Ken Harvey*

In the late 1970s, concerned about antibiotic-resistant germs, a group of microbiologists, clinical pharmacologists and others, asked the question: "Are antibiotics being used appropriately or inappropriately?" That was not an easy question to answer. If I walked up to my clinical colleagues and said: "I think you're prescribing wrongly", they turned around and said: "Who are you to tell me what to do? You are just a clinical microbiologist, I have got responsibility for this patient and it is my decision!" In short, we had a problem in agreeing on what was appropriate antibiotic therapy.

The way we solved the problem was to collect together a group of so-called experts from various disciplines and diverse teaching hospitals, lock them up in a room, and not let them out until they had come up with an agreement as to what constituted cost-effective treatment for a limited range of common conditions. The result was the first edition of Antibiotic Guidelines.

The booklet had two virtues: first, it became a teaching document for medical students and interns, and it rapidly proved its popularity there; but more importantly it became an audit standard against which we could compare what went on in the wards of our hospitals. Somewhat to our horror, we found that less than 50% of antibiotic treatment in the wards of Melbourne teaching hospitals complied with the recommendations of Antibiotic Guidelines.

Well, we scratched our heads and thought about this. Perhaps it just needed a few more editions, more time to teach it into the system, then surely things would get better. But what intrigued us, as we did repeated drug audits over the years, is that although some things got better, some things in fact got even worse.

When we thought it appropriate we advocated the use of older, cost effective, narrow-spectrum antibiotics, but increasingly our colleagues used newer, more expensive, heavily promoted drugs, and that intrigued us. So, in the early 1980s, a few of us decided that something more than Antibiotic Guidelines was required.

Our aim was to take a leaf out of the books of the pharmaceutical industry. They got their messages across with verve, vigour and nice imagery. We wanted to do the same. We decided to tackle the inappropriate use of what was then an expensive broad spectrum drug, amoxycillin, in particular, its intravenous use for the treatment of community-acquired pneumonia.

So we devised a marketing plan and a media campaign; a series of stimuli over time to see if we could change the behaviour of the doctors in our institution. We produced pads and pens containing our message: "This will cure one disease wonder drugs can't", referring to the disease of inappropriate prescribing, and the unnecessary use of broad spectrum antibiotics. Display stands positioned Antibiotic Guidelines where the doctors congregated, in the doctors' dining rooms next to the telephones, and proved a most effective way of distributing the good book. A poster was devised which asked our doctors to improve the treatment of community-acquired pneumonia, by using benzylpenicillin rather than intravenous amoxycillin.

We put this poster in the usual places you put posters, but also in less usual spots, such as in toilets and above urinals! Another poster showed a brown paper wrapped Antibiotic Guidelines booklet with a slogan: "The juicy bits are on page 14". This poster was quite effective in making people look at the book to see what on earth was on page 14.

We also got some free publicity from the Medical Journal of Australia which put another of our posters on their front cover. The picture showed a bruised, battered and well used Antibiotic Guidelines booklet. The slogan said "It is already working wonders". The advertising agency said that this was "positive reinforcement" and so in fact it proved to be.

The impact of this campaign was monitored by drug audits before and after the intervention. Prior to the campaign intravenous amoxycillin was prescribed for more than 50% of cases of community acquired pneumonia. Following the campaign the use of amoxycillin fell to 5% while the preferred treatment, benzylpenicillin, rose to 85%. This change in prescribing behaviour slowly faded away 12 months later especially when we got an influx of new residents from New South Wales and other places!

But, the results made the point that one needed to do more than simply produce a book in order to change prescribing habits. In the mid 1980s and with the fourth edition of the booklet, we consolidated this approach. The authority of the booklet had now spread to involve the Health Commission of South Australia, as well as Victoria. We moved the educational programme to the Victorian Medical Postgraduate Foundation (VMPF). We recruited more people to join the organizing committee. We convinced the Victorian Health Department to give us some start-up funds. Then we initiated a two-pronged campaign.

The first aim was to promote the distribution and sales of Antibiotic Guidelines booklets. We took out expensive double page advertisements in the Medical Journal of Australia telling the readers that our book had now "Cured the disease wonder drugs can't!" The advertisement had coupons that you could clip out and return with $3.50 to buy the book. So we are into privatisation and self-sufficiency. Like the pharmaceutical industry, we used a range of promotional techniques: direct mail, brochures and posters making the same modest claims for our product as the pharmaceutical industry makes for its. These were prominently displayed around Victorian hospitals.

We tracked the success of this distribution campaign by sales volume. In Victoria and South Australia, the two states that had embraced and owned the concept, we sold more booklets than there were doctors.

