Identifying where hospital smokers were being lost between screening and support​

Client and setting

An integrated care board, for tobacco dependence services in four acute and maternity trusts and five borough community services​

The challenge

The board had put Ottawa and CURE model services into its hospitals but could not say how well they were working. Each trust and borough kept its own data in its own format. Performance conversations relied on headline dashboards. Nobody could tell whether differences between sites came from different populations, different pathways, or simply different recording.​

What we did

  • Built data collection templates around inputs, processes and outputs, based on the Ottawa and CURE models, so acute, maternity and community services could report the same steps in the same way.​

  • Collected, cleaned and combined a full year of data from four trusts and five borough services, suppressing values under seven and flagging services with only part-year data.​

  • Compared the system with London and England using the national tobacco dependence dashboard, then tested each step of the pathway in turn: identification, referral, bedside support and four-week follow-up.​

  • Broke down outcomes by deprivation, ethnicity, age and sex to check whether services were reaching the people most harmed by smoking.​

The outcome

The system identifies more inpatient smokers than any other in London, around 11,600 in the year, but refers fewer than one in five to its own teams, against nearly four in five nationally. Quit rates once people are seen are 31.9% against 20.7%. The problem lay in referral and staffing, not in the quality of clinical care, and the board could act accordingly.​

Timescale and team

Eight weeks from commission to final report. Partner-led, with public health analyst support.​

Previous
Previous

Reviewed a 400-item national learning catalogue for removal or retention​

Next
Next

We showed the stop smoking service worked, cost less than most, and where more referrals could come from​