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.

Drug therapy for treating women with uncomplicated genital chlamydia trachomatis (CT).

Treatment.

Cost-effectiveness analysis.

Hypothetical cohorts of non-pregnant women of child bearing age who had uncomplicated cervical CT infections, which had been confirmed by laboratory testing.

The setting was not explicitly stated. The economic study was carried out in Seattle, Washington State, Denver, Colorado, and Philadelphia, Pennsylvania, USA.

The effectiveness and resource data were derived partly from literature published between 1975 and 1993 and partly from experts consulted by the authors.1993 prices were used.

Synthesis of previous studies and expert opinion.

A decision analytical tree model was used to calculate health outcomes and costs.

The primary outcomes assessed in the review were the probabilities ofmajor and minor complications resulting from the use of both DC and AZ to treat CT. The probability of major chlamydia sequelae following the use of DC and AZ were also assessed in the review.

The particular study designs included in the review were not specifically stated by the authors. No inclusion/exclusion criteria were stated by the authors.

Not stated.

Not stated.

Not stated.

43 studies were included in the review. Not all study types were specifically stated, however studies included randomised control trials, comparative studies, decision analysis studies and a cohort study.

A modified Delphi method was used to combine studies with findings from expert opinion in order to estimate health outcomes.

Not stated.

A summary of base estimate probabilities for various clinical events associated with the treatment options was as follows ( Plausible Range' in parentheses):

Sequelae of persistent cervical chlamydia:

Symptomatic pelvic inflammatory disease 0.15 (0.08 -0.20);

Outpatient treatment 0.86 (0.80 - 0.92);

Inpatient treatment 0.14 (0.08 -0.20);

Sequelae of symptomatic pelvic inflammatory disease:

Infertility 0.17 (0.08 -0.20);

Ectopic pregnancy (0.08 (0.05 - 0.10);

Chronic pelvic pain 0.12 (0.08-0.16);

Newly infected partners per woman with persistent chlamydia 1.00 (0.75 - 1.25);

Sequelae for infected male partners:

Symptomatic urethritis 0.50 (0.35-0.65);

Epididymitis 0.02 (0.01-0.04);

Outpatient treatment 0.90 (0.85-0.95);

Inpatient treatment 0.10 (0.05-0.15).

Sequelae for children born to infected mothers:

Neonatal conjunctivitis 0.20 (0.15-0.25);

Neonatal Pneumonia 0.10 (003-0.16);

Outpatient Treatment 0.80 (0.70-0.90);

Inpatient Treatment 0.20 (0.110 -0.30).

Antibiotic effectiveness:

DC 0.86 (0.80-0.90);

AZ 0.96 (0.94-0.98).

Adverse reactions to oral DC:

Minor reactions 0.16 (0.10-0.20);

Major reactions 0.001 (0.0001-0.001).

Adverse reactions to oral AZ:

Minor reactions 0.13 (0.10-0.20);

Major reactions 0.0001 (0.0001-0.001).

A survey of the expert opinion of thirteen physicians skilled in the management of CT was undertaken.

Estimates of epidemiological and clinical variables were derived. The experts were also asked for information on the use of medical services. These were then combined with information from the literature review to producebest estimates' of various outcomes.

Major and minor persistent cervical chlamydia complications avoided.

Direct hospital costs, outpatients costs and physician services were estimated from information from Colorado Blue Cross and Blue Shield. The manufacturers' recommended retail price was used to estimate the cost of pharmaceuticals. Costs for specific components of treatment were derived by aggregating the various costs elements (which were partly based on expert opinion). Specifically costs were obtained for :

Sequelae of persistent cervical chlamydia: outpatient treatment of pelvic inflammatory disease, inpatient treatment ofpelvic inflammatory disease.

Sequelae of symptomatic pelvic inflammatory disease: Infertility, ectopic pregnancy, chronic pelvic pain.

Treatment of partner, sequelae for infected male partners: Symptomatic urethritis, outpatient treatment of epididymitis, inpatient treatment of epididymitis,

Sequelae for children born to infected mothers: Neonatal conjunctivitis, outpatient treatment of neonatal pneumonia, inpatient treatment of neonatal pneumonia. Minor and major antibiotic reactions. Cost of drug therapy AZ 1g, DC 100mg for 7 days and the antibiotic cost differential.

