The association between public health team engagement in local alcohol licensing and public health and crime in England and Scotland: a timeseries analysis

Background
Public health teams (PHTs) in England and Scotland engage to varying degrees with local alcohol licensing processes to try to reduce harm to individuals and communities. Evidence is scarce on whether this engagement is effective. We aimed to quantify the effects of PHT engagement in alcohol licensing on selected health and crime outcomes.

Methods We matched 20 higher activity PHTs and 19 lower activity PHTs, defined on the basis of initial discussions with local authorities, in England and Scotland using propensity score methods. We developed a semi-quantitative metric based on structured interviews, documentary analyses, and expert consultation to provide an estimate for the intensity of each PHT’s engagement between April 2012 to March 2019. Public health outcomes were ambulance call outs, accident and emergency attendances, alcohol-related hospital admissions, alcohol-related and alcohol-specific mortality. Crime outcomes were violent and sexual crimes and public order offences. Data were analysed using negative binomial mixed-effects models, adjusted for area-level deprivation, age, season, country, time trend, population density, and local alcohol action area. Correlations were assessed between each outcome and PHT scores (range 0–35 points) and cumulative scores (range 0–364) for unlagged and 6-month lagged periods.

Findings
There was no evidence of associations between the primary exposure metric ‘18-months average PHT engagement scores’ and public health or crime outcomes. However, lagged and unlagged cumulative PHT engagement was correlated with A&E admissions for alcohol (–0·001 [95% CI –0·01 to –0·000] per unit exposure). Using the point estimates, had all areas had maximum PHT engagement across the whole period 318 500 (unlagged) or 227 620 (lagged) A&E admissions for alcohol (equivalent to 14–13%), might have been averted.

Interpretation
An important limitation of this analysis is that it is not based on a randomised controlled trial, nor does it use an exposure change (ie a natural experiment), which restricts the strength of causal inferences. However, we report what, to our knowledge, is the first evidence of an association between PHT engagement in alcohol licensing and subsequent effects on public health in England and Scotland. Our findings provide evidence that PHT engagement in licensing might have a measurable positive effect, suggesting allocation of PHT resources to alcohol licensing could be beneficial.