The contrast and convergence of Bayesian and frequentist statistical approaches in pharmacoeconomic analysis

Blood Pressure-Lowering Effect of Fimasartan Versus Comparators: A Cross-Inference With a Systematic Review and Meta-Analysis Through a Quality Management System

PURPOSE

Fimasartan, one of the newest angiotensin receptor blockers (ARBs) available worldwide, has been investigated extensively since its initial development. Our study group conducted a systematic review and meta-analysis of randomized controlled trials (RCTs) evaluating fimasartan and comparators for their blood pressure (BP)-lowering effect. Moreover, we employed a cross-inference (frequentist and Bayesian inference) system, which has never been used in the medical field, to confirm the results of our study. In addition, a quality management system was integrated throughout the study for data quality.

METHODS

PubMed, EMBASE, the Cochrane Central Register of Controlled Trials, ClinicalKey, and ClinicalTrial.gov were searched for RCT studies from March 1998 to March 2022. In each study, the mean differences (MD^s) and 95% CIs were identified for reductions in clinic sitting systolic and diastolic BP (SiSBP/SiDBP) or 24-hour mean systolic BP and diastolic BP by ambulatory BP monitoring (ASBP/ADBP) from baseline between the fimasartan and comparator groups, followed by meta-analysis. A subsequent meta-analysis was performed with frequentist and Bayesian inference as a tool in the cross-checking system.

FINDINGS

Eleven RCTs with a total of 2459 subjects were included in the study. The clinic SiSBP/SiDBP-lowering effect of fimasartan was significantly greater relative to those of comparators (MD for clinic SiSBP, -2.58 mm Hg [95% CI, -4.35 to -0.81; P = 0.004]; MD for clinic SiDBP, -2.13 mm Hg [95% CI, -2.96 to -1.30; P = 0.00001]). The ASBP/ADBP-lowering effect of fimasartan was also significantly greater relative to those of comparators (MD for ASBP, -3.58 mm Hg [95% CI, -5.74 to -1.43; P = 0.001]; MD for ADBP, -1.99 mm Hg [95% CI, -3.34 to -0.63; P = 0.004]).

IMPLICATIONS

Fimasartan seems to be more effective in lowering BP than its comparators, including other ARBs. Although there is a limited amount of data and a minuscule number of study subjects available, the results of cross-inference (frequentist + Bayesian) were fairly consistent with the meta-analysis results through our quality management system.

Differences between Frequentist and Bayesian inference in routine surveillance for influenza vaccine effectiveness: a test-negative case-control study

Background

Routine influenza vaccine effectiveness (VE) surveillance networks use frequentist methods to estimate VE. With data from more than a decade of VE surveillance from diverse global populations now available, using Bayesian methods to explicitly account for this knowledge may be beneficial. This study explores differences between Bayesian vs. frequentist inference in multiple seasons with varying VE.

Methods

We used data from the United States Influenza Vaccine Effectiveness (US Flu VE) Network. Ambulatory care patients with acute respiratory illness were enrolled during seasons of varying observed VE based on traditional frequentist methods. We estimated VE against A(H1N1)pdm in 2015/16, dominated by A(H1N1)pdm; against A(H3N2) in 2017/18, dominated by A(H3N2); and compared VE for live attenuated influenza vaccine (LAIV) vs. inactivated influenza vaccine (IIV) among children aged 2–17 years in 2013/14, also dominated by A(H1N1)pdm. VE was estimated using both frequentist and Bayesian methods using the test-negative design. For the Bayesian estimates, prior VE distributions were based on data from all published test-negative studies of the same influenza type/subtype available prior to the season of interest.

Results

Across the three seasons, 16,342 subjects were included in the analyses. For 2015/16, frequentist and Bayesian VE estimates were essentially identical (41% each). For 2017/18, frequentist and Bayesian estimates of VE against A(H3N2) viruses were also nearly identical (26% vs. 23%, respectively), even though the presence of apparent antigenic match could potentially have pulled Bayesian estimates upward. Precision of estimates was similar between methods in both seasons. Frequentist and Bayesian estimates diverged for children in 2013/14. Under the frequentist approach, LAIV effectiveness was 62 percentage points lower than IIV, while LAIV was only 27 percentage points lower than IIV under the Bayesian approach.

Conclusion

Bayesian estimates of influenza VE can differ from frequentist estimates to a clinically meaningful degree when VE diverges substantially from previous seasons.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-021-10543-z.

Hepatitis C Direct-Acting Antiviral Treatment Selection, Treatment Failure, and Use of Drug-Drug Interactions in a State Medicaid Program

BACKGROUND

Newer hepatitis C virus (HCV) treatments often provide high success rates with fewer adverse events, although the extent of all potential drug interactions is not fully known.

OBJECTIVE

To assess outcomes of receiving HCV treatment and subsequent sustained virologic response (SVR) based on patient and clinical characteristics, including direct-acting antiviral (DAA) drug-drug interactions (DDIs), in Medicaid members with chronic HCV.

