Trend of case detection and leprosy elimination in Brazil

Authors


Corresponding Author Maria Lucia Fernandes Penna, Instituto de Medicina Social, Universidade do Estado do Rio de Janeiro, Rua S. Francisco Xavier 524, 7 andar 20250040 Rio de Janeiro, RJ, Brazil. Tel.: 55 21 92327390; Fax: 55 21 22641142; E-mail: mlfpenna@terra.com.br

Summary

Only six countries did not meet the leprosy elimination target during 2005, amongst them Brazil. In 2006, the Brazilian Ministry of Health announced a reduction of the detection rate of 24% or 10 900 cases from 2004 to 2005. A negative binomial parabolic regression model was adjusted to the detection rate historical series from 1980 to 2004, in order to predict the 2005 detection rate and its 95% confidence interval. This analysis showed that the number of new leprosy cases for 2005 could not be predicted from the previous behaviour of the data what calls for an epidemiological or operational explanation hypothesis. The hypothesis that this drop in detected case number is due to operational change, as a reduction in diagnosis or a modification in the reporting routine, is more likely. Recent change in prevalence case definition turned the prevalence ratio a function of only one variable, the detection rate, as the duration of the diagnosed disease became fixed. In the early nineties, based on epidemiological data evaluation, the BMoH recognized the impossibility of reaching the elimination goal, but it committed to seek leprosy control. This position changed after some years. Leprosy Elimination is a strategy supported by the national and international public opinion. As a one for all recipe, it may cause unwanted effects for it is not flexible enough to deal with different epidemiological behaviours and public health traditions.

Abstract

Seuls six pays n'ont pas atteint le seuil d’élimination de la lèpre en 2005, parmi elles, le Brésil. En 2006, le Ministère brésilien de la S anté a annoncé une réduction de 24% du taux de dépistage soit 10900 cas, de 2004 à 2005. Un modèle parabolique binomial négatif de régression a été ajusté sur les séries historiques de taux de dépistage de 1980 à 2004, afin de prévoir le taux de dépistage de 2005 et son intervalle de confiance à 95%. Cette analyse a montré que le nombre de nouveaux cas de lèpre pour 2005 ne pouvait pas être prévu à partir du comportement précédent des données, ce qui requiert une hypothèse épidémiologique ou opérationnelle d'explication. L'hypothèse que cette baisse du nombre de cas détectés est due à un changement opérationnel, tel que la réduction du diagnostic ou la modification de la routine de report, est plus probable. La récente modification de la définition de la prévalence des cas a changé le rapport de prévalence en une fonction à une seule variable, le taux de détection, comme la durée de la maladie diagnostiquée devenait constante. Dans les années 90, basées sur l'analyse de données épidémiologiques, le Ministère brésilien de la Santé a reconnu l'impossibilité d'atteindre l'objectif d’élimination mais s'est promis de travailler sur le contrôle de la lèpre. Cette situation a changé après quelques années. L’élimination de lèpre est une stratégie soutenue par l'opinion publique nationale et internationale. En tant qu'une recette unique pour tous, elle peut causer des effets indésirables car elle n'est pas assez flexible pour traiter différents comportements épidémiologiques et traditions de santé publique.

Abstract

Durante el 2005 solo seis países no alcanzaron las metas para la eliminación de la lepra: uno de ellos fue Brasil. En el 2006, el Ministerio de Sanidad de Brasil anunció una reducción en la tasa de detección del 24% o 10,900 casos del 2004 al 2005. Con el fin de predecir la tasa de detección del 2005 y su intervalo de confianza del 95%, se ajustó un modelo de regresión parabólica binomial negativa para la tasa de detección de series históricas entre 1980 y 2004. Este análisis mostró que el número de casos nuevos de lepra para el 2005, no podía predecirse a partir del comportamiento anterior a los datos, requiriéndose de una hipótesis explicativa epidemiológica u operacional. La hipótesis más probable es que esta disminución en casos detectados sea debida a un cambio operacional, como una reducción en el diagnóstico o la modificación de la rutina al reportar. Cambios recientes en la definición de la prevalencia de casos cambió la razón de prevalencia a una función con una sola variable, la tasa de detección, a medida que la enfermedad diagnosticada quedaba fijada. A comienzos de la década de los noventa, basándose en una evaluación de datos epidemiológicos, el Ministerio de Sanidad del Brasil reconoció la imposibilidad de alcanzar el objetivo de eliminación, pero se comprometió a lograr el control de la lepra. Esta posición cambió después de algunos años. La eliminación de la lepra es una prioridad apoyada por la opinión pública nacional e internacional. Una estrategia de ‘‘todos para uno’’ puede traer efectos no deseados, ya que no es lo suficientemente flexible a la hora de manejar diferentes comportamientos epidemiológicos y tradiciones de la sanidad pública.

