Cost-effectiveness analysis of tuberculosis control policies in Ivanovo Oblast, Russian Federation. Ivanovo Tuberculosis Project Study Group.

Cost-effectiveness analysis of tuberculosis control policies in Ivanovo Oblast, Russian Federation G.B. Migliori,1 A.G. Khomenko,2 V.V. Punga,3 M. Ambrosetti,4 1. Danilova,5 L.N. Ribka,6 M. Grzemska,7 H. Sawert,8 M.C. Raviglione,9 & the Ivanovo Tuberculosis Project Study Group10 Many of the current tuberculosis control programmes in the Russian Federation are based on costly strategies which are underfunded and use long, individualized treatment regimens. This article compares, using a cost- effectiveness analysis, the new WHO strategy implemented in the Ivanovo Oblast (case-finding among symptomatic patients (SCF) and shorter regimens) and the old strategy (active screening of the asymptomatic population (ACF) and longer regimens). The cost per case cured was calculated at different levels of cure rate (45-95%/o) using three scenarios to describe the new WHO strategy (use of WHO-recommended regimens and three options at increasing rates of admission) and a fourth scenario to describe the old strategy (all patients admitted for the whole treatment and longer regimens). The cost per case detected was determined by calculating the following: yield of the new and old strategy (number of examinations necessary to diagnose one case); cost of the diagnostic process; multiplying yield per cost according to the three scenarios describing the new WHO strategy and a fourth scenario describing the old strategy. In the Ivanovo Oblast the cost per case cured, at 85% cure rate level, ranged from US$ 1197 (new strategy, scenario 1 without food) to US$ 6293 (old strategy, scenario 4) the cost per case detected ranged from US$ 1581 (new strategy, scenario 1) to US$ 4000 (old strategy, scenario 4). Significant savings can result from shifting towards the new WHO strategy. Decision-makers and health administrators should be responsible for re-investing the financial and human resources mobilized by the adoption of cost-effective strategies within the TB control programme. Introduction In 1995 the Russian Federation reported over 85 000 new tuberculosis (TB) cases (57.4 cases per 100000 population), a 69% increase over a minimum-ever incidence of 34 cases per 100000 in 1991. With a mortality rate of 14.7 cases per 100000 in 1994, the Russian Federation has the highest TB mortality rate in Europe and 40% of these deaths occur among patients <39 years of age (1-4). The present situa- tion has been attributed to a combination of delayed diagnosis and ineffective treatment (3), largely aris- ing from difficulties in the transition period towards a market economy. Many of the TB programmes currently in place in the former Soviet Union are based on costly strategies (mass screening proce- dures and extensive hospitalization), are under- funded, and use unnecessarily long treatment 1 Consultant Chest Physician and Statistician/WHO Temporary Advisor, Fondazione Salvatore Maugeri, Clinica del Lavoro e della Riabilitazione, Care and Research Institute, Tradate, Italy. Requests for reprints should be sent to Dr Migliori at the following address: Fondazione Salvatore Maugeri, Clinica del Lavoro e della Riabilitazione, Care and Research Institute, Via Roncaccio 16, 21049 - Tradate, Italy. 2 Director, Central Tuberculosis Research Institute, Moscow, Russian Federation. 3 Director, Ivanovo Project, Central Tuberculosis Research Insti- tute, Moscow, Russian Federation. 4 Medical Officer/WHO Temporary Advisor, Fondazione Salvatore Maugeri, Clinica del Lavoro e della Riabilitazione, Care and Research Institute, Tradate, Italy. Reprint No. 5886 Ivanovo Project Coordinator, Ivanovo Tuberculosis Dispensary, Russian Federation. 6 Deputy Director, Central Tuberculosis Research Institute, Moscow, Russian Federation. 7 Medical Officer, Global Tuberculosis Programme, World Health Organization, Geneva, Switzerland. 8 Medical Officer, World Health Organization Country Office, Nonthaburi, Thailand. 9 Medical Officer and Scientist, Global Tuberculosis Programme, World Health Organization, Geneva, Switzerland. '° Dr 1. Danilova, Ms L.V. Filipova, Dr T.A. Grishina, Dr L.S. Ivanova, Prof. A. Khomenko, Prof. V.V. Punga, Dr L.N. Ribka, Dr M.B. Stoyunin, Dr M. Ambrosetti, Dr M. Grzemska, Dr G.B. Migliori, Dr M.C. Raviglione, Dr H. Sawert, and Prof. L. Trnka. Bulletin of the World Health Organization, 1998, 76 (5): 475-483 © World Health Organization 1998 475 G.B. Migliori et al. regimens (2, 3). WHO advocates a strategy of TB control based on rapid case detection, largely through case-finding among symptomatic patients who self-report to health services, and the supervised administration of standardized short-course chemo- therapy. Government committment to the imple- mentation of a national TB programme, the regular supply of drugs, and a monitoring system for pro- gramme supervision and evaluation are essential elements of the WHO policy (5). In the present