ijsmr

Abstract: Combining Survival and Toxicity Effect Sizes from Clinical Trials: NCCTG 89-20-52 (Alliance)

ijsmr logo-pdf 1349088093

Combining Survival and Toxicity Effect Sizes from Clinical Trials: NCCTG 89-20-52 (Alliance) - Pages 137-146

Brittny T. Major-Elechi, Paul J. Novotny, Jasvinder A. Singh, James A. Bonner, Amylou C. Dueck, Daniel J. Sargent, Axel Grothey and Jeff A. Sloan

https://doi.org/10.6000/1929-6029.2018.07.04.4

Published: 16 November 2018

Abstract: Background: How can a clinician and patient incorporate survival and toxicity information into a single expression of comparative treatment benefit? Sloan et al. recently extended the ½ standard deviation concept for judging the clinical importance of findings from clinical trials to survival and tumor response endpoints. A new method using this approach to combine survival and toxicity effect sizes from clinical trials into a quality-adjusted effect size is presented.

Methods: The quality-adjusted survival effect size (QASES) is calculated as survival effect size (ESS) minus the calibrated toxicity effect sizes (EST) (QASES=ESS-EST). This combined effect size can be weighted to adjust for the relative emphasis placed by the patient on survival and toxicity effects.

Results: As an example, consider clinical trial NCCTG 89-20-52 which randomized patients to once-daily thoracic radiotherapy (ODTRT) versus twice-daily treatment of thoracic radiotherapy (TDRT) for the treatment of lung cancer. The ODTRT vs. TDRT arms had median survival time of 22 vs. 20 months (p=0.49) and toxicity rate of 39% vs. 54%, (p<0.05). The QASES of 0.18 standard deviations translates to a quality-adjusted survival difference of 5.7 months advantage for the ODRT arm over the TDRT treatment arm (22(16.3) months), p<0.05). Similar results are presented for the four possible case combinations of significant/non-significant survival and toxicity benefits using completed clinical trials.

Conclusions: We used a novel approach to re-analyze clinical trial data to produce a single estimate for each treatment that combines survival and toxicity data. The QASES approach is an intuitive and mathematically simple yet robust approach.

Keywords: Survival, toxicity, quality of life, effect size, quality-adjusted life years, QALY.

Download

Abstract: Bayesian Model Averaging for Selection of a Risk Prediction Model for Death within Thirty Days of Discharge: The SILVER-AMI Study

ijsmr logo-pdf 1349088093

Bayesian Model Averaging for Selection of a Risk Prediction Model for Death within Thirty Days of Discharge: The SILVER-AMI Study - Pages 1-7
Terrence E. Murphy, Sui W. Tsang, Linda S. Leo-Summers, Mary Geda, Dae H. Kim, Esther Oh, Heather G. Allore, John Dodson, Alexandra M. Hajduk, Thomas M. Gill and Sarwat I. Chaudhry

https://doi.org/10.6000/1929-6029.2019.08.01

Published: 05 April 2019

Abstract: We describe a selection process for a multivariable risk prediction model of death within 30 days of hospital discharge in the SILVER-AMI study. This large, multi-site observational study included observational data from 2000 persons 75 years and older hospitalized for acute myocardial infarction (AMI) from 94 community and academic hospitals across the United States and featured a large number of candidate variables from demographic, cardiac, and geriatric domains, whose missing values were multiply imputed prior to model selection. Our objective was to demonstrate that Bayesian Model Averaging (BMA) represents a viable model selection approach in this context. BMA was compared to three other backward-selection approaches: Akaike information criterion, Bayesian information criterion, and traditional p-value. Traditional backward-selection was used to choose 20 candidate variables from the initial, larger pool of five imputations. Models were subsequently chosen from those candidates using the four approaches on each of 10 imputations. With average posterior effect probability ≥ 50% as the selection criterion, BMA chose the most parsimonious model with four variables, with average C statistic of 78%, good calibration, optimism of 1.3%, and heuristic shrinkage of 0.93. These findings illustrate the utility and flexibility of using BMA for selecting a multivariable risk prediction model from many candidates over multiply imputed datasets.

Keywords: Risk prediction, AMI, Bayesian model averaging, AIC, BIC, backward-selection.

