M Murray, P L Padfield, S R J Maxwell*
Department of Medical Sciences, University of Edinburgh and *Clinical Pharmacology Unit, University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh
Correspondence to: Dr S Maxwell, Clinical Pharmacology Unit, Western General Hospital, Edinburgh EH4 2XU
E mail: s.maxwell@ed.ac.uk
SMJ 2005 50(2): 76-79
Abstract
Aim: To assess the extent to which prescribing of cardiovascular medications in a busy medical unit deviates from the local joint primary and secondary care drug formulary guidelines. Method: A retrospective audit of the case notes, prescription charts and discharge summaries of 150 randomly selected emergency medical admissions over a 4 month period. Results: No patient receiving a non-formulary cardiovascular drug on admission had the choice reviewed in line with formulary recommendations. One third of new cardiovascular medications commenced in hospital were not compliant with formulary recommendations. Decisions about drug therapy were rarely justified in the written hospital record. Conclusions: Our results demonstrate that in a busy acute medical admissions’ unit there is a clear failure to amend or query non-formulary prescribing at the time of admission and a tendency to exacerbate it during the inpatient period. This potentially undermines the purpose of a joint drug formulary as a guideline for safe, evidence-based and cost-effective prescribing.
Key words: Formulary, prescribing, cardiovascular, drugs
Introduction
The pressure on junior prescribers in the NHS is growing as a result of the increase in available medicines, the number of medicines being taken by patients, and the need to contain the cost of medicines.1,2 For this reason most hospitals have a formulary or approved drug list that offers guidance on safe and cost-effective prescribing choices. The Lothian Joint Formulary (LJF), developed jointly in 2001 by hospital doctors, general practitioners and pharmacists on behalf of the Lothian Area Drug and Therapeutics Committee, provides guidance based on clinical efficacy and expense on appropriate pharmacological therapy for a typical majority of patients.3 It is subject to continuous review. The aim of this study was to investigate the extent to which prescribing deviated from a recently introduced formulary in a busy medical service, reasons for deviation from these choices and the extent to which prescribers amended medication choices in line with formulary recommendations.
In 2002-3, four of the top ten most commonly dispensed drugs in the community by both cost and volume were for management of cardiovascular disease.4 The Scottish Health Service Analysis of Inpatients in Acute Specialties in Lothian for the year ending March 2003 presented the direct pharmaceutical cost per case per inpatient week: general medicine was the third most expensive specialty, closely followed by cardiology.5 Therefore, based on both frequency and cost, we decided to focus our investigation of compliance with formulary recommendations on cardiovascular prescriptions.
Methods
The study involved a retrospective audit of 150 randomly selected patients admitted as an emergency to the Acute Receiving and Admissions Unit (ARAU) of the Western General Hospital, a 550 bed teaching hospital based in North Edinburgh, between May and August 2003. The ARAU receives all adult patients who either self-present or are referred by their GP with acute medical problems. Initial assessment, investigations and management are undertaken by senior house officers and consultant ward rounds occur twice daily.
All patients included in the audit were formally admitted to and discharged from the acute medical ward. Each individual had a completed admission unitary record, including a full admission drugs list documented as having been verified with either the patients’ own drugs, GP practice or dispensing pharmacy by either a junior house officer or the ward pharmacist. A completed admission drug prescription chart and immediate discharge summary and prescription were also necessary for inclusion in the study.
Cases were excluded from the audit if the patient was either on no regular medications on admission or subsequently followed separate Trust protocols for management of specific conditions.
For each subject, the admission medications were scrutinised and non-formulary prescriptions identified. The initial hospital drug prescription sheet was examined for non-formulary medications that had been prescribed and supplied. Drugs initiated in hospital were identified from discharge script. All patient records relating to the admission were examined for evidence of reasoning and motivation in cases where non-formulary medications had been commenced. A single observer (MM) reviewed all documentation, including the combined multi-disciplinary case notes, drug prescription charts and immediate discharge summary scripts. The following analysis relates only to cardiovascular drugs.
Results
A total of 1,857 patients were admitted to the acute medical ward during a twelve week period from May-August 2003. From this number 150 cases were randomly selected to enter the study. A total of 35 were subsequently excluded as they were on no regular medication prior to admission and their management subsequently followed separate trust protocols for acute ethanol withdrawal (23) or treatment of infection (12).
