Search

Background: Cardinality matching (CM), a novel matching technique, finds the largest matched sample meeting prespecified balance criteria thereby overcoming limitations of propensity score matching (PSM) associated with limited covariate overlap, which are especially pronounced in studies with small sample sizes. The current study proposes a framework for large-scale CM (LS-CM); and compares large-scale PSM (LS-PSM) and LS-CM in terms of post-match sample size, covariate balance and residual confounding at progressively smaller sample sizes.Methods: Evaluation of LS-PSM and LS-CM within a comparative cohort study of new users of angiotensin-converting enzyme inhibitor (ACEI) and thiazide or thiazide-like diuretic monotherapy identified from a U.S. insurance claims database. Candidate covariates included patient demographics, and all observed prior conditions, drug exposures and procedures. Propensity scores were calculated using LASSO regression, and candidate covariates with non-zero beta coefficients in the propensity model were defined as matching covariates for use in LS-CM. One-to-one matching was performed using progressively tighter parameter settings. Covariate balance was assessed using standardized mean differences. Hazard ratios for negative control outcomes perceived as unassociated with treatment (i.e., true hazard ratio of 1) were estimated using unconditional Cox models. Residual confounding was assessed using the expected systematic error of the empirical null distribution of negative control effect estimates compared to the ground truth. To simulate diverse research conditions, analyses were repeated within 10 %, 1 and 0.5 % subsample groups with increasingly limited covariate overlap.Results: A total of 172,117 patients (ACEI: 129,078; thiazide: 43,039) met the study criteria. As compared to LS-PSM, LS-CM was associated with increased sample retention. Although LS-PSM achieved balance across all matching covariates within the full study population, substantial matching covariate imbalance was observed within the 1 and 0.5 % subsample groups. Meanwhile, LS-CM achieved matching covariate balance across all analyses. LS-PSM was associated with better candidate covariate balance within the full study population. Otherwise, both matching techniques achieved comparable candidate covariate balance and expected systematic error.Conclusions: LS-CM found the largest matched sample meeting prespecified balance criteria while achieving comparable candidate covariate balance and residual confounding. We recommend LS-CM as an alternative to LS-PSM in studies with small sample sizes or limited covariate overlap. [ABSTRACT FROM AUTHOR]

Climate models have provided driving data for impact studies for decades. However, the uncertainties related to the use of such data have typically not been sufficiently considered. We investigate how CORDEX climate simulations, which were corrected for bias based on MESAN reanalysis data for the period of 1989–2010, match the gridded observational data set E‐OBS. Furthermore, we investigate whether the bias‐corrected simulations contain significant residual bias (RB), which we defined as the bias exceeding the range of the observational uncertainty (Uobs) that emerges from differences between the two data sets MESAN and E‐OBS. Because the reference period selected in climate change impact studies often differs from the period used for bias correction, we further investigated whether the RB and other performance metrics of the periods 1989–2010 and 1961–1990 differ. We conducted this assessment for whole Europe and for biogeographical zones. Most of the used performance metrics show a good match of the simulations with MESAN in the period of 1989–2010. The comparison against E‐OBS yields worse results, indicating a significant difference between the two observational data sets. Minor bias exceeding the range of Uobs (RB) occurred over large land areas of Europe in this period. Based on the analysis conducted for the period of 1961–1990, the RB is several times larger than that of the period of 1989–2010; the other metrics also show worse performances. Our findings imply that both the selection of the reference climate data set and reference period warrant greater attention in impact studies. In particular, we recommend researchers to use a bias correction period as reference period in their studies. Alternatively, a new bias correction should be applied if any different period is to be used as a reference. This study pointed out several aspects of disagreement between the bias‐corrected EURO‐CORDEX climate simulations and observational data sets E‐OBS and MESAN. The testing was performed separately for the bias correction period of 1989–2010 and the testing period of 1961–1990. Our findings imply that both the selection of the reference climate data set and reference period warrant greater attention in impact studies. [ABSTRACT FROM AUTHOR]