This paper briefly introduces the definition, classification and significance of effect modification in epidemiological studies, summarizes the difference between effect modifier and confounders, and analyze the influence as well as the role of effect modification in epidemiological studies and Meta-analysis. In this paper, the possible scenarios of effect modification and related analysis strategy in Meta-analysis are indicated by graphics, aiming to arouse researchers' attention to effect modification. This paper also demonstrates how to identify and deal with effect modification in Meta-analysis through a study case of "Efficacy of sodium-glucose cotransporter 2 inhibitors in patients with type 2 diabetes", and shows the analysis process and interpretation of results of subgroup analysis and Meta-regression methods respectively. The advantages and disadvantages of these two methods are summarized to provide reference for the method selection of future research.

This paper briefly introduces the characteristics, research significance, and global reporting status of effect modification in network Meta-analysis, demonstrates the heterogeneity caused by effect modification in network Meta-analysis, and emphasizes the importance of exploring effect modification in network Meta-analysis. This paper also summarizes the normalized description and analysis strategies of effect modification in network Meta-analysis. Finally, by the case of "comparison of efficacy of three new hypoglycemic drugs in reducing body weight in type 2 diabetes patients", this paper demonstrates the realization of subgroup analysis and network Meta-regression in exploring effect modification, summarizes the advantages and disadvantages of the two methods, to provide references for future researchers.

Objective:Differences between randomized controlled trial (RCT) results and real world study (RWS) results may not represent a true efficacy-effectiveness gap because efficacy-effectiveness gap estimates may be biased when RWS and RCT differ significantly in study design or when there is bias in RWS result estimation. Secondly, when there is an efficacy- effectiveness gap, it should not treat every patient the same way but assess the real-world factors influencing the intervention's effectiveness and identify the subgroup likely to achieve the desired effect.Methods:Six databases (PubMed, Embase, Web of Science, CNKI, Wanfang Data, and VIP) were searched up to 31 st December 2022 with detailed search strategies. A scoping review method was used to integrate and qualitatively describe the included literature inductively. Results:Ten articles were included to discuss how to use the RCT research protocol as a template to develop the corresponding RWS research protocol. Moreover, based on correctly estimating the efficacy-effectiveness gap, evaluate the intervention effect in the patient subgroup to confirm the subgroup that can achieve the expected benefit-risk ratio to bridge the efficacy-effectiveness gap.Conclusion:Using real-world data to simulate key features of randomized controlled clinical trial study design can improve the authenticity and effectiveness of study results and bridge the efficacy-effectiveness gap.

This paper briefly introduces the unique advantages, overall analysis ideas and existing analysis methods of individual patient data Meta-analysis in terms of effect modification. In addition to Meta-regression and subgroup analysis, this paper also introduces the analysis methods based on part of individual patient data integrated with aggregated data and summarizes the current reporting of the above mentioned methods. In addition, the application and results interpretation of the above mentioned methods in individual patient data Meta-analysis are presented in this paper by taking "Effects of sodium-glucose cotransporter 2 inhibitors on SBP in patients with type 2 diabetes" as an example and by introducing their advantages and limitations.

Objective:Randomized controlled trials (RCT) usually have strict implementation criteria. The included subjects' characteristics of the conditions for the intervention implementation are quite different from the actual clinical environment, resulting in discrepancies between the risk-benefit of interventions in actual clinical use and the risk-benefit shown in RCT. Therefore, some methods are needed to enhance the extrapolation of RCT results to evaluate the real effects of drugs in real people and clinical practice settings.Methods:Six databases (PubMed, Embase, Web of Science, CNKI, Wanfang Data, and VIP) were searched up to 31 st December 2022 with detailed search strategies. A scoping review method was used to integrate and qualitatively describe the included literature inductively. Results:A total of 12 articles were included. Three methods in the included literature focused on: ①improving the design of traditional RCT to increase population representation; ②combining RCT Data with real-world data (RWD) for analysis;③calibrating RCT results according to real-world patient characteristics.Conclusions:Improving the design of RCT to enhance the population representation can improve the extrapolation of the results of RCT. Combining RCT data with RWD can give full play to the advantages of data from different sources; the results of the RCT were calibrated against real-world population characteristics so that the effects of interventions in real-world patient populations can be predicted.