A total of 115 cases (62 females, 53 males) with mean age 60.5 years (range 32yrs-92yrs) were included in the final analysis. The mean duration of hospital stay was 1.1 nights (range 1-3). At the time of admission, the patients were receiving a total of 368 drugs on admission (mean 3.2 drugs per patient), of which a subset of 63 patients were receiving a total of 172 cardiovascular drugs (mean 2.7 per patient). The patients’ cardiovascular drugs were considered on the basis of the classification used by the LJF. (Fig.1)
Twenty-three patients on cardiovascular medication (36.5%) were receiving at least one non-formulary cardiovascular drug at time of their admission to ARAU. The total number of non-formulary cardiovascular drugs was 27 (15.7%). (Table I) In all of these cases, the drug was written onto the hospital prescription sheet on admission and subsequently prescribed amongst the discharge medicines. No documentation was found in any of the notes to indicate that the prescription choice had been reviewed.
A total of 51 cardiovascular drugs were initiated during admission. 17 (33.3%) of these were non-formulary. (Table II) There was no documentation found in any of these cases to substantiate therapeutic selection or reason for deviation from the formulary.
Discussion
This study highlights the complexities of adhering to a joint formulary. It demonstrates that non-formulary drugs commenced in the community are invariably continued into secondary care without any apparent review. Furthermore, hospital admission, however brief, seems to serve as a further source of non-formulary prescribing.
Hospital formularies have evolved with the rapid increase in available medicines, a strengthening evidence base to support rational prescribing and the ubiquitous emergence of cost pressures in modern healthcare.6,7,8 Joint primary and secondary care formularies encourage safe, consistent and efficient prescribing throughout the health care team.9,10,11 However, concerns remain that drug selection is determined purely by cost.12 It is therefore essential that rigorous standards be met for drug inclusion: proven efficacy and a favourable risk:benefit ratio.13,14
The failure of the Lothian formulary to have an impact on prescribing practice in this audit may be attributed to several factors. First, the setting was in a busy receiving unit where patient turnover is high. In this circumstance, it is often easier and arguably safer to continue a stable prescription15 rather than to review and amend nonformulary choices. Second, the majority of prescriptions are written by junior doctors who may be less likely to question prior choices made by more senior clinicians. Third, many of the prescribers involved may remain ignorant about the existence and objectives of the local formulary while a minority may be hostile to a process that may be perceived as imposing restrictions on clinical freedom.
In this regard, the main areas of non-compliance were all readily identifiable as sources of controversy where individual clinicians might have personal and justifiable views.
Our audit was limited by both the small population sample and its restriction to one area of the hospital. The cohort of patients with such brief stays in hospital is selective and may not be reflective of patients admitted for longer periods of time or those attending outpatient clinics.16 We have no way of knowing whether the non-formulary prescribing, either at the point of admission or during admission, was justified on the basis of clinical indications. However, given the documented rates of non-class-related adverse effects to the main non-formulary drugs involved it seems unlikely that this could be an explanation.
A final important finding in this study was the virtually complete absence of comment in the records of a major teaching hospital that would serve to explain or justify therapeutic practice. This meant that, even after a thorough review of individual case notes, drug charts and discharge summaries it was usually impossible to analyse the decision making process that underpinned prescribing choices. Although many clinical decisions are so routine that they do not deserve comment, it might have been expected that some explanation would accompany an active decision to deviate from local guidelines. This finding should be of concern at a time when legal cases surrounding prescribing of medicines are increasing and good record keeping is a key component of the NHS risk management strategy.
In conclusion, this retrospective survey of prescribing practice in an acute medical unit demonstrates that nonformulary prescribing is neither amended nor questioned at the point of admission. Furthermore, new prescriptions commonly deviate from the agreed formulary with no documentation of why these decisions were made. Based on these findings, and previous studies,12,16,17 we would suggest the following recommendations for successful implementation of a hospital formulary
• The formulary should be readily accessible at the point of use. We have now made the formulary accessible on all unit computers and via the web (http:// www.ljf.scot.nhs.uk). It has also been made available in a format for palm held computers (PDAs), commonly carried by junior staff. These approaches are more successful than written documents which tend to be lost.
