Fig. 4.

Proportion of patients self-harming at baseline and post intervention split by evening and non-evening period. The figure shows the proportion of patients self-harming per month between 1st June 2016 and 31st November 2019 after removing outliers. Evening refers to 3-11pm and non-evening period refers to any time of day excluding 3-11pm. The intervention was introduced on 1st June 2018 indicated by the vertical time. The solid red line indicates the segmented regression analysis conducted at baseline and at post-intervention. The red dashed line indicates the counterfactual scenario (i.e. projected proportion of patients self-harming if the intervention had not been conducted). The intervention therefore achieved its main effect during the evening periods