Introduction
There are unexplained large variations in rates and costs of prescribing. International concern exists about the effectiveness and cost-effectiveness of current prescribing, and doctors, particularly general practitioners, are facing 'downwards pressure' on their prescribing.

One factor that may contribute to rising costs is the large number of new drugs launched each year. There is limited information on the market penetration of new drugs¹ and on their rational and safe prescribing. Existing guidance often fails to recognise the complex trade-offs that have to be made by the busy general practitioner during an average 8-minute consultation.² Prescribers are now required to consider not only balancing risks and benefits, but also costs, and how to involve the patient in a 'full therapeutic partnership'.

New drugs
The June/July 1998 version of MIMS (Monthly Index of Medical Specialties) lists 87 new products, 54 new brands, 41 additional presentations and 101 deleted products in the previous 12 months. Many new drugs have implications for general practice and they represent a large increase in therapeutic choice in a relatively short space of time. The cost of new drugs can also be controversial e.g. losartan and donepezil.

Prescription of new drugs in general practice
A review of research into prescribing in primary care³ found that early users of new drugs appeared to undergo a gradual phase of increased awareness, facilitated mainly by pharmaceutical company representatives, and then a stage of legitimisation facilitated by colleagues. Later adopters prescribed more confidently and moved from awareness to legitimisation more quickly. Other factors influencing the adoption of new drugs were hospital specialists and the degree of perceived risk of the drug. A qualitative study confirmed a gradual process of change, but also found that, for individual general practitioners, initial changes were perceived as precarious and needed reinforcement.⁴ This reinforcement usually came from the effects of the drug on a few chosen patients.

Observational studies reveal large variations in the prescribing of newly marketed drugs, which cannot be explained by medical need. In a study of 28 402 general practitioners in England, 85% had prescribed one of 27 newly launched drugs to up to 15 patients.⁵ However, 251 (1%) general practitioners had used these new drugs in over 45 patients each, and 19 general practitioners had prescribed them to over 120 patients each in the immediate post marketing phase. The patients of this latter group of doctors may be at serious risk if regularly changed from new drug to new drug. Another study found rapid increases in the prescribing of certain new drugs, without evidence of increases in specific indications for them.¹ The authors of both these studies^1,5 suggested that marketing strategies were mainly responsible for the large and variable increases in the use of new drugs. Marketing methods include widespread advertising campaigns, representative visits and persuading general practitioners to take part in company-sponsored post marketing studies (which have been shown to be of limited scientific value⁶, but which may be useful in promoting sales). These may be no more than 'seeding studies' and doctors should be wary about becoming involved.⁷

Individual general practitioner prescribing, however, is relatively stable over time.⁵,⁸ One study reported that adoption of new drugs occurred in only one per 1000 face-to-face consultations.⁸ Doctors who were more influenced by commercial sources had higher rates of new prescribing, adopted a wider range of new drugs and were less concerned with issues of economy.

Between 1990 and 1996, the use of antidepressants greatly increased in Australia, mostly due to a huge increase in prescribing of selective serotonin reuptake inhibitors (Fig. 1). Between 1990 and 1996, the prescription of proton pump inhibitors rapidly increased, but the use of H₂ antagonists did not greatly decrease (Fig. 2). The cost-effectiveness of such changes in prescribing remains unclear.

In contrast, aspirin is a highly cost-effective and safe drug recommended for use as a prophylactic agent in patients with arterial disease. However, only 35% of patients with stable angina may be regularly taking aspirin.⁹ There appears to be sub-optimal prescribing in many other areas where there is evidence supporting the cost-effectiveness of treatments. Examples include atrial fibrillation (warfarin) and the treatment of hypercholesterolaemia (statins) as part of secondary prevention of cardiovascular disease.

