Search

To assess any disparities in the initiation of second-line antidiabetic treatments prescribed among people with type 2 diabetes mellitus (T2DM) in England according to ethnicity and social deprivation level.This cross-sectional study used linked primary (Clinical Practice Research Datalink) and secondary care data (Hospital Episode Statistics), and the Index of Multiple Deprivation (IMD). We included people aged 18 years or older with T2DM who intensified to second-line oral antidiabetic medication between 2014 and 2020 to investigate disparities in second-line antidiabetic treatment prescribing (one of sulphonylureas [SUs], dipeptidyl peptidase-4 [DPP-4] inhibitors, or sodium-glucose cotransporter-2 [SGLT2] inhibitors, in combination with metformin) by ethnicity (White, South Asian, Black, mixed/other) and deprivation level (IMD quintiles). We report prescriptions of the alternative treatments by ethnicity and deprivation level according to predicted percentages derived from multivariable, multinomial logistic regression.Among 36 023 people, 85% were White, 10% South Asian, 4% Black and 1% mixed/other. After adjustment, the predicted percentages for SGLT2 inhibitor prescribing by ethnicity were 21% (95% confidence interval [CI] 19-23%), 20% (95% CI 18-22%), 19% (95% CI 16-22%) and 17% (95% CI 14-21%) among people with White, South Asian, Black, and mixed/other ethnicity, respectively. After adjustment, the predicted percentages for SGLT2 inhibitor prescribing by deprivation were 22% (95% CI 20-25%) and 19% (95% CI 17-21%) for the least deprived and the most deprived quintile, respectively. When stratifying by prevalent cardiovascular disease (CVD) status, we found lower predicted percentages of people with prevalent CVD prescribed SGLT2 inhibitors compared with people without prevalent CVD across all ethnicity groups and all levels of social deprivation.Among people with T2DM, there were no substantial differences by ethnicity or deprivation level in the percentage prescribed either SGLT2 inhibitors, DPP-4 inhibitors or SUs as second-line antidiabetic treatment.

Stated preference (SP) survey responses may not predict actual behavior, leading to hypothetical bias. We developed an approach that harnesses large-scale routine data to help SP surveys provide more accurate estimates of revealed preferences (RPs), within a study which elicited preferences for alternative changes to the blood service in England. The SP survey responses were used to predict the mean number of annual whole blood donations. Ex ante, the iterative survey design estimated hypothetical bias by contrasting pilot SP survey responses (N = 1254), with individually linked data on RPs, to inform the main SP survey design (N = 25,187). Ex post, the analysis recognized mediation of the relationship between SP and RP when blood donation is deferred. The pilot survey reported that donors' intended donation frequency of 3.2 (men) and 2.6 (women) times per year, exceeded their actual frequency by 41% and 30% respectively. Choice scenario attributes for the main SP survey were then modified, and over-prediction subsequently decreased to 34% for men and 16% for women. The mediating effect of deferrals explained 29% (men) and 86% (women) of the residual discrepancy between SP and RP. Future studies can use this approach to reduce hypothetical bias, and provide more accurate predictions for decision-making.

Background Diabetes mellitus (DM) is associated with increased cardiovascular (CV) risk. We compared health‐related quality of life (HRQoL), healthcare resource utilization (HRU), and clinical outcomes of stable post‐myocardial infarction (MI) patients with and without DM. Hypothesis In post‐MI patients, DM is associated with worse HRQoL, increased HRU, and worse clinical outcomes. Methods The prospective, observational long‐term risk, clinical management, and healthcare Resource utilization of stable coronary artery disease study obtained data from 8968 patients aged ≥50 years 1 to 3 years post‐MI (369 centers; 25 countries). Patients with ≥1 of the following risk factors were included: age ≥65 years, history of a second MI >1 year before enrollment, multivessel coronary artery disease, creatinine clearance ≥15 and

Nonignorable missing data poses key challenges for estimating treatment effects because the substantive model may not be identifiable without imposing further assumptions. For example, the Heckman selection model has been widely used for handling nonignorable missing data but requires the study to make correct assumptions, both about the joint distribution of the missingness and outcome and that there is a valid exclusion restriction. Recent studies have revisited how alternative selection model approaches, for example estimated by multiple imputation (MI) and maximum likelihood, relate to Heckman-type approaches in addressing the first hurdle. However, the extent to which these different selection models rely on the exclusion restriction assumption with nonignorable missing data is unclear. Motivated by an interventional study (REFLUX) with nonignorable missing outcome data in half of the sample, this article critically examines the role of the exclusion restriction in Heckman, MI, and full-likelihood selection models when addressing nonignorability. We explore the implications of the different methodological choices concerning the exclusion restriction for relative bias and root-mean-squared error in estimating treatment effects. We find that the relative performance of the methods differs in practically important ways according to the relevance and strength of the exclusion restriction. The full-likelihood approach is less sensitive to alternative assumptions about the exclusion restriction than Heckman-type models and appears an appropriate method for handling nonignorable missing data. We illustrate the implications of method choice for inference in the REFLUX study, which evaluates the effect of laparoscopic surgery on long-term quality of life for patients with gastro-oseophageal reflux disease.

