Roshan Bastani
Research Interests
Make your next move.
Prepare your application
Review drafts before you send them.
Explore related searches
Articles (12)
Year:
2026
Year:
2026
Year:
2026
Objective We assessed the performance of International Classification of Diseases (ICD) codes to identify patients with hepatocellular carcinoma (HCC) in a large academic health system and determined whether employing an algorithm using a combination of ICD codes could deliver higher accuracy and precision than single ICD codes in identifying HCC cases using electronic health record (EHR) data. Results The use of a single ICD code entry for HCC (ICD-9-CM 155.0 or ICD-10-CM C22.0) in our cohort of 1,007 established ambulatory care patients with potential HCC yielded 58% false positives (not true HCC cases) based on chart reviews. We developed an ICD code-based algorithm that prioritized positive predictive value (PPV), F-score, and accuracy to minimize false positives and negatives. Using manual chart reviews as the gold standard, the highest performing algorithm required at least 10 ICD code entries for HCC and the sum of ICD code entries for HCC to exceed the sum of ICD code entries for non-HCC malignancies. The algorithm demonstrated high performance (PPV 97.4%, F-score 0.92, accuracy 94%), which was internally validated (PPV 92.3%, F-score 0.90, accuracy 91%) using a separate sample of 285 cancer registry cases. Our findings support the need to assess the accuracy and precision of ICD codes before using EHR data to study HCC more broadly.
Year:
2025
Year:
2025
Background Screening, brief intervention, and referral to treatment (SBIRT) is an evidence-based approach to identify and initiate treatment for alcohol and substance use in primary care settings. Among 22 public hospitals incentivized to implement SBIRT as part of a value-based Medicaid waiver program over five years, this study examined trajectories, strategies, and challenges in standardizing SBIRT within primary care. Methods This study utilized data from narrative reports completed by hospital leadership, obtained from the evaluation of the Public Hospital Redesign and Incentives in Medi-Cal (PRIME) program in California. Following the Multi-Level Health Outcomes Framework, template analysis was used to characterize SBIRT implementation. Content analysis was used to catalogue implementation strategies using the Expert Recommendations for Implementing Change Framework. To assess trajectories (i.e., longitudinal implementation outcomes) of SBIRT implementation, we categorized standardized adoption of sequential SBIRT processes (screening only; screening and brief intervention; screening, brief intervention, and referral to treatment) and reach (limited vs. full primary care population). Results Hospitals used a wide variety of measures, personnel, platforms, and workflows in screening for substance use within primary care settings. Brief intervention was conducted by primary care or behavioral health care team members who had received targeted training. Hospitals implemented a wide range of treatment options to address substance use, including referral to co-located or contracted/partnered behavioral health providers. By the end of the first implementation year, only one hospital had standardized screening processes, and none had standardized brief intervention or referral. At the end of the fifth year, 20 of 22 hospitals had standardized screening, 15 had standardized brief intervention, and 12 had standardized referral among their full primary care populations. Strategies and challenges in planning, education, and restructuring processes (e.g., integration of screening processes within electronic health records and clinical workflows) were particularly influential in facilitating implementation. Conclusions This study highlighted significant progress made by public hospitals in implementing standardized SBIRT processes among their primary care populations within a value-based program. However, hospitals experienced delays and challenges, highlighting key areas in which additional support or investment may be needed to sustain and promote long-term progress in SBIRT implementation.
Year:
2025
Missing outcome data are common in cluster randomized trials (CRTs), which can complicate inference. Further, the missingness can occur due to dropout of individuals, termed sporadically missing data, or dropout of clusters, termed systematically missing data, and these two types of missingness could have potentially different missing data mechanisms. We aimed to develop a well‐performing and practical approach to handle inference in CRTs when outcome data may be both sporadically and systematically missing. To this end, we first examined the performance of several multilevel multiple imputation (MI) methods to handle sporadically and systematically missing CRT outcome data via a simulation study. Specifically, we examined performance under a multilevel covariate‐dependent missingness assumption. Our findings indicated that several full conditional specification (FCS) methods designed for missingness in linear mixed models performed well under various scenarios, while an FCS approach using a two‐stage estimator often performed poorly. We then developed methods for conducting sensitivity analysis to test the robustness of inferences under different missing at random (MAR) and missing not at random (MNAR) assumptions. The methods allow for different MNAR assumptions for cluster dropout and individual dropout to reflect that they may arise from different missing data mechanisms. We used graphical displays to visualize sensitivity analysis results. Our methods are illustrated using a real data application.
