Feedback about prescribing began in Australia in 1993 with a program run by the Department of Veterans' Affairs. Doctors are contacted about specific issues relating to the treatment of individual veterans. Prescribing profiles using aggregate data have been made available to practitioners by the Health Insurance Commission (HIC) since 1994. Both styles of feedback have their advantages and disadvantages. They each provide doctors with useful information when considering their prescribing. The HIC prescribing feedback is currently being updated and will include, in selected cases such as potential drug dependency, visits to doctors and patients.
Index words: prescribing, evaluation, general practice.

History of national prescriber feedback in Australia
The first feedback program that provided doctors with prescribing information was started in 1993 by the Department of Veterans' Affairs (DVA). This program identifies individual patients who are being prescribed particular, potentially hazardous drugs or drug combinations and then contacts the prescribing doctor. The program also looks for prescribing patterns that indicate poor patient compliance and other problems.

In 1994, the Health Insurance Commission (HIC) started feeding back prescribing profiles for general practitioners. This compared their prescribing of certain groups of drugs to that of their peers. It was designed as a trial to see if sending this type of information resulted in any change to prescribing.

An additional HIC feedback program took place in 1996. Doctors who were relatively high prescribers of a particular antibiotic were contacted by letter and reminded of the potential adverse effects of this drug in certain patients. The feedback involved overall prescribing rates, but did not identify individual patients. This approach clearly alerted the particular practitioners to an area that the HIC considered a problem.

The HIC has recently commenced an additional program, the Doctor Shopping Program. This program is aimed at minimising the poor health outcomes, and resulting cost, associated with patients who attend large numbers of doctors. The program involves HIC personnel, mainly pharmacists, visiting doctors and patients. As part of the program, a `principal' doctor is chosen by the patient and, with the agreement of the patient, all prescribing information relating to that patient is provided to that doctor by the HIC.

What works?
In the current Medicare feedback program, practitioners are sent details of services rendered and ordered by them, individual patients are not identified and no particular course of action is recommended. The intention is to enable practitioners to compare themselves to their peers. This feedback resulted in a significant reduction in the ordering of pathology services which was maintained for several years, but its effect has diminished lately. The initial impact of the program may have been related to some extent to the perceived policing role of the HIC. With the recent acceptance that the HIC has an educational role, the feedback program needs to be adapted.

The approach of the DVA generally involves messages about individual patients. Evaluations show that when a specific patient and a specific prescribing problem are identified, and there is advice that is clear and easy to implement, such as ceasing to prescribe a particular drug combination, the result is a reduction of approximately 20% in the target activity.

The HIC prescribing feedback trial has now been completed. Early indications are that the sending of untargeted prescribing feedback that enables peer comparison, but without clearly identified concerns, or suggested actions, is relatively ineffective.

In contrast, the 1996 HIC antibiotic feedback clearly identified a concern and had a significant impact on prescribing. An additional aim of this feedback was to discourage the use of broad-spectrum antibiotics as first-line therapy. However, the feedback may have had the unintended consequence of shifting prescribing to other broad-spectrum antibiotics instead of narrow-spectrum drugs. The lesson here is that this type of feedback can be very effective, but the action that is being promoted must be very clear.

The Doctor Shopping Program, like the DVA's strategy, involves an individual patient focus. It has not yet been evaluated; however, a trial in 1995 showed that a similar intervention reduced drug use by 18%.

The conclusion that could be reached from the programs of the HIC and the DVA is that well-established education principles have to be adopted if change is to be effected. There is probably a role both for prescribing profiles and for feedback about individual patients. In both approaches, the messages must be well thought out and the recommended course of action must be clear and easy to implement.

The future for prescribing feedback
The current HIC feedback program is being updated to benefit from the lessons learned. The resulting program will soon be operational and is a multilevel program, involving the sending of general information to all doctors and specific information to selected doctors.

As more practices become computerised, it may also be possible to request and receive tailored feedback electronically. This will enable doctors to decide just what information they require and the detail of the information. In addition, important issues such as adverse drug combinations may be highlighted. It is also proposed to provide feedback to Divisions of General Practice about prescribing in their area.