Propensity score methods belong to an analytical approach by incorporating the measured covariates and mimicking randomization to enhance the comparability between groups, hence reducing the impact of potential confounding in observational studies. Propensity score methods have been increasingly used in observational studies. This paper illustrates the principle and the methods based on the propensity score, in combination with its application in observational studies. It also compares results from propensity score methods with those from multivariable regression and randomized controlled trials. It was found that currently there has been a lack of recommendations for the selection of propensity score methods. Differences may exist when comparing results from propensity score methods with findings from typical regression analyses and randomized controlled trials.
ObjectivesTo compare the survival outcomes between hepatocellular carcinoma and hepatic angiosarcoma, and to develop and validate a nomogram predicting the outcome of hepatic angiosarcoma.MethodsThe Surveillance, Epidemiology and End Results (SEER) database was electronically searched to collect the data of hepatic angiosarcoma patients and hepatocellular carcinoma patients from 2004 to 2016. Propensity score matching (PSM) was used to match the two groups by the ratio of 1:3. Cox regression analysis was used to compare the survival outcomes between hepatic angiosarcoma and HCC. In the angiosarcoma group, population was divided into training set and validation set by 6:4. Nomograms were built for the prediction of half- and one- year survival, and validated by concordance index (C-index) and calibration plots.ResultsA total of 210 histologically confirmed hepatic angiosarcoma patients and 630 hepatocellular carcinoma patients were included. The overall survival of HCC was significantly longer than angiosarcoma (3-year survival: 18.4% vs. 6.7%, median survival: 5 months vs. 1 month, P<0.001), and the nomogram achieved good accuracy with an internal C-index of 0.751 and an external C-index of 0.737.ConclusionsThe overall survival of HCC is significantly longer than angiosarcoma. The proposed nomograms can assist to predict survival probability in patients with hepatic angiosarcoma. Due to limitation of the data of included patients, more high-quality studies are required to verify above conclusions.
ObjectiveTo evaluate the safety, efficacy, and cost-effectiveness of different uses of oxidized regenerated cellulose (ORC) in video-assisted thoracoscopic surgery (VATS) for lung cancer resection to provide a reference for the selection, clinical use, and rational utilization of absorbable hemostatic materials. MethodsA retrospective analysis of relevant data from inpatients who underwent VATS for lung cancer resection at a tertiary hospital from July 2019 to January 2020 and from July 2020 to December 2020 was conducted. Patients were divided into two groups based on the use of ORC: 1) combined use group (ORC and collagen sponge) and 2) sole-use group (ORC). Safety, efficacy, and economic outcome indicators were compared between the two groups. ResultsThe main analysis included a total of 904 patients, with 466 in the combined use group and 438 in the sole-use group. Compared to the combined use group, the sole-use group had a significantly longer hospital stay, used fewer hemostatic drugs, had a lower average cost of hemostatic materials, and a lower median total hospitalization cost (P<0.05). No statistically significant difference was found between the two groups in terms of intraoperative blood loss volume, massive blood loss rate, perioperative transfusion rate, reoperation rate, postoperative 48-hour drainage volume, bloody drainage fluid rate, or postoperative laboratory test indicators. ConclusionThere was no significant difference in the safety or efficacy of VATS for lung cancer resection between the sole use of ORC and the combined use of ORC, but the sole use of ORC was associated with a lower cost of hemostatic materials and a lower total hospitalization cost. The sole use of hemostatic gauze in VATS for lung cancer resection may be a more cost-effective choice.
ObjectiveTo explore the difference between minimally invasive direct and conventional thoracotomy off-pump coronary artery bypass surgery (CABG). MethodsWe selected 276 patients underwent off-pump CABG surgery in our hospital from June 2005 through June 2014. There were 55 patients with minimally invasive off-pump CABG surgery and 221 patients conventional thoracotomy surgery. By using the method of peopensity score matching, we selected 55 conventional thoracotomy patients as a control group in our study. There were 41 males and 14 females at age of 60.8±10.5 years with minimally invasive off-pump CABG surgery, 44 males and 11 females at age of 60.6±12.5 years with conventional thoracotomy. ResultsThere was no statistical difference in surgery time, stay in the intensive care unit (ICU) time between conventional thoracotomy surgery and minimally invasive off-pump CABG. Compared with conventional thoracotomy surgery, minimally invasive off-pump CABG patients had statistical improvement in post-operative hospital stay time (7.3±3.1 d vs. 8.8±3.9 d, P=0.01), postoperative drainage (684(0-2 790)ml vs. 739(50-4 460)ml, P=0.03), perioperative blood transfusion (1.91(0-20)U vs. 6.62(0-20)U, P=0.00), surgery incision length (5.6±1.1 cm vs. 26.3±4.5 cm, P=0.00). ConclusionOverlooking the learning curve, minimally invasive direct off-pump CABG surgery has more advantages than conventional thoracotomy surgery. It is a safe and effective procedure.
