Changing futures: premature discharges of alcohol or opioid detoxification in-patients, service improvement

Aims and method We retrospectively audited premature discharges of 
in-patients undergoing alcohol or opioid detoxification. Recommendations for 
good practice aimed at reducing premature discharge rates were implemented, 
following which a prospective audit was completed. Results The retrospective phase of the audit cycle showed a 
premature discharge rate of 30.8%, compared with a rate of 13.2% for the 
prospective phase. The difference in these rates is both clinically and 
statistically significant at the 5% level, with Fisher’s exact test 
producing a two-sided P = 0.0119. Clinical implications Implementing good practice guidelines improved 
outcomes, with more patients successfully completing detoxifications and a 
more effective use of resources.

In-patient detoxification is usually reserved for those who need increased supervision and support provided by a ward to safely and successfully withdraw from drugs, alcohol or prescribed maintenance treatments. In many psychiatric services, the only available beds for such in-patients are on general psychiatric wards. This is often not ideal as the planned course of treatment may be terminated prematurely, with patients being discharged before achieving the desired outcome of a completed detoxification with a consequent waste of effort and resources. Indeed, the Specialist Clinical Addiction Network (SCAN) report 1 recommended that dedicated National Health Service (NHS) specialist services are the optimal setting for the delivery of in-patient treatment to service users with severe alcohol or drug problems, but as this option is not available in many areas, the challenge of improving the care in existing facilities remains. We chose a target of 80% treatment completion rate which sets a comparable standard to the highly specialised dedicated in-patient units described in the SCAN report, 1 which estimates that 70-85% of admissions to such units achieve their in-patient assisted withdrawal goals.
A premature discharge rate can be calculated to measure the failure to achieve completed detoxification and is therefore also a measure of the treatment success of an in-patient unit. As treatment outcomes are related to the process of care, premature discharge rate can also be used as a proxy measure of the quality of care provided.

Study setting
Our service serves a population of 360 000 and utilises two beds on a 24-bed general psychiatric emergency admission ward. We decided to systematically examine the perceived high rate of premature discharge for in-patients in our service using clinical audit, with the aim of increasing the number of successful detoxifications.
The audit aimed to: . measure the extent of premature discharges . record the reasons stated for premature discharge . produce recommendations for reducing premature discharges and monitor their implementation.
We use a chlordiazepoxide reducing regimen for alcohol detoxification, and lofexidine combined with other remedies such as mebeverine, loperamide, quinine for akathisic symptoms, and simple analgesics and benzodiazepines for symptoms occurring during withdrawal from opioids, with the aim of reducing discomfort as much as possible and thereby minimising cravings and the risk of substance use.
Premature discharge was defined as an admission lasting less than 7 days because this is the time it takes for a patient to either complete a planned admission standard detoxification on our ward or to allow successful completion in the community following discharge. A criterion that patients should stay for 7 days was agreed with a standard that 80% of admissions should meet this Aims and method We retrospectively audited premature discharges of in-patients undergoing alcohol or opioid detoxification. Recommendations for good practice aimed at reducing premature discharge rates were implemented, following which a prospective audit was completed.

Results
The retrospective phase of the audit cycle showed a premature discharge rate of 30.8%, compared with a rate of 13.2% for the prospective phase. The difference in these rates is both clinically and statistically significant at the 5% level, with Fisher's exact test producing a two-sided P = 0.0119.
Clinical implications Implementing good practice guidelines improved outcomes, with more patients successfully completing detoxifications and a more effective use of resources.
Declaration of interest None.

CURRENT PRACTICE
Changing futures: premature discharges of alcohol or opioid detoxification in-patients, service improvement Mark Parry, 1 Nicholas Woodthorpe, 2 Priyanthi Gunawardena 3 target, which allows comparison with the 70-85% estimated completion rates in highly specialised in-patient units. 1 The audit had two phases: a retrospective phase (prerecommendations), and a prospective phase which followed the implementation of recommendations, thus completing the audit cycle.

Retrospective phase
A proposal form for the retrospective phase was sent to the clinical governance coordinator for approval and ethical clearance. A list of in-patients detoxified between September 2003 and January 2006 was obtained from the Medical Information Department at Prospect Park Hospital. The information obtained included the length of the admission and patient demographic data. It was then possible to calculate the premature discharge rate (admissions 57 days compared with those 57 days). Then, all the available notes were collected from the medical records department and analysed by a junior doctor to identify reasons for the premature discharge.

Service improvement recommendations
After the retrospective phase was completed, a meeting was held between the ward and substance misuse service staff and hospital management (focus group) to discuss how the process of care might be improved. The information obtained from the case notes formed a starting point for discussion, but as it was often incomplete, it needed to be supplemented by reflecting on the accumulated experience of past difficulties in the process of care and how these might be remedied. The focus group was fundamental in forming a cohesive approach between and within staff teams and generated a sense of ownership in tackling issues which was necessary for achieving service improvement for those undergoing an in-patient detoxification. The recommendations, with their underlying themes of clear aims, purposes and methods of providing care, addressing problems proactively and improved communication and support for staff and patients, were communicated to all involved in the in-patient detoxification programme, and were as follows.

Prospective phase
The prospective phase ran from March 2006 to October 2007 and ended when we had a sample size almost as large as the retrospective phase. As with the initial retrospective phase, the Medical Information Department were contacted and a list of those who had undergone an in-patient detoxification was obtained. The premature discharge rate was calculated using the same criterion and standard set previously (i.e. admissions 57 days compared with those 57 days), but it also included premature discharges of 7 days or longer for those patients who did not complete their detoxification programme. A junior doctor examined all the patient notes, to check adherence with the recommendations and the recorded reasons for any premature discharges. A form was created to make relevant information collating easier. Once all the audit data had been collated, a report was sent to the clinical governance coordinator and to all those involved in the in-patient detoxification programme.

