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.

Screening with a tuberculin test and chemoprophylaxis to control tuberculosis (TB) in high-risk positive reactor patients was compared with current standard care.

Screening and secondary prevention.

Cost-effectiveness and cost-utility analysis.

The study population consisted of elderly patients who were newly admitted to an institution. At admission, 27% had tuberculin reactors. Of these, 3% had old healed fibrotic lesions and 9% had medical risk factors.

The setting was an institution. The economic study was carried out in Montreal, Canada.

The references for the clinical model parameters dated from 1956 to 1995. The price year was 1992.

The effectiveness data were gathered from a non-systematic review of completed studies and the authors' assumptions.

A Markov model was developed to trace the health status of a cohort of patients over 15 years following their admission to an institution. No reason was given for why the disutility of treating false positives was not modelled.

The outcomes assessed in the review were:

the transition probabilities for specific health states and the diagnosis probability given TB;

the probabilities of fatal TB in diagnosed and undiagnosed cases;

the efficacy of isoniazid (INH);

the probability of INH discontinuation;

the annual conversion rate;

the residual probability of TB after INH and the year INH was taken; and

the probabilities of side effects other than hepatitis causing INH discontinuation.

Not stated.

Not stated.

Not stated.

Not stated.

Approximately 23 studies were included in the review.

Not reported.

Not stated.

The following results are presented as point estimates with a sensitivity analysis range given in brackets.

The diagnosis probability given TB was 25% (10 - 75).

The probability of fatal TB was 15% in diagnosed and treated cases, and 15% in undiagnosed cases (15 - 30).

The efficacy of INH in preventing TB was 65% (INH 6 months) or 75% (INH 12 months).

The probability of INH discontinuation was 8.6%. The probability of death from INH hepatitis was 1.9% (0.7 - 3.0).

The annual conversion rate in institutionalised patients with negative tuberculin tests was 0.6% (0.6 - 1.2) with no INH and 0.09% (0.09 - 0.18) with INH.

The probability of tuberculosis the year INH was taken was 0.37% (12-month course) and 0.28% (6-month course).

The probability of side effects other than hepatitis causing INH discontinuation was 6.0% (12-month course) and 4.6% (6-month course).

The measures of benefits were the life-years (LY) and quality-adjusted life-years (QALY). The authors adopted the quality of life values used by Sackett and Torrance (see Other Publications of Related Interest). The measures of benefit were discounted at a rate of 5% in the base-case.

The quantity/cost boundary adopted was that of the health care system. The direct costs were for tuberculin tests, INH chemoprophylaxis, INH complications and TB-related events. The cost of tuberculin tests included nursing time, supplies and transportation. The cost of INH complications included medical examinations, hospitalisation, laboratory tests and nursing time. The cost of TB-related events included treatment, hospitalisation, investigation and follow-up. The costs were discounted at a rate of 5%. The quantities were estimated from actual data. The cost estimates were derived from the Verdun General Hospital, the 1992 health ministry salary scale and Quebec physicians' fee scale. The unit costs were reported, whereas the quantities were not. The price year was 1992.

No statistical analysis of the costs was carried out.

No indirect costs were included in the analysis.

Canadian dollars (Can$).

A series of one-way sensitivity analyses was carried out to assess the effect of changes in baseline values on the results. The variables investigated were the conversion rates, discount rates, diagnosis rates, distributions of risk factors and the probabilities of events. A multivariate sensitivity analysis with extreme range values was also performed to produce a best-case and a worst-case combination of variables for screening (proportion of tuberculin reactors and risk factors for TB).

Screening yielded an additional 0.0049 LY and 0.0061 QALY per admission. This translated into 1.13 LY per case avoided and 2.78 LY per TB-related death avoided.

The annual costs per 1,000 institutionalised patients were reported.

The annual cost per 1,000 institutionalised patients was Can$6,100 with screening and Can$3,204 without screening.

The incremental annual cost of screening was Can$2,896.

The costs and benefits were combined by calculating an incremental cost-effectiveness ratio (additional cost required per additional LY or QALY gained) (ICER).

The incremental cost of screening was Can$3,437 per LY saved and Can$2,746 per QALY gained.

Three parameters in the sensitivity analysis had significant effects on the ICERs.

When no nosocomial transmission occurred (zero conversion rate), screening had smaller but still positive effects. The incremental cost also increased. The change in the ICER was not reported.

When the probability of diagnosing TB was increased from 25 to 75%, the gain in LY decreased from 0.0049 to 0.0029 per admission. The ICER of screening was Can$201 per LY gained.

When the LY values were discounted at rates of up to 15% and the discount rate for the costs was kept at 5%, the resulting ICER was Can$6,924 per LY gained.

When variables producing the worst-case scenario for screening were combined, this yielded positive health effects at low costs. The ICER was Can$3,784 per QALY gained.

Screening plus chemoprophylaxis for high-risk reactors was more cost-effective than case-finding approaches for controlling tuberculosis (TB).

The reason for the choice of the comparator was clear. The comparator was chosen because it represented the current standard practice of care to control TB in the authors' setting. You should consider whether this is a widely used technology in your own setting.

The authors did not state that a systematic review of the literature had been undertaken. The authors used data from the available studies selectively. Given the consequent uncertainty surrounding the parameter estimates, the reader may wish to appraise their applicability to his/her own setting.

The estimation of benefits was modelled. The decision-analysis model used to derive the LY saved was appropriate. Quality of life estimates were adapted from utilities perceived by the general population. Disutilities from treating false positives were not accounted for.

The authors reported that the costs were estimated from the perspective of the health care system (equivalent to the societal perspective in Canada). Although some relevant costs were omitted from the analysis (indirect costs, out-of-pocket costs for patients and families), these were unlikely to affect the authors' conclusions given that all of the patients were institutionalised. The costs and the quantities were not reported separately. No sensitivity analyses of the quantities and prices were carried out. This may limit the interpretation of the study findings.

The authors made appropriate comparisons of their findings with those from other studies. The issue of generalisability to other settings was not specifically addressed. The authors did not present their results selectively. The authors reported further limitations to their study, for instance, the difficulty of accurately reflecting nosocomial transmission.

Screening with the tuberculin skin test plus INH chemoprophylaxis was an attractive approach. TB is preventable in institutions and at a quite modest cost. The study also demonstrated that INH chemoprophylaxis can be offered to a wider high-risk population.

None stated.

Rose DN, Schechter CB, Fahs MC, et al. Tuberculosis prevention: cost-effectiveness analysis of isoniazid chemoprophylaxis. American Journal of Preventive Medicine 1998;4:102-9.

Fitzgerald JM, Gafni A. A cost-effectiveness analysis of the routine use of isoniazid prophylaxis in patients with a positive Mantoux skin test. American Review of Respiratory Disease 1990;142:848-53.

Sackett DL, Torrance GW. The utility of different health states as perceived by the general public. Journal of Chronic Diseases 1978;31:697-704.