• There should be an accompanying formulary education package. We have appointed a part-time Formulary Implementation Pharmacist with the aim of regularly highlighting the importance of the formulary adherence to NHS Lothian amongst our prescribers. The LJF is also highlighted during the education of all local medical students and also at the induction and subsequent education of all junior staff.
• There should be regular feedback from ward-based clinical pharmacists. The acute medical unit now benefits from the continuous review of all prescription charts by a clinical pharmacist who can highlight non-formulary prescribing. This intervention should prompt discussion about the reasons for the prescribing choice and encourage appropriate documentation is the case notes.
• A regular bulletin highlighting formulary matters should be sent to prescribers. The Lothian Prescribing Bulletin is now being sent to all local prescribers on a 2 monthly basis to highlight important formulary sections, amendments and the availability of new medicines.
• Peer-comparison in relation to formulary adherence. This process serves to identify unexpected prescribing patterns (individual clinicians or clinical groups). We have established a drug utilisation review group that regularly monitors formulary adherence.
• There should be regular audit of medical case records. This should be part of the normal governance process of all units but should review specifically records relating to therapeutic decisions.
• The formulary must receive the support of clinical management. It is axiomatic that without unambiguous support of those charged with leading clinical governance and risk management none of the above or the formulary itself can be implemented.
REFERENCES
1 Mucklow JC. How can we afford costly medicines? Journal of the Royal College of Physicians of London 2000;34:124-7
2 Griffin JP. An historical survey of UK government measures to control the NHS medicines expenditure from 1948 to 1996. Pharmacoeconomics 1996;10:210-24
3 The Lothian Joint Formulary - http://www.ljf.scot.nhs.uk
4 Update of NHS Scotland Prescribing Information on Scottish Health on the Web (SHOW)- information up to March 2003 http://www.isdscotland.org/isd/ info3.jsp?pContentID=2229&p_applic=CCC&p_service=Content.show& Murray, Padfield, Maxwell Hospital formulary and prescribing practice
5 Scottish Health Service Costs (Costs Book): Specialty Group Costs - Inpatients in Acute Specialties Report 040X http://www.isdscotland.org/isd/ files/Report040x_2003.xls
6 Parish PA. Drug prescribing - the concern of all. Journal of the Royal Society of Health 1973;4:213-7
7 Barber N. What constitutes good prescribing? Br Med J 1995;310:923-925
8 Steger JW, Wolverton SE, Greenberg RD, Ling MR, Resnick SD, James WD. Dermatologic drug formulary: An American Academy of Dermatology white paper. J Amer Acad Dermatol 1996;34:99-109.
9 Giaquinta D. Drug Formularies - Good or Evil? A View from a Managed Care Provider. Cardiology 1994;85(suppl 1):30-35
10 Duerden M, Walley T. Prescribing at the interface between primary and secondary care in the UK. Towards joint formularies? Pharmacoeconomics 1999;15:435-43
11 De Vries CS, van Dieper NM, Tromp TFJ, de Jong-van den Berg ITW. Auditing GPs’ prescribing habits: cardiovascular prescribing frequently continues medication initiated by specialists. Eur J Clin Pharmacol 1996;50:349- 352.
12 Woodhouse KW. Drug Formularies- Good or Evil? The Clinical Perspective. Cardiology 1994;85(suppl 1):36-40
13 Martin DK, Hollenberg D, MacRae S, Madden S, Singer P. Priority setting in a hospital drug formulary: a qualitative case study and evaluation. Health Policy 2003;66:295-303
14 Sanchez LA. Pharmacoecomics and formulary decision making. Pharmacoeconomics 1996;9 (Suppl 1):16-25
15 Fijn R, Lenderink AW, Egberts ACG, Brouwers JRBJ, De Jong-Van DenBerg LTW. Assessment of indicators for hospital drug formulary non-adherence. Eur J Clin Pharmacol 2001;57:677-684
16 Franklin GA. The driving force in hospital formularies: economic versus efficacy. Am J Surg 2003;186 5A :55S-60S
17 Feely J, Chan R, Cocoman Lelia, Mulpeter K, O’Connor P. Hospital formularies: need for continuous intervention. Br Med J 1990;300:28-30.