Effectiveness of Postmigration Screening in Controlling Tuberculosis Among Refugees: A Historical Cohort Study, 1984–1998

Abstract

Objectives. This study assessed the effectiveness of postmigration screening for the control of tuberculosis (TB) among refugee migrants.

Methods. We conducted a historical cohort study among 24 610 predominantly Southeast Asian refugees who had arrived in Sydney, Australia, between 1984 and 1994. All had been screened for TB before arrival and had radiologic follow-up for 18 months after arrival. Incident cases of TB were identified by record linkage analysis with confirmatory review of case notes.

Results. The crude annual incidence rate over 10-year follow-up was 74.9 per 100 000 person-years. Only 29.6% of the cases were diagnosed as a result of routine follow-up procedures.

Conclusions. Enhanced passive case finding is likely to be more effective than active case finding for the control of TB among refugees.

Migration from countries with a high incidence of tuberculosis (TB) is having a major effect on the epidemiology of TB in countries with lower incidence, such as Canada, the United States, and Australia,1– 4 and has led to calls for better efforts to prevent and control the disease among immigrants.3 State health authorities in Australia have taken an active approach to case finding among refugees settling in Australia. The primary aim of this study was to assess the effectiveness of a program of postarrival screening for active case finding in a population of refugees screened for TB before migration. A further aim was to establish whether any criterion, determined at initial screening after arrival in Australia, identified a subgroup in whom subsequent follow-up for active case finding was likely to be more effective.

METHODS

We conducted a historical cohort study among refugees who were screened for TB after arrival in Sydney, Australia, during the period 1984 through 1994.

All refugees 16 years and older, as well as symptomatic children, had a chest x-ray performed before departure for Australia. Those in whom currently or recently active TB was suspected began receiving antituberculous drug therapy at that time. Soon after their arrival in New South Wales, Australia, all refugees had further screening for TB at the Refugee Screening Unit or at Liverpool Chest Clinic. The results of this initial postmigration screening, which comprised a tuberculin skin test, examination for a BCG vaccine scar, and a chest x-ray, were recorded in a logbook. Tuberculin skin test reaction size was recorded for all reactions 10 mm or greater. Tuberculin skin test reaction sizes in the range of 0 to 9 mm were recorded as “negative”. Chest x-rays were recorded as “satisfactory” or “not satisfactory”. These records formed the basis of the database of screened refugees referred to below.

All refugees were then followed up by their local area chest clinics at 6 months and 18 months after initial postarrival screening. This follow-up screening comprised annual chest x-rays and, in certain cases, clinical examinations. For children younger than 14 years, chest x-rays were performed only if the tuberculin skin test reaction was 10 mm or greater. Refugees who did not attend follow-up appointments were sent reminder letters, telephoned, and, when necessary, visited at home to encourage them to attend. At the discretion of the attending medical officer, the period of screening continued beyond 18 months after arrival for some refugees.

Cases of TB arising in the cohort were identified by a record linkage analysis that used a database of screened refugees and a database of cases of TB reported to the New South Wales Department of Health during the period 1984 through 1998.5,6 We reviewed notification data, case notes, and x-ray films for the notified cases of TB identified from the refugee cohort to confirm or refute the diagnosis of active TB.5,6

The source of discovery of the case (i.e., referral because of symptomatic presentation or detection by screening procedures) was ascertained from case records and notification details.

To be certain that cases of TB arising in this cohort would have been reported to the New South Wales Department of Health, we sought to confirm that the members of the cohort were predominantly alive and living in New South Wales throughout the study period. We attempted to trace the current location of 200 cohort members selected at random from the study population.

The sensitivity and specificity of various cutpoints for a positive tuberculin skin test reaction (10, 15, and 20 mm) and an abnormal chest x-ray at initial postarrival screening for predicting the subsequent development of TB were calculated. Number needed to screen to detect a case was calculated as the inverse of the incidence rate. This analysis was limited to the cases arising within 18 months of arrival—that is, the period of routine postarrival screening.

