Alcohol Withdrawal Management for acute admissions to hospital

When the initial project plan was set up in March 2015, its aims were as follows:

1. Optimise medical and psychosocial care for inpatients with alcohol dependence in the acute trust
2. Increase recovery rates and reduce readmissions

The following objectives and timescales were proposed:

First 6 months - SOFT LAUNCH

1. Baseline evaluation of existing treatment, including establishing baseline and follow-up cohort
2. Collect feedback from patients on existing experience of treatment; invite patients to participate in design of treatment guidelines
3. Design alcohol care bundle and training programme based on evidence-based practice guidance (The forms were trialled and changed quickly according to feedback from ward staff)
4. Develop pathways into aftercare in partnership with CNWL, LNWH, patients and community organisations.

Second 6 months - Recruitment phase

5. Introduce finalised alcohol care bundle and pathways to the acute medical unit. Iteratively develop and monitor their use
6. Iteratively evaluate effect on clinical practice and establish further follow-up cohort

Final 6 months

7. Evaluation and dissemination
8. Roll out to other ward areas and throughout the Trust (LNWH)

The project was submitted to CLAHRC NWL - with a view to applying quality improvement methods using the models they had developed. By the time the project had been accepted and launched with their support (October 2015 to January 2016), the aims had been reviewed as below.

Project aims April 2016

To improve the experience of care and long-term well-being of Northwick Park Hospital in-patients who were identified as alcohol dependent.

To develop a protocol for interventions that ward staff and medical colleagues could carry out to ensure that trust guidelines are followed in a timely and consistent way.

The protocol included an adaptation of a validated alcohol withdrawal screening tool, to enable nurses to report more easily why additional “as required” drugs had been given.

It was intended that learning from this project would be used to review and improve the trust’s alcohol management guideline.

The Alcohol Care Bundle we started with was a 14 point checklist of all clinical actions we expected would take place during the course of an in-patient detox (see supporting attached document).

London North West University Hospital (LNWUH) Trust looks after 4 hospitals providing care for around 900,000 people in Ealing, Harrow and Brent. Employing 4,000 nurses and 1,200 doctors. We receive over 70,000 emergency admissions per year at two main acute sites. This project focussed on acute admissions at Northwick Park Hospital which is located on the border between Harrow and Brent.

Baseline hospital data (2014/15) showed around 20 people per month with primary alcohol diagnoses (see below) and between 60 to 80 people with secondary diagnoses. Usually the ICD10 codes for diagnosis were allocated when alcohol withdrawal treatment was carried out.

Central and North West London (CNWL) NHS Trust provides liaison psychiatry services for Northwick Park, including an alcohol nurse specialist. CNWL also provides two alcohol nurse specialists at Ealing Hospital.

Liaison psychiatry at Northwick Park had adopted the role of supporting colleagues working in acute settings with management of alcohol withdrawals. Our focus was improving care in a non-specialist setting, as we knew that symptoms were often missed during admissions therefore delaying alcohol care.

Regularly referrals were made to liaison psychiatry for agitated, anxious and often delirious people, under-treated for alcohol withdrawals.

Rather than opportunistic detoxification from alcohol (which can leave patients unsupported) we wanted hospital admissions to progress more smoothly for patients and medical colleagues.

Doctors were not confident to prescribe sufficient doses, and nurses were often unaware of the signs and symptoms.

The early stages of admission were the key time to start treatment, as early higher dosing (front-loading) can diminish severe withdrawal symptoms and help the anxious drinker to know that they are safe from incidents of delirium tremens or seizures.

Staff were trained to recognise risks and symptoms of alcohol withdrawal earlier, and to prevent severe problems for patients, including delaying important clinical procedures and prolonging admission to hospital. We also hoped to increase knowledge among the admitting wards’ doctors and nurses, reduce distress to their patients, and ensure better medical care for people with alcohol dependence. We anticipated this would reduce the incidence of those severe withdrawal states and perhaps reduce length of stay in hospital for some patients.

Quality improvement methods were used, for which project group members received mentoring from the NW London CLAHRC. There was matched funding for 18 months project of £163,000 from  CLAHRC NW London and CNWL to support staffing of the project lead (Band 7 nurse specialist), one session per week time of senior consultant psychiatrist and hepatology and acute medicine leads plus sessional input from colleagues, e.g. pharmacy. This funding included a factor to cover expenses and payment for service user reps. There was also funding to support meetings – mostly per session for other clinical staff plus admin and paperwork for the project.