Objective:To construct a diabetes foot prediction model for adult patients with type 2 diabetes based on retrospective cohort study using data from a regional health data platform.Methods:Using Yinzhou Health Information Platform of Ningbo, adult patients with newly diagnosed type 2 diabetes from January 1, 2015 to December 31, 2022 were included in this study and divided randomly the train and test sets according to the ratio of 7∶3. LASSO regression model and bidirectional stepwise regression model were used to identify risk factors, and model comparisons were conducted with net reclassification index, integrated discrimination improvement and concordance index. Univariate and multivariate Cox proportional hazard regression models were constructed, and a nomogram plot was drawn. Area under the curve (AUC) was calculated as a discriminant evaluation indicator for model validation test its calibration ability, and calibration curves were drawn to test its calibration ability.Results:No significant difference existed between LASSO regression model and bidirectional stepwise regression model, but the better bidirectional stepwise regression model was selected as the final model. The risk factors included age of onset, gender, hemoglobin A1c, estimated glomerular filtration rate, taking angiotensin receptor blocker and smoking history. AUC values (95% CI) of risk outcome prediction at year 5 and 7 were 0.700 (0.650-0.749) and 0.715(0.668-0.762) for the train set and 0.738 (0.667-0.801) and 0.723 (0.663-0.783) for the test set, respectively. The calibration curves were close to the ideal curve, and the model discrimination and calibration powers were both good. Conclusions:This study established a convenient prediction model for diabetic foot and classified the risk levels. The model has strong interpretability, good discrimination power, and satisfactory calibration and can be used to predict the incidence of diabetes foot in adult patients with type 2 diabetes to provide a basis for self-assessment and clinical prediction of diabetic foot disease risk.

Objective:To summarize the characteristics of pharmacoepidemiologic research involving diabetes patients, which were published in recent years, in terms of study design and analysis, and develop an identification process for time-related biases in pharmacoepidemiologic research.Methods:PubMed, Embase, CNKI and Wanfang were used for a systematical literature retrieval of relevant study papers published between January 1,2012 and September 26, 2022. Literature screening and data extraction were performed independently by two reviewers. Based on the mechanisms of different time-related biases and the characteristics of included study papers in terms of study design and analysis methods, an identification process for all types of time-related biases was developed.Results:A total of 281 study papers were included, of which 58 (20.64%) specifically mentioned certain time-related biases considered in the study. Based on the scoping review results, key points to identify time-related biases were summarized, involving data source, study design, control selection, comparator drugs, matching the duration of diabetes, identification of the washout period, identification of the induction/latency period, identification of the initiation of follow-up, identification of time window, statistical analysis methods, sensitivity analysis, and other design and analytical elements, in the identification process for time-related biases in pharmacoepidemiologic research.Conclusions:Time-related biases are common in pharmacoepidemiologic research and might significantly impact the study results. Based on scoping review results, this study further developed an identification process for time-related biases in pharmacoepidemiologic research, which will help researchers identify and avoid time-related biases and improve the reliability of related evidence in pharmacoepidemiologic research.

Objective:To develop a prediction model for the risk of diabetic retinopathy (DR) in patients with newly diagnosed type 2 diabetes mellitus (T2DM).Methods:Patients with new diagnosis of T2DM recorded in Yinzhou Regional Health Information Platform between January 1, 2015 and December 31, 2022 were included in the study. The predictor variables were selected by using Lasso-Cox proportional hazards regression model. Cox proportional hazards regression models were used to establish the prediction model for the risk of DR. Bootstrap method (500 resamples) was used for internal validation, and the performance of the model was assessed by C-index, the receiver operating characteristic curve and area under the curve (AUC), and calibration curve.Results:The predictor variables included in the final model were age of T2DM onset, education level, fasting plasma glucose, glycated hemoglobin A1c, urinary albumin, estimated glomerular filtration rate, and history of lipid-lowering agent and angiotensin converting enzyme inhibitor uses. The C-index of the final model was 0.622, and the mean corrected C-index was 0.623 (95% CI: 0.607-0.634). The AUC values for predicting the risk of DR after 3, 5, and 7 years were 0.631, 0.620, and 0.624, respectively, with a high degree of overlap of the calibration curves with the ideal curves. Conclusion:In this study, a simple and practical risk prediction model for DR risk prediction was developed, which could be used as a reference for individualized DR screening and intervention in newly diagnosed T2DM patients.

