# Agglomerative hierarchical cluster analysis and temporal trend of leprosy indicators in Brazilian states, 2012-2022

## TL;DR

This study uses clustering to analyze leprosy trends in Brazilian states from 2012 to 2022, identifying clusters with distinct epidemiological and socioeconomic patterns.

## Contribution

The novel use of agglomerative hierarchical clustering with multiple variables to analyze leprosy dynamics in Brazil.

## Key findings

- Five clusters of leprosy were identified, with Cluster 4 showing the highest new case detection rate and stable trends.
- Clusters 1 and 3 had the highest grade 2 disability rates, indicating late diagnosis.
- Cluster 4 had the highest proportion of multibacillary cases and Diforma clinical form.

## Abstract

Leprosy, a neglected tropical disease caused by Mycobacterium leprae, presents significant public health challenges in Brazil due to its slow progression, dermato-neurological manifestations, and potential for disability. Understanding leprosy’s spatial distribution and temporal trends is important for effective control and elimination strategies.

This study aimed to identify clusters of leprosy in Brazilian states using agglomerative hierarchical clustering and to analyse their temporal trends from 2012 to 2022.

An ecological study was conducted using data from the National System of Notifiable Diseases (SINAN). The agglomerative hierarchical clustering method was used to group states using the new case detection rate (NCDR) of leprosy per 100,000 inhabitants, the proportion of new cases of leprosy with grade 2 physical disability at the time of diagnosis (G2R), and the Gini index, a measure of socioeconomic inequality. Temporal trends within the clusters were assessed using Prais-Winsten regression analysis.

In the period 2012-2022, 293,030 new cases of leprosy were reported in Brazil. Five distinct clusters were identified. Cluster 4, comprising Mato Grosso and Tocantins, had the highest NCDR and stable temporal trends (APC: 3.2%, 95% CI: -0.1%, 6.7%). Clusters 1 and 3 had the highest proportions of grade 2 disability, indicating late diagnosis. Clusters 4 and 5 had the lowest percentages of individuals with incomplete/complete higher education (7.6% and 7.4%, respectively). Cluster 4 had the highest percentage of individuals with the Diforma clinical form (69.8%) and with cases classified as multibacillary (84.5%).

The use of agglomerative hierarchical clustering, a novel application of a non-supervised algorithm in this context, highlighting the integration of multiple epidemiological and socioeconomic variables for a better understanding the dynamics of leprosy transmission in Brazil. Significant variations in the spatial distribution and temporal trends of leprosy were observed across Brazilian states. To improve leprosy surveillance and control in Brazil, targeted interventions are needed, particularly in high-endemicity regions with late diagnosis.