This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.

Serology test (Helisal test) for Helicobacter pylori followed by eradication therapy for patients testing positive, plus endoscopy and further treatment as necessary for patients testing negative and for those with persistent symptoms following H. pylori eradication.

Diagnosis; treatment.

Cost-effectiveness analysis.

Patients newly presenting to their GP with dyspepsia, heartburn or excess acid, with no previous diagnosis and having had no previous therapy except antacid.

Primary care. The economic study was carried out in Scotland, UK.

Data on testing and treatment for H. pylori were collected for patients presenting with symptoms during 1995/1996 and followed for a 6 month period thereafter. Data on the comparator were based on studies published in 1996 and 1990. The cost of drugs was taken from the British National Formulary,March 1996. No date was given for other costs.

Effectiveness data for the intervention were based on a single study and for the comparator ona synthesis of previously completed studies.

Cost data were based on the same patient sample used in the effectiveness study. It was not clear whether data were collected prospectively or retrospectively.

No reasons were given for the choice of sample size and the methods of sample selection were not given. It was not stated whether any patients refused to participate nor whether any subjects were excluded from the study.57 patients presenting to their GP with dyspepsia at an urban fundholding practice agreed to be tested for H. pylori. There was no comparator group in the single study.

The study was a prospective case series. The study was single centred. Follow-up was for 6 months. No loss to follow-up was reported.

It was not stated whether the analysis was based on intention to treat or treatment completers only. Primary health outcomes were the proportion of patients who tested positive for H. pylori, the proportion of these in whom eradication therapy was successful at either the 1st or 2nd attempt, proportion who remained symptom free for the rest of the 6 month follow-up period and therefore the proportion of patients in the sample who could be cured of dyspepsia by diagnosis and treatment of H. pylori infection. There was no comparator group in the single study and the authors did not show that the sample in the study was comparable to the samples in the previously published studies. No demographic details were given for the sample.

36 patients (63%) tested positive for H. pylori. In 28/36 (78%) therapy was successful at the first attempt and in 25 of those (89%) symptoms did not return during the follow-up period. In 6/36 (17%) therapy was successful at the second attempt and in all 6 symptoms did not return during the follow-up period. Thus 31/57 (54%) were successfully treated for dyspepsia by treating H. pylori infection.

A high proportion of patients infected with H. pylori respond successfully to eradication therapy with no need of further medication.

Decision trees were used to model the costs and effectiveness of the intervention and comparator strategies over five years and over the expected lifespan of the patient.

Effectiveness information on the proportion of patients under 45 needing maintenance therapy after initial treatment was derived from a review of studies. Also the proportion of patients over 45 in the population of new dyspeptic patients and the proportion of patients over 40 needing maintenance therapy after endoscopy diagnosis were derived from published studies.

Not stated.

Not stated.

Not stated.

Not stated.

Two studies were included.

Studies were not combined. Each provided different inputs for the model.

Not investigated.

From the first study, in patients under 45 healing rates of80% were achieved following a 4 week course of cimetidine at a dose of 800 mg/day. In the second study 4,500 patients out of 109,000 were aged over 40. Derived from this the proportion of patients over 45 was assumed to be 41%. In patients over 40, 75% required maintenance therapy after endoscopy diagnosis.

The authors stated that "while it was recognised that the outcomes ... were not identical, a cost minimisation approach was considered appropriate". They therefore did not use benefits in the economic analysis.

Mean costs were given for the items listed below. Some quantities were given for the items costed, but prices per unit were not given. A price date for all costs was not given but drugs costs were based on the British National Formulary,March 1996. Long term maintenance therapy was discounted at 5% and other costs were assumed to fall within a short time span. Mean costs included in the model were: the serology test, a 15 minute pharmacy appointment, triple therapy using omeprazole, amoxycillin and metronidazole, the urea breath test (BSIA), nursing time, endoscopy, dual therapy using omeprazole and clarithromycin, administration time and maintenance therapy using 400 mg cimetidine daily (the cost of an alternative, 150 mg ranitidine daily was used in the sensitivity analysis). The cost boundary was that of the health service.

Costs were not analysed stochastically.

Not included.

UK pounds sterling ().

Simple one-way analyses were carried out on each parameter of the model. Values were varied within a range to test uncertainties in the data. Costs of triple therapy eradication were varied between 20.82 and 102.98, of dual therapy eradication between 18.11 and 29.53, of one year maintenance therapy between 56.21 and 169.66, of UBT between 10 and 250 and of endoscopy between 45 and 900. The discount rate was varied between 3% and 10%. The probability of a patient being under 45 was varied between 40 - 85%, of maintenance being needed after endoscopy in the comparator between 50 - 90% and of maintenance being needed if H. pylori was negative and scoped between 10 - 50%.

Not applicable.

Using the 5 year model total expected costs per patient would be 113.89 for the testing and treatment of H. pylori strategy and 155.54 for the conventional strategy. Using the lifetime model the testing and treatment strategy would cost 199.94 per patient and the conventional strategy 383.88.

Not applicable.

In the long term, considerable cost savings could be made by H. pylori eradication in patients presenting with dyspepsia for the first time. This model was based on a small sample of patients and further research should be done in this area.

It is a weakness of this study that the single study had no comparator and that no demographic details of the sample group were given. An indication, from experiment or review, of the proportion of symptomless carriers of H. pylori in the general population would have been relevant and a control group who were tested, but not treated, except with antacids, could have been considered (if this was ethically appropriate). It is not clear that the results of the comparators were valid because details of the methods of searching for relevant studies and the criteria for judging primary studies were not given.

In the discussion the authors mentioned that many patients"reported they could eat better and generally felt better after H. pylori eradication." However, the economic analysis did not consider such benefits and this was a weakness of the study. No account was taken of the possible delay in treatment of serious illness, such as gastric cancer, due to referral for H. pylori testing before endoscopy which would be an adverse effect of the intervention strategy.

It is not clear that all costs were included, for instance overheads and costs of further GP visits, particularly in the long term costs of therapy where only drug costs seem to have been covered. The 6 month follow-up period may have been too short for costing the intervention therapy and the assumption that no further costs would arise following H. pylori eradication compared with lifetime maintenance may be a false one.

A more reliable assessment of the relative benefits would come from a randomised controlled trial.

Funded by a research grant from the Scottish Office Home and health Department.