CKD prevalence and risk factors Jamaica
Description
Variables used in the analysis and code book to create variables
Files
Steps to reproduce
Main exposure variables: Hypertension as defined as self-reported hypertension (if the respondent indicated medication use for hypertension) or if mean systolic pressure of three readings 140 mmHg or diastolic blood pressure 90 mmHg. Diabetes mellitus was defined as a fasting glucose >7.0 mmol/L or self-reported diabetes (if the respondent indicated medication use for diabetes). Body mass index (BMI) was calculated and then categorized in World Health Organization weight categories.(14, 20) Sickle cell trait (SCT) and disease were defined using haemoglobin genotyping. Responses to highest education level attained was used to determine education level. Responses were categorized as “less than secondary school” if the response included no education, primary, or junior secondary school (less than grade 10); “Secondary school” if responded with high school education (grades 10-13) and “More than Secondary school” for beyond a high school education. For household socioeconomic status (SES), participants’ responses to ownership of a list of 22 household assets were used. Terciles of the number of household assets were created as follows: Low household SES persons with ≤9 household assets, middle household SES with 10–12 items, and high household SES with 13–22 items.Smoking status was determined by the participant response to a question on tobacco use, and categorized as “Never Smoked” if response was never smoked, “Former smoker” if response was former smoker and “Current smoker” if response was yes –not every day or yes-daily. Data were analysed using STATA software (version 17BE ; StataCorp LP).). Sample-based estimates of means and standard deviation for continuous variables, and proportions and frequencies for categorical variables were obtained. Associations for explanatory variables and the outcome of interest using chi-squared tests for proportions and t-test, ANOVA. Application of sampling weights yielded nationally representative CKD prevalence estimates. The sampling weights were based on the probability of selection of dwellings and EDs, with adjustments for item non-response, and calibration using the Jamaican population distribution in parish- and sex-specific 5-year age bands. These base weights were multiplied by a non-response adjustment factor to produce survey weights adjusted for unit non-response. Bivariable logistic regression was performed on exposure variables and CKD as the outcome variable. Exposure variables with p-values ≤0.20 on bivariable analyses or a-priori association with CKD were included the multivariable logistic regression model. Hosmer–Lemeshow goodness-of-fit testing was done to assess differences between observed and expected results in the subgroups of the model population. A complete case analysis was performed.