Two-year follow-up of a clustered randomised controlled trial of a multicomponent general practice intervention for people at risk of poor health outcomes

Table 2 Hospital service use for the entire cohort

Data are the number and mean (standard deviation [SD]) per patient of hospital emergency department (ED) presentations, admissions and night stays
^aThe intervention effect (incidence rate ratio [IRR]) is calculated from a multilevel negative binomial regression model for the difference between the control and intervention groups over ^bone or ^ctwo years. The dataset comprises 1028 (98.5%) of 1044 patients who were matched to SA Health hospital records. Baseline, the 12-month period prior to the intervention; 12 months, the 12-month intervention period; 24 months, the 12-month period following the intervention period; CI, confidence interval

ISSN: 1472-6963