Comparison of Cost-Effectiveness of Tuberculosis ... - ATS Journals

Comparison of Cost-Effectiveness of Tuberculosis Screening of Close Contacts and Foreign-Born Populations KABERI DASGUPTA, KEVIN SCHWARTZMAN, ROBERT MARCHAND, TERRY NAN TENNENBAUM, PAUL BRASSARD, and DICK MENZIES Respiratory Epidemiology Unit, McGill University; Montreal Chest Institute, Direction de la Santé Publique, Régie Régionale de la Santé et des Services Sociaux de Montreal-Centre, Montreal; and Centre Hospitalier Hôtel-Dieu d’Amos, Amos, Quebec, Canada

Although tuberculosis (TB) screening of immigrants has been conducted for over 50 yr in many industrialized countries, its costeffectiveness has never been evaluated. We prospectively compared the yield and cost-effectiveness of two immigrant TB screening programs, using close-contact investigation and passive case detection. Study subjects included all immigration applicants undergoing radiographic screening, already arrived immigrants requiring surveillance for inactive TB, and close contacts of active cases resident in Montreal, Quebec, Canada, who were referred from June 1996 to June 1997 to the Montreal Chest Institute (MCI), a referral center specializing in respiratory diseases. For all subjects seen, demographic data, investigations, diagnoses, and therapy were abstracted from administrative data bases and medical charts. Estimated costs of detecting and treating each prevalent active case and preventing future active cases, based on federal and provincial health reimbursement schedules, were compared with the costs for passively diagnosed cases of active TB. Over a period of 1 yr, the three programs detected 27 cases of prevalent active TB and prevented 14 future cases. As compared with passive case detection, close-contact investigation resulted in net savings of $815 for each prevalent active case detected and treated and of $2,186 for each future active case prevented. The incremental cost to treat each case of prevalent active TB was $39,409 for applicant screening and $24,225 for surveillance, and the cost of preventing each case was $33,275 for applicants and $65,126 for surveillance. Close-contact investigation was highly cost effective and resulted in net savings. Immigrant applicant screening and surveillance programs had a significant impact but were much less cost effective, in large part because of substantial operational problems.

For centuries, migration has played a key role in the epidemiology of tuberculosis (TB). During the 18th and 19th centuries, the prevalence and incidence of TB in northern Europe were high. Migration of northern Europeans to Africa, Asia, and the Americas introduced TB into highly susceptible populations, resulting in TB epidemics in those populations (1). Over the past 50 yr, the previous population migration has reversed (1). Although the burden of TB has fallen dramatically in industrialized countries, it has remained high in developing countries in Africa, Asia, and the Americas, (1). As a result, most adults who emigrate from developing countries carry la-

tent tuberculosis infection (LTBI) at the time of immigration (2) and so are at increased risk of reactivation of the disease for many years after their arrival (3–5). In industrialized countries, active TB among the foreignborn is important for three reasons: (1) As a result of a declining incidence of TB in native-born persons and the large number of immigrants from countries in which TB is endemic, the proportion of cases of TB among the foreign-born has increased over the past 15 yr, from 22% to 39% in the United States (5, 6) and from 30% to 62% in Canada (7). The absolute number of cases has also increased, but less dramatically. (2) In 1991, total expenditures for TB exceeded $700 million in the United States (8), of which approximately $195 million was spent on provision of care to the foreign-born. As the number and proportion of TB cases among the foreign-born increases, it can be anticipated that expenditures for their care will also increase. (3) TB infection may be transmitted to the native-born population in industrialized countries. This transmission may account for anywhere from 2% (9) to as much as 17% of all cases in the native-born population (10). (4) Given the high risk of active TB, which persists for more than 20 yr after immigration (3, 11), and in view of current immigration trends, the objective of TB elimination in certain industrialized countries will never be achieved unless substantially greater efforts are made to screen for and prevent TB in their immigrant populations. Because of these concerns, many industrialized countries have implemented TB screening programs for immigrants. These may be conducted overseas, as for applicants to the United States (12), or at ports of entry such as at Heathrow Airport in London, UK (13), or after arrival, as in Switzerland (14), or at both of the latter sites, as in Canada (15). In the study reported here, the efficiency and cost-effectiveness of immigrant applicant screening and of surveillance of newly arrived immigrants were compared with investigation of close contacts active cases of TB, and all three methods were compared with a policy of passive case detection.

METHODS Setting and Study Cohorts

(Received in original form January 28, 2000 and in revised form July 14, 2000) Supported by the Fonds de la Recherche en Santé du Québec. Dr. Menzies was the recipient of a medical scientist award from the Medical Research Council of Canada. Dr. Kevin Schwartzman was the recipient of a Chercheur Boursier Clinicien award from the Fonds de Recherche en Santé du Québec. Correspondence and requests for reprints should be addressed to Dr. Dick Menzies, Respiratory Epidemiology Unit, McGill University, 1110 Pine Avenue West, Room 103, Montreal, QC, H3A 1A3 Canada. E-mail: [email protected]. ca Am J Respir Crit Care Med Vol 162. pp 2079–2086, 2000 Internet address: www.atsjournals.org

The study compared two TB screening programs among foreign-born populations with screening of close contacts of active cases of TB. Some administrative activities, most of the screening activities, and all of the clinical activities in the study were conducted at the TB clinic of the Montreal Chest Institute (MCI), a university hospital specializing in chest diseases. Some of the screening and administrative activities were conducted at the infectious diseases unit of the public health department responsible for the island of Montreal. The health division of Citizenship and Immigration Canada was responsible for administrative aspects of the screening of applicants for permanent residence in Canada. The study was approved by an ethics committee of the MCI Research Centre.

