Logistic Regression Propensity Score Research Articles

Statin intake can decrease acute pancreatitis severit

Background/ObjectivesAcute Pancreatitis (AP) is one of the leading gastrointestinal conditions requiring hospitalization. It has been suggested that statins could exert a protective role in the natural history of AP; however, their influence is not entirely clear. Our objective was to investigate the relationship between statin intake and AP. MethodsRetrospective analysis of a prospective registry of patients diagnosed with AP. Statin intake on admission as well as clinical, analytical, demographic and radiological data were recorded. Outcome parameters: Severity of AP, SIRS development, organ failure, local complications, intensive care admission, collection drainage, hospital length of admission, and death. Univariate and multivariate analyses as well as a propensity score logistic regression were conducted. ResultsFrom March 2014–October 2018 we studied 356 patients. 101 patients (28%) were taking statins. 55 (15%) suffered from moderate/severe pancreatitis. Multivariate analysis showed a 50% less risk of suffering from moderate/severe AP (OR 0.50 95% CI 0.22–1.0, p 0.50) and 33% less risk of developing local complications (OR 0.33 95%CI 0.15–0.80, p 0.014) among statin consumers, with a tendency towards less SIRS. Propensity score analysis confirmed that patients on statins suffering from AP had a lower risk to have a moderate/severe episode (OR 0.409 95%CI 0.192–0.872, p 0.031), to develop local complications (OR 0.47 95%CI 0.20–1.06, p 0.11) and SIRS (OR 0.516 95% CI 0.28–0.93, p 0.041). ConclusionsPatients taking statins who suffer from an episode of AP are more likely to follow a mild course and have a lower risk of developing local complications and SIRS.

Avascular Necrosis Is Associated With Increased Transfusions and Readmission Following Primary Total Hip Arthroplasty.

Avascular necrosis (AVN) may confer an increased risk of complications and readmission following total hip arthroplasty (THA). However, current risk-adjustment models do not account for AVN. A total of 1706 patients who underwent THA for AVN from 2011 to 2013 were selected from the American College of Surgeon's National Surgical Quality Improvement Program database and matched 1:1 to controls using a predetermined propensity score algorithm. Rates of 30-day medical and surgical complications, readmissions, and reoperations were compared between cohorts. Propensity-score logistic regression was used to determine independent associations between AVN and outcomes of interest. Patients with AVN had a higher rate of medical complications than those without AVN (20.3% vs 15.3%, respectively; P<.001). Bleeding transfusion was the most common medical complication, occurring at a significantly higher rate in patients with AVN than those without AVN (19.6% vs 13.9%, respectively; P<.001). Patients with AVN were also twice as likely to experience a readmission after THA (odds ratio, 2.093; 95% confidence interval, 1.385-3.164). Avascular necrosis of the femoral head is an independent risk factor for transfusion up to 72 hours postoperatively and readmission up to 30 days following total hip replacement. [Orthopedics. 2017; 40(3):171-176.].

Clinical effect of Kudiezi injection on renal function based on propensity score

To explore the effect of Kudiezi injection on renal function in the real world, in order to provide the basis for the clinical medication safety. Patient aged between 18-80 were selected from 18 large hospitals information system (HIS) databases established by clinical research institute for basic traditional Chinese medicine of China academy of Chinese medical sciences. The patients who were treated with Kudiezi injection (24 225 cases) were defined as the exposed group, whereas those who were not treated with Kudiezi injection (14,191 cases) were defined as the non-exposed group. The propensity score method was used to balance the confounding factors. Classic logistic regression, GBM weighted propensity score logistic regression, GBM propensity score weighted logistic regression with covariate and sensitivity analysis were adopted to study the effect of Kudiezi injection on renal function. The results showed no significant difference in the possibility in abnormality in serum creatinine (Scr) (P = 0.940, 0.679, 0.834) and urea nitrogen (BUN) (P = 0, 0.045, 0.164) between both groups. Therefore, the existing data indicated no damage of Kudiezi injection on renal function. Because this study is a retrospective study based on the real world, there may be unknown confounding factors and potential bias. Therefore, further studies shall be conducted to monitor whether Kudiezi injection causes damage on renal function, in order to ensure the clinical medication safety.

PIN15 - Comparative Effectiveness of Clostridial collagenase ointment for the treatment of Venous Leg Ulcers In outpatient Care Settings

 Retrospective, de-identified electronic medical records from 2007-2013 were extracted from the Intellicure Limited Data Set (I-LDS).  The I-LDS extracts records from 96 hospital-based outpatient wound centers and is part of the U.S. Wound Registry [NCT02280733].  A propensity score logistic regression was performed to test for treatment effect and adjust for covariance between wounds that received CCO and wounds treated with undifferentiated wound dressings [Figure 1].  The average number of CCO tubes used by patients was 2.6 (SD = 1.2) [Figure 4].

Data analysis of electronic medical recored clinical changes in indexs of liver and kidney function when salvianolate injection is parenterally administered over extended period

To understand the impact of salvianolate injection treatment of liver and kidney function using different from the hospital information system nationwide 18 large three hospitals (hospital information system, HIS) to extract using salvianolate age 18-80 years-old patient, a total of 10 470 cases, depending on the treatment used to have two times before and after treatment 7 d aspartate aminotransferase (AST), alanine aminotransferase patients (ALT), serum creatinine (Cr) and blood urea nitrogen (BUN) measurement indicators grouped according salvianolate continuous use different treatment patients were divided into two groups, continuous medication time > 14 d were defined as the observation group, ≤ 14 d were defined as the control group, continuous medication longer than 31 d were not included in the analysis. Each index number of the observation group and the control group were: ALT (268/1 465), AST (270/1 585), Cr (278/1 582), BUN (278/1 611). After using propensity score methods to balance two groups of covariates, based on unweighted logistic regression logistic regression propensity score weighting combined with propensity score weighting to adjust for covariates logistic regression of liver and kidney function in the two groups were analyzed. The results showed: three logistic regression analysis showed no likelihood of ALT, AST, Cr targets two groups of patients with abnormal statistically significant difference, continuous medication time > 14 d may increase the risk of abnormal BUN indicators, comprehensive analysis still can not explain use different treatment of patients salvianolate cause liver and kidney toxicity damage, still large prospective randomized controlled trials further study.