RESULTS

Between 1984 and 1994, 24 652 refugees were screened after arrival in New South Wales. Of these, 42 were taking anti-TB drug treatment before arrival in Australia and were excluded from this study cohort. The members of the cohort were young (median age = 22.1 years), and there was a slight predominance of men (56.5%). The overwhelming majority (84.1%) originated from Southeast Asia, including 64.1% from Vietnam and 11.7% from Cambodia. Tuberculin skin test reactions were 10 mm or greater in 45.5%, and 6.2% had abnormal chest x-rays at the initial postarrival assessment.

Of the random sample of 200 cohort members whose whereabouts were traced in 1998, 171 (85.5%) were living in New South Wales.

By June 30, 1998, after an average follow-up interval of 10.3 years, 290 members of this cohort had been reported as a TB case. Of these, 189 (65.2%) were confirmed as cases of TB, representing an overall incidence rate of 74.9 per 100 000 person-years. Sixty-seven percent of the active cases were pulmonary; of these, 50% were smear positive on sputum examination, and 25% were smear negative but culture positive.

Fifty-five cases, 29.1% of all observed cases, occurred within 1 year of arrival, and 124 (65.6%) occurred within 5 years of arrival. The incidence rate in the first year was 223.8 per 100 000, and the average annual incidence rate over the first 5 years after arrival was 102.1 per 100 000.

Two thirds of the cases (126) were identified as a result of symptomatic presentation, usually through primary care. Most of these arose after the 18-month period of routine screening (Table 1 ▶). Fifty-six cases (29.6%) were diagnosed and reported as a result of the routine screening program. The remaining 7 cases (3.7%) were identified as a result of contact tracing investigations.

TABLE 1—

Proportion of Diagnostic Classes of Tuberculosis Detected by the Screening Program, 1984–1998

	Symptomatic Presentation
	Within 18 Months of Initial Postarrival Screen		>18 Months After Initial Screen		All		Screen Detected, All
Diagnostic Class	n	%	n	%	n	%	n	%
All cases (n = 189)	22	11.6	104	55.0	126	66.7	63a	33.3
Pulmonary
All (n = 127)	8	6.3	64	50.4	72	56.7	55	43.3
Smear positive (n = 64)	5	7.8	42	65.6	47	73.4	17	26.6
Smear negative, culture positive (n = 32)	1	3.1	16	50.0	17	53.1	15	46.9
Culture negative (n = 31)	2	6.5	6	19.4	8	25.8	23	74.2
Extrapulmonary (n = 62)	14	22.6	40	64.5	54	87.1	8	12.9

^aIncludes 7 cases that were detected as result of contact screening.

Table 2 ▶ shows the sensitivity and specificity of various tuberculin skin test cutpoints, the presence of an abnormal chest x-ray, and the combination of these findings for predicting subsequent development of TB. No single finding or combination of findings had both high sensitivity and high specificity.

TABLE 2—

Predictive Factors for Tuberculosis (TB) in a Cohort of Refugees Screened After Arrival, 1984–1994 (N = 24 610)

			Pulmonary
	All TB Cases		All		Smear Positive		Smear Negative, Culture Positive		Culture Negative		Extrapulmonary		No TBa
Findings at Initial Postarrival Screening	Rateb	%c	Rate	%	Rate	%	Rate	%	Rate	%	Rate	%	Specificity, %
TST ≥ 10 mm	122.0	79.7	78.8	77.1	38.9	76.3	19.9	79.3	19.9	76.7	43.3	84.7	52.1
TST ≥ 15 mm	160.3	62.1	104.9	61.0	51.0	59.3	24.8	58.6	29.1	66.7	55.4	64.4	71.2
TST ≥ 20 mm	191.7	35.0	117.5	32.2	61.8	33.9	30.9	34.5	24.7	26.7	74.2	40.7	86.3
Abnormal chest x-ray	245.4	19.3	182.0	20.9	55.4	12.5	47.5	21.4	79.1	38.5	63.3	15.7	93.9
TST ≥ 10 mm and abnormal chest x-ray	318.0	16.0	228.9	17.0	63.6	9.3	50.9	14.8	114.5	36.0	89.0	14.0	96.1
TST ≥ 15 mm and abnormal chest x-ray	391.2	13.5	279.4	14.2	93.1	9.3	55.9	11.1	130.4	28.0	111.8	12.0	97.3
TST ≥ 20 mm and abnormal chest x-ray	232.5	3.9	155.0	3.8	116.2	5.6	0	0	38.4	4.0	77.5	4.0	98.7