article the results are reported of an economic analysis of a pilot TB control project based on the WHO strategy currently being imple- mented in Ivanovo Oblast. The cost-effectiveness of two alternative approaches was compared: the new WHO strategy (three different scenarios) based on case-finding among symptomatic patients self- reporting to health services and shorter regimens; and the old strategy based on active screening among asymptomatic persons and use of longer regimens. Both the costs and consequences of health pro- grammes or treatments are examined (6). The four scenarios of two programmes were compared in terms of the following end-points: cost per TB case cured; and cost per TB case detected (6). Methods Setting A pilot project of TB control based on the WHO strategy was implemented in 1995 in the Ivanovo Oblast (population: 1.27 million; economy: industrial (mainly textile) and agricultural), located 300km NE of Moscow. In 1994 the TB notification rate was 44.4 per 100000 and the TB mortality rate, 12.2 per 100000. All patients have free access to all care facilities for diagnosis and treatment ofTB and other diseases. General principles A separate analysis was carried out for each scenario to determine the end-points. Only direct costs were included in the model. The future consequences of a wider application of outpatient care (reduced need for buildings and personnel time) were not evaluated in monetary terms. An exchange rate of US$ 1 = 5000 rubles was applied (7). All costs were adjusted for inflation as of 30 June 1996; the inflation rates applied were derived from the consumer price, based on the chained Laspeyres index, calculated using weights from the previous year and including 288 representative items in metropolitan areas and towns (7). All data used, TB case definitions, and regimen abbreviations were obtained from WHO (8). The role and cost of BCG vaccination and chemoprophylaxis in preventing TB were not included in the analysis. Cost per case cured The regimen categories applied in the model (see Tables 1 and 2) were as follows: category I: new cases of smear-positive pulmonary TB and other newly diagnosed seriously ill patients with severe forms of TB; category II: relapse and failure smear-positive TB patients; category III: new cases of smear- negative pulmonary TB and other newly diagnosed patients with TB not included in category I (7). Four scenarios were used (three for the new strategy and one for the old). In scenario 1, 10% of patients were admitted for the intensive phase and WHO- recommended regimens were used. In scenario 2, all patients were admitted for the intensive phase and the same WHO-recommended regimens were used as in scenario 1. In scenario 3, all patients were admitted for the whole treatment and the same WHO-recommended regimens were used. In sce- nario 4, all patients were admitted for the whole treatment and longer regimens were used. The same intensive phase was applied for the four scenarios (Table 2). In the new strategy, the continuation phase is intermittent when prescribed on an ambula- tory basis (e.g., category I: 4 H3R3) and daily when administered on a hospital basis (e.g., category I: 4 HR). In the old strategy the continuation phase is administered daily (e.g., category I: 7 HR). Since individualized regimens are prescribed in the old strategy, a conservative estimate of longer regimens was applied in the model (Table 2), based on existing recommendations (3). The cost of regimens applied in the model was derived from WHO estimates and Table 1: Distribution of TB cases in Ivanovo Oblast and regimens used in the model, by WHO category, 1996 No. Pulmonary smear-positive 288 Pulmonary smear-negative 380 Extrapulmonary 40 Total 708 Regimena Category 1 377 (53.3)b Category II 38 (5.4) Category III 293 (41.3) a See text for details. b Figures in parentheses are the percentages of cases who re- ceived the given regimen. WHO Bulletin OMS. Vol 76 1998476 Cost-effectiveness analysis of TB control in Ivanovo Oblast Table 2: Cost of TB chemotherapy regimens recommended by WHO, according to 1994 FOB drug prices, by treatment category Total cost Category Strategy/patienta Initial phaseb Cost (US$) Continuation phasec Cost (US$) (US$) A (US$)d Ns/IP 2 SHRZ 20.6 4 HR 15.4 36.0 +7.3 Ns/OP - 4 H3R3 8.08 28.7 0 Os/IP - 7 HR 26.9 47.5 +18.8 11 Ns/IP 2 SHREZ 31.7 5 HRE 24.7 56.4 +10.2 Ns/OP 1 HREZ - 5 H3R3E3 14.5 46.2 - Os/IP - - 9 HRE 44.5 76.2 +26 III Ns/IP 2 HRZ 13.6 2 HR 7.7 21.3 +3.7 Ns/OP - - 2 H3R3 4.04 17.6 - Os/IP - - 6 HR 23.1 36.7 +19.1 a NS = new strategy; Os = old strategy. IP = inpatient; OP = outpatient. b S = streptomycin, E = ethambutol, H = isoniazid, R = rifampicin, Z = pyrazinamide. c H,R, = isoniazid and rifampicin administered n times per week; H,R,E, = isoniazid + infampicin + ethambutol administered n times per week. d Additional cost compared with intermittent, fully ambulatory continuation phase. published data reporting free on board (FOB) prices (9). Because eight categories