment settings for clozapine initiation within a first-episode psychosis (FEP) cohort attending an early intervention in psychosis service.

STUDY DESIGN: Secondary analysis of a retrospective cohort study of 1220 young people presenting with FEP to the Early Psychosis Prevention and Intervention Centre (EPPIC) in Melbourne between 2011 – 2017.

STUDY RESULTS: Ninety-one cases of TRS were identified and included in the analysis, with 70 commencing clozapine, of whom 67 had a commencement setting identified. Over half (n = 36, 53.7%) commenced clozapine in the community. When compared to the hospital initiation group, the community initiation group were less likely to have had a hospital admission at baseline (odds ratio (OR) 0.26, 95%CI, 0.09-0.87) or an involuntary admission during the 2 year episode of care with EPPIC (OR 0.25, 95%CI, 0.09-0.70). The community initiated group had presented with less severe delusion scores on short form Scale for Assessment of Positive Symptoms at baseline (mean 3.08 vs 3.94, P = .031). First generation migrants were less likely to initiate clozapine in the community (OR 0.29, 95%CI, 0.09-0.97). The community initiation group also had reduced odds of clozapine discontinuation until discharge from EPPIC (OR 0.22, 95%CI, 0.06-0.76).

CONCLUSION: Community initiation provides an alternative route to clozapine treatment and may be associated with a reduced rate of clozapine discontinuation.

PMID:42104793 | DOI:10.1093/schbul/sbag071

Health Serv Res. 2026 Jun;61(3):e70126. doi: 10.1111/1475-6773.70126.

ABSTRACT

OBJECTIVE: To examine rural-urban disparities in telehealth utilization during the post-pandemic period and assess whether these disparities persist after adjusting for individual-level characteristics.

STUDY SETTING AND DESIGN: We used multivariable logistic regression and propensity score matching to estimate differences in telehealth use by rurality and examined self-reported reasons for non-use.

DATA SOURCES AND ANALYTIC SAMPLE: We analyzed 2022 and 2024 Health Information National Trends Survey (HINTS) data, a nationally representative survey of noninstitutionalized US adults. The analytic sample included 11,106 respondents after excluding missing observations.

PRINCIPAL FINDINGS: Overall, 38.7% of adults reported telehealth use in the past 12 months. After adjusting for covariates, rural residents were significantly less likely to use telehealth than urban core residents; remote rural residence was associated with a 10-percentage point lower probability (95% CI, -16.2 to -2.8; p < 0.01). Propensity score analyses yielded similar results (-7.7% points; 95% CI, -16.2 to -2.8; p < 0.01). Among non-users, rural respondents were more likely to report not being offered telehealth.

CONCLUSIONS: We observed significant rural-urban disparities in telehealth use in the post-pandemic period. Rural non-users were more likely to report not being offered telehealth, indicating delivery-side barriers.

PMID:42104788 | DOI:10.1111/1475-6773.70126

Gynecol Endocrinol. 2026 Dec 31;42(1):2670809. doi: 10.1080/09513590.2026.2670809. Epub 2026 May 9.

ABSTRACT

BACKGROUND: Female infertility is multifactorial, with adiposity and regional fat distribution hypothesized as contributors, though evidence using detailed fat measures is limited. This study aims to examine the association between fat distribution indicators and female infertility in a nationally representative sample.

METHODS: This retrospective cross-sectional study analyzed NHANES 2013-2018 data from 2,531 women aged 20-45. Infertility was defined by self-reported difficulty conceiving ≥ 12 months or seeking fertility care. Exposures included body mass index (BMI) and DXA-based measures: total percent fat (TPF), android percent fat (APF), gynoid percent fat (GPF), android fat/gynoid fat ratio (AGR), visceral fat/total fat (VPF), subcutaneous fat/total fat (SPF), and visceral fat/subcutaneous fat ratio (VSR). Multivariable logistic regression was used to assess associations, and sensitivity analyses were performed to evaluate robustness.

RESULTS: In multivariable-adjusted models, TPF, APF, AGR, and BMI were modestly associated with higher odds of infertility (TPF: OR = 1.02, 95%CI: 1.00-1.05; APF: OR = 1.03, 95%CI: 1.01-1.04; AGR: OR = 1.02, 95%CI: 1.01-1.03; BMI: OR = 1.02, 95%CI: 1.01-1.04). Smooth curve fitting suggested a generally monotonic positive pattern for these associations. Associations were broadly similar across subgroups, although some subgroup interactions were observed.

CONCLUSION: In this analysis, TPF, APF, AGR, and BMI showed modest associations with infertility, which should not be interpreted causally. Although associations were generally consistent across subgroups, subgroup-specific heterogeneity cannot be excluded.

PMID:42104773 | DOI:10.1080/09513590.2026.2670809