Year:
2025
Year:
2025
Background: Various population-based studies have shown Hispanic/Latino ethnicity is a risk factor for worse survival in patients with gastric cancer linked to disparate access to care. We aimed to address whether Hispanic patients treated within safety-net hospital systems continue to experience this survival deficit compared to non-Hispanic patients. Methods: We performed a retrospective cohort study comparing survival between Hispanic and non-Hispanic patients diagnosed with gastric adenocarcinoma between January 1, 2016, and December 31, 2020, within Los Angeles County’s safety-net hospital system. Gastric cancer–specific survival was compared between the two cohorts using the Kaplan–Meier estimate and Cox proportional-hazards regression model. Results: In total, 448 patients who received care from five medical centers were included; 348 (77.7%) patients self-identified as Hispanic and 100 (22.3%) as non-Hispanic. Mean follow-up time was 2.0 years (median 0.91 years, IQR, 0.34–2.5 years). Hispanic patients were found to be diagnosed at a younger age (55.6 vs. 60.7 years, P P P = 0.04). After adjusting social and oncologic variables, Hispanic ethnicity remained an independent risk factor for worse survival [HR 1.56, (95% CI, 1.06–2.28); P = 0.02]. Conclusions: Hispanic patients treated within a large, multicenter safety-net hospital system experience worse survival compared to non-Hispanic patients. This suggests ethnic disparities exist within safety-net hospital systems, independent of known clinicopathologic factors. Impact: Improving outcomes for Hispanic patients with gastric cancer requires future efforts aimed at defining and addressing these unidentified barriers to care. See related In the Spotlight, p. 12
Year:
2025
Background Gastric cancer disproportionately impacts populations in resource‐limited settings. Within a safety‐net network, we assessed the utility of computed tomography (CT) as a single staging modality. Methods We utilized a clinical database of gastric cancer patients treated within the Los Angeles County safety‐net hospital system from 2016 to 2023 in conjunction with retrospective imaging review by certified radiologists. We assessed agreement between clinical and pathological staging for patients who underwent curative gastrectomy using the Kappa coefficient. Results Of 107 patients with available CT imaging, 43.9% ( n = 47) were staged with CT as a single modality. Most tumors displayed infiltrating (75%) or diffuse (28%) morphology, 41% displayed adequate gastric distention and regional lymphadenopathy was common (68%). Twenty‐nine patients underwent curative gastrectomy. Overall agreement was minimal ( κ = 0.29, 95% CI [0.071−0.51], p = 0.022), weak for T3/T4 tumors ( κ = 0.50, 95% CI [0.17−0.82], p < 0.01), and weak for Hispanic/Latino patients ( κ = 0.47, 95% CI [0.19−0.76], p < 0.01). Conclusions There was minimal agreement between clinical and pathologic staging when assessing clinical stage by CT imaging alone, suggesting that CT is not adequate as a single modality staging tool. While every effort should be made to obtain multimodal staging, larger studies are warranted to improve CT imaging protocols for staging in resource‐limited settings.
Year:
2024
Year:
2024
Year:
2022
Collaborators (6)
Carrie Wong
University of California, Los Angeles
Mariam Thomas
Olive View-UCLA Medical Center
Joann G. Elmore
Professor
University of California
Brian Kadera
Assistant Professor of Surgery
David Geffen School of Medicine at UCLA
Samir Gupta
Professor of Medicine
University of California, San Diego
Michael Ong
Staff Physician / Associate Chief of Staff for Research
VA Greater Los Angeles Healthcare System
Make your next move.
Prepare your application
Review drafts before you send them.