of TB-dedicated facili- ties are currently operating in the oblast (TB cabinet; TB unit at Central Raion Hospital, with and without beds; TB dispensary, with and without beds; oblast TB hospital, oblast TB dispensary with beds; and TB sanatorium), costing procedures were applied to the various levels of health units and laboratories in the oblast. Costing was based on the official 1995-96 budget of the following health facilities in Ivanovo Oblast: oblast TB dispensary (separately for in- and outpatient departments); new oblast TB dispensary; TB sanatorium no. 1; oblast TB hospital, Ivanovo; and six Raion dispensaries (Tenkovo, Rodniki, Kinesmah, Shuia, Furmanov, and Vichuga). Costs were divided into fixed (building, diagnostic facili- ties, salaries, and overhead) and variable (food, TB and non-TB drugs, and examinations). All fixed and variable costs were calculated per bed-day and per outpatient visit in different level health units to determine their average value and range (Table 3). The results were used to perform a sensitivity analysis (6). The budget for the new oblast TB dispensary, which was completed at the end of 1996, was used to determine the cost of buildings. The equivalent annual cost was calculated, annuitizing the initial capital outlay over a useful life of 30 years (build- ings), 10 years (diagnostic facilities), and 5 years (ve- hicles) (6). The gross salaries of health personnel were derived from the budget of each health facility evaluated. Overhead costs (heating, telephone, elec- tricity, and other examinations and services), calcu- lated from the budget of each health facility, were allocated based on the floor area of the departments/ laboratories concerned (6). Food costs were calcu- lated from the budget of the oblast TB dispensary. Since not all the outpatient health facilities provided food, the analysis was repeated twice including/ omitting food for outpatient care. As discussed above, the costs of TB and non-TB drugs were calcu- lated separately in the model. The cost of examina- tions per bed-day was calculated by multiplying the unit cost of examinations by the number of such examinations per bed-day. The average number of examinations per patient admitted to different health units per bed-day was determined as follows: laboratory tests: 0.1; chest X-ray: 0.08; miniature ra- diography: 0.02; tuberculin test (PPD): 0.01; three direct smears for alcohol-acid-fast bacilli (AAFB): 0.02; and cultures: 0.02. The following examinations were not included in the model to keep the estimate conservative: surgical operations (0.002), endoscopic procedures (0.01), physiotherapy (0.29), and blood transfusions (0.02). The same procedure was used to determine the cost of examinations per outpatient visit. Over a treatment period of 158.4 days (aver- aged for category I, II and III patients) the following examinations were included in the model: two labo- ratory tests (0.03 per outpatient visit), two chest X- rays, three AAFB smears (0.08 per outpatient visit), and three cultures. For each scenario, fixed and variable costs were calculated separately for category, I, II, and III pa- tients. Sensitivity analyses were carried out on the variables when a result was uncertain (6), and, be- cause of the difficulties in determining cure rates for both strategies in Ivanovo Oblast, also on cure rates. All fixed and variable costs (except those of TB drugs) that determined the cost per outpatient visit in the most cost-effective scenarios (1 and 2) were WHO Bulletin OMS. Vol 76 1998 477 G.B. Migliori et al. progressively increased (and those of the less cost- effective scenario, the old strategy, decreased) until a similar cost-effectiveness was obtained at the 85% cure-rate level. The following cure rates (patients cured + treatment completed according to theWHO definition) were applied in the model: 95%, 85%, 75%, 55%, and 45% (5, 10). A final projection was carried out to evaluate the potential savings of scenario 1 and 2 compared with scenario 4 (old strategy) in Ivanovo Oblast. Cost per case detected The analysis compared yield and cost of the new WHO strategy (detection of TB cases among patients with symptoms who attended a health centre, scenarios 1-3) with the old strategy (active screening, i.e. deliberate search for suspects by house-to-house enquires or mass screening in schools and workplaces, scenario 4). The costs of the following diagnostic items, as previously calculated, were included in the model: New strategy = medical examination + chest X-ray + AAFB + culture; Old strategy = miniature radiography with and without medical examinations. The cost per case detected was calculated as follows: - determine the yield of the new and old strategy in terms of number of examinations necessary to diagnose one case; - calculate the cost of the diagnostic process; and - multiply the yield per cost according to the four scenarios (1 = 100% of patients diagnosed by case finding among symptomatic population; 2 = 10% of patients