Evaluation of effect of social determinants and of impact of social programs on incidence, frequency of disability and treatment outcomes of leprosy

Lead Research Organisation: London School of Hygiene & Tropical Medicine
Department Name: Epidemiology and Population Health

Abstract

Leprosy continues to be a disease of public health importance in Brazil and many other low and middle income countries. Not much is known about the social epidemiology of leprosy nor the effectiveness of social policies on this poverty-related disease.

We propose a new way to investigate wich aspects of poverty increase the risk of leprosy, and how much Bolsa Familia ( a conditional cash transfer programme) introduced by the Brazilian government helped to decrease the risk of leprosy, reduced the risk of disability, and increased the proportion of cases of leprosy that completed treatment.

We are going to do this by using data already collected for administrative purposes. The main database is "Cadastro Único para Programas Sociais" (CadÚnico), used for screening applications for social programmes, which contains data on a range of socioeconomic indicators for 100 million people. We will link this to two databases: of benefits received, and to notification of leprosy. This will be done protecting the identification of individual data with strict security measures approved by the ethical committee. The unprecedented size of this cohort will allow the investigation of effects of social conditions (education, income, housing, family structure, ethnicity, water supply, sanitation) and of impact of interventions (income received, duration of support) on different aspects of leprosy in subpopulations defined by type of disease, age, geography, income, gender, housing, etc..

The project will produce evidence of the impact of social interventions and social deprivation on leprosy. The study will advance our understanding of leprosy and will strengthen the case for implementation of social interventions- in addition to biomedical interventions-to promote effective control of leprosy and pave the way for evaluation of impact on other poverty related diseases.

Technical Summary

Leprosy remains a public health problem in Brazil. The epidemiology of leprosy is not fully understood: little is known about risk factors for infection, for disease, for delayed diagnosis and for treatment interruption. Large social inequalities and poverty are historical characteristics of Brazil; major improvements in the last 2 decades, with considerable expansion of programs in education, housing, social security and social development, make this an ideal setting for investigation.
We propose a novel strategy to investigate the effect of social determinants and the impact of the Brazilian cash-transfer program "Bolsa Familia" (BFP), on incidence, frequency of disabilities and treatment outcomes of leprosy by linking records from the database, "Cadastro Único para Programas Sociais" (CadÚnico), of applications for social programs, with detailed socio-economic information on over 100 million individuals, to two databases: individual payments of BFP, and the leprosy component of the National Notification Database, with information on notifications, disability, treatment and outcomes for all cases of leprosy. Statistical methods will include standard statistical methods for longitudinal data to estimate the effect of social determinants and regression discontinuity design and propensity scores to investigate the impact of BFP on notifications, disability, treatment and outcomes for all cases of leprosy.

The project will produce strong and specific evidence of the impact of social interventions and social deprivation on leprosy, advance our understanding of leprosy and will strengthen the case for implementation of social interventions- in addition to biomedical interventions- to promote effective leprosy control. There is a strong drive for structural social protection interventions for poverty related infectious diseases and evidence to support these policies can play an essential role in progress towards implementation.

Planned Impact

The study will assess the effect of social and economic factors and the impact of the conditional transfer programme "Bolsa Familia" (BFP) on incidence, frequency of disabilities and treatment outcomes of leprosy in Brazil, overall and in sub populations. The study is based on routinely collected data. The Cadunico, the Brazilian database of applications for social programmes, containing detailed socio-economic information on over 100 million individuals, will be linked to the records of all individual payments of BFP, and the leprosy component of the National Notifiable Diseases Database (SINAN), with detailed information on all cases of leprosy. By producing specific and scientifically sound evidence of the effects of social deprivation and of the impact of social interventions on leprosy burden in Brazil, this study i) will advance our understanding of leprosy and ii) will strengthen the case for implementation of social interventions- in addition to, and in synergy with, biomedical interventions-, and provide information to support ( or adjust) policies aimed at socially vulnerable segments of the population, and their impact on effective leprosy control.

The institutions involved, LSHTM, Oswaldo Cruz Foundation (FIOCRUZ-Brasilia), Tropical Medicine Group of University of Brasilia (NMT/UNB) and Institute of Collective Health of Federal University of Bahia (ISC/UFBA), are recognized institutions in the field of public health research with international reputation for excellence. FIOCRUZ, UNB and ISC/UFBA signed a technical-scientific cooperation agreement with the Brazilian Ministry of Social Development (MSD), to develop a platform for impact evaluation of BFP and other social public policies on a wide spectrum of outcomes in health, education, work.