Introduction

Although current elimination strategies have greatly reduced the known prevalence of leprosy, the new case detection rate was not affected in most endemic countries due to ongoing transmission of Mycobacterium leprae; this means that the control of leprosy needs permanent efforts to assure early diagnosis and treatment (Durrheim & Speare 2003; Lockwood & Suneetha 2005).

In 1991, 10 years after the introduction of multi-drug therapy (MDT), WHO proposed at the 44th WHA the elimination to be achieved by the year 2000, reflecting an optimism similar to that which followed the discovery of dapsone in the 1940s. The leprosy elimination strategy was proposed based on the assumption that the cure of the known leprosy cases through MDT would greatly reduce transmission in addition to reducing the burden on health systems. Leprosy is considered eliminated if the known case prevalence is <1 per 10 000 inhabitants. Point prevalence is a frequency measure that does not include the time dimension behaviour of the disease. Its choice as the main measurement of impact certainly contributed to the lack of reasoning about how to maintain the low prevalence scenario after elimination. The inclusion of leprosy as a topic in the health policy is not feasible after its elimination. Who would support public expenditure on an eliminated disease?

After many years of use of MDT, there is no evidence of its impact on transmission. Better knowledge of transmission mechanisms is still needed (Britton & Lockwood 2004). Many parts of the world present high leprosy detection rate. The detected rate represents the flow of cases from the hidden prevalence to the known prevalence through diagnosis. It is then a function of the size of the hidden prevalence and the access to the diagnosis. The actual incidence rate and the detection rate are the main determinants of the hidden prevalence, as the leprosy cases mortality is close to the mortality of the general population. An operational reduction of leprosy cases detection will then increase the size of the hidden prevalence.

At the beginning of 2005, leprosy elimination objective was achieved in most of the world, with the exception of only nine countries, amongst them Brazil (WHO 2005). Early in 2006, the Brazilian Ministry of Health (BMoH) announced a reduction of 24.32% in the number of newly diagnosed leprosy cases over the previous 2 years, i.e. a reduction of 10 900 cases (MoH Brazil 2006a). The objective of this communication is to evaluate this reduction by studying the historical data on leprosy detection, and to discuss hypotheses that could explain this surveillance data.

Materials and methods

The Brazilian Ministry of Health provided the epidemiological data and the population estimates for the period (MoH Brazil 2006b). The numbers of newly diagnosed leprosy cases in every year from 1980 to 2004 were adjusted as a parabolic function of time in a negative binomial regression model, with the logarithm of the yearly population as an offset variable, using stata 9.0 (StataCorp. 2005). The observed number of newly diagnosed leprosy cases in 2005 was compared with the case number predicted by the model and its 95% confidence interval. The detection rate predict by the model was estimated until 2020.

Results

The estimates for the adjusted model can be seen in Table 1. For 2005, the model predicted 52 477.3 cases, with a 95% confidence interval of 57 544.9 to 47 856.3. The number of newly diagnosed leprosy cases computed by the BMoH was 38 410 or 73.2% of the predicted value and 80.3% of the 95% confidence interval the lower limit. The model predicted a detection rate per 100 000 inhabitants from 1980 to 2020; observed detection rates from 1980 to 2005 are presented in Figure 1.

Table 1.   Parameters estimated for the negative binomial regression
Variable| CoefficientSEzP > |z|[95% CI]
  1. Likelihood-ratio test of α = 0: chibar2(01) = 3272.43 P ≥ chibar2 = 0.000.

t0.06088820.00818857.440.0000.0448389, 0.0769375
t2−0.0009770.0003073−3.180.001−0.0015793, −0.0003747
_constant−9.0859240.0461312−196.960.000−9.176339, −8.995508
ln (population)(offset)    
ln (α)−5.2757190.2848699  −5.834054, −4.717385
α0.00511430.0014569  0.0029262, 0.0089385
Figure 1.