diagnosed by active screening and 90% by case finding among symptomatic persons; 3 = 50% of patients diagnosed by active screening and 50% by case-finding among symp- tomatic population; and 4 = 100% of patients diagnosed by active screening). The yields for the new and old strategies were determined for 1995-96. The analysis was repeated twice, including/omitting the cost of medical exami- nations in scenarios 2-4 in addition to miniature radiography. The potential savings using different scenarios were calculated for Ivanovo Oblast. Sensitivity analysis was performed, increasing the yield of miniature radiography and decreasing the cost of active screening (scenario 4) until the cost per case detected became equal to that of scenario 1 and 2. Results Cost per case cured The average fixed and variable costs per bed-day and per outpatient visit are summarized in Table 3 over the interval range -7.5% and +7.5%. The cost per case cured (Table 4 and Table 5), at an 85% cure rate level, ranged from US$ 1197 (scenario 1 without food) to US$ 6293 (scenario 4). The potential annual savings to be gained by using scenario 1 rather than scenario 4 are US$ 2.8 million and US$ 2.3 million for scenario 2 versus scenario 4. According to the sensitivity analysis, the cost per case cured using scenario 4 is the same as that using scenario 2, decreasing fixed and capital costs per bed-day by 61.5%, as well as of scenario 1, decreasing these same costs by 74.5%. Furthermore, the cost per case cured using scenario 1 and scenario 2 is equal to that of scenario 4, increasing the fixed and capital costs per outpatient visit by 322% and 420.5%, respectively. The cost per case cured using scenario 1 is equal to that of scenario 2, increasing the fixed and capital costs per outpatient visit by 55%. Cost per case detected The unit cost of the examinations used in the model was as follows: AAFB: US$ 4; culture: US$ 6; chest X-ray: US$ 2.8; miniature radiography: US$ 2; and medical examination by a chest physician: US$ 8. The yield of active screening in 1994 was 0.046% in Ivanovo Oblast (263 cases detected using 571832 miniature radiographs) and 0.05% in Ivanovo Oblast dispensary (144 cases detected using 287912 min- iature radiographs). A 0.05% yield was used in the model (2000 miniature radiographs to diagnose 1 Table 3: Fixed and variable costs per bed-day and per outpatient visit, Ivanovo Oblast, 1996 Cost (US$) Bed-day- Bed-day Outpatient Outpatient Fixed costs Buildings/equipment 9.6 3.3 6.3 Personnel 4.7 1.4 3.3 Overhead 0.9 0.4 0.5 Variable costs Food 2.4 2.4 0 (2.4)8 Drugs (non-TB) 1.8 0.3 1.5 Examinations 0.8 0.3 0.5 a For food the costs are shown including (not including) food for outpatient care. WHO Bulletin OMS. Vol 76 1998478 Cost-effectiveness analysis of TB control in Ivanovo Oblast Table 4: Cost per case of tuberculosis cured, by cure rate, in the four treatment scenarios Scenario 1 (US$)a Scenario 2 (US$)a Cure rate Food No food Food No food Scenario 3 (US$) Scenario 4 (US$) 95% (673)b 1454.98 1070.68 2160.53 1 916.18 3397.47 5629.1 85% (602) 1 626.58 1196.97 2415.34 2142.17 3898.17 6292.99 75% (531) 1 884.07 1 357.02 2738.29 2428.60 4306.02 7134.43 55% (389) 2517.22 1 852.38 3737.88 3315.13 5877.88 9738.77 45% (319) 3026.59 2258.86 4558.10 4042.59 7167.7 11 875.81 Total cost 979200 720576 1454034.6 1 289586.6 2286496.8 3788382.7 a Outpatient costs are shown with and without food. b Figures in parentheses are the number of TB patients cured. Table 5: Cost per case cured, by treatment scenario, Ivanovo Oblast, 1996 New strategy Old strategy Scenario 1 a Scenario 2a Scenario 3 Scenario 4 (US$) (US$) (US$) (US$) Cure rate Cost Ab Cost Ab Cost Ab Cost Ab 95% 1454.98 -705.55 2160.53 0 3397.47 +1 236.9 5629.1 +3468.57 85% 1626.58 -788.76 2415.34 0 3898.17 +1 482.8 6293 +3877.66 75% 1 884.07 -854.22 2738.29 0 4306.02 +1 567.7 7134.4 +4396.11 55% 2517.22 -1 220.66 3737.88 0 5877.88 +2140 9738.8 +6000.92 45% 3026.59 -1 531.51 4558.10 0 7167.7 +2609.6 11875.8 +7317.7 a Outpatient costs were calculated including food. bCost difference compared with scenario 2. Table 6: Cost per case detected, by treatment scenario, Ivanovo Oblast, 1996, including and not including the cost of the medical examination in scenarios 2-4 in addition to miniature radiography as a basis for comparison New strategy Old strategy Scenario 1 Scenario 2 Scenario 3 Scenario 4 (US$) (US$) (US$) (US$) Medical examination Cost A Cost A8 Cost A Cost A8 Not included 1 580.8 -1 209.6 1 823.4 -967 2790.4 0 4000 +1 209.9 Included - - 3427.9 -7362.5 10790.5 0 20000 +9209.6 a Cost difference compared with scenario 3. case). The yield of case-finding among asymptomatic persons in 1994-95 in three different health units was 2.08%, 1.54%, and 0.92%, with a total yield of 1.32% (24 cases diagnosed at 1823 visits; 76 visits to diagnose 1 case) being used in the model. The cost per case detected according to the dif- ferent scenarios is shown in Table 6. The potential annual savings with scenario 2 versus scenario 