2080

AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE

Since shortly after the second World War, all applicants for permanent residence in Canada, as well as individuals coming to Canada on work or study visas for longer than 6 mo, have undergone medical evaluation, including a chest radiograph done primarily for screening for active TB. If radiographic abnormalities are detected, the affected individual is referred, usually to a chest specialist, for further evaluation, including additional radiographic, microbiologic (e.g., sputum acid-fast bacilli), and tuberculin tests when judged appropriate. Medical evaluation is performed wherever the application is made, in the applicant’s country of origin or in Canada (15). Since mid-1995, all immigrants from within the greater Montreal area who have applied for permanent residence have undergone screening chest radiography at the MCI. Those with significant abnormalities have automatically undergone further evaluation by chest specialists. When they arrive in Canada, immigrants or long-term visa-holders in whom inactive TB is detected (equivalent to the U.S. Immigration and Naturalization Service classification of “B2” [12]) during the medical evaluation done as a part of preimmigration screening are referred by Citizenship and Immigration Canada to provincial health authorities for medical surveillance (15). Since 1994, the Montreal Public Health Department is notified of all immigrants arriving in Quebec who require surveillance for inactive TB, and it refers those who are resident on the island of Montreal to the MCI. In the present study, close-contact identification and tuberculin screening were performed by staff members at the MCI for active cases of TB diagnosed there and by staff members of the Montreal Public Health Department for contacts of active cases diagnosed at other Montreal hospitals. Positive reactors (⬎ 4-mm induration in response to 5 tuberculin units of purified protein derivative-T [Connaught Laboratories, Toronto, ON, Canada]) were referred for chest radiography, chest specialist evaluation, and all further management to the MCI.

VOL 162

2000

The three study cohorts consisted of all immigration applicants undergoing radiographic screening, all newly arrived immigrants with inactive TB requiring surveillance, and all close contacts of patients with active contagious TB who were referred to the MCI between June 1, 1996 and May 30, 1997.

Evaluation of Costs and Outcomes Administrative data bases at the MCI and Montreal Public Health Department were used to identify members of the cohorts. Hospital records of all patients seen at the MCI were reviewed to abstract demographic data and data for clinic visits, chest radiography, tuberculin tests, microbiologic tests, other investigations, final diagnoses, disposition, follow-up, and treatment. The initial evaluation was considered to end with establishment of a diagnosis and management plan. Follow-up included all hospitalizations and activities associated with therapy for TB disease or infection. Major investigations and duration of hospitalization were recorded for hospitalized patients. Compliance measures included pill counts, punctuality of attendance, and urine testing for isoniazid (INH) metabolites (INH Test Strips; Difco Laboratories, Detroit, MI). Costs for screening chest radiography, clinic visits, investigations, and hospitalization were based on reimbursement paid by the federal government of Canada for refugees’ health services; these costs were established in 1996 on the basis of actual costs at the MCI. Costs for drugs, as well as pharmacists’ and physicians’ fees, were based on the fee schedule of the Quebec provincial government. Costs for administrative activities related to the study cohorts were determined as follows: Interviews were held with the individuals responsible for the respective administrative programs at the MCI, Montreal Public Health Department, and Citizenship and Immigration Canada to identify all personnel involved and their salaries, tasks, and time spent on the

TABLE 1 SUMMARY OF ESTIMATED COSTS AND OUTCOMES FOR DETECTING CASES OF TUBERCULOSIS Item Costs for passively diagnosed cases Hospitalized cases, total cost Initial workup cost Hospitalization costs -Hospital stay -Charges -Outpatient therapy cost Nonhospitalized cases, total cost Total cost per case % Hospitalized Hospitalized portion (14,274 ⫻ 0.76) Nonhospitalized portion (1,006 ⫻ 0.24) Total average cost per passively diagnosed case Costs for contacts Investigation, per index case Management of tuberculin-positive contact Outcomes Active disease incidence TB-infected, no other risk factors (baseline) Granuloma Fibronodular disease (inactive TB) Contact Infected within prior 2 yr Infected more than 2 yr earlier Risk reduction from treatment of latent TB infection for: ⬍ 6 mo and/or poor compliance 6–8 mo with good compliance 9–10 mo with good compliance 11–12 mo with good compliance

Study Value*

Source Used for Study

$14,274 $391

Observed

19 days $13,141 $742 $1,006

(16)† MSSS‡ Observed Observed

76% $10,848 $242 $11,090

(16)†

19d (17), 19.9d (8), 20d (18, 19) $30,406 (8), $31,772 (19), $33,328 (18) $731 (19), $3,440 (8)

79% (8), 86% (17), ⬎ 100% (18)

$28,240 (8), $51,740 (19), $78,895 (18)

$694 $534

Observed Observed

0.1% 0.2% 0.6%

(27) (24, 29) (25)

2.5% 0.1%

(38) (27)

0% 69% 90% 93%

Values from Other Studies (Reference)§

$318 (8) $316 (31)

0.05% (24), 0.2% (25) 1.9% (20)

(32) (38, 39) (32)

Definition of abbreviations: MSSS ⫽ Ministere de Sante et de Services Sociaux du Quebec. * Study value: all dollar values are in $CAN (0.67 $US). † Results of a survey of all adults with active treated in Montreal-area hospitals other than the Montreal Chest Institute in 1996 and 1997 [16]. ‡ Average per diem cost for acute care inpatient beds in the Montreal region (MSSS); observed ⫽ actual costs measured in study. § Values from other studies (all conducted in the United States): costs converted from $US to $CAN.

