Bayesian Hierarchical Models for Subgroup Analysis.

In conventional subgroup analyses, subgroup treatment effects are estimated using data from each subgroup separately without considering data from other subgroups in the same study. The subgroup treatment effects estimated this way may be heterogenous with high variability due to small sample sizes in some subgroups and much different from the treatment effect in the overall population. A Bayesian hierarchical model (BHM) can be used to derive more precise, and less heterogenous estimates of subgroup treatment effects that are closer to the treatment effect in the overall population. BHM assumes exchangeability in treatment effect across subgroups after adjusting for effect modifiers and other relevant covariates. In this article, we will discuss the technical details for applying one-way and multi-way BHM using summary-level statistics, and patient-level data for subgroup analysis. Four case studies based on four new drug applications are used to illustrate the application of these models in subgroup analyses for continuous, dichotomous, time-to-event, and count endpoints.

Evidence-based objective performance criteria for the evaluation of hip and knee replacement devices and technologies.

BACKGROUND: Objective performance criteria (OPC) is a novel method to provide minimum performance standards and improve the regulated introduction of original or incremental device innovations in order to prevent patients from being exposed to potentially inferior designs whilst allowing timely access to improvements. We developed 2-year safety and effectiveness OPC for total hip and knee replacement (THR and TKR). METHODS: Analyses of large databases were conducted using various data sources: a systematic literature review; a direct data analysis from The Functional Outcomes Research for Comparative Effectiveness in Total Joint Replacement and Quality Improvement Registry (FORCE-TJR) and the Kaiser Permanente Implant Registry (KPIR); and claims data analyses from longitudinal discharge data in New York and California states. The literature review included U.S. patients (≥18 years) who received THR or TKR for primary end-stage osteoarthritis and prospectively collected data on patient-reported outcome measures (PROMs) from at least 100 subjects and/or 2-year implant survival for at least 250 implants. Random effects models were used for meta-analysis. RESULTS: Data were available from a total of 951 100 patients. After screening of 7979 abstracts, 294 studies underwent full-text review and 31 studies contributed to the evidence synthesis (333 995 implants). Direct data analysis of FORCE-TJR contributed 9223 joint replacement patients to the construction of OPC for effectiveness; KPIR contributed 262 044 patients for the construction of OPC for safety. Claims database analysis contributed 345 838 patients to the construction of safety OPC. OPC for safety were constructed for cumulative incidences of 2-year all-cause and septic revision (THR/TKR 2.0%/1.6% and 0.6%/0.7%), and OPC for effectiveness were constructed based on four disease-specific and three general health-related quality of life PROMs (HOOS/KOOS 87.1/80.6; HSS/KSS function 94.4/90.6; SF-12/SF-36, PCS 46.5/41.9, EQ-5D 0.88/0.84). CONCLUSION: This study is the first to construct a 2-year OPC for the safety and effectiveness of THR and TKR based on U.S. real-world data. Based on these OPC, potential benchmarks for (single-arm study) evaluation of new device innovations are suggested for a regulated and safe introduction to the (commercial) market.

Retrieval and clinical analysis of distraction-based dual growing rod constructs for early-onset scoliosis.

BACKGROUND CONTEXT: Growing rod constructs are an important contribution for treating patients with early-onset scoliosis. These devices experience high failure rates, including rod fractures. PURPOSE: The objective of this study was to identify the failure mechanism of retrieved growing rods, and to identify differences between patients with failed and intact constructs. STUDY DESIGN/SETTING: Growing rod patients who had implant removal and were previously enrolled in a multicenter registry were eligible for this study. PATIENT SAMPLE: Forty dual-rod constructs were retrieved from 36 patients across four centers, and 34 of those constructs met the inclusion criteria. Eighteen constructs failed due to rod fracture. Sixteen intact constructs were removed due to final fusion (n=7), implant exchange (n=5), infection (n=2), or implant prominence (n=2). OUTCOME MEASURES: Analyses of clinical registry data, radiographs, and retrievals were the outcome measures. METHODS: Retrievals were analyzed with microscopic imaging (optical and scanning electron microscopy) for areas of mechanical failure, damage, and corrosion. Failure analyses were conducted on the fracture surfaces to identify failure mechanism(s). Statistical analyses were performed to determine significant differences between the failed and intact groups. RESULTS: The failed rods fractured due to bending fatigue under flexion motion. Construct configuration and loading dictate high bending stresses at three distinct locations along the construct: (1) mid-construct, (2) adjacent to the tandem connector, or (3) adjacent to the distal anchor foundation. In addition, high torques used to insert set screws may create an initiation point for fatigue. Syndromic scoliosis, prior rod fractures, increase in patient weight, and rigid constructs consisting of tandem connectors and multiple crosslinks were associated with failure. CONCLUSION: This is the first study to examine retrieved, failed growing rod implants across multiple centers. Our analysis found that rod fractures are due to bending fatigue, and that stress concentrations play an important role in rod fractures. Recommendations are made on surgical techniques, such as the use of torque-limiting wrenches or not exceeding the prescribed torques. Additional recommendations include frequent rod replacement in select patients during scheduled surgeries.

Bayesian approaches in medical device clinical trials: a discussion with examples in the regulatory setting.

Challenging statistical issues often arise in the design and analysis of clinical trials to assess safety and effectiveness of medical devices in the regulatory setting. The use of Bayesian methods in the design and analysis of medical device clinical trials has been increasing significantly in the past decade, not only due to the availability of prior information, but mainly due to the appealing nature of Bayesian clinical trial designs. The Center for Devices and Radiological Health at the Food and Drug Administration (FDA) has gained extensive experience with the use of Bayesian statistical methods and has identified some important issues that need further exploration. In this article, we discuss several topics relating to the use of Bayesian statistical methods in medical device trials, based on our experience and real applications. We illustrate the benefits and challenges of Bayesian approaches when incorporating prior information to evaluate the effectiveness and safety of a medical device. We further present an example of a Bayesian adaptive clinical trial and compare it to a traditional frequentist design. Finally, we discuss the use of Bayesian hierarchical models for multiregional trials and highlight the advantages of the Bayesian approach when specifying clinically relevant study hypotheses.