4 are US$ 1.5 million without visits and US$ 11.7 million if the cost of the medical examinations is included. The potential savings with scenario 3 versus scenario 4 are US$ 0.8 million without the cost of medical examinations and US$ 6.5 million if the cost of such examinations is included. Sensitivity analysis indicates that the cost per case detected in scenario 4 without including the cost of medical examination becomes the same as that of scenario 2 if the yield of active screening is increased by 220% (580% with examinations included) and that of scenario 1 if the yield is increased by 260% WHO Bulletin OMS. Vol 76 1998 479 G.B. Migliori et al. without and 1260% with examination. The cost of active screening in scenario 4 with and without medical examinations should be decreased by 55% and 83%, respectively, to become the same as that of scenario 2, and by 60% and 89%, respectively, to become the same as that of scenario 1. Discussion The aim of the study was to compare the new WHO strategy of TB control and the old strategy by means of cost-effectiveness analysis, with the cost per case cured and the cost per case detected at different cure rate levels being the end-points of the analysis. The cost per case cured was determined under four scenarios. The results of our study show that the new strategy (scenarios 1-3) is significantly more cost-effective than the old strategy (scenario 4) at all levels of cure rate. Compared with scenario 2 at a cure rate level of 85%, the unit savings of scenario 1 are US$ 788.8, and the extra costs of scenarios 3 and 4 are US$ 1482.8 and US$ 3877.7, respectively (Table 5). The cost per case detected was also described under four scenarios. The results of the analysis show that the new strategy (scenarios 1-3) is signifi- cantly more cost effective than the old strategy (sce- nario 4). Compared with scenario 3, excluding the cost of the medical examinations, the unit savings with scenario 1 and 2 are, respectively, US$ 1209.6 and US$ 967 (including the cost of medical examina- tions in scenario 2: US$ 7362.5), and the extra costs of the old strategy (scenario 4), US$ 1209.9 (US$ 9209.6 including medical examinations, Table 6). Cost per case cured Scenario 2 corresponds to the present level of imple- mentation of the WHO strategy of TB control in Ivanovo Oblast. The 85% cure rate level, selected as baseline for comparisons, represents the WHO target in the oblast (5, 11). Currently in the oblast, patients are hospitalized for the intensive phase of treatment. Scenario 1, under which there is admis- sion during the intensive phase for the most serious cases (10%), is a reasonable target for the oblast. Scenario 3 (admission of all patients for the duration of treatment and use of WHO-recommended regimens) was selected as intermediate between scenario 2 and the old strategy (scenario 4: all pa- tients admitted for the whole treatment and use of longer regimens). The current policy of case holding is not standardized in the Russian Federation. Most patients are hospitalized, and longer regimens are used; no universal regimens corresponding to those recommended by WHO and International Union against Tuberculosis and Lung Disease (IUATLD) (8, 12) are used, although specialist centres in Mos- cow, St Petersburg, and Novosibirsk have produced their own recommendations (3). For example, in 1992 the Central Tuberculosis Research Institute in Moscow proposed 4-8-month regimens, which cor- respond more closely to WHO/IUATLD-recom- mended regimens, classifying patients according to the disease-related group (DRG) system (13). Other documents recommend that 9-12-month chemo- therapy regimens should be prescribed, but TB spe- cialists are free to modify them in instances of real or perceived drug side-effects, drug shortage, and a widespread belief that rifampicin and pyrazinamide have an additional hepatotoxicity if combined (3). Thus, the duration of treatment and the related costs in scenario 4 are conservative, as are the estimated costs of examinations performed during hospital ad- mission. Adjunct therapies and surgical procedures are still considered important in the Russian Federa- tion (3); for example, in 1993, a total of 5-10% of new TB patients underwent surgery (usually seg- mental resection or lobectomy) when a cavity per- sisted after 6 months of treatment (3). In our model the surgery-related costs were not included nor were the costs for physiotherapy procedures (intrapleural, intrabronchial and intracavitarian administration of drugs, galvanization and electromagnetic therapies, etc) and those for autotransfusion of irradiated blood (3). In terms of fixed and variable costs (Table 3), the main determinants of the difference between bed-day and outpatient visit costs are the fixed costs for discounting buildings and equipment and person- nel salaries. It should be noted that, since the costs of new buildings were allocated using floor area, over- head costs were allocated using the same method and the