Objective:To construct a risk prediction model for diabetes kidney disease (DKD).Methods:Patients newly diagnosed with type 2 diabetes mellitus (T2DM) between January 1, 2015, and December 31, 2022, were selected as study subjects from the Yinzhou Regional Health Information Platform in Ningbo City. The Lasso method was used to screen the risk factors, and the DKD risk prediction model was established using Cox proportional hazard regression models. Bootstrap 500 resampling was applied for internal validation.Results:The study included 49 706 subjects, with an median ( Q1, Q3) age of 60.00 (50.00, 68.00) years old, and 55% were male. A total of 4 405 subjects eventually developed DKD. Age at first diagnosis of T2DM, BMI, education level, fasting plasma glucose, glycated hemoglobin A1c, urinary albumin, past medical history (hyperuricemia, rheumatic diseases), triglycerides, and estimated glomerular filtration rate were included in the final model. The final model's C-index was 0.653, with an average of 0.654 after Bootstrap correction. The final model's area under the receiver operating characteristic curve for predicting 4-year, 5-year, and 6-year was 0.657, 0.659, and 0.664, respectively. The calibration curve was closely aligned with the ideal curve. Conclusions:This study constructed a DKD risk prediction model for newly diagnosed T2DM patients based on real-world data that is simple, easy to use, and highly practical. It provides a reliable basis for screening high-risk groups for DKD.

Objective:To assess the safety of sitagliptin added to metformin on cardiovascular adverse events in real world patients with type 2 diabetes mellitus(T2DM).Methods:Real world data from Yinzhou Regional Health Care Database were used to select T2DM patients with diagnosis and treatment records in the platform from January 1,2017 to December 31,2022.According to drug prescription records,the patients were divided into metformin plus sitagliptin group(combination group)and metformin monotherapy group(monotherapy group).A series of retrospective cohorts were constructed according to the index date.Finally,full retrospective cohorts were constructed according to propensity score model,including baseline covariates that might be related to outcomes,to match the subjects in the combination group and monotherapy group for the purpose of increasing the comparability of baseline characteristics.The participants were followed up from the index date until the first occurrence of the following events:Diagnosis of outcomes,death,or the end of the study period(December 31,2022).Cox proportional risk model was used to estimate the hazard ratio(HR)and 95％confidence interval(CI)of sitagliptin added to metformin on 3-point major adverse cardiovascular events(3P-MACE)combination outcome and secondary cardiovascular outcomes.Results:Before propensity score matching,the proportion of the pa-tients in combination group using insulin,α glucosidase inhibitors,sodium-glucose transporter 2 inhibi-tors(SGLT-2I)and glienides at baseline was higher than that in monotherapy group,and the baseline fasting blood glucose(FBG)and hemoglobin A1c(HbA1c)levels in combination group were higher than those in monotherapy group.After propensity score matching,5 416 subjects were included in the combination group and the monotherapy group,and baseline characteristics were effectively balanced be-tween the groups.The incidence densities of 3P-MACE were 6.41/100 person years and 6.35/100 per-son years,respectively.Sitagliptin added to metformin did not increase or decrease the risk of 3P-MACE compared with the metformin monotherapy(HR=1.00,95％CI:0.91-1.10).In secondary outcomes analysis,the incidence of cardiovascular death was lower in the combination group than in the monothera-py group(HR=0.59,95％CI:0.41-0.85),and no association was found between sitagliptin and the risk of myocardial infarction and stroke(HR=1.12,95％CI:0.89-1.41;HR=0.99,95％CI:0.91-1.12).Conclusion:In T2DM patients in Yinzhou district of Ningbo,compared with metformin alone,sitagliptin added to metformin may reduce the risk of cardiovascular death,and do not increase the inci-dence of overall cardiovascular events.The results of this study can provide real-world evidence for post-marketing cardiovascular safety evaluation of sitagliptin.