Demographic data
Retrospective phase Of a total of 81 admissions, 75% of individuals were admitted for alcohol detoxification; 75% were male; 83% were unemployed, 12% were employed and the remainder were retired or studying. The majority of those admitted (99%) were White; 16% were married, 64% single and the remainder divorced, separated or widowed. Two-thirds were aged 30-49 years. The 25 patients prematurely discharged did not differ demographically from those completing treatment; one in three opioid and one in three alcohol patients were discharged prematurely.

Prospective phase
Out of a total of 76 admissions, 72% of individuals were admitted for alcohol detoxification, 25% for opioid detoxification and 3% for both alcohol and opioid detoxification; 66% were male. The majority (72%) were unemployed, 18% were employed and the remainder were retired or information was not available; 99% were White; 17% were married, 57% were single and the rest were divorced, separated or widowed; 64% were aged 30-49 years. The 13 patients prematurely discharged did not differ demographically from those completing treatment: one in four of opioid and one in seven of alcohol patients were discharged prematurely.

Premature discharge in both phases
Length Figure 1 shows the length (in days) of the premature discharges for the retrospective (n = 81) and prospective (n = 76) phases. In the retrospective phase, eight case notes were not available and in the prospective phase two case notes were not available. If the case notes were not available then the discharge summary was used, if completed, to ascertain whether it was a premature discharge. The median length of the premature discharge for the retrospective phase was 5 days and for the prospective phase 4.5 days.

Rate
The rate of premature discharge fell from 30.8% in the retrospective phase to 13.2% in the prospective phase. A comparison of proportions for both phases using Fisher's exact test yielded a two-sided P = 0.0119, a statistically significant difference in the rate of premature discharge between the two phases. Table 1 lists the causes of premature discharges for both phases.

Implementation of the recommendations in the prospective phase
In total, 74 case notes were available for assessing the implementation of the recommendations (case notes were unavailable for two in-patients). Table 2 records the extent to which the recommendations were implemented during the prospective phase. In general, better adherence to the recommendations seems to have occurred for those patients who had a 7-day or longer admission compared with those discharged prematurely. For example, only 30% of those prematurely discharged had a pre-admission visit compared with 59% staying 7 days or longer, and only 30% of those prematurely discharged received occupational therapy compared with 59% staying 7 days or longer. The recommendation for pre-formulated contingency plan for premature discharge was not implemented.

Discussion
Premature discharge rates fell from 30.8 to 13.2%, a reduction especially significant in alcohol patients, from one in three to one in seven. This very large reduction was a surprise, particularly as the recommendations were implemented to variable extent and despite the fact that the prospective phase of the audit started only 10 weeks after the recommendations were drawn up and started being implemented. None of the patients discharged early had formal contingency plans drawn up, possibly because they seemed superfluous with the improved communication and support in place. Some interventions took longer to implement, for example, ward visits by a link worker which required redeployment of staff. Some recommendations were very successfully implemented, for example postdischarge plans and pre-admission assessments by the link worker. It is evident that some recommendations require more coordination, effort and planning, and some are relatively resource-intensive and thus harder to implement.
It proved impossible to measure changes associated with some recommendations, for example improved communication. Also, as Balogh & Bond 2 point out, the action stage of audit is not a single discrete event, but a process occurring over time so that in clinical practice the components of the process of change frequently occur at differing speeds and with varying success.
We did not quantify the relative importance of the recommendations but the staff felt that most important in preventing premature discharge was pre-admission visits to the ward, link worker visits during admission and occupational therapy. We suspect there was also a powerful non-specific benefit from increased patient-staff contact.
Our improved results compare well with the SCAN report 1 on in-patient treatment, which estimated that 70-85% of in-patients achieve withdrawal; Martinez-Raga et al 3 quoted a rate of 32.8% for unplanned discharges for alcohol detoxification, whereas Scherbaum et al 4 and Gossling et al 5 reported rates of 50% for in-patient opioid detoxification.
Detoxification and subsequent abstinence are different processes 6 but they form a continuum 7,8 and thus a better process of detoxification may help reduce subsequent relapse rates. A systems approach is needed to reflect the complexity of the needs as well as circumstances of individual patients and their treatment; 9 such an approach would lead to better use of resources. 10 Factors in premature discharge are discussed by Day 7 and Martinez-Raga et al. 3 Identifying those at risk of leaving early may allow preventive measures. Some detoxification medication regimes are better than others, although Scherbaum et al 4 found that the severity of craving and withdrawal symptoms in heroin addicts were not the primary reason for the premature termination of detoxification, and Reoux & Miller 11 in comparing routine regimes with symptom-triggered management for alcohol patients did not find much to distinguish them. Rü esch & Hättenschwiler 8 found that a lapse during detoxification may be overcome if in-patients are allowed to stay but that no patient discharged subsequently achieved abstinence.
The research findings discussed above challenge our existing ideas and clearly we have much to learn about improving the processes of care. We conclude that using existing knowledge will improve outcomes in the meantime, and therapeutic pessimism is unjustified. Pre-admission assessment by ward named nurse 10 0 1 Visit by substance misuse link worker during admission 60 73 72 Occupational therapy during admission 30 59 55 Post-discharge plan 100 100 100 Contingency plan for premature discharge 0 0 0 a. Percentages rounded to nearest whole number.