Problem

Smoking disproportionately affects people suffering with mental illnesses with rates being 3–4 times higher for those with schizophrenia or bipolar disorder.1 First episode psychosis is associated with decreased life expectancy2 due to cardiovascular risk factors and for the schizophrenia subgroup there is an increased morbidity and mortality.3 Alongside the illness, treatment with antipsychotics and other psychotropic medication can lead to metabolic syndrome which increases the risk.

The first episode of psychosis is typically preceded by subtle signs and prodromal symptoms. Early assessment and treatment are critical to maximise longer-term outcomes for the mental health and physical health of these patients. Early and Quick Intervention in Psychosis (EQUIP) is a community mental health team in City and Hackney; the latter is one of the most deprived boroughs of England until 2015.4 It provides care for adults (18–65 years old) in the first 3 years of illness.

There is a significant focus on improving quality of physical healthcare in order to bridge the gap between physical and mental health services and it is a strategic priority for East London NHS Foundation Trust (ELFT). Smoking cessation is associated with significant positive outcomes for both mental health and physical health5 6 in the longer term. Smoke-free policies have been implemented across services in the National Health Service (NHS), including mental health settings7 despite difficulties in implementation.8 9 The ELFT Smokefree Vision strategy is in line with recommendations from the National Institute for Health and Care Excellence (NICE), Public Health England and the Department of Health.

Background

The incidence of smoking is higher in groups of people with serious mental illness (SMI).10 Tobacco is a leading cause of death and a key modifiable factor for several physical health morbidities. Vulnerable groups such as people with low socioeconomic background, co-morbid alcohol and substance misuse, homeless and minority populations are over-represented among those with SMI. Given the overwhelming benefits of smoking cessation for physical11 and mental health12 this is a major potential area for intervention and improvement.

Historically patients with SMI engage poorly with smoking cessation services. Pragmatic approaches are required to address this13 and teams are faced with multiple barriers along the way.14 Patients report smoking to gain autonomy, relieve boredom, manage stress, gain pleasure and enjoyment, to develop social relations and as a form of social pastime, as a form of self-medicating, to manage or counteract the unwanted effects of some antipsychotic medications. Nicotine and co-morbid cannabis addiction, lack of motivation, stress, peer pressure and the severity of mental health symptoms are perceived barriers to cessation despite the patients’ concerns about the adverse physical health effects, the peer pressures to quit, the concerns about the economic cost. Nicotine replacement therapy, smoking cessation, pharmacological interventions, behavioural interventions, smoking cessation advisors and monetary incentives may be effective interventions.15

Members of staff hold mixed perceptions about smoking. Some consider it a patient right and hold misconceptions about patient inability to quit or even actively discourage patients from quitting.9 16 Staff may also feel that smoking is an important issue, ascribing variable urgency to this matter with other health complaints taking precedence.

The smoking status of staff is a known barrier to smoking cessation especially in inpatient settings.17 Other barriers include lack of formal training in offering smoking cessation and motivational interviewing, increased workload and other competing treatment priorities and lack of set targets for offering support and treatment.18 Given the complexity of systemic factors, barriers and facilitators relating to smoking and smoking cessation, the task of promoting smoking cessation is complex. A pragmatic approach is required when implementing strategies of change on a local level.

All EQUIP staff undertake level 1 smoking cessation training.19 Two members have completed level 2 smoking cessation training and are leads for the in-house smoking cessation programme. They provide concurrent supportive interventions for four patients at any time. Patients seeking smoking cessation are also referred to national smoking cessation (NSC) services including the Hackney Stop Smoking Service, Stop Smoking London, Stop Smoking clinics and pharmacies in the borough.

Baseline measurement

Baseline data in April 2017 indicated that out of 89 smokers on a caseload of 240 patients, only one had a smoking cessation assessment completed. The low completion rate was suggestive of a need to promote awareness of smoking cessation to staff and we used the percentage of completed smoking status assessments as a process measure.

We set the overall aim of the project at increasing the number of smoking cessation referrals through the national pathway, as recorded in the Trust’s patient electronic records (RIO), to a total of 15% of the number of smokers in the team caseload in 6 months. The target was decided as a ‘realistic’ but ambitious goal compared with the Clinical Commissioning Group’s expectation for 7% reduction in the number of smokers. Accounting for the insufficient baseline data, we safeguarded against the risk of an overinflated cumulative result: those who relapsed into smoking after successful cessation or transitioned from the service were removed from the outcome numbers.

All patients on the caseload were required to have a smoking assessment as part of the Trust’s physical health-related Commissioning for Quality and Innovation (CQUIN) target: lifestyle assessment. This assessment classifies patients as current smokers, ex-smokers and non-smokers and if in the former category eligibility for interventions and consent for referral to NSC services is recorded. Staff ought to undertake a paper, email or over the phone referral for any interested patients. An automatic electronic national referral option was introduced in electronic records in September 2018, however, full functionality was established in another couple of months.

For measurement, smoking cessation referrals were counted by feedback through various sources: Trust’s performance data, confirmation from smoking cessation providers of referrals received and immediate feedback from the care coordinators.

