Conducting Trial-Based Economic Evaluations Using R: A Tutorial.

Trial-based economic evaluations are increasingly being conducted to support healthcare decision-making. When analysing trial-based economic evaluation data, different methodological challenges may be encountered, including (i) missing data, (ii) correlated costs and effects, (iii) baseline imbalances and (iv) skewness of costs and/or effects. Despite the broad range of methods available to account for these methodological challenges in effectiveness studies, they may not always be directly applicable in trial-based economic evaluations where costs and effects are analysed jointly, and more than one methodological challenge typically needs to be addressed simultaneously. The use of inappropriate methods can bias results and conclusions regarding the cost-effectiveness of healthcare interventions. Eventually, such low-quality evidence can hamper healthcare decision-making, which may in turn result in a waste of already scarce healthcare resources. Therefore, this tutorial aims to provide step-by-step guidance on how to combine appropriate statistical methods for handling the abovementioned methodological challenges using a ready-to-use R script. The theoretical background of the described methods is provided, and their application is illustrated using a simulated trial-based economic evaluation.

Mapping Oswestry Disability Index Responses to EQ-5D-3L Utility Values: Are Cost-Utility Results Valid?

OBJECTIVES: To develop and validate approaches for mapping Oswestry Disability Index responses to 3-level version of EQ-5D utility values and to evaluate the impact of using mapped utility values on cost-utility results compared with published regression models. METHODS: Three response mapping approaches were developed in a random sample of 70% of 18 692 patients with low back pain: nonparametric approach (Non-p), nonparametric approach excluding logical inconsistencies (Non-peLI), and ordinal logistic regression (OLR). Performance was assessed in the remaining 30% using R-square (R2), root mean square error (RMSE), and mean absolute error (MAE). To evaluate whether MAEs and their 95% limits of agreement (LA) were clinically relevant, a minimally clinically important difference of 0.074 was used. Probabilities of cost-effectiveness estimated using observed and mapped utility values were compared in 2 economic evaluations. RESULTS: The Non-p performed the best (R2 = 0.43; RMSE = 0.22; MAE = 0.03; 95% LA = -0.40 to 0.47) compared with the Non-peLI (R2 = 0.07; RMSE = 0.29; MAE = -0.15; 95% LA = -0.63 to 0.34) and OLR (R2 = 0.22; RMSE = 0.26; MAE = 0.02; 95% LA = -0.49 to 0.53). MAEs were lower than the minimally clinically important difference for the Non-p and OLR but not for the Non-peLI. Differences in probabilities of cost-effectiveness ranged from 1% to 4% (Non-p), 0.1% to 9% (Non-peLI), and 0.1% to 20% (OLR). CONCLUSIONS: Results suggest that the developed response mapping approaches are not valid for estimating individual patients' 3-level version of EQ-5D utility values, and-depending on the approach-may considerably affect cost-utility results. The developed approaches did not perform better than previously published regression-based models and are therefore not recommended for use in economic evaluations.

The handling of missing data in trial-based economic evaluations: should data be multiply imputed prior to longitudinal linear mixed-model analyses?

INTRODUCTION: For the analysis of clinical effects, multiple imputation (MI) of missing data were shown to be unnecessary when using longitudinal linear mixed-models (LLM). It remains unclear whether this also applies to trial-based economic evaluations. Therefore, this study aimed to assess whether MI is required prior to LLM when analyzing longitudinal cost and effect data. METHODS: Two-thousand complete datasets were simulated containing five time points. Incomplete datasets were generated with 10, 25, and 50% missing data in follow-up costs and effects, assuming a Missing At Random (MAR) mechanism. Six different strategies were compared using empirical bias (EB), root-mean-squared error (RMSE), and coverage rate (CR). These strategies were: LLM alone (LLM) and MI with LLM (MI-LLM), and, as reference strategies, mean imputation with LLM (M-LLM), seemingly unrelated regression alone (SUR-CCA), MI with SUR (MI-SUR), and mean imputation with SUR (M-SUR). RESULTS: For costs and effects, LLM, MI-LLM, and MI-SUR performed better than M-LLM, SUR-CCA, and M-SUR, with smaller EBs and RMSEs as well as CRs closers to nominal levels. However, even though LLM, MI-LLM and MI-SUR performed equally well for effects, MI-LLM and MI-SUR were found to perform better than LLM for costs at 10 and 25% missing data. At 50% missing data, all strategies resulted in relatively high EBs and RMSEs for costs. CONCLUSION: LLM should be combined with MI when analyzing trial-based economic evaluation data. MI-SUR is more efficient and can also be used, but then an average intervention effect over time cannot be estimated.

Effectiveness and Cost-effectiveness of Minimal Ovarian Stimulation in-vitro Fertilization versus Conventional Ovarian Stimulation in Poor Responders: Economic Evaluation Alongside a Propensity Score Adjusted Prospective Observational Study.