Download

Abstract: Italian Version of the Risk Assessment and Prediction Tool: Properties and Usefulness of a Decision-Making Tool for Subjects’ Discharge after Total Hip and Knee Arthroplasty

ijsmr logo-pdf 1349088093

Italian Version of the Risk Assessment and Prediction Tool: Properties and Usefulness of a Decision-Making Tool for Subjects’ Discharge after Total Hip and Knee Arthroplasty - Pages 8-16
Marco Monticone, Luca Frigau, Cristiano Sconza, Calogero Foti, Francesco Mola and Stefano Respizzi

https://doi.org/10.6000/1929-6029.2019.08.02

Published: 05 April 2019

Abstract:
Background: Growing attention is being given to standardized outcome measures to improve interventions for total hip arthroplasty (THA) and total knee arthroplasty (TKA). We culturally adapt and validate the Italian version of the Risk Assessment and Prediction Tool (RAPT-I) to allow its predictive use after THA and TKA.

Methods: The RAPT-I was adapted by forward–backward translation, a final review by an expert committee and a test of the pre-final version to establish its correspondence with the original version. The psychometric testing included test–retest reliability (intraclass correlation coefficient, ICC). The RAPT score was used to predict the subjects’ destination (<6: rehabilitation unit; 6-9: additional intervention before discharging home or >9: discharge directly at home) by comparing the actual discharge destination with the predicted destination. The predictive effects of RAPT items on the discharge destination were further described by a logistic regression model (repeated leave-one-out bootstrap procedure).

Results: The questionnaire was administered to 78 subjects with THA and 70 subjects with TKA and proven to be acceptable. The questionnaire showed excellent test–retest reliability (ICC = 0.839; with 95% confidence interval (CI) of 0.725–0.934 for THA; ICC = 0.973, with 95% CI of 0.930–0.997 for TKA). The RAPT-I overall predictive validity was 87.2%, and the discharge destination was directly related to living condition (odds ratio (OR) = 2.530), mobility (OR = 2.626) and age (OR = 1.332) and inversely related to gait aids (OR = 0.623) and gender (OR = 0.474).

Conclusions: The RAPT-I was successfully adapted into Italian and proven to exhibit satisfactory properties, including predictive validity in determining discharge destination.

Keywords: RAPT, cross-cultural adaptation, predictive validity, logistic regression, repeated leave-one-out bootstrap.

Download

Abstract: Improvement in Heart Rate Variability Following Spinal Adjustment: A Case Study in Statistical Methodology for a Single Office Visit

ijsmr logo-pdf 1349088093

Improvement in Heart Rate Variability Following Spinal Adjustment: A Case Study in Statistical Methodology for a Single Office Visit - Pages 17-22
John Hart

https://doi.org/10.6000/1929-6029.2019.08.03

Published: 11 May 2019

Abstract: Introduction: Statistical analysis is typically applied at the group level. The present study analyzes data during a single office visit as a novel approach providing real-time feedback to the clinician and patient regarding efficacy of an intervention. In this study, heart rate variability (HRV) was analyzed before versus after a chiropractic spinal adjustment.

Methods: The patient is an adult female who signed a consent form for the study. HRV was measured twice before a chiropractic adjustment and once afterwards using app-based technology. The three HRV values (two pre and one post) were then statistically analyzed using an online calculator for outliers using Grubbs test.

Results: The two pre-adjustment HRV (rMSSD) readings were consistently low: pre 1 = 16.0 milliseconds [ms] and pre 2 = 16.2 ms. The low HRV was an indicator that the patient’s nervous system was not functioning optimally. The patient’s atlas (C1) vertebra was palpated to be slightly out of alignment. These two findings (low HRV and vertebral misalignment) indicated the presence of a chiropractic subluxation (of the atlas vertebra). The subluxation was adjusted and within minutes the HRV increased (improved) to 27.5 ms. This improvement was calculated to be a statistically significant outlier (p < 0.05).

Conclusion: This study is an example of how statistical methods can be applied to the level of an individual patient during one office visit to assess neurological effectiveness of a chiropractic adjustment. Since this is a case study, the results may not apply to all patients. Therefore, further studies in other patients, and for longer follow-up times, are reasonable next steps.

Keywords: Chiropractic adjustment, heart rate variability, biostatistics, Grubbs test.

Download