Note. TST = tuberculin skin test. Rates and percentages are calculated excluding subjects with missing data for the characteristic.

^aNo TB: column shows percentage of those not reported as TB cases in whom this characteristic was absent (specificity).

^bRate = incidence rate per 100 000 person-years in persons with this characteristic.

^c% = percentage of cases with this characteristic (sensitivity).

During the 18-month period in which routine postarrival screening was implemented, 65 cases of TB were diagnosed in the cohort of 24 610 refugees who received screening. If the screening program had detected all these cases, 379 persons would have been screened for each case detected during this period. In fact, only 43 of these cases were detected by screening; hence, the number of persons screened per detected case was 572. The screening program detected 11 cases of smear-positive pulmonary TB, representing 2237 persons screened per detected case during this period.

Restriction of follow-up screening to population subgroups selected on the basis of a positive initial tuberculin skin test or abnormal chest x-ray findings would have reduced the number screened per detected case (data not shown), but the number of cases actually detected would be reduced proportionately.

DISCUSSION

We have shown that although the incidence of TB was moderate to high in this cohort of refugees from high-prevalence countries,6 short-term screening for active case finding identified fewer than one third of the cases destined to occur within a decade of arrival and required follow-up of several hundred individuals for each case found and several thousand individuals for each infectious case found. Although the specificity of screening can be enhanced by limiting its coverage to refugees who have, at initial assessment, a positive tuberculin skin test result and an abnormal chest x-ray, this criterion has a low sensitivity and would be expected to identify only 16.0% of the incident cases.

The validity of this method of assessing the effectiveness of screening is based on the assumption that most of the cases of TB arising in this cohort represented reactivation of TB that was latent at the time of arrival in Australia. Although transmission, and hence infection or reinfection, may have occurred in Australia or during visits back to the country of origin, this transmission timing probably has been even less common in the Sydney environment than it was in San Francisco, Calif.7,8

A strength of this study was the sensitivity and specificity of case ascertainment. Evidence indicates that the overwhelming majority of the patients treated for TB in New South Wales were reported to the Department of Health and were identified as cases in our study.9,10 This is, in part, attributable to the fact that the great majority of TB treatment (for all ages and for all sites) is given free of charge through chest clinics in New South Wales. The validity of this database as a complete, and thus sensitive, record of cases of TB occurring in New South Wales is supported by the observation that recent intensified surveillance, including tracing through records of sale of antituberculous drugs, did not result in any detectable increase in the identification rate. The specificity of the diagnosis of TB cases in this analysis was enhanced by review of case records of reported cases and reassessment of the diagnosis based on the criteria specified in the Methods section of this report.5,6 Finally, the finding that the cohort has remained intact, within New South Wales, during the period of follow-up supports the likelihood of complete ascertainment of incident cases.

Even if the period and intensity of postmigration screening were increased, it is unlikely that active case finding would be a cost-effective strategy for the control of TB in this population. Enhancing the capacity for passive case finding by increasing awareness of the diagnosis, particularly among primary care physicians attending the refugee and migrant communities, and improving access to specialist services for the diagnosis and management of TB is likely to be a more effective strategy. Preventive therapy with isoniazid in tuberculin skin test–positive refugees may have a role in the control of TB in situations in which passive case finding is ineffective and there are long delays to diagnosis. However, to prevent a substantial proportion of incident cases, this strategy would be costly because it would require the administration of chemoprophylaxis to, and medical review of, a large proportion of incoming refugees.