The well established collaboration between these institutions, LSHTM, and the Brazilian Ministries of Social Development and of Health, by periodic meetings and a continuous flow of communications, will maintain the steady interest of the policy makers on this research, and effectively contribute to the debate on type and extent of social interventions to be implemented in vulnerable communities in order to relieve the burden of leprosy.The multidisciplinary team is composed of researchers from LSHTM, UNB, FIOCRUZ, ISC/UFBA and UFF, including epidemiologists and statisticians with experience in epidemiology of infectious diseases and specifically leprosy, and design, management and interpretation of big data studies, and researchers experts in programming models and tools for high performance and data-intensive (big data) computing applied to Bioinformatics and Health. Post-doctoral researchers in epidemiology, statistics and information technology will be trained in the project. Other researchers, policy makers, directors of leprosy control programs and patients will benefit from the evidence produced.
Periodical meetings between UK and Brazilian researchers will be held, during which steps and achievements of the Project will be discussed and evaluated. Technical reports on state of progress of the Project and updated findings will be released each year ,and disseminated through governmental and non-governmental media.

This study will contribute to the advancement of the on going debate in the scientific community on the relationship between distributive social policies and infectious diseases of poverty around the world, with a specific focus on leprosy.
The results will be presented to the Brazilian National Leprosy Control Programme coordination and the coordination of MSD, with the aim of promoting a discussion at national level on the drive for integrating social programmes and leprosy control programmes, and to provide evidence for supporting/informing decision making related to social support and social programmes impact on leprosy, in order to ensure better life conditions and a better health for people vulnerable to or affected by leprosy.

Publications

10 25 50

publication icon
De Oliveira GL (2021) Estimating underreporting of leprosy in Brazil using a Bayesian approach. in PLoS neglected tropical diseases

 
Description This project strengthened the collaboration between LSHTM and CIDACS-Fiocruz, supported training in causal inference and large-scale data analysis, and provided new insight into the epidemiology of Hansen's Disease in Brazil. Some of our key achievements included:
1. A systematic review (https://doi.org/10.1371/journal.pntd.0006622) of 39 articles describing socioeconomic risk markers of leprosy in high-burden countries, which found evidence of associations between leprosy risk and increased age, poor sanitary and living conditions, lower educational levels, and food insecurity. A further meta-analysis indicated that leprosy risk was elevated among males, manual laborers, individuals suffering from food shortages, being a household contact of a leprosy patient, and living in a crowded household (=5 per household).
2. An analysis (https://doi.org/10.1016/S2214-109X(19)30260-8) of the social determinants of leprosy in Brazil using data on 23,899,942 individuals including 18,518 patients with leprosy, which demonstrated risk of leprosy varied geographically (i.e., individuals residing in regions with the highest poverty in the country (central-west, north, and northeast regions) had a risk of leprosy incidence five-to-eight times greater than did other individuals) and in relation to socioeconomic conditions (i.e., decreased levels of income and education and factors reflecting unfavourable living conditions were associated with an up to two-times increase in leprosy incidence).
3. An analysis (https://doi.org/10.1371/journal.pntd.0007714) investigating geographic and socioeconomic factors associated with leprosy treatment default among 20,063 new leprosy cases followed as part of the 100 Million Brazilian Cohort between 2007 and 2014. In total, 5.0% of the leprosy patients defaulted from multidrug therapy. Among the associated factors, we found that having residency in the North and Northeast of Brazil, black ethnicity, low familial income, lack of formal electricity, and a high household density were associated with higher odds of leprosy treatment default.
4. A quasi-experimental study (https://doi.org/10.1016/S1473-3099(19)30624-3) of the effect of a conditional cash transfer programme on leprosy treatment adherence and cure. Our results suggest that being a beneficiary of the conditional cash transfer programme, which facilitates cash transfers and improved access to health care, is associated with greater leprosy multidrug therapy adherence and cure, especially among patients with multibacillary leprosy.
Exploitation Route Overall, our research highlights how socioeconomic development and social protection policies have the potential to improve the control of neglected tropical diseases, such as Hansen's Disease.

Our findings contribute to the evidence base for policymakers. Specifically, our research suggests that strategies targeting high-risk populations to reduce leprosy transmission and prevent progression towards potentially stigmatising disabilities should prioritise individuals and families living in precarious situations with low income. In addition, it shows that programmes that mitigate poverty might bolster leprosy control and should be considered essential tools for helping countries to achieve the goals outlined in the WHO Global Leprosy Strategy 2016-2020.