Objective To compare the ability of three propensity score weighting methods to balance the covariates and the advantages and disadvantages to estimate the treatment effects when dealing with multiple treatment data under different sample sizes. Methods Monte Carlo simulation was used to generate data sets and the advantages and disadvantages of balancing covariates and estimating the treatment effects of three propensity score weighting methods, Logistic-IPTW, Logistic-OW and GBM-OW were compared. The evaluation index of covariate equilibrium level was the absolute standard mean difference. The evaluation indexes of effect estimation included the point estimate of treatment effect, root mean square error and confidence interval coverage. Results Compared with Logistic-IPTW and Logistic-OW, GBM-OW was better in effect estimation and had a smaller root mean square error in five scenarios where covariates were related to treatment factors and outcome variables with different varying degrees of complexity. In terms of covariate equilibrium, all three methods had good effects. GBM-OW method performed better when the overlap of propensity score distribution of multiple treatment data was relatively low and covariables had increasingly complex nonlinear relationships with treatment factors and outcome variables. Conclusion When dealing with multiple treatment data, GBM-OW method has advantages over the other two methods when there is nonlinearity and/or interaction between covariates and treatment factors and outcome variables. Using this method, the effect estimation is closer to the real value, which is a better choice.
With the continuous progress of national medical insurance strategic purchasing and value-based healthcare, pharmacoeconomic evaluation, serving as a technical tool for assessing the cost-effectiveness of healthcare interventions, has played an important role in policy decision support. Comparative efficacy evidence is the core data source for pharmacoeconomic evaluation, and also the foundation for conducting pharmacoeconomic research. In recent years, the number of innovative drugs approved based on single-arm trial has been increasing. Most existing randomized controlled clinical trials are also placebo-controlled or compared with traditional treatments, unable to directly meet the need for efficacy evidence of comparisons with conventional or standard treatments in pharmacoeconomic evaluations. In the absence of direct comparative efficacy evidence, exploring indirect comparison methods for efficacy has become a cutting-edge direction in pharmacoeconomic evaluation. Through a comprehensive literature review and systematic analysis, this study focuses on five indirect comparison methods based on individual patient data for population adjustment, including match adjusted indirect comparison (MAIC), simulated treatment comparison (STC), propensity score matching (PSM), inverse probability of treatment weighting (IPTW) and network meta regression (NMR), and discussing their basic concepts, advantages and disadvantages and application comparisons. Finally, it provides methodological suggestions on how to choose an indirect comparison method for efficacy, with the aim of promoting the generation of higher-quality indirect comparison evidence for efficacy and advancing pharmacoeconomic evaluation to provide high-quality evidence references for healthcare policy decision-making.
Objective To analyze the effects of hyperuricemia (HUA) on the prevalence of dyslipidemia in the elderly. MethodsA total of 5 990 elderly people with complete and important variables from the China Health and Retirement Longitudinal Study (CHARLS) public database in 2015 were extracted. Their blood lipids, related physiological and biochemical indices, and basic demographic information were collected. The effects of HUA on the prevalence of dyslipidemia in the elderly were analyzed using the probit model, and empirical analysis was performed using the propensity score matching method (PSM). Results Among the 5 990 subjects, 13.6% of the elderly had HUA and the prevalence of dyslipidemia was 37.5%. After correcting the endogeneity among variables, the probability of dyslipidemia in elderly patients with HUA increased by 9.5%-11.7% (P<0.01), in which the probability of high triglyceridemia (TG), high total cholesterol (TC), high low-density lipoprotein cholesterol (LDL-C), and low high-density lipoprotein cholesterol (HDL-C) increased by 10.4%-11.5% (P<0.01), 2.7%-3.8% (P<0.01), 1.7%-2.3% (P<0.05), and 4.3%-4.9% (P<0.05), respectively. Conclusion HUA is associated with various types of dyslipidemia, among which its relationship with high TG and low HDL-C is strong. Targeted interventions should be taken for elderly HUA patients, aiming to reduce the rate of dyslipidemia and promote the goal of "healthy ageing" in China.
