Screening for leprosy in immigrantsða decision analysis model

Lepr Rev 2003) 74, 240±248 Screening for leprosy in immigrantsða decision analysis model RICHARD TAYLOR*, KATHLEEN KING**, PETER VODICKA*, JOHN HALL* & DAVID EVANS* *School of Public Health, Faculty of Medicine, University of Sydney, Sydney, Australia **Department of Immigration and Multicultural Affairs, Federal Government, Australia Accepted for publication 3 April 2003 Summary Almost all leprosy cases reported in industrialized countries occur amongst immigrants or refugees from developing countries where leprosy continues to be an important health issue. Screening for leprosy is an important question for governments in countries with immigration and refugee programmes. A decision analysis framework is used to evaluate leprosy screening. The analysis uses a set of criteria and parameters regarding leprosy screening, and available data to estimate the number of cases which would be detected by a leprosy screening programme of immigrants from countries with different leprosy prevalences, compared with a policy of waiting for immigrants who develop symptomatic clinical diseases to present for health care. In a cohort of 100,000 immigrants from high leprosy prevalence regions 3 6/10,000), screening would detect 32 of the 42 cases which would arise in the destination country over the 14 years after migration; from medium prevalence areas 0 7/10,000) 6 3 of the total 8 1 cases would be detected, and from low prevalence regions 0 2/10,000) 1 8 of 2 3 cases. Using Australian data, the migrant mix would produce 74 leprosy cases from 10 years intake; screening would detect 54, and 19 would be diagnosed subsequently after migration. Screening would only produce signi cant case-yield amongst immigrants from regions or social groups with high leprosy prevalence. Since the number of immigrants to Australia from countries of higher endemnicity is not large routine leprosy screening would have a small impact on case incidence. Introduction Almost all leprosy cases reported in industrialized countries occur amongst immigrants or refugees from developing nations, 1±5 where leprosy continues to be an important health issue. 6 There are intermittent case reports of leprosy in developed nations 1,2,7±9 and recommendations for screening immigrants and refugees for leprosy. 4,10,11 Correspondence: R. Taylor, School of Public Health, Faculty of Medicine, Building A27, The University of Sydney, NSW 2006, Australia 240 0305-7518/03/064053+9 $1.00 q Lepra

Screening for leprosy in immigrants 241 Leprosy is a candidate disease for screening, as it has a detectable early clinical phase that may not prompt self referral for health care, which can be successfully treated. 12 Further, early treatment can prevent the considerable morbidity associated with the disease. 13,14 However, the long latent phase of leprosy, lasting from 2 to 15 years, between infection with Mycobacterium leprae and onset of detectable clinical disease 13 has implications for the proportion of cases detected in a screening programme. 15 Screening for leprosy requires skin inspection for suspicious lesions and examination and palpation of some peripheral nerves. 16 Although quick and inexpensive, it requires trained and experienced clinicians. The most reliable tests are slit-skin smears or, preferably, skin biopsy, but negative results do not exclude paucibacillary disease. 17 The lepromatin skin test, which is useful in classifying leprosy according to cell-mediated response, is of no value in screening. 18 Surveillance studies of leprosy suspects, 19 screening of household and dwelling contacts for leprosy, 20 and eld surveys for detecting leprosy 20 in developing countries have used these techniques. Although a need for an immigrant leprosy screening programme in some developed countries has been put forward, 4,10,11 there has been no explicit consideration of such a programme's usefulness. This article derives data estimates for use in the analysis of leprosy screening of immigrants to developed countries with extremely low leprosy prevalence, and examines the usefulness of screening in a decision analysis framework The decision analysis compares a screening programme for immigrants, from countries with different leprosy prevalences, with a policy of waiting for immigrants who develop symptomatic clinical disease to present for health care in the usual manner. Materials and methods A summary of the main assumptions and numerical estimates used in the decision models is given in Table 1. PREVALENCE ESTIMATES Prevalence estimates were based upon recent WHO data, which de ned prevalent cases as those needing chemotherapy. 6 High prevalence countries were considered to be those with prevalences of 1 case/10,000 population or greater, medium prevalence countries those with prevalences of 0 5 to < 1 case/10,000 population, and low prevalence countries those with prevalences of 0 1 to < 0 5 cases/10,000 population. The mean prevalence for each group was used as baseline estimates Table 1). An upper limit prevalence of 30 cases/10,000 population and the prevalence of India 6/10,000), the highest prevalence country which also contributes signi cantly to international immigration, were also tested. A prevalence of 0 01 cases/10,000 population was used for industrialized nations. INCIDENCE ESTIMATES Linear regression of case-detection rates on prevalence rates both log transformed) from 1998 WHO data suggested that the annual incidence rate was between 35 and 50% of prevalence for low to high prevalence countries. Using data from the Global Burden of Disease study, 22 a similar analysis gave an annual incidence of about 20% of prevalence.

