Evaluation of a national smoke-free prisons policy using medication dispensing: an interrupted time-series analysis

Abstract

Background: Internationally, smoking prevalence among people in prison custody (ie, people on remand awaiting trial, awaiting sentencing, or serving a custodial sentence) is high. In Scotland, all prisons implemented a comprehensive smoke-free policy in 2018 after a 16-month anticipatory period. In this study, we aimed to use data on medication dispensing to assess the impact of this policy on cessation support, health outcomes, and potential unintended consequences among people in prison custody.

Methods: We did an interrupted time-series analysis using dispensing data for 44 660 individuals incarcerated in 14 closed prisons in Scotland between March 30, 2014, and Nov 30, 2019. We estimated changes in dispensing rates associated with the policy announcement (July 17, 2017) and full implementation (Nov 30, 2018) using seasonal autoregressive integrated moving average models. Medication categories of primary interest were treatments for nicotine dependence (as an indicator of smoking cessation or abstinence attempts), acute smoking-associated illnesses, and mental health (antidepressants). We included antiepileptic medications as a negative control.

Findings: A 44% step increase in dispensing of treatments for nicotine dependence was observed at implementation (2250 items per 1000 people in custody per fortnight, 95% CI 1875 to 2624) due primarily to a 42% increase in dispensing of nicotine replacement therapy (2109 items per 1000 people in custody per fortnight, 1701 to 2516). A 9% step decrease in dispensing for smoking-related illnesses was observed at implementation, largely accounted for by respiratory medications (-646 items per 1000 people in custody per fortnight, -1111 to -181). No changes associated with announcement or implementation were observed for mental health dispensing or antiepileptic medications (control).

Interpretation: Smoke-free prison policies might improve respiratory health among people in custody and encourage smoking abstinence or cessation without apparent short-term adverse effects on mental health dispensing.

Funding: National Institute of Health Research Public Health Research programme, Scottish Government Chief Scientist Office, and UK Medical Research Council.

Figure 1

Medications of interest included in the analysis and their grouping for analytical purposes BNF=British National Formulary. COPD=chronic obstructive pulmonary disease. *The BNF is the pharmaceutical reference book used in the UK National Health Service.

Figure 2

Time-series analysis of fortnightly dispensing rate for medications for nicotine dependence in Scottish prisons during the study period Dashed lines indicate policy announcement and implementation dates.

Figure 3

Time-series analysis of fortnightly dispensing rate for medications for smoking-associated illnesses in Scottish prisons during the study period (A) All medications for smoking-associated illnesses. (B) Medications for respiratory illnesses. (C) Medications for cardiovascular illnesses. (D) Medications for gastrointestinal illnesses. (E) Medications for sensory illnesses. Dashed lines indicate policy announcement and implementation dates.

Figure 4

Time-series analysis of fortnightly dispensing rate for antidepressant medications in Scottish prisons during the study period Dashed lines indicate policy announcement and implementation dates.

Figure 5

Time-series analysis of fortnightly dispensing rate for antiepileptic medications in Scottish prisons during the study period Dashed lines indicate policy announcement and implementation dates.