Estimating clinical trial hazard functions

Clinical Trials · 2026-04-24

BACKGROUND: Although the analysis of event-based clinical trials commonly relies on assumptions about the underlying hazard functions, in practice it is rare to see estimates of those functions. METHODS: I describe conventional and novel methods for estimating the hazard function using discrete and discretized continuous survival models. The conventional approach involves parametric modeling; the novel approach applies Bayesian model averaging to flexible modeling by splines or fractional polynomials. I evaluate the methods in a Monte Carlo study and illustrate them in the analysis of three historical clinical trials. RESULTS: Although flexible models can capture features of the hazard functions-such as multimodality-that parametric models miss, they are not foolproof. Spline modeling was generally the most reliable, in the sense of yielding good coverage probabilities for the mean and median with modest loss of efficiency. In the examples, the discreteness of the measurements-days, weeks, or months-had little effect on the shape of estimated hazard functions. All three data sets showed some evidence of departure from the proportional hazards assumption, but in only one did a test for proportionality detect this departure. CONCLUSION: Flexible parametric models, estimated in the Bayesian model averaging framework, offer a robust approach to recovering the shape of the hazard function. Analyses of three clinical trial databases suggest that visualization of the hazard function can be a valuable adjunct to conventional survival analysis.

Factors influencing the selection of an SGLT2i vs. a GLP-1RA as cardioprotective agent in patients with type 2 diabetes

Frontiers in Cardiovascular Medicine · 2025-05-23 · 1 citations

Purpose: Sodium glucose co-transporter-2 inhibitors (SGLT2is) and glucagon-like peptide-1 receptor agonists (GLP-1RAs) have demonstrated cardioprotective effects in people with type 2 diabetes and atherosclerotic cardiovascular disease (ASCVD). In this patient group, there is treatment equipoise, from the standpoint of cardiovascular effect between these medication classes; however, factors associated with prescribing are poorly characterized. Methods: We performed a retrospective real-world analysis by creating an electronic health record registry of people with type 2 diabetes and ASCVD (without additional indications for a specific cardioprotective class) who received a prescription for either an SGLT2i or GLP-1RA. We analyzed patient-, provider-, and clinical encounter-related predictors of being prescribed an SGLT2i or GLP-1RA using univariable and multivariable logistic regression analysis. Results: = 299) between January 2019 and October 2024. Care in cardiology (OR = 4.78; 95% CI, 2.53-9.04) strongly predicted SGLT2i prescription. Care in endocrinology (OR = 0.40; 95% CI, 0.23-0.68), higher BMI (OR = 0.92; 95% CI, 0.88-0.95, per BMI unit), and a higher recent estimated glomerular filtration (OR = 0.98; 95% CI, 0.96-0.99, per eGFR unit) predicted GLP-1RA prescription. The area under the receiver operating characteristic curve of the model was 0.78. Conclusion: Prescriber's specialty strongly determined the selection of cardioprotective agents. Treatment guidelines should provide more specific guidance regarding patient selection and consider the holistic benefits of each drug class beyond their cardiovascular protective effects.

777-P: Predictors of Selection of a GLP1RA vs. SGLT2i in People with Type 2 Diabetes and Atherosclerotic Cardiovascular Disease

Diabetes · 2024-06-14

Introduction: In people with T2D and preexisting ASCVD, either SGLT2i or GLP1RA are indicated by treatment guidelines to reduce MACE. We evaluated predictors of prescription of SGLT2i vs GLP1RA in a population eligible for either. Methods: An electronic health record (EHR) based registry was created to identify people with T2D and ASCVD who were indicated either a GLP1RA or SGLT2i for cardiorenal protection within a large, academic health system. Data pertaining to demographics, lab and imaging results, ICD9/10 diagnoses, prescriptions, provider and clinic characteristics were extracted. Eligible encounters occurred in a primary care, endocrinology, cardiology, or nephrology clinic between January 1, 2019 and August 23, 2023. For each eligible encounter where a drug was prescribed, the first treatment type (GLP1RA or SGLT2i) was determined based on medication history. We estimated a logistic regression using stepwise variable selection to identify a best-predicting model and forced the variables of age, sex, and race into the model. Results: A total of 315 patients with T2D and ASCVD were eligible for either treatment and were prescribed one of these medications: 142 were prescribed a GLP1RA and 173 were prescribed SGLT2i. Lower BMI was associated with use of SGLT2i (OR = 0.91, 95% CI 0.87-0.96), as was being an established patient (OR 2.32, 95% CI 1.14-4.72). Compared to treatment in a primary care setting, treatment in a cardiology clinic was strongly associated with prescription of SGLT2i (OR = 7.77, 95% CI 3.18-19.04), whereas treatment in endocrinology clinic was strongly associated with prescription of a GLP1RA (OR = 0.35, 95% CI 0.18-0.68). Area under the receiver operating characteristic curve for the model was 0.82. Conclusion: In a real-world dataset from a large academic center, the selection of guideline directed therapy for patients with T2D and ASCVD was strongly determined by the provider’s specialty, highlighting an important opportunity for education. Disclosure S. Agarwal: None. M.A. Basit: None. M.E. Bowen: Research Support; Boehringer-Ingelheim. D. Heitjan: Consultant; Bluejay Diagnostics, Medcognetics, Sebela, Abbott, Macrogenics, Guardant, Bristol-Myers Squibb Company, Gilead Sciences, Inc. C. Mai: None. K. Marble: None. Z. Xiang: None. I. Lingvay: Consultant; Altimmune, Astra Zeneca, Bayer, Biomea, Boehringer-Ingelheim, Carmot, Cytoki Pharma, Eli Lilly, Intercept, Janssen/J&J, Mannkind, Mediflix, Merck, Metsera, Novo Nordisk, Pharmaventures, Pfizer, Sanofi. Research Support; NovoNordisk, Sanofi, Mylan, Boehringer-Ingelheim. Consultant; TERNS Pharma, The Comm Group, Valeritas, WebMD, and Zealand Pharma. Funding This study was supported by Boehringer Ingelheim Pharmaceuticals, Inc. (BIPI) and Lilly USA, LLC. The authors meet criteria for authorship as recommended by the International Committee of Medical Journal Editors (ICMJE) and were fully responsible for all aspects of the trial and publication development.

