|
The cost-effectiveness of evidence-based guidelines and practice for screening and prevention of tuberculosis |
MacIntyre C R, Plant A J, Hendrie D |
|
|
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 Contact tracing through the use of skin test or mass chest radiographs (CXR) in the detection of tuberculosis (TB).
Economic study type Cost-effectiveness analysis.
Study population The study population was made up of people who had come into contact with a case of infectious TB (about 20% of exposed contacts).
Setting Institution. The study was conducted in Victoria, Australia.
Dates to which data relate The effectiveness data were collected in 1991 and the resource data were based on direct health costs in 1991. The price year was 1991.
Source of effectiveness data Effectiveness data were derived from a single study as well as the authors' assumptions.
Link between effectiveness and cost data The screening records for identified contacts of all cases notified in 1991 were reviewed and the information retrieved was then used to determine costs. The cost data were derived (for model 1) from the same patient sample as that used in the effectiveness results. Costing was undertaken retrospectively.
Study sample 1,142 contacts were traced for model 1, of which 35% were skin tested, 57% had initial CXR and 24% had repeat CXR. Model 2 had 1,142 contacts of which 62% aged less than 35 years and had skin tests, 94% were aged more than 35 years and had an initial CXR and 64% of all contacts with skin test results greater than 5mm had repeat CXR. Model 3 had 1,613 contacts of which 100% were skin tested and 21% of the contacts had a positive skin test or were symptomatic and had initial CXR, and less than 10% had repeat CXR. The median age of contacts was 33 years but no other demographic details were given of the study sample. No power calculations were used to determine sample size.
Study design This was a retrospective cohort study at a single centre.
Analysis of effectiveness The analysis of the clinical study was based on treatment completers only. The outcome measures were prevention of cases of TB, cases of TB found and the number of contacts screened.
Effectiveness results The presence or absence of infection was identified in 35% of contacts in model 1. In model 2, 62% of contacts were identified and in model 3, 100% of contacts were identified.
Model 1 had 3 cases prevented, model 2 had 14 cases prevented and model 3 had 27 cases prevented (of the 80% treated).
Preventive therapy was given to 38 (3%) in model 1, 152 (13%) in model 2 and 271 (17%) patient in model 3.
Clinical conclusions Model 3 prevented the most number of cases (27), followed by model 2 (14) and model 1 prevented the least number of cases (3). Therefore, without full compliance with the guidelines, fewer cases were prevented.
Modelling Three models were used to estimate costs and benefits. Model 1 was based on the current practice, model 2 considered a scenario that applied what would have happened if existing contact screening guidelines had been followed, and model 3 was a hypothetical model based on internationally approved practice of screening 9 contacts per case of infectious disease.
Methods used to derive estimates of effectiveness Results of a previous study were used to obtain the incidence of active TB in identified contacts, and data on the expected lifetime incidence of TB were obtained from published sources. The authors also made assumptions to derive estimates of effectiveness.
Estimates of effectiveness and key assumptions The authors assumed that case finding would be based on the fact that the majority of cases would be symptomatic, or if not symptomatic, would warrant CXR because of a positive skin test. The authors assumed that the number of cases found would be the same in the three models (4) and this was supported by the evidence that mass CXR screening does not increase case finding. The maximum incidence of secondary cases was estimated using 2-year follow-up data from 1991-1993 and the figure for lifetime incidence of TB of 10% of infected contacts.
Measure of benefits used in the economic analysis The benefit measures were the cases of TB prevented, number of TB cases found, and number of contacts screened.
Direct costs The main costs included were the individual components of contact tracing, screening and preventive therapy. Costs and quantities were reported separately. These were salaries of the nurse manager, nurse and clerical staff, non-salary operational budget (costs of travel, vehicles, cleaning, office supplies, electricity, heating, postal and telephone expenses, electronic data processing costs (incurred by the TB programme), skin test (syringes, needles, swabs and tuberculin), CXR, physician fee, medicine costs (isoniazid, ethambutol, rifampicin, pyrizinamide, pyridoxine), liver function tests and BCG vaccines. The cost of hospital care for TB was calculated by studying all TB cases in the financial year 1991/1992 documented in the Victoria Inpatient Minimum Database (VIMD). Average hospitalisation weighted costs were estimated by using Australian National Diagnosis Related Groups (AN-DRGs) by costs and frequencies assigned to TB. Outpatient costs were calculated from estimates of individual components of treatment. Discounting was not appropriate due to the short period of cost analysis (less than 1 year). The perspective adopted was that of the provider.