But New South Wales, Queensland and Tasmania did not want much to do with a Victorian/South Australian booklet, and over in the west they wanted nothing to do with it whatsoever. This was an interesting reflection on the nature of Australian society and showed how important it is to have local involvement in ownership of such a project. The second thing we did with the fourth edition was to run another education campaign. Since we had concentrated on physicians and the treatment of pneumonia for our first campaign, we thought we would aim at surgeons for this campaign. In particular, the use of prophylactic antibiotics in surgery.

We divided Victoria up into matched pairs of 12 public hospitals ranging from the inner city to the outer country. The aim was to do drug audits in all of them, to give a controlled dose of education intervention in half, and then to reaudit the use of antibiotics. Drug audits are hard work. But they are undoubtedly the beginning of wisdom. Without collecting data about what is going on in our institutions, one has no idea about the standards of prescribing. Our initial data showed us that the standard of surgical prophylaxis was not good.

We specifically compared the timing and duration of antibiotic prophylaxis with recommendations in Antibiotic Guidelines. In inner city teaching hospitals it was about 60 to 70% appropriate, but the further we got out towards the periphery the worse the results became. In some far-flung centres, less than 10% of surgical prophylaxis was appropriate. Whether this was a reflection of the hardy individualism of people who practise in those locations, or whether it was a reflection that educators rarely got that far out from the centre, remains uncertain. However, the good news for these country centres was that they had more room for improvement than anyone else!

Our education campaign had two messages: first, antibiotic prophylaxis should be timed so that high levels of the drug are achieved at the time of surgery and second, that surgical prophylaxis should not be like an operatic aria which goes on and on forever, but one or two shots was all that was required. These messages were encapsulated in posters for display in the intervention hospitals. We also used the traditional visiting lecturer going at great expense to each centre. The problem with this mode of education is that often only three people out of 30 turn up, and those are the same three people that you saw last year and the year before!

We were anxious to reach the busy doctors who did not have the time to attend postgraduate education. Once again we decided to adopt the promotional approach of the pharmaceutical industry. We recruited an academic detailer, a pharmacist chosen not just for her intellectual skills, but for her ability to knock on the door of busy surgeons at a time convenient to them and talk to them about our messages, as distinct from drug company messages. We also made a short video to reinforce our messages.

The results were very encouraging. One outcome measure was the cost of prophylaxis. In the control hospitals, in the first year, costs rose, as always, but in the intervention hospitals costs fell. Left to themselves in the second year, the intervention hospitals started to go back to their old habits, but when we crossed over and gave a dose of intervention in the second year to the control hospitals, the costs in the control hospitals also fell. These differences were statistically significant.

Now around this time there were a number of groups producing Antibiotic Guidelines. There was a recognition that it was a lot of hard work for each group to keep their books up-to-date and there were advantages and economies of scale in coming together and producing a common document. By the fifth edition, this process was complete and groups in most States and Territories were assisting in up-dating the common project.

For the fifth edition we moved for the first time to address inappropriate antibiotic use in general practice. The target we chose was the treatment of tonsillitis. If an antibiotic was required, Antibiotic Guidelines recommended penicillin or erythromycin. However, there were good data suggesting that many doctors were prescribing more expensive broad spectrum drugs, such as amoxycillin, which were more likely to produce side effects and were no more efficacious.

We devised a brochure with the following messages: "For tonsillitis it is best to be narrow minded"; "phenoxymethylpenicillin works"; "ecologically sound"; "less side effects and less to pay". Once again, our messages were introduced by an academic detailer and direct mail in a controlled trial design. The result was a significant rise in the proportion of treatments that coincided with Antibiotic Guidelines recommendations, from 60% compliance to 85% compliance in our intervention group.

That brings us to the present. New editions have continued, as have new versions of guidelines covering other therapeutic groups. The publicity has also continued, as have the drug audits and the feedback.

Feedback is an important aspect. Let me give you a relevant example. In earlier editions of Antibiotic Guidelines, we recommended phenoxymethylpenicillin four times a day for sore throats. When we asked GPs why they didn't use phenoxymethylpenicillin, they said, "It's not practical. It is all very well in hospitals where you have nurses shoving things down patients' throats four times a day, but out here we give co-trimoxazole, because it is given twice a day or amoxycillin because it's given three times a day".

Now, that made us go back to the books and realise that, in fact, clinical trials showed you could give phenoxymethylpenicillin twice a day. That was important in terms of patient compliance in the real world. Subsequently the recommendations in Antibiotic Guidelines were changed accordingly. It was a nice example of the synergy and interaction between drug audits, doctors in the real world, and academics sitting in their ivory towers, and we need to continue it.