Comment on “Causal interpretation of the hazard ratio in randomized clinical trials” by Fay and Li

Clinical Trials · 2024-04-28 · 2 citations

Response to comments on ‘sensitivity of estimands in clinical trials with imperfect compliance’

The International Journal of Biostatistics · 2024-05-03 · 1 citations

Article Response to comments on ‘sensitivity of estimands in clinical trials with imperfect compliance’ was published on November 1, 2024 in the journal The International Journal of Biostatistics (volume 20, issue 2).

Predicting Hospital Readmission in Medicaid Patients With COPD Using Administrative and Claims Data

Respiratory Care · 2024-03-26 · 2 citations

BACKGROUND: The goals of this study were to develop a model that predicts the risk of 30-d all-cause readmission in hospitalized Medicaid patients diagnosed with COPD and to create a predictive model in a retrospective study of a population cohort. METHODS: was an admission for any condition (not necessarily COPD) that occurred within 30 d of a COPD discharge. We estimated a mixed-effects logistic model to predict 30-d readmission from patient demographic data, comorbidities, past health care utilization, and features of the index hospitalization. We evaluated model fit graphically and measured predictive accuracy by the area under the receiver operating characteristic (ROC) curve. RESULTS: Among 12,283 COPD hospitalizations contributed by 9,437 subjects, 2,534 (20.6%) were 30-d readmissions. The final model included demographics, comorbidities, claims history, admission and discharge variables, length of stay, and seasons of admission and discharge. The observed versus predicted plot showed reasonable fit, and the estimated area under the ROC curve of 0.702 was robust in sensitivity analyses. CONCLUSIONS: Our model identified with acceptable accuracy hospitalized Medicaid patients with a diagnosis of COPD who are at high risk of readmission. One can use the model to develop post-discharge management interventions for reducing readmissions, for adjusting comparisons of readmission rates between sites/providers or over time, and to guide a patient-centered approach to patient care.

Supplementary Methods from Pilot and Feasibility Trial Evaluating Immuno-Gene Therapy of Malignant Mesothelioma Using Intrapleural Delivery of Adenovirus-IFNα Combined with Chemotherapy

2023-03-31

<p>Additional Trial related information and methods</p>

Supplementary Methods from Modeling of the Growth Kinetics of Occult Breast Tumors: Role in Interpretation of Studies of Prevention and Menopausal Hormone Therapy

2023-03-31

<p>PDF file, 126KB.</p>

Supplemental Table 1 from CDK 4/6 Inhibitor Palbociclib (PD0332991) in Rb<sup>+</sup> Advanced Breast Cancer: Phase II Activity, Safety, and Predictive Biomarker Assessment

2023-03-31

<p>Supplemental Table 1. Biomarker Associations with Outcome</p>

Data from Modeling of the Growth Kinetics of Occult Breast Tumors: Role in Interpretation of Studies of Prevention and Menopausal Hormone Therapy

2023-03-31

<div>Abstract<p><b>Background:</b> Autopsy studies report a reservoir of small, occult, undiagnosed breast cancers in up to 15.6% of women dying from unrelated causes. The effective doubling times (EDT) of these occult neoplasms range from 70 to 350 days and mammographic detection threshold diameters from 0.88 to 1.66 cm. Modeling of the biologic behavior of these occult tumors facilitates interpretation of tamoxifen breast cancer prevention and menopausal hormone therapy studies.</p><p><b>Methods:</b> We used iterative and mathematical techniques to develop a model of occult tumor growth (OTG) whose parameters included prevalence, EDT, and detection threshold. The model was validated by comparing predicted with observed incidence of breast cancer in several populations.</p><p><b>Results:</b> Iterative analysis identified a 200-day EDT, 7% prevalence and 1.16 cm detection threshold as optimal parameters for an OTG model as judged by comparison with Surveillance Epidemiology and End Results (SEER) population incidence rates in the United States. We validated the model by comparing predicted incidence rates with those observed in five separate population databases, in three long-term contralateral breast cancer detection studies, and with data from a computer-simulated tumor growth (CSTG) model. Our model strongly suggests that breast cancer prevention with anti-estrogens or aromatase inhibitors represents early treatment not prevention. In addition, menopausal hormone therapy does not primarily induce <i>de novo</i> tumors but promotes the growth of occult lesions.</p><p><b>Conclusions:</b> Our OGTG model suggests that occult, undiagnosed tumors are prevalent, grow slowly, and are the biologic targets of anti-estrogen therapy for prevention and hormone therapy for menopausal women. <i>Cancer Epidemiol Biomarkers Prev; 21(7); 1038–48. ©2012 AACR.</i></p></div>