Dasgupta, Schwartzman, Marchand, et al.: Cost-Effectiveness of Three TB Screening Programs screening programs. Overhead costs for all involved personnel included costs for their office space, such as rental, heating, and insurance and security costs; costs of equipment and supplies used for the programs such as telephone, fax and computer costs; and general administration costs; such as salaries of senior management personnel prorated to the time the senior managers spent on such programs. The total administrative costs attributable to each program were based on the salaries, benefits, and overhead costs of all involved personnel, prorated according to the proportion of time they spent on administration of the programs. The cost for diagnosis and treatment of a passively diagnosed case of TB was estimated from information in Table 1. The proportion of subjects hospitalized, physician visits, major investigations, and median stay were taken from a survey of adults with active pulmonary TB diagnosed in 1996 and 1997 in Montreal hospitals other than the MCI (16) and were similar to the corresponding data in other studies (8, 17–19). The costs of outpatient management were taken from observed costs in the present study. All costs are expressed in Canadian dollars (0.67 $US).

Outcomes Active cases detected. Active TB was defined as the presence of cultures positive for Mycobacterium tuberculosis or radiographic improvement after at least 2 mo of therapy for active disease. This was judged by a board-certified respirologist who was unaware of the clinical diagnosis and the timing of radiographs. Active cases completely treated. Four months of therapy (2 HRZE/ 2 HR) was considered adequate for patients with culture negative disease, 6 mo of therapy (2 HRZE/4 HR) was considered adequate for patients with culture-positive, drug sensitive disease, and 18 mo of appropriate therapy was considered adequate for cases resistant to INH and rifampin. Patients were judged to have been compliant with therapy if they were less than 1 wk late for all clinic appointments, pill counts indicated that more than 80% of pills had been taken, and more than 80% of urine tests for INH metabolites were positive at the time of clinic visits (20). Cases prevented. Based on information summarized in Table 1, and the observed prescription, compliance with and completion of therapy in the three study cohorts, the expected number of cases of active TB over the next 20 yr was estimated by using a simple Markov model with 3% discounting, as shown in Figure 1 (Data version 3.0 for Windows; TreeAge Software, Williamstown, MA). The difference between this estimate and the number of cases that would have been expected in the same populations had there not been any screening or prevention programs was taken as representing the number of cases prevented.

2081

nosed patients (observed). The cost per extra infected contact, estimated from observed costs, plus all contact investigation costs for active cases prevented, were added to the cost per passively diagnosed case, and the cost-effectiveness of each program recalculated. Cost-effectiveness was estimated for subgroups of patients, such as applicants from countries with a high incidence of TB or immigrants with prescribed therapy for LTBI. Cost-effectiveness was also recalculated after varying the future risk of active TB in persons with TB infection (and no other risk factors) from 0.05% (24) to 0.2% (25) and 0.3%, or after varying the cost per passively diagnosed case from $5,000 to $20,000.

RESULTS Outcomes

As shown in Table 2, of 12,898 applicants screened radiographically, 722 (5.6%) were felt to have potentially significant lesions and were referred for further evaluation, as were 119 of 200 close contacts of active cases of TB because they had a positive tuberculin test. Among members of these two cohorts referred for chest specialist evaluation, more than 90% were seen in person. In contrast, only 64% of the surveillance cohort presented themselves, because the mailing address was missing for 10%, and for another 13%, either the mailing address was incorrect or the subject had already moved. These problems were compounded by delayed notification, with an average of 61 d between arrival in Canada and receipt of notification by the Montreal Public Health Department. In the three study cohorts, three smear-positive; 18 smearnegative, culture-positive; and six clinical cases of active TB were detected. All culture-confirmed cases among applicants and contacts were fully drug-sensitive. Of subjects with active TB, none of the applicants and only one contact were hospitalized, the latter for 5 d. On the other hand, of the four members of the surveillance cohort with active TB, two had multidrug

Data Analysis The cost per prevalent active case detected included administrative, screening, and evaluation costs, whereas the cost per prevalent active case treated included these same costs plus costs for treatment of TB disease. The total cost per future case prevented included all program costs. The marginal cost of treatment of LTBI included only those additional costs directly attributable to therapy for LTBI. This estimate assumed that the primary purpose of the three programs was the detection and treatment of active disease. Incremental costs (21) were calculated relative to the costs of a program relying only on passive case detection and treatment without screening (incremental cost ⫽ costs if no program ⫺ cost of program). This assumed that all prevalent active cases detected through screening, as well as all cases prevented, would be diagnosed passively later, and would incur the same costs as the passively diagnosed cases in Montreal in 1996 and 1997.

Sensitivity Analyses To account for the effects of secondary spread of infection, an estimate was made of the number of infected contacts of active cases of TB that would have occurred if the prevalent cases detected through screening had instead been diagnosed passively. The number of infected contacts was estimated as follows: (1) four contacts of each smear-positive case and one per smear negative case would be infected (22, 23); (2) 71% of passively diagnosed cases of pulmonary TB would be smear positive (16), compared with 11% of actively diag-

Figure 1. Model for future incident cases of tuberculosis (TB). For persons with latent tuberculosis infection (LTBI), future reactivation risks were simulated with a simple three-state Markov process. In any given year, infected persons may die with the probability, pDie, based on Canadian life tables. Those who survive may experience reactivation of TB with probability, pReactivate, which varies with the presence or absence of defined radiographic abnormalities (see text). Persons who neither die nor have reactivation of disease in a given year of the simulation (the complementary probability, denoted by a hatchmark) continue to face the corresponding annual risks of death or reactivation in each subsequent year. For those who complete INH prophylaxis, the annual reactivation probability is reduced, reflecting the protective effect of INH (INH Effect), which was assumed to remain constant throughout the 20-yr simulation.