OBJECTIVE: Information on the pregnancy rate after successive in-vitro fertilization (IVF) cycles and their associated costs is relevant for couples undergoing assisted reproduction treatments (ARTs). This study, therefore, sought to investigate the effectiveness and the cost-effectiveness of two ARTs, the minimal ovarian stimulation IVF (MS-IVF) compared to the conventional ovarian stimulation IVF (C-IVF) from the payer's perspective. METHODS: A 10-months follow-up prospective observational study was conducted in a sample of couples who sought ARTs in a private clinic in Southern Brazil. Women had to satisfy the Bologna Criteria and be older than 35 years. The effect outcome was pregnancy rate per initiated cycle. Medication costs were based on medical records. Costs and effect differences were estimated using seemingly unrelated regressions adjusted for the propensity score estimated based on women's characteristics. RESULTS: All 84 eligible women who agreed to participate received a total of 92 IVF cycles (MS-IVF, n=27[35 cycles]; C-IVF n=57[57 cycles]. The effect difference between MS-IVF and C-IVF was -5.1% (95%CI, -13.2 to 5.2). Medication costs of MS-IVF were significantly lower than C-IVF by €-1260 (95%CI, -1401 to -1118). The probabilities of MS-IVF being cost-effective compared to C-IVF ranged from 1 to 0.76 for willingness-to-pay of €0 to €15,000 per established pregnancy, respectively. CONCLUSIONS: Even though there were no positive effect differences between groups, MS-IVF might be cost-effective compared to C-IVF from the payer's perspective due to its relatively large cost savings compared to C-IVF. However, further investigation is needed to confirm these findings in a larger sample.

To what extent does the use of crosswalks instead of EQ-5D value sets impact reimbursement decisions?: a simulation study.

PURPOSE: Inconsistent results have been found on the impact of using crosswalks versus EQ-5D value sets on reimbursement decisions. We sought to further investigate this issue in a simulation study. METHODS: Trial-based economic evaluation data were simulated for different conditions (depression, low back pain, osteoarthritis, cancer), severity levels (mild, moderate, severe), and effect sizes (small, medium, large). For all 36 scenarios, utilities were calculated using 3L and 5L value sets and crosswalks (3L to 5L and 5L to 3L crosswalks) for the Netherlands, the United States, and Japan. Utilities, quality-adjusted life years (QALYs), incremental QALYs, incremental cost-effectiveness ratios (ICERs), and probabilities of cost-effectiveness (pCE) obtained from values sets and crosswalks were compared. RESULTS: Differences between value sets and crosswalks ranged from -0.33 to 0.13 for utilities, from -0.18 to 0.13 for QALYs, and from -0.01 to 0.08 for incremental QALYs, resulting in different ICERs. For small effect sizes, at a willingness-to-pay of €20,000/QALY, the largest pCE difference was found for moderate cancer between the Japanese 5L value set and 5L to 3L crosswalk (difference = 0.63). For medium effect sizes, the largest difference was found for mild cancer between the Japanese 3L value set and 3L to 5L crosswalk (difference = 0.06). For large effect sizes, the largest difference was found for mild osteoarthritis between the Japanese 3L value set and 3L to 5L crosswalk (difference = 0.08). CONCLUSION: The use of crosswalks instead of EQ-5D value sets can impact cost-utility outcomes to such an extent that this may influence reimbursement decisions.

Interpretation of trial-based economic evaluations of musculoskeletal physical therapy interventions.

BACKGROUND: As resources for healthcare are scarce, decision-makers increasingly rely on economic evaluations when making reimbursement decisions about new health technologies, such as drugs, procedures, devices, and equipment. Economic evaluations compare the costs and effects of two or more interventions. Musculoskeletal disorders have a high prevalence and result in high levels of disability and high costs worldwide. Because physical therapy interventions are usually the first line of treatment for musculoskeletal disorders, economic evaluations of such interventions are becoming increasingly important for stakeholders in the field of physical therapy, including physical therapists, decision-makers, and reseachers. However, economic evaluations are relatively difficult to interpret for the majority of stakeholders. OBJECTIVE: To support physical therapists, decision-makers, and researchers in the field of physical therapy interpreting trial-based economic evaluations and translating the results of such studies to clinical practice. METHODS: The design, analysis, and interpretation of economic evaluations performed alongside randomized controlled trials are discussed. To further illustrate and explain these concepts, we use a case study assessing the cost-effectiveness of exercise therapy compared to standard advice in patients with musculoskeletal disorders. CONCLUSIONS: Economic evaluations are increasingly being used in healthcare decision-making. Therefore, it is of utmost importance that their design, conduct, and analysis are state-of-the-art and that their interpretation is adequate. This masterclass will help physical therapists, decision-makers, and researchers in the field of physical therapy to critically appraise the quality and results of trial-based economic evaluations and to apply the results of such studies to their own clinical practice and setting.

Development and validation of a patient no-show predictive model at a primary care setting in Southern Brazil.

Patient no-show is a prevalent problem in health care services leading to inefficient resources allocation and limited access to care. This study aims to develop and validate a patient no-show predictive model based on empirical data. A retrospective study was performed using scheduled appointments between 2011 and 2014 from a Brazilian public primary care setting. Fifty percent of the dataset was randomly assigned to model development, and 50% was assigned to validation. Predictive models were developed using stepwise naïve and mixed-effect logistic regression along with the Akaike Information Criteria to select the best model. The area under the ROC curve (AUC) was used to assess the best model performance. Of the 57,586 scheduled appointments in the period, 70.7% (n = 40,740) were evaluated including 5,637 patients. The prevalence of no-show was 13.0% (n = 5,282). The best model presented an AUC of 80.9% (95% CI 80.1-81.7). The most important predictors were previous attendance and same-day appointments. The best model developed from data already available in the scheduling system, had a good performance to predict patient no-show. It is expected the model to be helpful to overbooking decision in the scheduling system. Further investigation is needed to explore the effectiveness of using this model in terms of improving service performance and its impact on quality of care compared to the usual practice.