An indicator of appropriate neuroleptic prescribing in nursing homes.
Oborne CA, Hooper R, Chi Li K, Swift CG and Jackson SHD;
Commented by , 16 Dec 2002
Aim of the Study
To derive and apply objective criteria to assess appropriateness of neuroleptic prescribing in nursing homes.
Method
A random sample of all nursing homes with >35 residents in the South Thames Region of England was selected. The study was essentially an audit of neuroleptic prescribing practice between July 1997 and June 1998.
The standard set for appropriate neuroleptic prescribing was the US OBRA guidelines (the Omnibus Reconciliation Act 1990) – see Figure 1 below. Data was collected from GP and nurse records.
Figure 1 – “Appropriate” Neuroleptic Prescribing – Based on OBRA Guidelines
| i) |
Psychotic disorder. |
| ii) |
Organic mental syndrome with behaviour presenting danger to the patient, others, or interfering with provision of care. |
| iii) |
Hiccough, nausea or vomiting (short-term use only). |
Inappropriate use included unspecified aggression, wandering, restlessness, agitation (that was not a danger), anxiety and unco-operativeness.
Objective and quantitative case note documentation was required to assess whether neuroleptic prescribing was appropriate. Data on attempted dose reduction was also collected.
The data was then analysed using logistic regression to see if there was any link between inappropriate prescribing and nursing home type (private, public) and size; or general practice type (teaching, single-handed).
Results
22 homes were studied, sized 40-120 beds (median 46.5 beds). From the original sample, 7 homes or their attending general practitioners refused to participate.
There were 934 residents aged 65-105, median 86 years, 74% female. Residents were prescribed a mean of 5.1 items. 229 (24.5%) residents were prescribed a total of 245 neuroleptics. Thioridazine comprised 51% of neuroleptic prescriptions, haloperidol 17% and chlorpromazine 7%. Atypicals only constituted 3.7% of prescriptions.
Clinical data on residents was collected for 234/245 prescriptions. 45.3% of prescriptions were initiated before nursing home admission.
The main finding was that an appropriate indication was documented for only 89/234 prescriptions (38%). Only 42/89 prescriptions had been reviewed in the previous 6 months. Levels of appropriate prescribing were similar in all homes.
The logistic regression suggested that none of the characteristics of homes or their attending general practices were significantly related to appropriate or inappropriate prescribing.
Discussion
These results broadly concur with the often quoted McGrath and Jackson Glasgow study of non-randomly selected nursing homes (BMJ 1996; 312 : 611-2). In that study, 24% of residents received regular neuroleptics, of which 88% of prescriptions were deemed inappropriate, based on OBRA guidelines.
But are the OBRA guidelines an appropriate yardstick to use? In an accompanying editorial, Jonathan Waite points out that when the guidelines were introduced, neuroleptic prescribing in American nursing homes was much higher.
It did reduce prescribing, but only to levels now seen in the UK and other Western countries. Arguably, the most beneficial effect of OBRA was to encourage physicians to recognise and treat depression.
What this audit shows, again, is that documentation of behavioural and psychological symptoms in dementia (BPSD) in nursing and general practitioner notes is not very specific. Clinical practice will have inevitably changed since 1997, with the withdrawal of thioridazine and an increasing evidence base for efficacy of atypical antipsychotics in BPSD.
For a detailed review, readers are referred to Lawlor, B. "Managing Behavioural and Psychological Symptoms in Dementia", British Journal of Psychiatry 2002; 181 (6); 463-465.