|
Screening for urinary tract infections in a gynaecological setting: validity and cost-effectiveness of reagent strips |
Preston A, O'Donnell T, Phillips C A |
|
|
Record Status 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. Health technology Screening for urinary tract infections in a gynaecological setting using Ames 8SG reagent strips to detect leukocyte esterase and nitrite.
Economic study type Cost-effectiveness analysis.
Study population The study population consisted of women presenting to a gynaecological ward. No exclusion criteria were reported.
Setting The study was set in a thirty-bed gynaecological unit in a NHS hospital trust, namely Leicester Royal Infirmary. The economic study was conducted in the UK.
Dates to which data relate The authors did not state the dates to which effectiveness and cost data related. The results of the study were published in 1999.
Source of effectiveness data The effectiveness data were derived from a single study.
Link between effectiveness and cost data Costing was undertaken on the same patient sample as that used in the effectiveness analysis; it was not clear whether costing was undertaken prospectively or retrospectively.
Study sample A convenience sample of 228 women presenting in a gynaecological ward were tested using Ames 8SG reagent strips to detect leukocyte esterase and nitrite, against the results of laboratory analysis for UTI. Power calculations to determine the sample size were not reported. Baseline characteristics were not reported, thus preventing comparison with the study population.
Study design This was a diagnostic test accuracy evaluation conducted in a single centre. The duration of the study was 6 months.
Analysis of effectiveness The analysis was based on intention to treat. The parameters used to compare the diagnostic value of the test strips with conventional microbiological methods were sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). The authors stated that sensitivity must be greater than or equal to 95% to ensure diagnosis of UTI; specificity for detection of true negatives should be greater than or equal to 70% in order to avoid over diagnosis; PPV should be greater than or equal to 50% to ensure reliability of diagnosis and NPV should be greater than or equal to 95% (ideally close to 100%) to measure the reliability of the negative result. Results from the standard microbiological techniques were defined as either 'no abnormality detected' or 'significant for UTI' (greater than 10^5 organisms/mL). No statistical analysis for confounding was carried out.
Effectiveness results Sensitivity, specificity, PPV and NPV were as follows:
nitrite: sensitivity 64.3%, specificity 99.0%, PPV 90.0%, NPV 95.3%;
leukocyte esterase: sensitivity 75%, specificity 88.5%, PPV 47.7%, NPV 96.2%; and
combined: sensitivity 96.4%, specificity 88.5%, PPV 54.0%, NPV 99.4%.
Clinical conclusions The use of Ames 8SG reagent strips to determine the presence of nitrite and leukocyte esterase as indicators of UTI provides levels of accuracy, which exceed the criteria given by the authors.
Measure of benefits used in the economic analysis The authors did not provide a summary measure of benefits, but opted for a cost-consequences approach. The reader is referred to the effectiveness results reported above.
Direct costs Direct health service costs (referred to by the authors as 'visible costs') were considered, namely cost of reagent strips and laboratory urinalysis. Quantities and costs were presented separately. Costs were not discounted due to the short duration of the study (6 months). The authors stated that costs such as nurse time, specimen distributor, and microbiologist, were not included in the analysis. The price source was not stated, although it can be assumed the hospital in which the study was conducted provided the data. The price year was not stated. The costs given were for reagent strips (0.18 each) and laboratory urinalysis (3 for a negative specimen and 15 for a positive specimen).
Statistical analysis of costs No statistical analysis was performed.
Indirect Costs Indirect costs were not considered.
Sensitivity analysis An analysis of percentage saving versus prevalence was presented as a graph.
Estimated benefits used in the economic analysis The authors did not provide a summary measure of benefits and, as such, a cost-consequences analysis was conducted; the reader is referred to the effectiveness results reported above.
Cost results The incremental cost was -495.31 (48.6%).
The graph of cost savings versus prevalence showed that savings would decrease with prevalence, as they are based on the fact that negative specimens are not subsequently sent to the laboratory for urinalysis. Though there would be cost savings if prevalence remained below 50%.
Synthesis of costs and benefits Costs and benefits were not combined due to the cost-consequences approach adopted in the analysis.
Authors' conclusions The results of this study suggest that the combined use of nitrite reductase and leukocyte esterase reagent strips provides a valid method for screening gynaecological patients, and, together with laboratory urinalysis, results in accurate diagnosis of UTI. Cost-effectiveness is dependent on the rate of infection in the patient population.
CRD COMMENTARY - Selection of comparators The reason for the choice of the comparator is clear, as both reagent strips and microbiological analysis were used in the authors' setting. You, as a database user, should consider if the same applies to your own setting.
Validity of estimate of measure of effectiveness The analysis was based on a convenient sample of patients presenting to a gynaecological ward. Although this approach is appropriate for this type of study, the authors did not back it up with appropriate statistical analyses to take account of potential biases and confounding variables. However, since each test was performed on the same sample, bias should be low. PPV and NPV must be interpreted with caution because they vary with prevalence. For example, PPV is directly, and NPV inversely related to prevalence. Therefore, the rate of false negatives will increase and false positives decrease with prevalence; the health implications of this were not accounted for or discussed by the authors. The study sample was representative of the study population, but no power calculations were used to determine the sample size.
Validity of estimate of measure of benefit The authors did not provide a summary measure of health benefit, and, as such, a cost-consequences analysis was conducted.
Validity of estimate of costs All categories of costs relevant to the perspective adopted were considered; costs such as nurse and microbiologist time were excluded from the analysis, but these are not likely to have affected the final conclusions. Quantities and costs were presented separately. Discounting was not performed due to the short duration of the study. No sensitivity analysis of quantities was conducted and this may limit how the study findings are interpreted. Also, costs were not statistically analysed, and their source and price year were not stated, thus limiting the validity of the study conclusions. The costs of delayed treatment due to false negatives were not mentioned.
Other issues The authors made extensive comparisons with other relevant studies in terms of sensitivity and specificity, indicating generalisability to other NHS settings. However, as the authors acknowledged, NPV and PPV generalisability do depend on prevalence, but not simply in terms of the costs reported. The authors do not appear to have presented their results selectively.
Implications of the study The authors stated that the use of reagent strips in a mass-screening programme in a gynaecological study proved both valid and cost-effective. This must be viewed within the caveats mentioned. The authors suggest that further research might be needed to test the accuracy of laboratory testing before recommending a change in practice.
Source of funding Financial support from Leicester Royal Infirmary Hospital Trust.
Bibliographic details Preston A, O'Donnell T, Phillips C A. Screening for urinary tract infections in a gynaecological setting: validity and cost-effectiveness of reagent strips. British Journal of Biomedical Science 1999; 56(4): 253-257 Other publications of related interest Stevens M. Screening urines for bacteriuria. Medical Laboratory Sciences 1989;46(3):194-206.
Indexing Status Subject indexing assigned by NLM MeSH Cost-Benefit Analysis; England; Female; Gynecology /methods; Humans; Mass Screening /methods; Reagent Strips; Urinary Tract Infections /diagnosis AccessionNumber 22000006236 Date bibliographic record published 31/12/2001 Date abstract record published 31/12/2001 |
|
|
|