Dr Tim Harlow describes how practice guidelines on upper gastrointestinal acid-related disease improved patient care and cut presecribing costs


As a practice we had been aware of our historically high prescribing costs for some time, and our PACT data showed that gastrointestinal tract (GIT) drugs were a major area of cost.

GIT drugs accounted for 17% of our total drugs budget, with histamine H2 antagonists and proton pump inhibitors (PPIs) making up the bulk of this group. 0ur total prescribing costs were 14% above the health authority average and 9% above the national figures.


Several factors combined to trigger change. For years we had procrastinated and hoped things would improve on their own, but it was clear that we had to grasp this nettle. Even as non-fundholders we had seen locality budgets in some form ahead. There was also a feeling among the partners that we were ready to look hard at this.

Our experience with, for instance, the sore throat protocol1 showed that we could deliver better care and still look at and modify our prescribing.

Some useful local guidelines had been produced for dyspepsia and gastro-oesophageal reflux disease (GORD), and we felt that we could fine-tune these.

Ready acceptance by the partners that individual doctors' prescribing and PACT data could be looked at and shared in a non-threatening way was vital. For a partnership of eight doctors, this step was fundamental to change.

When one's decisions are lost in a large amorphous budget it is easy to blame problems on everyone else. The realization that our individual practices would be subjected to friendly scrutiny was a great motivator. We had to own the changes individually as well as collectively. This demanded a high degree of trust.


The local prescribing advisers were involved from the start. They gave detailed information and sustained support in a helpful manner, which was a key factor in developing our guidelines.

We dispense to one-third of our patients and there was clear acceptance of the need not to undermine our dispensing income, which was crucial to the survival of our branch surgeries.

These issues could not be dealt with in the hurly-burly of a practice meeting. We arranged a prescribing half day out of the practice with locum cover, and the prescribing advisory team gave presentations to start each section.

With the cost of our GIT drugs at worst 31% higher than the health authority average, we focused hard on this area. It was clear that there were as many guidelines for the management of GIT problems as there were partners.


PPIs were first against the wall. We used almost all omeprazole and lansoprazole. We were presented with evidence that a lower dose (15mg) of lansoprazole was as effective in most cases as the full dose (20mg) of omeprazole, at much lower cost.

An early guideline decision was therefore made to use lansoprazole 15mg as first-line PPI and to switch patients from other preparations to this drug.

We agreed a principle to be applied to all changes, which helped carry along all partners. Each partner would be given a list of their patients on PPIs and could choose any who, for clinical or other reasons, it was felt inadvisable to attempt to change. Thus patients with severe Barrett's mucosa or stricture formation or those who might have trouble understanding the change were excluded from the attempt to change.

I drafted a letter to send to patients explaining all these changes, which could be modified for each drug change (Figure 1). In this letter we were honest about the main reason for the change and patients seemed to respect this approach.

Figure 1: Draft letter to send to patients affected by the changes
draft letter

Experience elsewhere suggested that around 10% would find the change unacceptable, although our figures were a little higher than this.

Generic substitution across the board for all GIT drugs was agreed. There were very few problems with this, although I did have one patient who found that only Zantac would control his dyspepsia and that ranitidine was not as effective.

This highlights a fundamental point about guidelines: if they are to succeed then we must accept that we live in an imperfect world and that there will be practical problems which any guideline must be flexible enough to allow for.


H2 blockers were next in the firing line. As often happens, the Hawthorne effect (just looking carefully at a problem changes it even before any action is agreed) was noticeable.

There were significant numbers of patients on ranitidine 600mg and clearly this could be reduced or changed to a PPI. It was agreed that cimetidine would be first-choice H2 blocker unless drug interaction was an issue. Nizatidine was next choice, with ranitidine third choice.

Where there seemed a need for a high-dose H2 blocker, we thought it best to check the diagnosis first and then consider a PPI. The issue of each doctor looking at a patient's medication critically and reconsidering the case was very useful.

We were struck by the way that an exercise which started off as cost cutting became one where the clinical management was improved with direct patient benefit.


Helicobacter pylori was obviously central to this whole exercise. As part of our guidelines for implementing these changes we looked at whether Helicobacter status had been checked, and whether a trial of eradication therapy had been carried out.

There were several different regimens for this and we agreed the local 7-day one (omeprazole 20g bd, clarithromycin 500mg bd, metronidazole 400mg bd).


The whole management of dyspepsia was woven into this discussion, so we produced practice guidelines to manage this problem in a coherent way.

We recognized that there were real difficulties with regard to upper GIT endoscopy supply locally, and that there was already pressure on barium meal investigations. This is a national problem and there was no point in automatically sending everyone for investigation. We followed the existing local guidelines, but adjusted them to suit our practice and organisation (Figure 2,below).

Figure 2: Practice guidelines for the management of dyspepsia and gastro-oesophageal reflux disease

Dyspepsia guidelines
  1. Patients aged less than 45, with no alarm symptoms/signs, and Helicobacter negative: Treat with antacids, or if severe/unresponsive treat with cimetidine 800mg a day for a maximum of 4 weeks. If better, no problem.
  2. If Helicobacter positive: Treat with eradication therapy as agreed; then if no response, or (i) above not better, refer for endoscopy.
  3. Over 45, consider 7 days cimetidine; if complete and prolonged resolution of symptoms, then no investigation needed.
  4. Over 45 or symptoms not gone in (iii) above, then refer for endoscopy.
  5. Warning signs/symptoms (dysphagia, anaemia, bleeding, weight loss, persistent vomiting, epigastric mass, previous gastric ulcer or gastric surgery, NSAID use), then refer at once irrespective of age.
  6. Endoscopy not needed if recently investigated for same symptoms and diagnosis clearly established or known duodenal ulcer with good response to treatment.
Gastro-oesophageal reflux guidelines
  1. Lifestyle advice: weight control, smoking cessation, alcohol, big meals, lying too flat after food.
  2. Simple antacids and alginates.
  3. If still symptomatic, consider H2 blocker, prokinetic agent or PPI.
  4. Refer for endoscopy if not effective or problem recurs after treatment.

Refer at once if there are any of the alarm symptoms/signs listed above in (5) in dyspepsia guidelines

Guidelines on GORD seemed easier to agree upon, but were occasionally harder to implement. An overweight, heavy drinking smoker who could eat what he liked so long as he had his PPI was going to be hard to change.

There is evidence too that in severe GORD with stricture risk, high dose omeprazole may be better than high-dose lansoprazole.


The GIT drug review was only a part, albeit a central one, of a review of several aspects of prescribing that we did at the same time.

Our cumulative expenditure this year was less in absolute terms than the previous year, despite an increase in list size. We are under budget at 94% of our target, compared with 111% of the target the year before. Our GIT drug costs dropped in the year by 34%.

There have been problems. Some patients found that the new medication was not as effective, and we had to spend time sorting this out and discussing it with patients. Our dispensing and secretarial team had to put in a huge effort to provide the information and carry out the changes. Around 15% of patients needed to go back on their previous treatment.

Overall it has worked well. We realise that we can cut costs and improve patient care by pragmatic application of guidelines. Patients too are prepared to help reduce NHS costs if they are told honestly what is happening.

This is not a static process; we cannot sit back and congratulate ourselves that the job is done. We need to work on other areas and ensure that things do not slip.


  1. Harlow T. Evolution of a practice-based sore throat protocol. Guidelines in Practice December 1998; 1: 14-15.

Guidelines in Practice, August/September 1999, Volume 2
© 1999 MGP Ltd
further information | subscribe