242 R. Taylor et al. Table 1. Summary of assumptions and data estimates No screening Prevalent casesðcases with disease at time of migration Present and treated in destination country Incident cases after migrationðnew cases after migration from Present and treated in destination country infection in country of origin Secondary cases after migrationðnew cases after migration from Not considered. infection in destination country Total cases after migrationðprevalent plus incident cases Screening Screen positive casesðtrue positive plus false positive cases Re-examined True positive case Diagnosed and treated in home country False positive cases Re-examined, cleared and migrate. Includes incubating cases Screen negative casesðtrue negative plus false negative cases Migrate True negative cases Migrate. Includes incubating cases False negative cases Migrate. Present and treated in destination country Data estimates Prevalence in country of origin Cases/10,000 population High prevalence 3 6 Medium prevalence 0 7 Low prevalence 0 2 Upper limit 30 India 6 Industrialized nations 0 01 Incidence in country of origin %prevalence Baseline 30% Lower limit 20% Upper limit 50% Cumulative incidence in migrant group after migration %incidence in country of origin Baseline 52% Sensitivity of screening Baseline 90% Lower limit 70% Upper limit 99% Speci city of screening Baseline 99% Given this range, an incidence rate of 30% of national prevalence was used as a baseline, but a range was tested Table 1). SCREENING ACCURACY There are no reports of the sensitivity and speci city of leprosy screening in the literature. Expert opinion was that the diagnosis may not be con rmed in about 15% of patients initially suspected of having leprosy, giving a positive predictive value of approximately 85% personal communication, Britton, Lloyd and Board). A false negative rate of about 10% was expected personal communication, Britton, Lloyd and Board), so sensitivity was assumed to be 90%, but a range was tested Table 1). Speci city was calculated from the estimated positive predictive value and sensitivity for different prevalences, and was consistently higher than 99%. A speci city of 99% was thus used in this analysis.

Screening for leprosy in immigrants 243 NEW LEPROSY CASES AFTER MIGRATION A synthetic cohort method was used to estimate cases infected before, but occurring as incident cases of disease after migration, owing to the 2±14 year asymptomatic period after infection. The cumulative leprosy incidence for any single year migrant intake over 14 years after migration was calculated as the average ofthe cumulative incidences ofa group of14 hypothetical immigrant cohorts, each cohort infected at 1, 2, 3,..., 14 years, respectively, before migration. This assumes that the proportion of immigrants infected each year before migration is approximately equal. Expert opinion suggested that the annual proportions ofan infected cohort developing clinical leprosy each year are 10% per year for the rst 6 years after the initial 2 asymptomatic years, 8 3% per year for the next 3 years, and around 5% per year for the last 3 years to 14 years 100%) personal communication, Britton, Lloyd and Board). Other new cases of leprosy occurring after migration will be secondary cases from domiciliary contact with multibacillary index cases. Secondary transmission is rare in developed nations and estimated at about one in 100 multibacillary cases or 1 in 300 leprosy cases personal communication, Britton, Lloyd and Board). No secondary transmission is considered in this analysis. LEPROSY TYPE Literature suggests that the distribution ofleprosy types is approximately 30% multibacillary and 70% paucibacillary. 13 DECISION MODEL STRUCTURE Decision tree models were developed using DATA 3.0 software TreeAge, MA, 1997). Figures 1 and 2 show the models without and with screening, respectively, and Table 1 summarizes the de nitions and actions followed in each tree. The number of cases detected in the home country, and in the destination country, was calculated for a hypothetical cohort of 100,000 immigrants from each population of given prevalence. For the model without screening Figure 1), the rst chance node divides the cohort into prevalent clinical leprosy or no leprosy at migration. Along the branch for immigrants with no leprosy on arrival, the second chance node separates those who remain without leprosy from those who develop incident disease. Figure 1. Decision tree for immigrant cohort without screening programme.