2082


VOL 162

2000

TABLE 2 SUMMARY OF OUTCOMES AMONG THE THREE STUDY COHORTS*

Initially identified Initially screened† Referred for evaluation† Seen by chest specialist Percent of subjects referred Active TB Cases detected Percent of all screened Percent of all evaluated Cases with completed treatment Latent TB infection including inactive TB Cases detected Eligible for therapy‡ Prescribed therapy Percent of subjects eligible for therapy Treatment discontinued by physician (side effects) Moved/care transferred§ Refused/dropped out/noncompliant Completed therapy§ Percent of subjects seen by chest specialist Cases of active TB prevented

Preimmigration Screening

Postarrival Surveillance

Close Contacts

— 12,898 722 658 91

828 — 622 401 64

244 200 119 109 92

17 0.13 2.6 16¶

4 N/A 1.0 4

6 3.0 5.8 6

353 332 190 58 0 0 45 145 22 7.85

191 119 66 55 1 4 11 49 14 1.58

94 91 88 97 4 12 15 57 58 3.21

* Outcomes known for all members of preimmigration cohort. For surveillance cohort, results were unknown for 28 subjects evaluated at other Montreal hospitals and 206 not resident in the Montreal area. Results were not available for five contacts evaluated at other Montreal hospitals. ‡ Initial screening with chest radiography for preimmigration and postarrival cohorts; those with abnormal chest radiography were referred and tuberculin skin test was performed later. Close contacts screened with tuberculin skin test and positive reactors referred for chest radiography and other workup. ‡ Subjects considered eligible for therapy if not treated previously and who were close contacts (all ages), or had an abnormal chest radiograph consistent with inactive TB (all ages), or who had normal chest radiography, age less than 36 yr and positive tuberculin skin test. § Completion of therapy and costs of transferred cases assumed to be similar to those who remained. ¶ One preimmigration applicant returned to country of origin.

resistant (MDR) strains of M. tuberculosis and required hospitalization for a total of 166 d. Therapy was prescribed for 88 of 91 (97%) previously untreated, tuberculin-positive close contacts of cases of active TB, for only 58% of applicants with LTBI who appeared eligible for therapy, because of barriers such as inadequate health insurance coverage or perceived difficulties in obtaining work permits. Among subjects in the surveillance cohort who were evaluated at the MCI, four (1%) had active TB; 191 (51%) had LTBI, including inactive TB and previously treated active TB; 111 (30%) had normal chest radiographs; and 67 (18%) had other diseases unrelated to TB. Of the 191 subjects with LTBI, 72 (19% of the entire cohort) had previously been adequately treated. This left 119 subjects eligible for therapy, of whom 30 (25%) were judged too old (average age: 53 yr), another 23 (19%) were not treated for a variety of reasons, and 66 in whom therapy was begun (55% of eligible subjects and 18% of the entire cohort). Of 344 subjects prescribed treatment for LTBI, it was stopped by physicians in five; another 16 were transferred elsewhere, and 71 (21%) refused therapy, dropped out, or were poorly compliant. Cost-Effectiveness

Among adults with active pulmonary TB diagnosed in 1996 and 1997 in other Montreal hospitals than the MCI, 76% were hospitalized for a median of 19 d (mean: 30 d), 71% were smear-positive, and 13 (8%) died (16) (Table 1). Based on a per diem hospitalization cost of $668, plus all other inpatient and outpatient costs, a passively diagnosed, hospitalized case of TB was estimated to cost $14,274. Since 76% of passively diagnosed patients would be hospitalized and would incur

these costs, and 24% of cases would not be hospitalized and would incur costs of only $1,006, the overall average cost for a passively diagnosed case was estimated to be $11,090. Based on this, the incremental cost exceeded $20,000 for each prevalent case of active TB treated, or for each future active case prevented in the applicant and postarrival surveillance cohorts, whereas direct-contact investigation resulted in net savings (Table 3). If the three TB detection programs examined in the study had existed simply to detect and treat active cases of TB, then all costs for initial identification, screening, and evaluation would have had to be spent in any case for the detection of active cases. The observed expenditures of $452–$507 for provision of therapy for LTBI could then be considered separately from all other costs of these programs. When considered as a marginal cost, therapy of latent TB infection was highly cost effective (Table 3), with net savings among applicants and contacts of active cases. Sensitivity Analyses

As shown in Figure 2 and Tables 4 and 5, the cost-effectiveness and impact of the applicant screening program would have been considerably improved if therapy for LTBI had been prescribed to 90% of eligible candidates. On the other hand, restricting screening to applicants from areas endemic for TB would have had much less benefit. As shown in Figure 3, the surveillance program evaluated in this study would have resulted in cost savings only if the cost per passively diagnosed case exceeded $40,000. Cost-effectiveness would have been somewhat better if the prescription rate had been higher, and considerably better if stricter referral criteria were applied. All programs would have been substan-

