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.

The following treatment strategies in the treatment of children with dyspepsia were compared:

(1) Immediate endoscopy (EGD) with biopsy for Helicobacter pylori (H. pylori) infection;

(2) immediate EGD without biopsy;

(3) H. pylori serology and treatment for H. pylori if serology is positive;

(4) empiric treatment with antisecretory agents alone;

(5) empiric treatment with both antisecretory agents and antibiotics for H. pylori.

Treatment and diagnosis.

Cost-effectiveness analysis.

Children with dyspepsia not concurrently taking non-steroidal anti-inflammatory drugs.

Hospital. The economic study was carried out in Ann Arbor, Michigan, USA.

The data for the effectiveness analysis were based mainly on studies published between 1986 and 1994. The data for the different resources used were not reported as belonging to any specific date. The date of the prices used was not reported.

Effectiveness data were derived from a review of previously published studies and expert opinion.

A decision tree was used to estimate costs and benefits of alternative therapies.

The outcomes assessed were as follows: incidence of ulcer disease, rate of H. pylori infection if ulcer, rate of H. pylori infection if no ulcer, ulcer healing rate after antisecretory therapy, H. pylori infection eradication after antibiotics (including compliance), symptoms if ulcer persists after antisecretory therapy, symptoms if ulcer healed after antisecretory therapy, H.pylori serologic test sensitivity and specificity.

One prospective double-blind study was reported. No more details were presented apart from the conditions described in the patient population. The authors stated that "the hypothetical cohort was presumed to have a mix of clinical conditions as drawn from the published literature".

Medline was searched.

Not reported.

The methods used to judge relevance and validity were not reported. The data were extracted by means of summary statistics corresponding to the outcomes relevant to the decision tree used.

Twenty studies were included in the review.

Not combined. The ranges of the different values reported for the same summary statistics were used in the sensitivity analysis.

Not reported.

The results were reported as follows (ranges in parentheses):

incidence of ulcer disease 4% (range: 2 - 24);

H.pylori infection if ulcer, 77% (range: 33 - 100);

H. pylori infection if no ulcer, 23% (range: 4 - 60);

Ulcer healing rate after antisecretory therapy, 75% (range: 50 - 90);

H.pylori infection eradication after antibiotics, 80% (range: 50 - 90);

symptoms if ulcer persists after antisecretory therapy, 90% (range: 50 - 100);

symptoms if ulcer healed after antisecretory therapy, 10% (range: 0 -30);

H.pylori serologic test sensitivity and specificity, 95% (range: 50 - 100) each.

Estimates of effectiveness were also based on authors' assumptions and expert opinion.

Symptom rate after empirical therapy if no ulcer was assumed to be 60% (range: 30 - 90). Monthly ulcer recurrence rate if H.pylori positive was assumed to be 2.7% (range: 2 - 6.6); monthly ulcer recurrence rate if H.pylori positive, was assumed to be 0.6% (range: 0.1 - 2); serious antibiotic complication/dose, was 0.5 (range: 0.1 - 1).

The measure of benefits was ulcer cured at 1 year. A decision tree was used to deal with uncertainty and to represent the variety of general practice treatments.

Quantities were not fully reported separately. Cost items were reported separately. Operating costs were measured (ambulatory services, inpatient care and physicians' fees, endoscopy costs, costs of medications) as well as cost of complications. The boundaries adopted were third-party payer and patient. Some quantities were based on the institution's own experience, others were based on assumptions. The unit cost data was based on actual data. Patients' out of pocket expenses for medications were collected. The sources of unit costs were the Health Care Financing Administration for Medicare (endoscopy charges) and a "large private third-party payer covering approximately 3 million persons in the Eastern US". The prices implicit in the cost estimates were not dated. The costs incurred by those with symptoms, but with duodenal ulcer disease ruled out by endoscopy, were not included since they were thought to be common to the treatment strategies. Out-of-pocket expenses by the patients for medical service other than drugs were not included.

Not included.

US dollars ($).

The parameters varied in the sensitivity analysis were not clearly specified. The values used were those of the ranges reported in the published literature. One way and two-way simple sensitivity analyses and threshold analyses were carried out.

It was reported that the alternative strategies had virtually identical clinical outcomes. The percentage of patients with ulcers identified and cured was reported to be essentially the same for all strategies involved (>98%). The duration of benefits was one year after initiation of the intervention.

The total cost per patient was:

treated with antisecretory agents alone, $1,160;

antisecretory agents in combination with antibiotics, $1,164;

test and treat (serologic) strategy, $1,224;

endoscopy alone, $1,248;

endoscopy performed simultaneously with biopsy, $1,458.

The duration of costs was one year after initiation of intervention.

The cost per ulcer cured was reported. The ratio for the non-invasive (empiric) treatment strategies was $29,000 each. The corresponding figure for H. pylori serology screening prior to treatment was $30,600, whereas for immediate endoscopy alone it was $31,200. The treatment strategy of endoscopy with biopsy had, in turn, a figure of $36,450. The one-way sensitivity analysis showed that the results were sensitive only to the cost of endoscopy and the probability of recurrent symptoms in empirically treated patients having no ulcer. The threshold value of the cost of endoscopy, including professional and suite fees, was $950 (baseline value: $1,180), value below which EGD (immediate endoscopy) was the strictly preferred strategy. The recurrence rate of symptoms in patients treated with empiric antisecretory therapy had a threshold value of 65% (baseline value of 60%) above which the EGD strategy was strictly preferred to the rest of the strategies.

In children presenting with dyspepsia, empiric antisecretory therapy prior to endoscopy is a cost-effective, less-invasive alternative to immediate endoscopy. The sensitivity analysis showed that if the cost of endoscopy decreases, the cost-effectiveness of immediate endoscopy will improve. "Given the relatively small savings with empiric therapy and the current pressures to further lower the cost of EGD, we would project that EGD in children is and will become an even more cost-effective strategy. These diagnostic decisions should be balanced by patient preference pertaining to the value of information from the EGD compared with the costs in terms of perceived discomfort and inconvenience".

No specific strategy was chosen as the comparator.

The internal validity of the benefit results may be weakened by the lack of a systematic review of the literature and quality assessment of the primary studies included in the review. The 'discomfort' or 'inconvenience' experienced by patients and relatives were not valued.

Quantities were not fully reported separately. The sources of resource utilisation data were not clearly specified. The exclusion of indirect costs was recognised by the authors as important.

Given the lack of randomisation, systematic review of the literature and quality assessment of the primary studies, and statistical analysis of the costs, the results need to be treated with some caution. The issue of generalisability to other settings or countries was not addressed.

None stated.