gross salary of personnel for in- and out- patient care was the same, major biases should not affect our estimates. Furthermore, the low overhead bed-day and outpatient costs are due to the low costs of electricity, telephones, and heating in the Russian Federation. The cost of drugs was determined from accurate budgets for in- and outpatient services. The difference in the cost of non-TB drugs between bed-day and outpatient approaches is due to additional treatments (immunostimulants, anti- oxidants, liver tonics, etc.) still widely prescribed in the Russian Federation, particularly for hospitalized patients (3). As mentioned above, because of the non-stand- ardized approach to TB case-management (2,3), it is difficult to estimate the effectiveness of the old strat- egy; this is why we tested our hypothesis over a wide range of cure rates for all scenarios. The results of WHO Bulletin OMS. Vol 76 1998480 Cost-effectiveness analysis of TB control in Ivanovo Oblast the study are consistent, even in view of a higher level of cure rate among hospitalized patients (e.g. 95%) compared with the 85% cure rate in scenarios 1 and 2. Sensitivity analysis revealed that the hospital costs per bed-day should be decreased by >50% or the costs of outpatient visits increased by >300% to produce results in scenarios 1 and 2 similar to those in scenario 4. Although the consequences of future reorganization of health services emphasizing outpa- tient care were not evaluated in economic terms, additional benefits can be expected, including avail- ability of new spaces in existing buildings and per- sonnel time. Both these resources can be redirected towards TB control activities that need support (e.g. health education, supervision, training) or other health activities (e.g. prevention, treatment, and re- habilitation of chest diseases). It should be noted, however, that well-equipped TB units and special- ized staff are always necessary (e.g., scenario 1) to treat severe cases and to maintain expertise in control efforts when the incidence and prevalence of TB decline (14). Furthermore, even if we have not included indirect and intangible costs in our analysis, additional benefits resulting from lower hospitaliza- tion rates can be expected for society as a whole as well as for TB patients and their families. This consideration is strengthened by the increasing demand for outpatient TB care recently described in the Russian Federation (3). To keep the model simple, we have not estimated the potential economic impact of new TB cases originating from infectious patients attending outpatient serv- ices during the ambulatory intensive phase (new strategy, scenario 1), since the period when they are infectious during treatment is very short (on average 2 weeks) and the probability of infecting others is low. A rough estimate of the potential savings that can be obtained for the Russian Federation as a whole can be made by extrapolating the data for Ivanovo Oblast: the hypothetic annual savings result- ing from the adoption of the new strategy probably exceed US$ 290 million (data not shown). Cost per case detected Scenario 3 describes the present case-finding policy in Ivanovo Oblast, which is consistent with that used throughout the Russian Federation (2). Since TB and general health services are not integrated in the Russian Federation, due to low suspicion of clinical TB among general physicians, more reliance is still placed on active screening services (2). All adults should undergo miniature radiography at 2-year in- tervals (1 year if they are considered to be at high risk), and a normal miniature radiogram is still a precondition for employment. The coverage of active screening at the national level has fallen to 50% in the general population, with higher percent- ages in occupational risk groups (teachers, medical workers) (2). The yield of active screening has been estimated to be in the range 0.04-0.1% depending on the incidence ofTB (3). In our model we used a yield of 0.05%, based on local performance of miniature radiography. Sensitivity analysis revealed that the results are not modified by a significant increase in miniature radiography yield or a significant decrease in active screening costs. The analysis including the cost of medical examinations in scenarios with active screening produced significantly greater savings. This estimate is higher than the real one because not all screened patients undergo medical examination. The true situation is probably between the two proposed, with all patients with positive miniature radiograms and suspects undergoing medical exami- nation and additional tests. To keep the model simple, we did not examine the effects of adopting different case-finding policies on cure rates. Mass screening, though expensive, might detect more cases, with the potential to decrease the cost per case cured. The progressive decline observed in active screening coverage in Russia, particularly among high-risk groups (ex- prisoners, the