2083

Dasgupta, Schwartzman, Marchand, et al.: Cost-Effectiveness of Three TB Screening Programs TABLE 3 COMPARISON OF OBSERVED COSTS AND COST-EFFECTIVENESS IN THE THREE STUDY COHORTS

Outcomes (repeated from Table 2) Prevalent TB disease detected and treated Future TB disease prevented by treatment of LTBI Summary of costs Administrative costs Screening chest radiography Initial evaluation Treatment Of disease Of infection Total for program Cost/outcome Cost for TB disease detected and treated* Total cost for TB infection treated Total cost for TB disease prevented Incremental costs† Cost (savings) for prevalent TB disease treated* Total cost (savings) for future TB disease prevented Marginal costs‡ Cost for TB infection treated Cost for TB disease prevented Incremental cost (savings) for future TB disease prevented

Applicant

Surveillance

Contacts

16 7.85

4 1.58

6 3.21

61,177 283,756 145,718

51,600 — 65,538

39,536 — 14,396

12,031 71,172 573,854

105,774 23,514 246,426

7,718 33,470 95,120

31,418 3,958 73,125

55,728 4,739 155,729

10,275 1,441 29,668

20,328 39,409

24,225 65,126

(⫺815) (⫺2,186)

491 9,123 (⫺1,967)

452 14,860 3,770

507 10,427 (⫺663)

Definition of abbreviations: LTBI ⫽ latent tuberculosis infection; MDR-TB ⫽ multidrug-resistant TB. * Does not include costs for treatment of LTBI. † Cost to diagnose and treat each passively diagnosed case; cost for drug sensitive cases ⫽ $11,090; cost for MDR-TB ⫽ $51,912 (see text). ‡ Includes only costs for treatment of LTBI. Assumes that program already exists and expenditures are nevertheless for detection and treatment of active disease.

tially more cost-effective if the future risk of active TB was higher than our baseline estimate (Figure 4). After taking into account the additional costs related to secondary spread of infection, contact investigation would have resulted in net savings of more than $7,000, whereas the incremental cost to prevent each future case of active TB was $28,753 among applicants and $55,447 among the surveillance cohort.

Figure 2. Incremental cost per active case prevented for immigration applicants, showing effect of varying the cost per passively diagnosed case, as well as the efficiency of the applicant screening program. Dashed line with solid squares: applicant program. Dotted line with stars: screening of applicants from only countries with endemic TB. Dashed line with solid diamonds: screening of all applicants with prescription of INH to 90% of those eligible for preventive therapy. Solid line with solid triangles: actual cost of screening of close contacts of active cases of TB.

DISCUSSION In this study, investigation of contacts of active cases of TB had a high case detection rate and produced a high program efficiency, and so resulted in net cost savings. The incremental costs of the two immigrant screening programs could have been comparable if operational problems had been resolved and/or if costs per passively diagnosed TB patient had been similar to those reported elsewhere (8, 18, 19). Over the 1-yr study period, the three programs examined in the study detected 27 cases of active TB and prevented another 14. The total of 41 cases would have represented 28% of the 148 foreignborn cases of TB reported annually in Montreal from 1992 to 1995 (26). Only 11% of the cases detected through screening were smear-positive, as compared with more than 70% of passively detected cases (16, 17). Since smear-negative cases are much less contagious (22), significantly fewer secondary cases should have occurred if earlier diagnosis of TB in these cases prevented their progression to more advanced and contagious states. This means that the two immigrant screening programs examined in the study could have considerable public health benefit, especially if applied over many years and with improved program efficiency. Strengths of the study included direct ascertainment of all diagnostic, treatment, and follow-up activities to provide precise and realistic estimates of costs and outcomes for the three programs examined. This detailed information also allowed precise estimates of cost-effectiveness for subsets of patients, such as those from areas endemic for TB. The costs of the programs examined in the study included unforeseen events such as additional visits for education and reassurance, evaluation of side effects or new medical problems, or assistance with social problems, all of which are common in newly arrived immigrants. These events are not usually considered in theoretical analyses of cost effectiveness, but, in fact, increase the true costs of providing such programs. The cohorts in our study

2084


VOL 162

2000

TABLE 4 COMPARISON OF COST AND COST-EFFECTIVENESS FOR SCREENING AND THERAPY OF LATENT TUBERCULOSIS INFECTION IN THE THREE STUDY COHORTS IF 90% OF ELIGIBLE SUBJECTS HAD HAD THERAPY PRESCRIBED AND 80% HAD BEEN COMPLIANT

Outcomes Number eligible for therapy (repeated from Table 2) Number prescribed (at least 90%) Number completing therapy (at least 80%) Future active cases prevented Total cost for program Total cost/outcome Cost for TB infection treated Cost for future TB disease prevented Incremental costs‡ Total cost (savings) for future TB disease prevented Marginal costs‡ Cost for TB infection treated Cost for future TB disease prevented Incremental cost (savings) for future TB disease prevented

Applicant

Surveillance

Contacts*

332 299 239 13.69 620,031

114 103 82 2.49 259,976

91 88 66 3.21 95,120

2,594 45,291

3,170 104,408

1,441 29,668

21,240

42,714

(⫺2,186)

491 8,572 (⫺2,518)

452 14,885 3,796

507 10,427 (⫺663)

* Contact included for comparison although figures are unchanged from Table 3 because 97% had prescription of therapy and 80% were compliant. † Cost to diagnose and treat each passively diagnosed case; cost for drug sensitive cases ⫽ $11,090; cost for MDR-TB ⫽ $51,912. ‡ Includes only costs for treatment of LTBI. Assumes that program already exists and expenditures are nevertheless made for detection and treatment of active disease.