Health economics studies with missing data are increasingly using approaches such as multiple imputation that assume that the data are "missing at random." This assumption is often questionable, as-even given the observed data-the probability that data are missing may reflect the true, unobserved outcomes, such as the patients' true health status. In these cases, methodological guidelines recommend sensitivity analyses to recognise data may be "missing not at random" (MNAR), and call for the development of practical, accessible approaches for exploring the robustness of conclusions to MNAR assumptions. Little attention has been paid to the problem that data may be MNAR in health economics in general and in cost-effectiveness analyses (CEA) in particular. In this paper, we propose a Bayesian framework for CEA where outcome or cost data are missing. Our framework includes a practical, accessible approach to sensitivity analysis that allows the analyst to draw on expert opinion. We illustrate the framework in a CEA comparing an endovascular strategy with open repair for patients with ruptured abdominal aortic aneurysm, and provide software tools to implement this approach.

The long-term progression of coronary artery disease as defined by the natural disease course years after a myocardial infarction (MI) is an important but poorly studied area of clinical research. The long-Term rIsk, clinical manaGement, and healthcare Resource utilization of stable coronary artery dISease in post-myocardial infarction patients (TIGRIS) study was designed to address this knowledge gap by evaluating patient management and clinical outcomes following MI in different regions worldwide. TIGRIS (ClinicalTrials.gov Identifier: NCT01866904) is a multicenter, observational, prospective, longitudinal study enrolling patients with history of MI 1 to 3 years previously and high risk of developing atherothrombotic events in a general-practice setting. The primary objective of TIGRIS is to evaluate clinical events (time to first occurrence of any event from the composite cardiovascular endpoint of MI, unstable angina with urgent revascularization, stroke, or death from any cause), and healthcare resource utilization associated with hospitalization for these events (hospitalization duration and procedures) during follow-up. Overall, 9225 patients were enrolled between June 2013 and November 2014 and are being followed in 369 different centers worldwide. This will allow for the description of regional differences in patient characteristics, risk profiles, medical treatment patterns, clinical outcomes, and healthcare resource utilization. Patients will be followed for up to 3 years. Here we report the rationale, design, patient distribution, and selected baseline characteristics of the TIGRIS study. TIGRIS will describe real-world management, quality of life (self-reported health), and healthcare resource utilization for patients with stable coronary artery disease ≥1 year post-MI.

Missing observations are common in cluster randomised trials. The problem is exacerbated when modelling bivariate outcomes jointly, as the proportion of complete cases is often considerably smaller than the proportion having either of the outcomes fully observed. Approaches taken to handling such missing data include the following: complete case analysis, single-level multiple imputation that ignores the clustering, multiple imputation with a fixed effect for each cluster and multilevel multiple imputation. We contrasted the alternative approaches to handling missing data in a cost-effectiveness analysis that uses data from a cluster randomised trial to evaluate an exercise intervention for care home residents. We then conducted a simulation study to assess the performance of these approaches on bivariate continuous outcomes, in terms of confidence interval coverage and empirical bias in the estimated treatment effects. Missing-at-random clustered data scenarios were simulated following a full-factorial design. Across all the missing data mechanisms considered, the multiple imputation methods provided estimators with negligible bias, while complete case analysis resulted in biased treatment effect estimates in scenarios where the randomised treatment arm was associated with missingness. Confidence interval coverage was generally in excess of nominal levels (up to 99.8%) following fixed-effects multiple imputation and too low following single-level multiple imputation. Multilevel multiple imputation led to coverage levels of approximately 95% throughout. © 2016 The Authors. Statistics in Medicine Published by John Wiley & Sons Ltd.

Cost-effectiveness analysis is well established for pharmaceuticals and medical technologies but not for evaluating variations in clinical practice. This paper describes a novel methodology--risk adjusted cost-effectiveness (RAC-E)--that facilitates the comparative evaluation of applied clinical practice processes. In this application, risk adjustment is undertaken with a multivariate matching algorithm that balances the baseline characteristics of patients attending different settings (e.g., hospitals). Linked, routinely collected data are used to analyse patient-level costs and outcomes over a 2-year period, as well as to extrapolate costs and survival over patient lifetimes. The study reports the relative cost-effectiveness of alternative forms of clinical practice, including a full representation of the statistical uncertainty around the mean estimates. The methodology is illustrated by a case study that evaluates the relative cost-effectiveness of services for patients presenting with acute chest pain across the four main public hospitals in South Australia. The evaluation finds that services provided at two hospitals were dominated, and of the remaining services, the more effective hospital gained life years at a low mean additional cost and had an 80% probability of being the most cost-effective hospital at realistic cost-effectiveness thresholds. Potential determinants of the estimated variation in costs and effects were identified, although more detailed analyses to identify specific areas of variation in clinical practice are required to inform improvements at the less cost-effective institutions.