What information should general practitioners consider when prescribing new drugs to patients?
The quality use of medicines requires prescribing to be judicious, appropriate, safe, effective and economic. A prescriber should aim to maximise effectiveness, minimise risks, minimise costs and respect patients' choices.²

Safety
Although pre-marketing studies are useful in determining efficacy, information on the safety of new drugs is limited to experience in a median of 1500 patients.¹⁰ Widespread use in the 'real world' may uncover hazards that were not detected in carefully selected patients recruited for clinical trials e.g. the recent withdrawal of mibefradil. Spontaneous reporting systems, such as sending a blue card to the Adverse Drug Reactions Advisory Committee, can provide early signals of problems with drugs. However, there is under-reporting of adverse drug reactions, and heavy prescribers of new drugs appear to be the worst offenders.⁵

There is very little formal testing of drug safety in children and pregnant women. The product information may therefore state that a drug is contraindicated in these groups because of lack of data. However, chronic disabling diseases requiring effective treatment occur in children and expectant mothers, and prescribers are placed in a difficult situation.

Effectiveness and cost-effectiveness
When a drug is first licensed, its efficacy (usually measured by surrogate outcomes e.g. blood pressure) is known. There is less information on its effectiveness in real-world clinical practice, and on its impact on clinical outcomes such as morbidity and mortality. Without these data, it is often difficult to justify the increased costs compared to cheaper, established drugs of proven effectiveness. Inefficient prescribing of new, expensive drugs utilises scarce resources that could have produced greater benefit for other patients. Only when expensive drugs are shown to be cost-effective is their use justified by the additional benefits they bring.

One such group, the new angiotensin receptor antagonists, is relatively expensive. Short-term clinical trials suggest that valsartan, for example, is of similar efficacy to thiazide diuretics, calcium channel blockers and ACE inhibitors. However, only diuretics and beta blockers have ever been proven to reduce cardiovascular morbidity and mortality. Diuretics and beta blockers therefore remain a rational first-line choice for the majority of hypertensive patients.

However, there are examples of expensive drugs where evidence of effectiveness exists. These include inhaled anti-inflammatory drugs which improve morbidity in asthma and ACE inhibitors which decrease morbidity and mortality in heart failure and delay the progression of diabetic nephropathy. The benefits in terms of morbidity and mortality are likely to justify their increased costs over existing treatments.

Patient's perspective
For ethical and practical reasons, it is important to take account of patients' views on their treatment. Facilitating informed choice is part of good prescribing, and is likely to improve compliance. However, complying indiscriminately with patient choice is often cited as a characteristic of poor prescribing. An example includes prescribing antibiotics for uncomplicated upper respiratory tract infections in response to perceived patient demand. Controversy exists over the prescribing of antibiotics in sore throats because they have little effect on morbidity, are associated with adverse effects and encourage increased general practitioner visits. Nevertheless, new expensive antibiotics are often heavily marketed to general practitioners for these indications.

Promoting cost-effective prescribing
Australia and Canada have developed guidelines for the economic evaluation of new pharmaceuticals. In the U.K., general practitioners hold either fund holding or indicative prescribing budgets, although these schemes tend to emphasise cost-containment rather than cost-effectiveness. Feedback prescription data are provided to Australian general practitioners to help them assess their prescribing. Evidence about good prescribing needs to be effectively disseminated¹¹; unfortunately, in Australia, cost-effectiveness data are not open to public scrutiny.¹²

Some authors favour a probation period for newly licensed drugs until they have been shown to be at least as effective and safe as existing alternatives.¹³ Randomised controlled trials in primary care would establish cost-effectiveness in the situations in which these drugs are actually used. The true costs and consequences of drug administration and adverse effects would be revealed.

Conclusion
New alternatives to existing treatments should be used only if these bring additional benefits commensurate with their cost. Doctors have a duty to prevent patients being unnecessarily exposed to medicines where the extent of effectiveness and risk in general use are not known.

Acknowledgement

I am grateful to Dr J.S. Dowden and Professor J. Marley for advice on drug utilisation in Australia.