represented all applicants and all new immigrants requiring surveillance from a large geographic area for a 1-yr period. Members of all three cohorts were concurrently treated in the same setting by the same three chest radiologists, 10 chest specialists, and three TB clinic nurses, and by the same small group of public health physicians and nurses. This should have strengthened the internal validity of the study findings for the

comparison of the three screening programs, although this could have reduced the generalizability of the study. The estimate of future TB cases prevented was based on the incidence of tuberculin-reactive individuals in the general population of Ontario (27), although sensitivity analyses in our study (Figure 4) used lower estimates from a Danish cohort (24) and higher estimates from a cohort of refugees flee-

TABLE 5 COMPARISON OF COST AND COST-EFFECTIVENESS IN THE TWO IMMIGRATION COHORTS IF CERTAIN CHANGES TO THE CURRENT PROGRAMS WERE MADE Immigrant Applicant Screening

Screening of Subjects Only from TB-Endemic Regions Outcomes Number of persons screened Number of persons evaluated Prevalent TB disease detected and treated Prevalent TB infection treated Future TB disease prevented Total costs for programs Cost/outcome Cost for prevalent TB disease detected and treated§ Total cost for TB infection treated Total cost for future TB disease prevented Incremental costs‡ Cost (savings) for prevalent TB disease detected and treated§ Total cost (savings) for future TB disease prevented Marginal costs储 Cost for TB infection treated Cost for future TB disease prevented Cost (savings) for future TB disease prevented

Postarrival Surveillance

Screening of Subjects Only from TB-Endemic Regions and If 90% of Those Subjects Had Had If Stricter Referral Therapy Prescribed for LTBI Criteria Were Applied*

If Stricter Referral Criteria Were Applied and 90% of Referrals Were Seen

8,793 483 11 116 6.10 397,279

8,793 483 11 229 12.04 452,762

N/A 123 4 52 1.58 167,944

N/A 184 6† 73 2.22 241,018

30,938 3,425 65,132

30,938 1,977 37,596

36,108 3,230 106,115

34,670 3,302 108,496

19,848 34,039

19,848 16,383

4,607 15,454

3,169 12,339

491 9,337 (⫺1,753)

491 9,337 (⫺1,753)

452 14,882 3,792

452 14,882 3,792

Definition of abbreviations: LTBI ⫽ latent tuberculosis infection; MDR-TB ⫽ multidrug-resistant tuberculosis. * Includes all subjects with prevalent or past active TB who were prescribed therapy for latent TB infection or were given chest radiographic surveillance. † Assumes that one of the two additional active cases detected and treated would be MDR-TB. ‡ Cost to diagnose and treat each passively diagnosed case; cost for drug sensitive cases ⫽ $11,090; cost for MDR-TB ⫽ $51,912. § Does not include costs for treatment of latent TB infection. 储 Includes only costs for treatment of latent TB infection.

Dasgupta, Schwartzman, Marchand, et al.: Cost-Effectiveness of Three TB Screening Programs

Figure 3. Incremental costs per active case of TB prevented for surveillance cohort. Effect of varying the cost per passively diagnosed case and the efficiency of the surveillance screening program. Dashed line with solid squares: actual cost for surveillance cohort. Dotted line with solid circles: surveillance program with prescription of INH to 90% of subjects eligible for preventive therapy. Dotted line with solid diamonds: surveillance program with referral criteria restricted to subjects who needed to be seen. Dotted line with stars: surveillance program with referral criteria restricted to subjects who needed to be seen, 90% of whom were evaluated. Solid line with solid triangles: actual cost for screening close contacts of active cases of TB.

ing wartime conditions (25), which may magnify the risk of disease (28). The estimated incidence of TB-positive individuals was increased if there were radiographic abnormalities consistent with “inactive tuberculosis” (24, 25, 29). This is a poorly defined term, but more precise information on relative risk associated with specific radiographic abnormalities is not available. The cost for each passively diagnosed case of TB was not measured directly in this study but was available from a population-based study of all passively diagnosed cases in Montreal in 1996 and 1997 (16). The proportion and duration of hospitalization among these cases, which are major determinants of costs (8), were remarkably similar to those in studies of patients treated at seven Montreal-area tertiary care hospitals between 1992 and 1995 (17), in a Chicago hospital in 1993 (19), in New York City hospitals in 1990 (18), and in a United States national survey in 1991 (8). These studies in the United States, which have been cited in other cost-effectiveness analyses (30, 31), reported much higher total costs because of higher per diem hospitalization costs. On the other hand, if the costs for passively diagnosed cases of TB were lower because fewer such patients were hospitalized or were hospitalized for a shorter duration, the cost-effectiveness of the prevention programs examined in the present study would have been lower. The cost-effectiveness of these screening programs may have been under estimated because: (1) The costs of outpatient therapy for passively diagnosed cases of TB were assumed to be the observed average cost of $1,006, which was much lower than the $2,305 reported in a United States national survey (8). (2) Testing of applicants and new arrivals for human immunodeficiency virus (HIV) was rarely performed. This testing is not required by Immigration Canada, and both providers and patients are reluctant to perform such testing unless it is essential for patient care. If HIV infection was present among members of the study cohorts, the risk of dis-

2085

Figure 4. Incremental cost per active case of TB prevented. Effect of variation of future risk of disease for a tuberculin-reactive subject with normal chest radiography. Solid line with solid diamonds: close contacts of active cases of TB. Dashed line with solid squares: applicant program. Dotted line with stars: surveillance program.