Design

We formed a local quality improvement (QI) team including doctors, care-coordinators and frontline staff. ELFT is partnered with the Institute for Healthcare Improvement (IHI), the project design followed the IHI model, and all project leads had received prior QI training.20 Service user involvement was sought and achieved from August 2018 onwards. Links with the inpatient service were created by the shared appointment of the senior trainee psychiatrist across inpatient and outpatient pathways. The initial step included the creation of a driver diagram (figure 1) with obstacles and facilitators to change, hypothesising a major role for staff attitudes and education. Consistently with this hypothesis, we set our project measures aimed at staff performance: our target was to increase smoking cessation referrals to 15% out of the total number of smokers and we aimed to achieve this in 6 months. Smoking status assessments, another performance indicator, were used as a process measure.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Driver diagram. EQUIP, Early and Quick Intervention in Psychosis; PDSA, plan-do-study-act; QI, quality improvement.

Strategy: plan-do-study-act (PDSA) cycles

Lessons and limitations

This project aimed to address a significant population health problem at a local team level in the challenging area of mental health. We used a rigorous QI PDSA methodology, in one site, with sequential staggered introduction of interventions some of which run concurrently. We adopted a pragmatic approach and chose outcome measures aiming primarily at staff attitudes, as directed by our driver diagram. Slow initial progress is indicative of the complexity of the issue and a long latency period for the first two PDSAs. Effective leadership and collaborative team work, persistence in the data collection despite shortcomings and commitment in the face of difficulty in progression was vital for the project’s progress.

Patient participation was particularly important and the third and fourth PDSAs were co-produced. The service user was an expert by experience, knowledgeable in the smoking cessation interventions and with a gift of passion and commitment to smoking cessation that was inspiring. This created a stimulating dialogue with QI members who had personal experience of smoking cessation.

The project team were able to manage with the impact of staff change-over although this affected data collection in June and September 2018. It can be hypothesised that such a change may have otherwise led to an early closure of the incomplete project. Effective leadership and patient participation was essential in regaining momentum.

The drop-in session change idea proved unsuccessful compared with the implementation of staff education measures. This indicated that the major driver in change is the day-to-day relationship between front-line staff and their patients, consistently with our primary hypothesis. We also neglected to collect measurable feedback on the perceived quality of the educational intervention (third PDSA) which makes the evaluation vulnerable to bias.

The Trust had established QI processes and methodologies in place and technical support from the QI team was invaluable. An improvement in the electronic lifestyle assessment form and referral pathway, leading to transition from manual and paper referral methods to e-referral processes was important but may also be a confounding factor.

Positive reinforcement of progress was given to staff on a regular basis by the project lead in the form of the QI project update email and healthy mutual support between staff to improve practice developed over time.

The cost effectiveness of this project was not studied. It is likely that this was a low cost intervention that made good use of already existing processes. However, the longer time required for implementation of change may be a limitation to future replication.

The target for 90% completion rate for smoking status assessments at all times was not achieved. This relates to (1) a lag between new referrals and first assessment completion and (2) expiring 6 monthly assessments that were not renewed in a timely manner. Challenges in this area were not overcome with the PDSAs used, and are suggestive of limitations related to time-commitments and suboptimal attitudinal change through interventions used (eg, in the inpatient site and primary source of many new referrals).

A hypothesis of the project that was not articulated from the outset was that access to smoking cessation pathways may lead to a reduction in the number of smokers. The initial design of the project did not aim to evaluate the true effectiveness of the interventions offered, by measuring dynamic variations in smoking rates. We consider this a limitation. However, we retrospectively identified that 25.7% of the total number of smokers had been recorded as having stopped smoking during the course of this project, which is important in the context of the complexity of the task.21 Sustainability of outcomes was studied in the short team and there is a possibility that results may diminish with time. We aim to repeat measurements after 6 months from the completion and closure of the project to revisit the sustainability of positive outcomes. The aim is also to extend this project by measuring actual changes in smoking status rather than completed referrals.

The findings of this project were developed into an intervention package and this was disseminated to other teams within the Trust wishing to carry out smoking cessation work. There are some limitations to the generalisability of interventions from results in one site, as EQUIP have specific structures and processes that differ from other community mental health teams (eg, a reduced caseload, quick access to psychology). Contextual factors that may facilitate replicability are the patient participation and the focus on staff education and performance. Gaps in data collection may be addressed in future trials by anticipating staff changeover and absences.

It is possible that change took place in ways that were not readily observable or measured, which is an identified risk for the pragmatic QI approach we chose.22 In keeping with the nature of the methodology, causality was not established, but an effective approach was developed and embedded to the team’s practice.

The collateral benefit of an increased number of smokers who did actually quit smoking is encouraging. Increasing awareness of research evidence on smoking and SMI, developing a wide range of interventions and developing a culture to challenge misconceptions about the intractable nature of smoking patterns in the mental health population contributed to change that may have been indirectly associated with the implemented change ideas.