METHODS

Comprehensive medical and pharmacy claims and prior authorization data were collected for HCV patients requesting treatment between January 2014 and June 2015. Outcomes of receiving treatment with DAAs and treatment failure based on SVR were analyzed according to demographics, prior/current HCV treatment, severity of DDIs, advancing liver disease, and comorbidities. Multivariable generalized linear models were employed, including a Bayesian sensitivity analysis.

RESULTS

Among 3,412 Medicaid members with HCV, 13.6% received DAAs (n = 464), averaging 53.6 ± 10.0 years, with 52.8% female. Multivariable analyses indicated that higher odds of DAA treatment initiation were associated with older age, prior HCV treatment, and advancing liver disease. Some 4.8% of treatment failures occurred among 168 patients with reported SVRs, wherein a 3.218 times higher adjusted odds of treatment failure was associated with concomitant use of medications with DDIs classified as significant or potentially clinically significant by the University of Liverpool HEP Drug Interactions resource (P = 0.001).

CONCLUSIONS

In a cohort of state Medicaid members with chronic HCV, a markedly higher adjusted odds of treatment failure was independently associated with DDIs classified as significant or potentially clinically significant, warranting continued inquiry and potential alternate treatments concerning conditions that require their use.

DISCLOSURES

This research was funded by an unrestricted research grant by Gilead Sciences. During the course of this study, all authors were either employed by the Oklahoma HealthCare Authority or engaged in contractual work for this employer. Keast, Holderread, and Skrepnek report unrelated research grants from AbbVie, Otsuka, and Amgen. Keast and Skrepnek acknowledge funding from Purdue Pharma for an unrelated research fellowship grant. Posters based on this work were presented at HepDart 2015 on December 6-10, 2015, in Grand Wailea, HI, and at Academy of Managed Care Nexus 2015 on October 26-29, 2015, in Orlando, FL.

Increased Statistical Efficiency in a Lognormal Mean Model

Within the context of clinical and other scientific research, a substantial need exists for an accurate determination of the point estimate in a lognormal mean model, given that highly skewed data are often present. As such, logarithmic transformations are often advocated to achieve the assumptions of parametric statistical inference. Despite this, existing approaches that utilize only a sample’s mean and variance may not necessarily yield the most efficient estimator. The current investigation developed and tested an improved efficient point estimator for a lognormal mean by capturing more complete information via the sample’s coefficient of variation. Results of an empirical simulation study across varying sample sizes and population standard deviations indicated relative improvements in efficiency of up to 129.47 percent compared to the usual maximum likelihood estimator and up to 21.33 absolute percentage points above the efficient estimator presented by Shen and colleagues (2006). The relative efficiency of the proposed estimator increased particularly as a function of decreasing sample size and increasing population standard deviation.

Controlling Influenza A (H1N1) in China

This article discusses two approaches to controlling the newly identified influenza A (H1N1): via Bayesian and frequentist statistical reasoning. This study reviewed the measures implemented in China as an example to illustrate these two approaches. Since May 2009, China has deployed strict controlling mechanisms based on the strong prior Bayesian assumption that the origin of influenza A (H1N1) was from outside China and as such strict border control would keep the virus from entering China. After more than 4 months of hard work by Chinese health professionals and officials, the number of confirmed influenza A (H1N1) has increased steadily against the expectations of the general public. Taking into account the great financial burden in maintaining strict measures, Chinese health authorities have slowly adjusted their policies of controlling influenza A (H1N1) in China according to frequentist reasoning since July 2009.

Handling uncertainty in economic evaluations of patient level data: A review of the use of Bayesian methods to inform health technology assessments

Objectives: Due to potential advantages (e.g., using all available evidence), Bayesian methods have been proposed to assist healthcare decision making. This review provides a detailed description of how Bayesian methods have been applied to economic evaluations of patient level data. The results serve both as a reference and as a means by which to examine the appropriate application of Bayesian methods to inform decision making.

Methods: MEDLINE, EMBASE, and Cochrane Economic Evaluation databases were searched to identify studies, published up to November 2007, meeting three inclusion criteria: (i) the study conducted an economic evaluation, (ii) sampling uncertainty was incorporated using Bayesian methods, (iii) the likelihood function was informed by patient level data from a single source. Data were collected on key study characteristics (e.g., prior distribution, likelihood function, presentation of uncertainty).

Results: The search identified 366 potentially relevant studies, from which 103 studies underwent full-text review. Sixteen studies met the inclusion criteria. Half of the studies used uninformative priors; most studies incorporated the potential dependence between costs and effects, and presented cost-effectiveness acceptability curves. Results were sensitive to changes in the priors and likelihoods.

Conclusions: Limited use of informative priors, among the included studies, gives policy makers little guidance on one of the main benefits of Bayesian methods, the ability to integrate all available evidence to capture the uncertainty inherent in decision making.