Process Map – initial scoping of what happens when patients are admitted to hospital with risk or evidence of alcohol withdrawals. This was prepared with participation of nurses and doctors from acute medical wards, emergency department and pharmacy. We also had contributions from several patients who had undergone detox in the hospital.

Action Effect Diagram to outline our intended changes, and steps towards our over-arching aims (as per earlier section)

Stakeholder Meetings

Regular weekly meetings attended by core team: patient representative, clinical lead, pharmacy, and CLAHRC QI colleagues. Consultants and matrons who often could not attend meetings contributed support to the project through their managerial roles.

Monthly meetings with all stakeholders invited, and the project manager ensured that those people who could not attend were met with, so their views were addressed in the meetings.

Alcohol Care Bundle was designed and tested iteratively using Plan-Do-Study-Act cycles.

These were focussed on design of the care bundle form and the Clinical Institute Withdrawal Assessment of Alcohol (CIWA) versions which went through 12 and 6 revisions respectively in the first phase (see attached document)

At key times during the project, training was carried out to inform nurses in ED and acute wards, plus ED doctors – who are most likely to start the prescribing of detox regimens. We carried out a limited evaluation of nurses’ knowledge after these sessions, and found an increased awareness of alcohol withdrawals and recommended drinking limits afterwards. The survey also gave positive feedback about the use of the forms from nurses who use them.

See initial aims numbered above,

1. Our baseline tool showed that diagnostic coding was inaccurate for many patients. We used these findings to influence changes in practice – an unexpected gain.
2. Not enough patients could be contacted afterwards due to poor engagement with follow up. There was a recovery worker with personal experience of detox in the hospital and his conversations with patients proved useful in our PDSAs.
3. A simple version of the protocol was introduced to admissions unit (rather than implementing a care bundle). This was favourably received as ward staff felt involved in its introduction.
4. Our local services changed and links are now being re-established, but because our protocol is robust in stabilising withdrawals, we are hopeful that local agencies will take people for home based end of detoxes.
5. It has been established to use CIWA-M to set initial prescription. Compliance with the protocol increased and we aim to re-audit 12 months after its introduction.
6. From our surveys, staff showed more knowledge about the protocol and despite staff turnover this has been sustained. We continue to inform new medical staff about this at their four monthly rotation.
7. Statistical analysis of audit data was not conclusive.

Data was not collected as the project ran. Notes were audited against a set of criteria which had to be met to say that the care bundle was followed correctly. Not all cases during the testing phase had 100% completion of the care bundle, but those that did had a slight improvement in lengths of stay and detox. A key factor here appeared to be use of higher initial doses of chlordiazepoxide - as indicated by CIWA-M scores. However, statistical analysis was not conclusive. Although there were two series of sustained improvement in consecutive cases, this was not seen in weekly aggregated data.

AUDIT-C screening tool was not used in most cases, so instead we accepted records of alcohol use in notes with details of amount and frequency, i.e. the first two questions of AUDIT-C could be completed.

8. The protocol is now standard practice throughout inpatient wards, and subsequent roll out of the trust protocol has coincided with better coding and an apparent reduction in length of stay for dependent drinkers (see run charts in attached document).

The main change would be to continuously collect more information about prescribing from individual drug charts. Because the demands of development and education fell to one project lead, and it became clear that case notes were not always available when we wanted to audit them. Therefore we did a retrospective audit of case notes and prescriptions. Most people who started alcohol care bundles during the project had their notes audited (137/210). When recruitment and audit took longer than expected it was decided to focus on the first two phases.

A second change would be to agree follow up opportunities to evaluate experiences of the care bundle with patients who had completed detox while in hospital. It can be a difficult time for people to engage, and during the project, we did engage link workers with Brent and Harrow community alcohol recovery services. However, it was found that the majority of people referred did not attend their appointments.

Two key members of the team left early in the project, meaning that project lead role had to change. All of the ward “champions” - i.e. nurses who would take any developments back to their colleagues – moved on.  We would recommend a robust stakeholder engagement and management plan, highlighting key actions and corresponding responsibilities, including hand-over between staff, and in the event that a team member ceases to work with the team.

During the initial baseline collection, we found that use of alcohol related diagnostic codes (ICD10) was inconsistent. This meant that many cases who had alcohol dependence treated, had the wrong alcohol-related diagnostic code, or none at all.

Statistical evidence was not good, but significant support from our colleagues and subsequent improvements in care. This project brings learning that could contribute to more comprehensive research, but also shows a process that can be used to improve practice within a busy acute hospital