Sas version 9.4m5

Event-free survival (EFS) was defined as the time from diagnosis of PTLD to relapse, progression, retreatment (second-line therapy), or death due to any cause. Overall survival (OS) was defined as the time from diagnosis until death due to any cause. The EFS and OS analyses were performed using the Kaplan–Meier method. Cox proportional hazards models were used to assess the association of clinical factors with OS. Baseline clinical and pathological characteristics and management categories by EBV status were compared using the Chi-square test, and P values of <0.05 were considered significant. OS stratified by event within 24 months was compared to the general US population as previously described for newly diagnosed DLBCL.^{13 (link)} The study was approved by Mayo Clinic IRB and conducted according to the Declaration of Helsinki. All analyses were performed using R version 3.6.2 and SAS version 9.4M5 software.

We described baseline characteristics, including publicly paid drug claims, according to whether individuals made an insulin claim to compare those who were likely to have type 1 diabetes with those who had type 2 diabetes or did not have diabetes. We used all available HbA_1c data, and the unit of analysis was the date on which the HbA_1c level was measured.
We used segmented regression models to analyze temporal changes in HbA_1c and to test our hypotheses about the impact of OHIP+ on temporal trends in HbA_1c. Segmented linear regression is a method to analyze temporal data to evaluate the impact of a policy change. We estimated the models using generalized estimating equation methods with an autoregressive covariance structure to account for repeated measurements for individuals.19 (link),20 We excluded young people with missing values for any variable specified in the models from the analyses. We fit an adjusted model without an interaction between deprivation quintiles to test whether temporal trends changed overall between periods and an adjusted model with interaction terms to test whether the change in temporal trends between periods differed across SES strata (Appendix 1, available at www.cmajopen.ca/content/10/2/E519/suppl/DC1). We used SAS version 9.4 M5 to conduct the analyses.

We categorized FEV₁/FVC ratios and further analysed the association between the different FEV₁/FVC ratios and presence of any respiratory symptom, as well as cough with phlegm, dyspnoea and wheeze by using multivariable logistic regression models. All models included age, sex, smoking, pack‐years, BMI and asthma. We used cubic restricted splines with four knots placed at the 5th, 35th, 65th and 95th percentiles for BMI and pack‐years among ever‐smokers, respectively (Harrell, 2015). In an extended analysis, we treated the FEV₁/FVC ratios as a continuous variable using a spline with five knots placed at FEV₁/FVC = 0.50, 0.60, 0.675, 0.725, 0.80 and 0.90. We analysed the entire population, and performed additional separate analyses for never‐smokers and ever‐smokers. We also performed a sensitivity analysis excluding all individuals with RSP.
All analyses were performed using SAS version 9.4 M5 (SAS Institute Inc, Cary, NC, USA). All results from the logistic regression models are expressed as ORs with 95% confidence intervals (CI).

Based on previous data with OW s.c. semaglutide, a sample size of 20 subjects was considered sufficient to achieve an acceptably narrow confidence interval for the dose ratio [22 (link)]. Dose proportionality was assessed by the ratio of AUC_{0–168 h,SS} for 0.5 mg/1.0 mg.
To account for potential dropout of up to 20%, a total of 24 subjects were planned for active treatment, and 12 subjects were planned for treatment with placebo (total for both placebo doses), leading to a total number of 36 subjects planned to start in the trial.
The primary endpoint, AUC_{0–168 h,SS,} was derived from the concentration-time curves 0–168 h (1 week) after last semaglutide dose using the non-compartmental, linear trapezoidal method on the observed concentrations using actual time points. The endpoint was analysed by a linear normal model based on the log-transformed values and back-transformed to provide dose ratios alongside 95% confidence intervals. The model included dose group as a fixed factor.
Accumulation ratio (R_acc,DC) was calculated as: $$\frac{{\text{AUC}}_{0-168\text{h},\text{sema},\text{ss}}/\text{last}\text{dose}(\text{in}\text{mg})}{{\text{AUC}}_{0-168\text{h},\text{sema},\text{SD}}/\text{first}\text{dose}(\text{in}\text{mg})},$$ where ‘last dose’ is the steady-state dose level of interest (either 0.5 or 1 mg) and ‘first dose’ is the first dose of trial product (0.25 mg).
Analyses were conducted using non-compartmental methods in the statistical software SAS, version 9.4 M5.

Statistical analyses were performed using SAS Version 9.4 M5. Baseline characteristics and some outcome data that have not been statistically compared are presented descriptively. In some cases, group sizes were too small for outcomes to be compared statistically. Statistical comparisons of continuous variables were performed using t-tests. Statistical comparisons of incidence rates were performed using Poisson regression. P-values <.05 were considered significant.
Clinical outcomes were determined each year for up to 10 years of follow-up. Changes in clinical outcomes were compared statistically between older adults and middle-aged adults at 2 years. The 2-year period was selected because it was considered a long enough period for changes to become apparent and, after 2 years, the number of patients decreased steadily.
SARs, NSARs, and SAEs are presented as incidence rates per 1000 patient-years and as incidence rate ratios (IRRs) for older vs middle-aged adults.