244 R. Taylor et al. Figure 2. Decision tree for immigrant cohort with a screening programme. Proportion screen positive ˆ prevalence sensitivity) 1 prevalence) 1 speci city). Proportion false negative ˆ prevalence 1 sensitivity))/ prevalence 1 sensitivity)) speci city 1 prevalence))). Proportion true positive ˆ prevalence sensitivity)/ prevalence sensitivity) 1 prevalence) 1 speci city). For other de nitions, see Table 1. In the model including screening Figure 2), the rst chance node separates the cohort into those who screen positive and negative. Along the screened positive branch, the second chance node separates the group into true and false positives with subsequent management outlined in Table 1. As the false positive cases do not have leprosy at migration, they are at risk for incident disease after migration as represented by the third chance node along this branch. Along the screened negative branch, the second chance node divides the group into true and false negatives. The true negative cases may be incubating leprosy and are thus at risk of incident disease after migration, shown by the third chance node in this branch. Results Table 2 summarizes the cases detected in the country of origin and the destination country under programmes of no screening and screening. The results are for a hypothetical cohort of 100,000 immigrants from regions of different prevalences. In a cohort of immigrants from high leprosy prevalence regions 3 6/10,000), screening would detect 32 of the 42 cases that would arise in the destination country over the 14 years after migration. Of the 13 multibacillary cases which would occur in this group overall, screening would detect 10 cases. Similarly, amongst 100,000 immigrants from medium prevalence areas 0 7/10,000), screening would detect 6 3 1 9 multibacillary) of the total 8 1 cases 2 4 multibacillary). For immigrants from low prevalence regions 0 2/10,000) screening would detect 1 8 0 5 multibacillary) of a total 2 3 cases 0 7 multibacillary) in 100,000. Varying the estimate of incidence, as a proportion of prevalence, showed an appreciable change in the number of cases detected in the destination country after migration only amongst immigrants from high prevalence regions, reducing by two cases for the lower incidence and increasing by three cases for the higher incidence tested. Varying the sensitivity of the screening test showed an appreciable change in the number of cases detected after migration only amongst immigrants from high prevalence regions. With a screening sensitivity of 70%, an additional seven leprosy cases would be rst detected

Screening for leprosy in immigrants 245 Table 2. Cases of leprosy detected under protocols of screening and no screening in 100,000 immigrants No screening Cases detected in destination country Screening Cases detected in country of origin Cases detected in destination country Prevalence variation 1 Prevalence category Prevalence cases/10,000 population) High 3 6 42 32 10 Medium 0 7 8 1 6 3 1 8 Low 0 2 2 3 1 80 5 Upper limit 30 347 224 123 India 6 69 54 15 Industrialized nations 0 01 0 12 0 09 0 03 Incidence variation 2 Prevalence category Incidence % of prevalence) High prevalence 20% 40 32 8 50% 45 32 13 Medium prevalence 20% 7 7 6 3 1 4 50% 8 8 6 3 2 5 Low prevalence 20% 2 2 1 80 4 50% 2 5 1 80 7 Screening sensitivity variation 3 Prevalence category Screening sensitivity High prevalence 70% 42 25 17 99% 42 36 6 Medium prevalence 70% 8 1 4 9 3 2 99% 8 1 6 9 1 2 Low prevalence 70% 2 3 1 4 0 9 99% 2 3 2 0 0 3 1 Baseline analysis: see baseline assumptions Table 1). 2 Assumes baseline values for cumulative incidence after migration, screening sensitivity and screening speci city Table 1). 3 Assumes baseline values for incidence, cumulative incidence after migration and speci city of screening Table 1). after migration; with a sensitivity of 99%, three fewer cases would be detected in the destination country after migration. In Australia's case in the full year 1997±1998, the Australian migrant intake was around 78,000 including refugees). 23,24 Around 20% came from highly endemic countries, 20% from medium-low endemic countries, and 60% from non-endemic countries. Assuming average rates for these countries, and using the estimations above, this migrant mix would produce 74 leprosy cases from 10 years intake of this magnitude, and with no screening, all would present clinically in Australia. With screening, 54 would be detected in the country of origin, and 19 would be diagnosed subsequently in Australia after migration. Discussion This is the rst study to examine the effectiveness of screening for leprosy amongst immigrants. It shows that screening would detect about three quarters of the leprosy cases

