Sas proc surveyreg

All statistical analyses were conducted with the SAS software, V.9.4. Differences with p values below 0.05 were considered statistically significant. SAS/STAT survey analyses were conducted because the KNHANES had applied a stratified cluster sampling method when selecting participants from 2008 to 2012. Descriptive statistics have been presented as means (SEs) for categorical variables and weighted percentages (SEs) for continuous variables and frequencies. A Pearson’s χ² test was used to compare categorical variables, and an independent samples t-test was used to compare continuous variables. The SAS PROC SURVEYREG was used to analyse continuous variables, and the PROC SURVEYFREQ was used for categorical variables. A multiple logistic regression analysis was conducted by the SAS PROC SURVEYLOGISTIC, after adjusting for multiple variables, to clarify the association between CRS phenotypes and general and central obesity (BMI-related and WC-related, respectively). Models were run after adding age and sex for model 1; smoking and severe drinking for model 2; and family income, residence, hypertension, stroke, bronchial asthma, influenza vaccination, allergic rhinitis, and pulmonary tuberculosis for model 3.

Descriptive statistics were used to analyse the data. Analysis of longitudinal trends in mean Hb between 1999 and 2018 were conducted using SAS® PROC SURVEYREG (SAS, Cary, NC, USA), employing the relevant stratum, cluster and weighting variables. Weighted statistics were used to calculate the prevalence of anaemia, with weight factors used to take into consideration confounding variables and the complex survey sampling scheme that is conducted with NHANES. The 20-year weight was calculated starting with the 4-year NHANES weights provided for 1999–2002 and then combining the weights for each 2-year period between 2003 and 2018. Weighted prevalence values were converted to estimates of the corresponding numbers in the US population using 2015–18 US census data.
Regression analysis was used to examine the association between the mean Hb levels across all study participants or the weighted prevalence of anaemia or Hb <10 g/dL and CKD stage, age, sex, race, body mass index, smoking status, and the presence of concurrent diabetes or uncontrolled hypertension.

Considering its distribution and prior publication [27 (link)], duration of diabetes was considered in 5-year strata, allocating the participants into three groups: < 5 years, 5–9 years, and ≥ 10 years from the time of diagnosis. The survey sampling methods of the KNHANES (multistate, stratified, and clustered sampling) were considered in all statistical analyses. In the multivariable logistic regression models, odds ratios (OR) and 95% confidence intervals (CI) were calculated for the association between the duration of illness and the prevalence of hyperglycemia, hypertension, and dyslipidemia. Potential confounding factors and effect modification were considered and examined through literature review [7 (link)8 (link)9 (link)10 (link)11 (link)28 (link)29 (link)] and descriptive statistical analysis approaches. Using SAS PROC SURVEYREG (version 9.4, SAS Institute Inc., Cary, NC, USA), multivariable generalized linear regression analysis was conducted to estimate means and standard errors (SE) of the prevalent hyperglycemia for the three different groups of illness duration for each survey. The P for trend was calculated using the median value for the prevalent hyperglycemia for each survey round between the KNHANES I (1998) and VI (2014). All analyses were performed using SAS, and the critical value for P was set at α = 0.05 in two-tailed tests.