 Observed and predicted yearly detection rate of leprosy cases.

Discussion

The Brazilian leprosy detection rate has been showing a clear ascending trend since 1980. Our decision to adjust the historical series to a parabolic function allows for alterations of the ascending trend in time. The adjusted model shows that the increasing rate of the number of detected cases relative to the population size is been reduced in time (a positive estimate for t and a negative estimate for t2), but a decreasing trend is predicted only after 2010.

The negative binomial distribution is a Poisson distribution with over dispersion. The use of this statistical approach to deal with heath surveillance data is well established (Esteve et al. 1994; Simonsen et al. 1997). Our analysis shows that the number of new leprosy cases for 2005 is an outlier, i.e. it could not be predicted from the previous behaviour of the data. This calls for an epidemiological or operational explanation hypothesis.

It is difficult to attribute this fact to an epidemiological event, since leprosy is a chronic disease and its epidemiological behaviour is not expected to suffer a big change in a short period of time. A possible explanation for this sudden decrease of the detection rate could be the reduction of the hidden prevalence through a previous increase in diagnosis coverage. But this hypothesis is not consistent with the data, for it would be expected that the reduction would follow an increase in the detection rate.

The hypothesis that this drop in detected cases number is due to operational change, as a reduction in diagnosis and/or a modification in the reporting routine, seams more likely. The criteria for maintaining leprosy patients in the active registry have recently been changed by the BMoH, and are now based on the diagnosis date. Paucibacillary cases stay in the file for a maximum of 6 months and multibacillary cases stay for 12 months, after which period patients are discharged automatically (Datasus 2006). This change in prevalence case definition turns the prevalence ratio a function of only one variable, the detection rate, as the duration of the disease is fixed. Therefore the elimination goal can only be achieved through the reduction of the detection rate.

India also presented a reduction of the detection rate before reaching the elimination goal during 2005 what justified critics as ‘health authorities could meet that target only through artful manipulation of the definition of leprosy and by avoiding active case detection.’ (Mudur 2005). It is worth remembering that amongst the countries quoted by WHO in 2005 – Angola, Brazil, Central African Republic, Democratic Republic of the Congo, India, Madagascar, Mozambique, Nepal and Tanzania – as those not meeting the goal of elimination of leprosy as a public health problem, Brazil and India were those with the strongest economies and the ones with stable democratic governments. Certainly, a second failure of a public health challenge can have a bad repercussion in the public opinion.

In the early 1990, based on the available epidemiological data evaluation, the BMoH recognized the impossibility of reaching the elimination goal, but it committed to seek leprosy control using the political impact of the elimination target (MoH Brazil 1992). But policy transfer from the international level to national level uses strategies of marketing and brand sale (Ogden et al. 2003). As DOTS for tuberculosis control, ‘Leprosy Elimination’ is a brand with a strong power on international public opinion. As a one for all recipe, it may cause unwanted effects since it is not flexible enough to deal with different epidemiological behaviours and public health traditions.

The Global Strategy for Further Reducing the Leprosy Burden and Sustaining Leprosy Control Activities (WHO, 2006b)has evolved from the WHO Strategic Plan for the Elimination of Leprosy, and reintroduces the concept of leprosy control, broadens the indicators to be monitored, at the same time that encourages six countries – Brazil, the Democratic Republic of the Congo, Madagascar, Mozambique, Nepal and the Tanzania – to reach the elimination goal (WHO 2006). This document also emphasizes that new tools need to be developed to permit an eradication strategy and that early case detection and treatment remain the cornerstones of leprosy control.

The operational reduction of case detection as happened in Brazil causes concern. Not only can it jeopardize the epidemiological achievements of leprosy control, but it may also reduce the national and international political commitment to it. This sudden reduction in the detection rate, if not well evaluated, could be misinterpreted as an indicator of competence to quickly control leprosy in the process of or after elimination, resulting in a trade of the known prevalence for the hidden prevalence. This reduced political commitment is already reflected in the lack of research funding (Scollard 2005), although the mapping of M. leprae’s genome (Cole et al. 2001) had opened a research opportunity for answering scientific questions that remain unsolved, and for developing new intervention tools.

The objective of reducing the known prevalence through treatment/cure is ethically justified, but as other international recommendations (Naafs 2006), the exact target for prevalence ratio to be achieved must be based on solid scientific evidence.

Ancillary