246 R. Taylor et al. that would emerge over the 14 years after migration in an immigrant group, and indicates that leprosy screening may be worthwhile in migrant populations from highly endemic areas. This analysis uses data from developing countries and there are several potential sources of error. WHO prevalence data from 1998 were used, 6 which are lower than the prevalences in the Global Burden of Disease GBD) Study. 22,25 Recent WHO data were chosen as the GBD study may over-estimate current leprosy prevalence, being based on data from the late 1980s before the widespread removal of inactive leprosy cases from national registers. 14 In addition, WHO treatment recommendations in 1997 reduced treatment duration for multibacillary cases from 24 months to twelve months; this may have further reduced prevalence, as estimated from national registers, by reducing case-holding time compared with a decade ago. 14 However, the prevalence data used in this study may still be an underestimate, owing to undetected clinical cases in a country, or an overestimate, owing to the inclusion on the registers of treated patients under surveillance or receiving ongoing care for disability. The prevalence of leprosy in an immigrant group from a particular country may also not re ect the national prevalence. For example, ordinary immigrants may have a lower leprosy prevalence as they may come from higher socioeconomic groups, whereas refugees and other emergency entrants may have leprosy rates higher than the national prevalence. These possibilities are considered in the analysis by testing a range of prevalences. If a refugee group from a highly endemic country manifested a prevalence 10 times higher than the national average, their prevalence would be close to the upper limit tested Table 1). However, if predominantly middle class immigrants have a leprosy prevalence ten times lower than the national prevalence, this subgroup of immigrants from high prevalence countries would fall into the low prevalence range. The incidence of leprosy is dif cult to estimate, as it is based on WHO-reported case detection rates in the context of active disease control programmes that include both prevalent and incident cases. The incidence rate may be in ated because of increased case detection activity resulting from enhanced leprosy campaigns that were common during the above period. Regression of case detection rates on prevalence rates, after logarithmic transformation of both variables, from the GBD study gave a lower value than a similar analysis based on WHO data. As the GBD data may overestimate prevalence, an incidence between these was used as the baseline but a range was tested. There was little variation in the cases detected after migration by changing the incidence estimate in this analysis, suggesting that inaccuracies in estimating incidence are not a major concern. Screening was assumed in this study to consist of skin inspection and peripheral nerve examination or palpation, with appropriate follow up in specialist centres by skin smears or biopsy. Although diagnostic criteria, such as impaired sensation of skin lesions, would be used, there are no reports of the sensitivity and speci city of this approach. These parameters will depend upon the experience of the examining clinician, which may be highly variable. The increase in cases detected after migration if sensitivity is low is relatively large in this analysis and may result in missing an important number of cases amongst immigrants from high prevalence regions. Speci city was estimated to be 99%, but a range was not tested as all false positive cases were assumed to be referred and subsequently diagnosed correctly. However, false positive cases incur costs, both to the potential immigrant and to the health system. In particular, it may be possible to attain a high sensitivity but low speci city by setting a low threshold for investigation of suspicious skin lesions. In this case, the cost balance between less false negatives but more false positives will be relevant for policy.

Screening for leprosy in immigrants 247 Finally, this analysis considered the number of cases detected amongst a cohort of 100,000 immigrants. This is a considerable number of immigrants, but does apply to countries with a large migrant intake, noting that only a proportion would come from leprosy endemic areas. This analysis also assumed that all prospective immigrants would undergo normal screening procedures and thus did not include emergency entrants, such as refugees, who may be allowed entry to and require treatment in the destination country regardless of screening results. However, the early identi cation and treatment of active leprosy by screening would reduce the sequelae of the disease in these immigrants. There are several case reports in the literature of leprosy occurring amongst immigrants in developed nations. 1,2,4,7 Some authors support screening programmes to reduce `imported' leprosy, 4,10,11 whilst others acknowledge the dif culties arising from a long latency period and possibly low sensitivity of screening. 6 The number of leprosy cases identi ed before migration in high prevalence countries is comparable with the number of cases of active TB averted in screening models examining low TB prevalence populations, 26 but fewer than in models examining higher TB prevalence groups or when secondary transmission of TB is included. 26 Secondary transmission is not considered relevant in leprosy, as few multibacillary cases are predicted and transmission in industrialized counties appears very rare. 3 This analysis outlines a set of assumptions regarding leprosy screening and uses available data to estimate the number of cases that would be detected by a leprosy screening programme of immigrants. It indicates that screening would only produce signi cant case-yield amongst immigrants from regions or social groups with high leprosy prevalence. Based on this study, routine leprosy screening would have a very small impact on case incidence. Leprosy is not highly contagious and can now be effectively treated. The number of immigrants to Australia from countries of higher endemnicity is small. Acknowledgements The authors would like to thank the following leprosy experts for their valuable advice on the natural history, diagnosis and management of leprosy. Dr Warwick J. Britton, Associate Professor, Department of Medicine, Faculty of Medicine, The University of Sydney, and Mycobacterial Research Group, Centenary Institute for Cancer Medicine & Cell Biology, Sydney. Dr Andrew Lloyd, Associate Professor, In ammation Research Unit, School of Pathology, The University of New South Wales, and Department of Infectious Diseases, The Prince of Wales Hospital, Sydney. Mr Neville Board RN, NSW Health Department, Sydney. References 1 Ong AKY, Frankel RI, Maruyama MH. Cluster of leprosy cases in Kona, Hawaii: impact of the Compact of Free Association. Int J Lepr, 1999; 67: 13±18. 2 Zastrow KD, SchoÈneberg I. Lepraeinschleppungen nach Deutschland 1981±1992. Gesundh Wes, 1993; 55: 414±417. 3 Wolfe MS. Tropical diseases in immigrants and internationally adopted children. Med Clin N Am, 1992; 76: 1463±1479. 4 Younger B, Michaud RM, Fischer M. Our Southeast Asian refugee experience. Arch Dermatol, 1982; 118: 981±984.