Statistical analysis of costs No statistical analysis was undertaken.
Indirect Costs Indirect costs were not considered.
Currency Australian dollars (Aus$).
Sensitivity analysis Sensitivity analysis was carried out for model 3 and three scenarios were tested.
(1) only 65% of infected contacts were referred;
(2) all the infected contacts were referred but only 65% received isoniazid (INH);
(3) all infected contacts were referred, 80% received INH as model 3, but the efficacy of INH was 65% rather than 80%.
Sensitivity analysis for model 3 was performed using 6 private physicians' visits compared to the 3 used in the other models.
Estimated benefits used in the economic analysis The authors assumed that 4 cases of TB were identified in the three models considered. 3 cases were prevented in model 1, 14 cases in model 2, and 27 cases were prevented in model 3.
Cost results The total cost of model 1 was Aus$927,196, model 2 Aus$822,385 and model 3 Aus$872,835. The cost per contact screened for model 1 was Aus$812, model 2 Aus$720 and model 3 Aus$541.
Synthesis of costs and benefits Cost per case detected and cost per case prevented were the benefit measures used in the study. The cost per case prevented for model 1 was Aus$309,065, model 2 Aus$58,742 and model 3 Aus$32,210. The cost per case found in model 1 was Aus$231,799, model 2 Aus$205.596 and model 3 Aus$218,208. Despite preventing fewer cases of TB the total cost of model 1 is greater than that of models 2 and 3. Case finding was over Aus$200,000 per case found in all three models. Case prevention was more cost-effective than case finding in models 2 and 3. The incremental cost-effectiveness of model 3 compared to model 2 was Aus$107 per additional contact screened and Aus$3,881 per additional case prevented.
Authors' conclusions Case finding as a product of contact screening is not as cost-effective as case prevention in any of the three models examined, ranging from Aus$205,596 to Aus$231,799 per case found. Prevention of TB can be cost-effective as indicated by model 3, although it had not been so in 1991 because prevention had not been considered a priority. Physician adherence to guidelines and high rates of preventive therapy are essential to achieve cost-effectiveness. The key conclusion was that screening is not cost-effective unless TB cases are prevented.
CRD COMMENTARY - Selection of comparators The rationale for the comparators was clear as the study compared the current practice with what should have been: i.e., an ideal hypothetical scenario based on internationally accepted guidelines, and a situation where the existing Australian guidelines had been adhered to.
Validity of estimate of measure of benefit The estimate of measure of benefit was explicitly stated and was appropriate for the study. The retrospective nature of determining clinical outcomes may be open to some potential biases. The model used was hypothetical and reflected the ideal scenario, which may not be a realistic expectation.
Validity of estimate of costs The study reported estimates of costs thoroughly and in great detail. Quantities and costs of each step of contact tracing were reported separately. However, the authors indicated that their use of estimates for treatment of third generation cases of TB may affect the internal validity of the estimates as they were averaged AN-DRGs which are very broad disease categories. They attempted to address this limitation by weighting the costs according to the frequency of AN-DRGs under which TB cases were classified.
Other issues The authors indicated that the external validity of their study was affected to some extent by the types of screening programmes in different settings and highlighted that the results may not be applicable in different programme settings. Comparisons with other studies were made. The sample appears to have been representative of the patient domain being considered.
Implications of the study The study highlighted the fact that all TB screening programmes can only be cost-effective if TB cases are prevented through thorough adherence to guidelines. The study indicated that case finding as a product of screening was not cost effective in any of the models examined. The problem of translating evidence into practice and the challenge of implementation of guidelines were highlighted. The authors indicated that there is a need for change in ingrained cultures and patterns of practice, which point to the need for education, training and clinical ownership.
Bibliographic details MacIntyre C R, Plant A J, Hendrie D. The cost-effectiveness of evidence-based guidelines and practice for screening and prevention of tuberculosis. Health Economics 2000; 9: 411-421 Indexing Status Subject indexing assigned by NLM MeSH Antitubercular Agents /economics /therapeutic use; Contact Tracing; Cost of Illness; Cost-Benefit Analysis; Drug Costs; Evidence-Based Medicine; Humans; Isoniazid /economics /therapeutic use; Mass Screening /economics /standards; Models, Econometric; New South Wales; Practice Guidelines as Topic; Tuberculosis /diagnosis /drug therapy /economics /prevention & control AccessionNumber 22000008243 Date bibliographic record published 30/04/2001 Date abstract record published 30/04/2001 |
|
|
|