ease would have been greater, as would the benefits of a prevention program. Additionally, hospitalization costs might have been as much as 50% higher (18). (3) The estimated cost per passively diagnosed case did not account for possible rehospitalization, although this may be required for 18% (19), 29% (8), or more of cases (18). (4) Therapy taken for less than 6 mo, or with poor compliance was assumed to provide no benefit, even though there may be some (32). (5) We estimated that excess spread of infection from passively diagnosed cases would engender costs of only $1,607. However, if each passively diagnosed case will in the long run produce 1.2 further cases of active TB (as recently estimated [33]), the cost of each passively diagnosed case should have been doubled. INH was the only therapy used for LTBI, and will not be effective if the latent infection is INH resistant (34). However, the estimates of efficacy of INH therapy were based on results of clinical trials in which at least some of the failures were due to this problem. Moreover, among all cases of active TB diagnosed among foreign-born persons in Montreal between 1990 and 1993, only 6.3% of isolates were INH resistant (26). Therefore, INH-resistant LTBI should be uncommon and is unlikely to have significantly reduced the effectiveness of the programs we examined. Physician failure to prescribe therapy for LTBI is well recognized (35, 36), although the resultant impact on program cost-effectiveness seen in Tables 4 and 5 has not been described. In our study, the same physicians prescribed therapy appropriately to 97% of contacts of cases of active TB, suggesting that they were well aware of guidelines for therapy of LTBI. It would have been cheaper to provide health insurance to applicants who needed therapy for LTBI, because failure to prescribe for them actually resulted in higher costs in the long term. For many members of the surveillance cohort who were of advanced age, had minimal radiographic abnormalities, or had negative tuberculin skin tests, therapy was not indicated. Referral of such patients wasted clinical and diagnostic resources; if only those who truly required follow-up had been referred, the surveillance program would have been much more cost-effective. Although needless referral may occur under the best of circumstances, given the limitations of chest radiography in screening for TB (37), this is more likely to occur

2086


when chest radiographs are of poor technical quality, which is a potential problem with overseas screening. Therapy of LTBI accounted for only 10% to 12% of the total cost of the two immigrant screening programs examined in our study, and cost less than $500 per person treated. Our findings emphasize that provision of therapy to high-risk patients as soon as they are identified as having LTBI is not only good clinical practice, but also sound management of limited public health resources. We conclude that relative to direct-contact investigation, radiographic screening programs for immigrants were expensive, in large part because of substantial operational problems. If these problems can be avoided, and opportunities for provision of therapy for LTBI are fully realized, then radiographic screening of newly arriving foreign-born populations for TB could be cost-effective and have considerable individual and public health benefits. Acknowledgment : The authors acknowledge the support of the Fonds de Recherche en Santé du Québec for funding the study, and MCI Research Centre. The authors also thank the staff of the MCI, Montreal Public Health Department, and the Health Division of Citizenship and Immigration Canada for their help and collaboration in the conduct of the study. The authors thank Dr. Klaus Jochem for helpful comments on earlier drafts of this report.

References 1. Davies PDO. Tuberculosis and migration: The Mitchell Lecture 1994. J R Coll Physicians Lon 1995;29:113–118. 2. Menzies RI, Vissandjee B, Amyot D. Factors associated with tuberculin reactivity among the foreign-born in Montreal. Am Rev Respir Dis 1992;146:752–756. 3. Enarson D, Ashley MJ, Grzybowski S. Tuberculosis in immigrants to Canada: a study of present-day patterns in relation to immigration trends and birthplace. Am Rev Respir Dis 1979;119:11–17. 4. Enarson DA, Sjogren I, Grzybowski S. Incidence of tuberculosis among Scandinavian immigrants in Canada. Eur J Respir Dis 1980;61:139–142. 5. McKenna MT, McCray E, Onorato I. The epidemiology of tuberculosis among foreign-born persons in the United States, 1986–1993. N Engl J Med 1995;332:1071–1076. 6. U.S. Department of Health And Human Services. Reported tuberculosis in the United States, 1997. Atlanta, GA: Centers for Disease Control and Prevention; 1998. 7. Long R, Njoo H, Hershfield E. Tuberculosis: 3. Epidemiology of the disease in Canada. Can Med Assoc J 1999;160:1185–1190. 8. Brown RE, Miller B, Taylor WR, Palmer C, Bosco L, Nicola RM, Zelinger J, Simpson K. Health-care expenditures for tuberculosis in the United States. Arch Intern Med 1995;155:1595–1600. 9. Chin DP, DeRiemer K, Small PM, Ponce de Leon A, Steinhart R, Schecter GF, Daley CL, Moss AR, Paz EA, Jasmer RM, Agasino CB, Hopewell PC. Differences in contributing factors to tuberculosis incidence in U.S.-born and foreign-born persons. Am J Respir Crit Care Med 1998;158:1797–1803. 10. Borgdorff MW, Nagelkerke N, van Soolingen D, de Haas PEW, Veen J, van Embden JDA. Analysis of tuberculosis transmission between nationalities in the Netherlands in the period 1993–1995 using DNA fingerprinting. Am J Epidemiol 1998;147:187–195. 11. Medical Research Council Tuberculosis and Chest Diseases Unit. National survey of tuberculosis notifications in England and Wales 1978– 9. Br Med J 1980;281:895–898. 12. Binkin NJ, Zuber PLF, Wells CD, Tipple MA, Castro KG. Overseas screening for tuberculosis in immigrants and refugees to the United States: current status. Clin Infect Dis 1996;23:1226–1232. 13. Markey AC, Forster SM, Mitchell R, Larson E, Smith H. Suspected cases of pulmonary tuberculosis referred from port of entry into Great Britain, 1980–3. Br Med J 1986;292:378–379. 14. Bonvin L, Zellweger JP. Mass miniature x-ray screening for tuberculosis among immigrants entering Switzerland. Tuberc Lung Dis 1992;73: 322–325. 15. Thomas RE, Gushulak B. Screening and treatment of immigrants and

16.