Statistical methods have been developed for cost-effectiveness analyses of cluster randomised trials (CRTs) where baseline covariates are balanced. However, CRTs may show systematic differences in individual and cluster-level covariates between the treatment groups. This paper presents three methods to adjust for imbalances in observed covariates: seemingly unrelated regression with a robust standard error, a 'two-stage' bootstrap approach combined with seemingly unrelated regression and multilevel models. We consider the methods in a cost-effectiveness analysis of a CRT with covariate imbalance, unequal cluster sizes and a prognostic relationship that varied by treatment group. The cost-effectiveness results differed according to the approach for covariate adjustment. A simulation study then assessed the relative performance of methods for addressing systematic imbalance in baseline covariates. The simulations extended the case study and considered scenarios with different levels of confounding, cluster size variation and few clusters. Performance was reported as bias, root mean squared error and CI coverage of the incremental net benefit. Even with low levels of confounding, unadjusted methods were biased, but all adjusted methods were unbiased. Multilevel models performed well across all settings, and unlike the other methods, reported CI coverage close to nominal levels even with few clusters of unequal sizes.

Many cost-effectiveness analyses (CEAs) use data from observational studies. Statistical methods can only address selection bias if they make plausible assumptions. No quality assessment tool is available for appraising CEAs that use observational studies. We developed a new checklist to assess statistical methods for addressing selection bias in CEAs that use observational data. The checklist criteria were informed by a conceptual review and applied in a systematic review of economic evaluations. Criteria included whether the study assessed the 'no unobserved confounding' assumption, overlap of baseline covariates between the treatment groups and the specification of the regression models. The checklist also considered structural uncertainty from the choice of statistical approach. We found 81 studies that met the inclusion criteria: studies tended to use regression (51%), matching on individual covariates (25%) or matching on the propensity score (22%). Most studies (77%) did not assess the 'no observed confounding' assumption, and few studies (16%) fully considered structural uncertainty from the choice of statistical approach. We conclude that published CEAs do not assess the main assumptions behind statistical methods for addressing selection bias. This checklist can raise awareness about the assumptions behind statistical methods for addressing selection bias and can complement existing method guidelines for CEAs.

In cost-effectiveness analyses (CEA) that use randomized controlled trials (RCTs), covariates of prognostic importance may be imbalanced and warrant adjustment. In CEA that use non-randomized studies (NRS), the selection on observables assumption must hold for regression and matching methods to be unbiased. Even in restricted circumstances when this assumption is plausible, a key concern is how to adjust for imbalances in observed confounders. If the propensity score is misspecified, the covariates in the matched sample will be imbalanced, which can lead to conditional bias. To address covariate imbalance in CEA based on RCTs and NRS, this paper considers Genetic Matching. This matching method uses a search algorithm to directly maximize covariate balance. We compare Genetic and propensity score matching in Monte Carlo simulations and two case studies, CEA of pulmonary artery catheterization, based on an RCT and an NRS. The simulations show that Genetic Matching reduces the conditional bias and root mean squared error compared with propensity score matching. Genetic Matching achieves better covariate balance than the unadjusted analyses of the RCT data. In the NRS, Genetic Matching improves on the balance obtained from propensity score matching and gives substantively different estimates of incremental cost-effectiveness. We conclude that Genetic Matching can improve balance on measured covariates in CEA that use RCTs and NRS, but with NRS, this will be insufficient to reduce bias; the selection on observables assumption must also hold.

Economic evaluations must use appropriate costing methods. However, in multicentre cost-effectiveness analyses (CEA) a fundamental issue of how best to measure and analyse unit costs has been neglected. Multicentre CEA commonly take the mean unit cost from a national database, such as NHS reference costs. This approach does not recognise that unit costs vary across centres and are unavailable in some centres. This paper proposes the use of multiple imputation (MI) to predict those centre-specific unit costs that are not available, while recognising the statistical uncertainty surrounding this imputation.We illustrate MI with a CEA of a multicentre randomised trial (1014 patients, 60 centres), implemented using multilevel modelling. We use MI to derive centre-specific unit costs, based on centre characteristics including average casemix, and compare this to using mean NHS reference costs. In this case study, using MI unit costs rather than mean reference costs led to less heterogeneity across centres, more precise estimates of incremental cost, but similar estimates of incremental cost-effectiveness.We conclude that using MI to predict unit costs can preserve correlations, maximise the use of available data, and, when combined with multilevel modelling is an appropriate method for recognising the statistical uncertainty in multicentre CEA.