What conclusions can be drawn from this 17-year history of therapeutic guidelines? I believe we have clearly shown that therapeutic guidelines have proven their usefulness, because there have been seven editions of Antibiotic Guidelines. We have sold more books than there are medical practitioners, and companion volumes such as Psychotropic and Cardiovascular Drug Guidelines have also proliferated and sold well. But, what we have also shown is that a written distillation of knowledge by so-called experts is not enough. The users must also be involved through representation amongst the authorship, through drug utilization studies, through reflection on the discrepancy between drug utilization studies and guideline recommendations, and by feedback to authors.

In this context, there is still significant room for improvement. In Australia therapeutic guidelines were born in hospital practice, directed primarily to hospital doctors and interns. There is still not sufficient involvement of general practitioners and the Royal Australian College of General Practitioners. There has been slow recognition that GPs prescribe 80% of drugs, and hospital specialists only 20%, and that these GPs have their own expertise and needs. They are gradually being brought into the process. I think this trend can and should be accelerated. My own view is that if we were looking towards national guidelines, then much more extensive involvement of general practitioners is crucial. One way this could happen would be to farm out the hospital version to a group of general practitioners and ask them to distil the essence relevant to general practice. It is possible that this process could reduce six or eight hospital volumes to one GP version.

Following the 1992 National Formulary Workshop the Australian Government accepted that the regular production, promotion and evaluation of a comprehensive set of Therapeutic Guidelines was an essential component of the Australian National Drug Policy. The Victorian Medical Postgraduate Foundation Therapeutic Committee successfully tendered for this task and have now expanded their operations with financial support and the imprimatur of the Commonwealth Department of Health. The authorship of the various booklets has progressively widened to represent general practice better and to involve all the States and Territories of Australia. A network of general practitioners has been recruited to provide feedback on the content and user-friendliness of the booklets. The content of each booklet is updated every two years. Several groups in neighbouring countries have found the booklets helpful when developing their own therapeutic guidelines. Others who are interested in this approach are invited to contact the Victorian Medical Postgraduate Foundation Therapeutic Committee at the address below.

What has been achieved in Australia has also been of great help to our colleagues in neighbouring countries, where the resources in clinical pharmacology and microbiology are far less than Australia, but the needs are greater and the problems horrendous. It is not widely known that these booklets have been an exportable commodity and have proved of great value in the region.

The current guidelines' process is not perfect, and could benefit from a full time secretariat to review the literature, to answer queries and ease the task of the panel of authors. The literature sometimes is not sufficiently thoroughly reviewed. The pharmaceutical industry has complained that the committee has not got the time to answer some of their questions and queries. There is no doubt that the process could be better improved and streamlined. That is our challenge for the future.

Further reading

Harvey K, Stewart R, Hemming M. et al. Educational antibiotic advertising. Med J Aust 1986; 145: 28-32.

Landgren F, Harvey K, Mashford M. et al. Changing antibiotic prescribing by educational marketing. Med J Aust 1988; 149: 595-599.

Hemming M. The non traditional use of commercial advertising in prescribing education. Hosp Pharm 1991; 26: 693-696.

Primrose J. Australian National Formulary Workshop. Aust Prescr 1993; 16; 1-46 Suppl. 3.

De Santis G, Harvey K, Howard D, et al. Assuring the quality of medicinal drug use: antibiotic guidelines, audit and educational intervention in Victorian general practice. Med J Aust 1994, accepted for publication.

* Dr Ken Harvey, on behalf of the Victorian Medical Postgraduate Foundation Therapeutic Committee, 3/55 Flemington Rd., North Melbourne, Vic, 3051, Australia. Fax: +613 326 5632.

Two new titles...

Gastrointestinal and respiratory disorders are challenges faced every day by many doctors. Gastrointestinal Drug Outlines and Respiratory Drug Outlines provide valuable clinical information and set out the latest recommendations for drug treatment.

... and two new editions...

For more than 16 years Antibiotic Guidelines has provided an up-to-date and objective second opinion regarding the treatment of infections. The newly released 8th edition incorporates the developing knowledge and experience as regards drug efficacy, potential bacterial resistance, safety and cost. The new edition helps place drugs in perspective and contains enlarged sections on fungal and viral disease.

The second edition of Cardiovascular Guidelines came out in January 1995. It reflects the rapid evolutionary change in the treatment of cardiovascular disease in the four years since the book was first published.

Available from: Victorian Medical Postgraduate Foundation Therapeutic Committee, 3/55 Flemington Rd., North Melbourne, Vic, 3051, Australia. Price: Aus$20 each or Aus$90 for the set of six Guidelines.