ObjectiveTo study the direct economic burden of hospitalization in patients with carbapenem-resistant Enterobacteriaceae (CRE) infection.MethodsPatients with CRE detected in Jianyang People’s Hospital between January 2017 and June 2019 were divided into infection group and colonization group, and multiple linear regression analysis was used to analyze the confounding factors, and then propensity score matching method was used to match the confounding factors of the two groups, finally the direct economic burden of hospitalization was compared between the two groups.ResultsA total of 2 013 patients were enrolled, including 507 CRE-infected patients and 1 506 CRE-colonized patients. Multiple linear regression results showed that factors affecting the direct economic burden of hospitalization included mechanical ventilation, intravenous catheterization, blood transfusion or use of blood products, urinary intubation, tracheotomy, gender, surgery, intensive care unit stay, main diagnosis, and the type of medical insurance, together with CRE infection (P<0.05). After propensity score matching, these confounding factors were well balanced between the infection group (n=249) and the colonization group (n=249), and the differences were not statistically significant (P>0.05). The median of total hospitalization cost of CRE-infected patients was 1.29 times (15 589.23 yuan more than) that of CRE-colonized patients and the difference between the two groups was statistically significant (Z=–3.116, P=0.002). The top two types of hospitalization cost with largest differences in the medians were the cost of Western medicine (4 850.21 yuan; Z=–3.460, P=0.001) and the cost of laboratory diagnosis (2 613.00 yuan; Z=–3.529, P<0.001), respectively. The cost of antimicrobial drugs differed significantly between the two groups (Z=–3.391, P=0.001), and it was 1.54 times in patients with CRE infection what it was in patients with CRE colonization.ConclusionsThere are many factors affecting the direct economic burden of inpatients, and it is necessary to avoid the influence of confounding factors as far as possible during analysis. The economic burden of patients with CRE infection is significantly greater than that of CRE-colonized patients, so active measures should be adopted to prevent and control CRE infection.
ObjectivesTo explore the value of neural networks (NN) in estimating propensity score, and to compare the performance of propensity score methods based on both logistic regression (LR) and NN.MethodsData sets including ten binary or continuous covariates, binary treatment variable and continuous outcome variable were simulated by SAS 9.2 software, and 5 scenarios differing by non-linear and/or non-additive associations between treatment assignment and covariates were set up. The sample sizes 500, 1000, 2000, 5000 and 10000 were considered. Propensity scores were estimated using either LR or NN model using only partial covariates associated with the outcome (methods LR1, NN1), or all covariates associated with either outcome or treatment (methods LR2, NN2). The average treatment effect (ATE) estimates, standard error (SE), bias, and mean square error (MSE) of ATE among the different models were compared.ResultsThe 95% confidence intervals of the average treatment effect were narrower in NN than that in LR models. SE, bias and MSE increased with the increasing complexity of non-linear and/or non-additive associations between the treatment and covariates, and smaller SE, bias, and MSE were observed in LR1 than LR2, and in NN1 than NN2. NN generally produced less bias than LR under most scenarios when variables associated with the outcome were introduced. SE and MSE decreased with the increasing sample size for both LR and NN models.ConclusionsNN for estimating propensity scores may be less biased and produce more precise estimates for ATE than LR in a meaningful manner when the complex association between treatment and covariates exists.
ObjectiveTo explore the clinical efficacy of fascial manipulation (FM) treatment in patients with chronic ankle instability (CAI).MethodsThe clinical data of CAI patients who received rehabilitation treatment in the Department of Rehabilitation Medicine of the Second Hospital of Jilin University from October 2018 to December 2020 were retrospectively collected. According to different treatment methods, patients were divided into balance training (BT) group and FM group. The BT group received BT for 4 weeks, while the FM group received BT for 4 weeks after FM treatment. Propensity score matching (PSM) was used for 1∶1 matching to compare the effects of different treatment options on the dysfunction of CAI patients. Foot and Ankle Ability Measure (FAAM) [including FAAM-activity of daily living (FAAM-ADL), activity of daily living (ADL) self-scoring, FAAM-sports (FAAM-S), and sports self-scoring], center of pressure (COP), foot lift test (FLT) were used to evaluate the changes in balance function and symptoms pre-treatment and post-treatment.ResultsA total of 52 patients were included, including 24 cases in FM group and 28 cases in BT group. Finally, after PSM method, 34 patients were included, 17 cases in each group. Before intervention, there was no significant difference in FAAM, COP and FLT between the two groups (P>0.05). After the intervention, FAAM-ADL, ADL self score, COP and FLT in the FM group were better than those in the BT group (P<0.05); there was no significant difference between FAAM-S and exercise self score (P>0.05). Before and after the intervention, FAAM, COP and FLT were improved in both groups (P<0.05). The improvement of FAAM ADL, ADL self-score and the decrease of COP in FM group were higher than that of the BT group (P<0.05). Comparison of FAAM-S, exercise self score and FLT before and after intervention, there was no significant difference between the two groups (P>0.05).ConclusionBT can improve the function of patients with CAI, and the combination of FM is more effective in improving the ability of daily living and static balance.