Evidence for efficacy of established treatment guidelines for chronic hepatitis C virus (HCV) disease is based on multinational randomized controlled trials (RCTs). Strategies for managing HCV, however, require an assessment of the effectiveness of intervention in routine clinical practice. We report the outcomes of combination therapy in a large cohort of HCV-infected individuals in the UK. A total of 347 (113 genotype 1, 234 genotype non-1) patients were treated with pegylated interferon and ribavirin according to current guidelines. Forty-two (37.2%) of those with genotype 1 infection and 164 (70.1%) with genotype non-1 infection achieved sustained viral response (SVR). Thirty-nine (11%) patients withdrew from treatment. In addition to viral genotype, factors predictive of a response to therapy were age at start of treatment and disease stage on pretreatment liver biopsy. Multivariate regression analysis demonstrated that the effects of age [odds ratio 0.5; 95% confidence interval (0.31-0.82) per 10-year increment (P = 0.006)] were confined to genotype 1 disease. In order to further inform the management of the individual patient, a multivariate logistic model was used to predict the probability of SVR for subgroups defined by disease stage, genotype and age at commencement of therapy. This model revealed striking differences in predicted response rates between subgroups and provided a strong rationale for early treatment, particularly for those with genotype 1 disease. Our study demonstrates that results comparable with those of RCTs can be achieved in clinical practice, and suggests that prediction of response rates based on probability modelling will provide a valuable adjunct to individual patient management.

Objective To demonstrate cost-effectiveness analysis (CEA) for evaluating different reimbursement models.

Multilevel models (MLMs) have been recommended for estimating incremental net benefits (INBs) in multicentre cost-effectiveness analysis (CEA). However, these models have assumed that the INBs are exchangeable and that there is a common variance across all centres. This paper examines the plausibility of these assumptions by comparing various MLMs for estimating the mean INB in a multinational CEA. The results showed that the MLMs that assumed the INBs were exchangeable and had a common variance led to incorrect inferences. The MLMs that included covariates to allow for systematic differences across the centres, and estimated different variances in each centre, made more plausible assumptions, fitted the data better and led to more appropriate inferences. We conclude that the validity of assumptions underlying MLMs used in CEA need to be critically evaluated before reliable conclusions can be drawn.

This paper examines the synthetic control method in contrast to commonly used difference‐in‐differences (DiD) estimation, in the context of a re‐evaluation of a pay‐for‐performance (P4P) initiative, the Advancing Quality scheme. The synthetic control method aims to estimate treatment effects by constructing a weighted combination of control units, which represents what the treated group would have experienced in the absence of receiving the treatment. While DiD estimation assumes that the effects of unobserved confounders are constant over time, the synthetic control method allows for these effects to change over time, by re‐weighting the control group so that it has similar pre‐intervention characteristics to the treated group. We extend the synthetic control approach to a setting of evaluation of a health policy where there are multiple treated units. We re‐analyse a recent study evaluating the effects of a hospital P4P scheme on risk‐adjusted hospital mortality. In contrast to the original DiD analysis, the synthetic control method reports that, for the incentivised conditions, the P4P scheme did not significantly reduce mortality and that there is a statistically significant increase in mortality for non‐incentivised conditions. This result was robust to alternative specifications of the synthetic control method. © 2015 The Authors. Health Economics published by John Wiley & Sons Ltd.

Multinational economic evaluations often calculate a single measure of cost-effectiveness using cost data pooled across several countries. To assess the validity of pooling international cost data the reasons for cost variation across countries need to be assessed. Previously, ordinary least-squares (OLS) regression models have been used to identify factors associated with variability in resource use and total costs. However, multilevel models (MLMs), which accommodate the hierarchical structure of the data, may be more appropriate. This paper compares these different techniques using a multinational dataset comprising case-mix, resource use and cost data on 1300 stroke admissions from 13 centres in 11 European countries. OLS and MLMs were used to estimate the effect of patient and centre-level covariates on the total length of hospital stay (LOS) and total cost. MLMs with normal and gamma distributions for the data within centres were compared. The results from the OLS model showed that both patient and centre-level covariates were associated with LOS and total cost. The estimates from the MLMs showed that none of the centre-level characteristics were associated with LOS, and the level of spending on health was the centre-level variable most highly associated with total cost. We conclude that using OLS models for assessing international variation can lead to incorrect inferences, and that MLMs are more appropriate for assessing why resource use and costs vary across centres.