248 R. Taylor et al. 5 Janssens PG. La pathologie exotique importeâe par voyageurs et immigrants. Bruxelles-MeÂdical, 1970; 5: 361±369. 6 World Health Organization. Global leprosy situation in 1998. WHO 1999, http://www.who.int/lep/12c.htm. 7 Goldstein GB, Reid JC, Keo L. A review of refugee medial screening in New South Wales. Med J Aust, 1987; 146: 9±12. 8 Vanbunyen P, Leese J, Lockwood DNJ. Leprosy in England and Wales. Communicable Diseases and Public Health, 1999 2:119±122. 9 Lockwood DNJ, Reid AJC. The diagnosis of leprosy is delayed in the United Kingdom. Q J Med, 2001; 94: 207±212. 10 Lloyd AR, Boughton CR. Immigrant health screening. Med J Aust, 1987; 146: 609. 11 Wiessman AM. Preventive health care and screening of Latin American immigrants in the United States. JAm Board Fam Practice, 1994; 7: 310±323. 12 World Health Organization. Chemotherapy of leprosy for control programs. Technical report series no. 675. WHO, Geneva, 1982. 13 World Health Organization. 7th WHO expert committee on leprosy. WHO, Geneva, 1997. 14 World Health Organization. Progress towards leprosy elimination. WHO Wkly Epidemiol Rec, 1997; 23: 165±172. 15 Gellert GA. International migration and control of communicable diseases. Soc Sci Med, 1993; 37: 1489±1499. 16 Gupta SC, Humne AY, Ingole DL. A rapid survey technique for detection of leprosy. Ind J Lepr, 1990; 62: 488±491. 17 Department of Health & the Welsh Of ce. Memorandum on leprosy. Publication no. 11057 HP 10K 2P JUL97 SA. Department of Health, Wetherby, 1997. 18 Ridley D, Jopling W. Classi cation of leprosy according to immunity: a ve group system. Int J Lepr, 1966; 34: 255±273. 19 Ponnighaus J, Fine P, Sterne J et al. Long-term active surveillance of leprosy suspectsðwhat are the likely returns? Lepr Rev, 1993; 64: 25±36. 20 Fine P, Sterne J, Ponnighaus J et al. Household and dwelling contact as risk factors for leprosy in northern Malawi. Am J Epidemiol, 1997; 146: 91±102. 21 Ponnighaus J, Fine P, Sterne J et al. Incidence rates of leprosy in Karonga District, North Malawi: Patterns by age, sex, BCG status and classi cation. Int J Lepr Other Mycobact Dis, 1994; 62: 10±23. 22 Murray C, Lopez A. A compendium of incidence, prevalence and mortality estimates for over 200 conditions. Harvard University Press, Boston. World Health Organization, Harvard School of Public Health, World Bank, 1996. Global health statistics, Global Burden of Disease and Injury series, vol. II. called 1996b) 23 Department of Immigration and Multicultural Affairs. Population ows: immigration aspects. Economic and Environment Section, Migration Branch, Department of Immigration and Multicultural Affairs, Commonwealth of Australia, 1999. 24 Department of Immigration and Multicultural Affairs. Humanitarian program. Offshore program delivery. Refugee and Humanitarian Branch, Department of Immigration and Multicultural Affairs. Monthly Summary: March 1999. 25 Murray C, Lopez A. A comprehensive assessment of mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020. Harvard University Press, Boston. World Health Organization, Harvard School of Public Health, World Bank; 1996. Global health statistics, Global Burden of Disease and Injury series, vol. I. called 1996a). 26 Schwartzman K, Menzies D. Tuberculosis screening of immigrants to low-prevalence countries. Am J Respir Crit Care Med, 2000; 161: 780±789.