Costs were discounted at a rate of 5%. 1993 prices were used. 2 models were developed with discounted and undiscounted benefits, although only the undiscounted benefits are shown as the results were similar for both models.

US dollars ($)

Univariate sensitivity analysis was carried out on rates of chlamydia sequelae, antibiotic reactions, antibiotic effectiveness, costs of antibiotics and chlamydia sequelae. Multivariate analysis was used to combine the most extreme estimates for each chlamydia sequelae in order to maximise the cost-effectiveness of the standard treatment DC. Specifically the model used the highest rate of DC effectiveness (0.9), lowest rate of pelvic inflammatory diseases in patients with persistent cervical CT (0.08), hospitalization for such infections (0.08), pelvic inflammatory disease sequelae (0.08 for infertility, 0.05 for ectopic pregnancy and 0.08 for chronic pain), infected partner sequelae (0.35 for urethritis, 0.01 for epididymitis), infected neonatal sequelae (0.15 for conjunctivitis and 0.03 for pneumonia) and lowest number of infected partners per infected female patient (0.75).

Using baseline probabilities the number of persistent chlamydia infections using the AZ strategy was estimated to be 4000 per 100,000 compared with 14,300 per 100,000 for DC treatment. Using the AZ strategy led to 938 major complications compared with 3,330 in the DC group. The number of minor complications in the AZ and DC groups were 15,715 and 25,706 per 100,000 respectively. AZ continued to be more effective than DC as a strategy for preventing major complications, even after sensitivity analysis adjusted any of the probabilities through their plausible ranges. The result was also insensitive to a 30% decrease in the cost of treating chlamydia sequelae.

Cost results were as follows :-

Sequelae of persistent cervical chlamydia:

Outpatient treatment ofpelvic inflammatory disease $191.55,

Inpatient treatment ofpelvic inflammatory disease $5,259.03.

Sequelae of symptomatic pelvic inflammatory disease:

Infertility $5172.68, Ectopic pregnancy $4717.69, Chronic pelvic pain $3,809.11.

Treatment of Partner $50.39.

Sequelae for infected male partners: Symptomatic urethritis $82.39,

Outpatient treatment of epididymitis $206.70,

Inpatient treatment of epididymitis $3421.59.

Sequelae for children born to infected mothers:

Neonatal Conjunctivitis $81.80,

Outpatient treatment of neonatal pneumonia $225.80,

Inpatient Treatment of neonatal pneumonia $3023.80.

Minor and major antibiotic reactions $35.00 and $3747.50 respectively.

Cost of drug therapy AZ 1g $18.75, DC 100mg for 7 days $5.51 and the antibiotic cost differential $13.24.

A synthesis was only undertaken by the authors when AZ treatment was not the dominant strategy. In all scenarios AZ was more effective than DC treatment. However, when multivariate sensitivity analysis was carried out it was found that although AZ treatment led to 440 fewer major and 14,431 fewer minor complications than DC, it was a more costly treatment. The incremental cost-effectiveness of AZ treatment per additional complication prevented was $521. Even with multivariate analysis, AZ was the dominant strategy, if the cost differential between the two drugs was less than $9.80.

The authors concluded that their model used the best available data with which to compare AZ with DC. AZ was, in almost all cases, the dominant strategy, although with certain variations in the sensitivity analysis, it became more costly, although still more effective. The results of the study are not universal however, and the authors strongly recommend that further studies be carried out to confirm their findings. In particular the compliance rate of DC should be studied since, if compliance with this treatment is improved, then its efficacy might also be improved. More research should be conducted into the rate of adverse reactions with AZ as this study has had to assume the same rate as DC, which might not be the case.

In general an excellent study, in which the authors have carried out a wide range of sensitivity analyses to confirm their findings of the superiority of AZ treatment. Extensive information on costing is provided in the appendix. It would be helpful however to know exactly how the authors combined the results of the literature survey with the expert opinion survey and also what sources were searched for relevant literature. The inclusion/exclusion criteria for the literature survey could also have been stated. Although the authors have used a plausible range in order to perform sensitivity analysis, reporting confidence intervals would also have been useful.

Well designed prospective randomised control trials comparing the efficacy of AZ and DC are required. These should examine rates of DC compliance, the rate of AZ reactions and the effectiveness with regard to a wider population with CT infections.

Supported in part by an unrestricted investigator-initiated grant from Pfizer Inc.