17.

18.

19.

20. 21.

22. 23.

24.

25. 26. 27. 28. 29.

30.

31.

32.

33.

34.

35. 36.

37. 38. 39.

VOL 162

2000

refugees to Canada for tuberculosis: implications of the experience of Canada and other industrialized countries. Can J Infect Dis 1995;6: 246–255. Musana K, Menzies D, Tannenbaum T, Brassard P, Thibert L, Pollock S, FitzGerald JM, Kunimoto D, Schwartzman K. Molecular epidemiology of tuberculosis in 3 Canadian cities, 1995–97 [abstract]. Am J Respir Crit Care Med 1999;159:A904. Menzies D, Fanning A, Yuan L, FitzGerald JM, The Canadian Collaborative Group in Nosocomial Transmission of Tuberculosis. Tuberculosis in health care workers: a multicentre Canadian prevalence survey: preliminary results. Int J Tuberc Lung Dis 1998;2:S98–S102. Arno PS, Murray CJL, Bonuck KA, Alcabes P. The economic impact of tuberculosis in hospitals in New York City: a preliminary analysis. J Law Med Ethics 1993;21:317–323. Wurtz R, White RW. The cost of tuberculosis: utilization and estimated changes for the diagnosis and treatment of tuberculosis in a public health system. Int J Tuberc Lung Dis 1999;382–387. Sumartojo E. When tuberculosis treatment fails: a social behavioural account of patient adherence. Am Rev Respir Dis 1993;147:1311–1320. Drummond MF, O’Brien B, Stoddart GL, Torrance GW. Methods for the economic evaluation of health care programmes, 2nd ed. Oxford, UK: Oxford University Press; 1997. p. 59. Menzies D. Issues in the management of contacts of patients with active pulmonary tuberculosis. Can J Public Health 1997;88:197–201. Behr MA, Warren SA, Salamon H, Hopewell PC, Ponce de Leon A, Daleyand CL, Small PM. Transmission of Mycobacterium tuberculosis from patients smear-negative from acid-fast bacilli. Lancet 1999;353:444–449. Horwitz O, Wilbek E, Erickson PA. Epidemiological basis of tubeculosis eradication: longitudinal studies on the risk of tuberculosis in the general population of a low-prevalence area. Bull World Health Organ 1969;41:95–113. Nolan CM, Elarth AM. Tuberculosis in a cohort of southeast Asian refugees. Am Rev Respir Dis 1988;137:805–809. Rivest P, Tannenbaum T, Bédard L. Epidemiology of tuberculosis in Montreal. Can Med Assoc J 1998;158:605–609. Grzybowski S, Allen EA. The challenge of tuberculosis in decline. Am Rev Respir Dis 1964;90:707–720. Barr R, Menzies R. The effect of war on tuberculosis. Tuberc Lung Dis 1994;75:251–259. Grzybowksi S, McKinnon NE, Tuters L, Pinkus G, Philipps R. Reactivations in inactive pulmonary tuberculosis. Am Rev Respir Dis 1966;93: 352–360. Gourevitch MN, Alcabes P, Wasserman WC, Arno PS. Cost-effectiveness of directly observed chemoprophylaxis of tuberculosis among drug users at high risk for tuberculosis. Intl J Tuberc Lung Dis 1998;2:531– 540. Mohle-Boetani JC, Miller B, Halpern M, Trivedi A, Lessler A, Solomon SL, Fenstersheib M. School-based screening for tuberculous infection: a cost-benefit analysis. JAMA 1995;274:613–619. International Union Against Tuberculosis Committee on Prophylaxis. Efficacy of various durations of isoniazid preventive therapy for tuberculosis: five years of follow-up in the IUAT trial. Bull World Health Organ 1982;60:555–564. Salpeter SR, Sanders GD, Salpeter EE, Owens DK. Monitored isoniazid prophylaxis for low-risk tuberculin reactors older than 35 years of age: a risk-benefit and cost-effectiveness analysis. Ann Intern Med 1997; 127:1051–1061. Nolan CM, Aitken ML, Elarth AM, Anderson KM, Miller WT. Active tuberculosis after isoniazid chemoprophylaxis of Southeast Asian refugees. Am Rev Respir Dis 1986;133:431–436. Geiseler PJ, Nelson KE, Crispen RG. Tuberculosis in physicians: compliance with preventive measures. Am Rev Respir Dis 1987;135:3–9. Gemson DH, Elinson JE. Prevention in primary care: variability in physician practice patterns in New York City. Am J Prev Med 1986;2: 226–234. Toman K. Tuberculosis—case-finding and chemotherapy: questions and answers. Geneva: World Health Organization; 1979. Ferebee SH. Controlled chemoprophylaxis trials in tuberculosis. Adv Tuberc Res 1969;17:28–106. Comstock GM. How much isoniazid is needed for prevention of tuberculosis in immunocompetent adults. Int J Tuberc Lung Dis 1999;3: 847–850.