unemployed, alcoholics, etc), indicates that the previous comprehensive mass screening system cannot be sustained under the present eco- nomic situation (2). Active screening, as exemplified by scenario 2, can be performed cost-effectively among groups where the prevalence of TB is significantly higher than that among the general population (14). By extrapolating the data for Ivanovo Oblast, we estimated that, for the Russian Federation as a whole, the annual savings resulting from the adoption of the new strategy would probably exceed US$ 106 million. Conclusions Under the present programme conditions, the Russian Federation might be unable to respond to any new challenges to TB control, such as those presented by human immunodeficiency virus (HIV) infection, which might spread in the next few years, and multidrug-resistant TB, which is probably already increasing because of the shortage of some anti-TB drugs. The Russian programme dedicated to TB control needs to be provided with appropriate information, resources, drugs, and equipment. A new approach to TB control in Russia, based on the WHO strategy, may indeed have a profound impact, WHO Bulletin OMS. Vol 76 1998 481 G.B. Migliori et al. since it will prevent deaths and disabilities among the most productive age groups and make more effective use of scarce resources by reducing the length of hospitalization, number of beds, and other costly interventions. The expected benefit will result from rapid reduction in the pool of previously treated smear-positive cases and more rapid treat- ment and cure of new smear-positive cases. Ulti- mately there will be a permanent reduction in the incidence of TB. The results of our study show that substantial savings could be made in the Russian Federation by shifting towards scenario 1 (lowest cost per case cured and detected). Under the present situation (different funding for general and TB services) it is unlikely that the potential resources mobilized by cutting active screening and unnecessary hospitaliza- tion can be redirected to essential TB control activi- ties that need financial support (e.g. purchasing standardized regimens of proven bioavailability, increasing awareness about TB in general health services, and improving coverage and quality of smear microscopy). Further integration between TB and general health services and the provision of a common budget for health are necessary in the Russian Federation and in the republics of the former Soviet Union. Policians and health administrators should take the responsibility for re-investing the monetary resources mobilized through the adoption of cost- effective strategies within the TB control pro- gramme or within the health system if they exceed TB control needs. Acknowledgements We thank Dr J. Kutzin (WHO, Geneva), and Dr A. Spanevello and Dr M. Neri (Department of Pneumo- nology, Tradate, Italy) for their useful comments on the manuscript. The views expressed in this article are those of the authors and do not necessarily reflect those of WHO. Resume Analyse du rapport cout-efficacite des mesures de lutte contre la tuberculose dans la region d'lvanovo (Federation de Russie) La plupart des programmes actuels de lutte contre la tuberculose dans l'ex-Union Sovietique, y compris la Federation de Russie, reposent sur des strategies coOteuses qui manquent de ressources et ont recours a des sch6mas therapeutiques individualises de longue duree. La pr6sente etude se livre a une comparaison, au moyen d'une analyse des couts et de l'efficacite, entre la nouvelle strat6gie de l'OMS instauree dans la r6gion (Oblast d'lvanovo (d6pistage des cas symptomatiques et schemas therapeutiques de courte duree) et l'ancienne strategie (depistage actif de la population asymptomatique et traitements de longue duree). Le coGt par cas a ete calcule pour des taux variables de guerison (45-95%) en utilisant trois scenarios pour decrire la nouvelle strategie de I'OMS (utilisation des sch6mas th6rapeutiques recommandes par l'OMS et trois options avec des niveaux croissants d'admission) et un quatrieme scenario concernant l'ancienne strategie (tous les patients sont admis pour l'ensemble du traitement et des sch6mas therapeutiques plus longs). Le coOt par cas depiste a ete determine en calculant les elements suivants: le rendement des deux strategies (nombre d'examens n6cessaires pour diagnostiquer un cas); le coOt de la methode de diagnostic; la multiplication du rendement par le coOt selon les trois scenarios de la nouvelle strategie (proportion croissante de malades diagnostiqu6s lors du depistage des cas asympto- matiques) et le quatrieme pour I'ancienne (tous les malades diagnostiques par depistage actif de la population asymptomatique). A Ivanovo Oblast le coOt par cas gueri, avec un taux de guerison de 85%, s'est situe entre US$1197 (nouvelle strategie, scenario 1 sans nourriture) et US$6293 (ancienne strategie, scenario 4). Le coOt par cas depiste s'est situe entre US$1581 (nouvelle strategie, scenario 1) et US $4000 (ancienne strategie, scenario 4). Dans les conditions actuelles du programme, la F6d6ration de Russie pourrait ne plus etre capable de r6pondre a de nouveaux problemes dans la lutte antituberculeuse, comme l'infection par le virus de l'immunod6ficience humaine (VIH), qui pourrait se propager dans les prochaines annees, et la tuberculose resistante a la polychimiotherapie qui sans doute se propage d6ja en raison des penuries de certains medicaments antituberculeux. Le Pro- gramme russe de lutte antituberculeuse a besoin d'acceder a une information, des ressources, des m6dicaments et du mat6riel en quantites suffisantes. Une nouvelle m6thode de lutte contre la tuberculose en Fed6ration de Russie, reposant sur la strategie OMS, pourrait de fait avoir un effet profond car elle eviterait les deces et les incapacit6s dans les groupes d'ages les plus productifs et, dans le meme temps, elle permettrait une utilisation plus efficace des ressources limitees en reduisant la WHO Bulletin OMS. Vol 76 1998482 Cost-effectiveness analysis of TB control in Ivanovo Oblast dur6e des hospitalisations, le nombre de lits et d'autres interventions coOteuses. Les ben6fices attendus proviendraient de la reduction rapide du nombre de cas a frottis positifs deja trait6s auparavant ainsi qu'un traitement et une guerison plus rapides des nouveaux cas a frottis positifs. L'effet ultime serait une reduction permanente de l'incidence de la tuberculose. Les r6sultats de notre 6tude montrent que des economies considerables pourraient etre r6alis6es en passant au sc6nario 1 (coOt le plus faible par cas gueri et depist6). Dans la situation actuelle (financement diff6rent pour les services generaux et ceux consacres a la tuberculose), il est improbable que les ressources potentielles d6gag6es par I'arret du depistage actif et des hospitalisations inutiles pourraient etre r6orient6es vers des activit6s essentielles de lutte antituberculeuse n6cessitant un appui financier (par exemple l'achat de sch6mas th6rapeutiques standardises ayant une biodisponibilit6 6tablie, la sensibilisation des serv- ices de sante g6n6raux a la tuberculose, I'am6lioration de la couverture et de la qualite de l'examen microscopique des frottis). La F6d6ration de Russie et les republiques de l'ex-Union Sovi6tique ont besoin d'int6grer davantage la tuberculose dans les services de sant6 g6neraux et de fixer un budget commun pour ces activit6s. Les politiciens et les administrateurs de la sante doivent prendre la responsabilit6 de reinvestir les ressources financieres d6gag6es par I'adoption de strategies rentables dans les programmes de lutte antituberculeuse ou dans le systeme de sant6 si elles d6passent les besoins de la lutte contre la tuberculose. References 1. Raviglione MC et al. Tuberculosis trends in eastern Europe and the former USSR. Tubercle and lung disease, 1994, 75: 400-416. 2. Drobniewski F et al. Tuberculosis in Siberia: 1. An epidemiological and microbiological assessment. Tubercle and lung disease, 1996, 77: 199-206. 3. Drobniewski F et al. Tuberculosis in Siberia: 2. Diag- nosis, chemoprophylaxis and treatment. Tubercle and lung disease, 1996, 77: 199-206. 4. Tuberculosis - a global emergency: case notification update. Unpublished document WHO/TB/96.197 (available upon request from: Global Tuberculosis Programme, World Health Organization, 1211 Geneva 27, Switzerland). 5. Framework for effective tuberculosis control. Unpub- lished document WHO/TB/94.179 (available upon request from: Global Tuberculosis Programme, World Health Organization, 1211 Geneva 27, Switzerland). 6. Drummond MF, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes. Oxford, Oxford University Press, 1989. 7. International financial statistics. Washington, DC, International Monetary Fund, 1996: 498-501. 8. Treatment of tuberculosis. Guidelines for national programmes, 2nd edit. Unpublished document WHO/TB/97.220 (available upon request from: Global Tuberculosis Programme, World Health Organization, 1211 Geneva 27, Switzerland). 9. Chaulet P. The supply of antituberculosis drugs: price evolution. Tubercle and lung disease, 1995, 76: 261- 263. 10. Tuberculosis surveillance and monitoring. Unpub- lished document WHO/TB/91 .163 (available upon request from: Global Tuberculosis Programme, World Health Organization, 1211 Geneva 27, Switzerland). 11. Raviglione MC, Snider DE, Kochi A. Global epi- demiology of tuberculosis. Morbidity and mortality of a world-wide epidemic. Journal of the American Medical Association, 1995, 273: 220-226. 12. Committee on Treatment, International Union against Tuberculosis and Lung Disease. Antituberculosis regimens of chemotherapy. Bulletin of the International Union against Tuberculosis and Lung Disease, 1988, 63: 60-64. 13. Ministry of Health, Russian Federation. [Edict No. 324]. Moscow, 22.11.1995 (in Russian). 14. Clancy L et al. Tuberculosis elimination in the countries of Europe and other industrialised countries. European respiratory journal, 1991, 4: 1288-1295. WHO Bulletin OMS. Vol 76 1998 483