Our research may also stimulate further academic investigations. Our results show the public health potential of using large-scale linked administrative datasets to study the effect of social policies on the outcomes of rare diseases. Additionally, it highlights the need for further research to identify the specific mechanisms by which participation in social protection programmes improves infectious disease outcomes and whether this is generalizable to other diseases treated with long-term multidrug therapies.
Sectors Communities and Social Services/Policy,Healthcare,Government, Democracy and Justice

 
Description The findings from this project have the potential to improve the social, economic, and health outcomes of individuals affected by leprosy in Brazil. Using large-scale data from the 100 Million Brazilian Cohort, this project was able to elucidate how the risks of leprosy and of leprosy treatment default are closely associated with socioeconomic deprivation and to identify high-risk groups for targeted intervention. Encouragingly, the findings of this project also showed that investments in social protection policies, such as the Bolsa Familia conditional cash transfer program, have the potential to improve the treatment and cure rates following leprosy diagnoses. Improvements to leprosy treatment achieved through social development and welfare policies, in turn, reduces individuals' risks of developing potentially stigmatizing disabilities and improves their social and economic engagement, resulting in a virtuous cycle that improves the quality of life for those affected and reduces the risks of onward transmission.
Sector Communities and Social Services/Policy,Healthcare
Impact Types Societal,Economic

 
Description Estimating the public health burden of arbovirus-related complications inBrazil: A population-based big data project
Amount £119,042 (GBP)
Funding ID 527418645 
Organisation British Council 
Sector Charity/Non Profit
Country United Kingdom
Start 02/2020 
End 02/2022
 
Description Post doctoral funding - Cooperacao CAPES - FIOCRUZ - Brazil sem Miseria
Amount R$ 65,600 (BRL)
Funding ID Cooperacao Capes FIOCRUZ number 1658894 
Organisation Oswaldo Cruz Foundation (Fiocruz) 
Sector Public
Country Brazil
Start 11/2016 
End 04/2018
 
Title development of tested linkage alggorithyms 
Description We are testing two different linkage tools, specifically developed for this Project, to link cleaned SINAN-Leprosy database with CADU database from 2007 to 2015, containing around 114 million records. One of them, called Buscador, uses an indexing method on the larger database (CADU). Each registry from SINAN-Leprosy is searched in the indexed CADU based on the five identification variables, returning, for each individual in SINAN-Leprosy, a list of 1,000 similar individuals. A weight system is thn applied to generate a similarity measure, which is used it to select the most similar candidate. The other tool, called AtyImo, adopts a two-step probabilistic data linkage, and applies Bloom filters to ensure anonymization of sensitive data, blocking techniques to ensure that large datasets can be efficiently mapped to hardware with different memory and processing capacity, and Dice calculation to perform matching decisions. We performed tests of linkage of SINAN_Leprosy 2010 records from two Brazilian states, Bahia (3260 registries) and Sergipe (490 registries), with CADU database from 2007 to 2015. Using Buscador, approximately 49% leprosy registries of Bahia were linked (ROC curve area [Sensitivity/Specificity]: 0.942[0.846/0.891]), and, for Sergipe, approximately 39% registries were linked (ROC curve area [Sensitivity/Specificity]: 0.963 [0.913/0.921]). Using AtyImo, 55.3% of leprosy registries from Sergipe were linked, with a similar accuracy as found with Buscados (ROC curve area [Sensitivity/Specificity]: 0.969 [0.908/0.914]). We have just linked the entire SINAN-Leprosy database using Buscador and we expect to find approximately 40% of linked pairs, with a similar accuracy. Formal validation tests on both linkage procedures will be performed using proportional samples for all Brazilian states. Machine learning techniques with AtyImo are also being developed, to provide the complete accuracy results. An exploratory analysis on SINAN-Leprosy database alone (i.e., not linked to CADU, is being conducted, on socioeconomic and clinical characters of leprosy cases (all and new), by year, by type of outcome and by disability level separately at enrolment and discharge. 
Type Of Material Improvements to research infrastructure 
Year Produced 2017 
Provided To Others? Yes  
Impact This work is shared to all grous working in data linkage in CIDACS we plan to publish the code as open access 
 
Title a new database with over 300000 records of notifications of leprosy matched to applications/receipt of social benefits 
Description In November 2016 we obtained the individual databases of leprosy registries for the whole country (SINAN-Leprosy), from 2007 to 2015, from the Brazilian Ministry of Health. The databases report a full set of information concerning the patient's demographic and socio-economic characteristics, patient's mode of detection and discharge, clinical and bacteriological data, and data on treatment and disabilities. Cleaning procedures were applied to the original 378,215 SINAN-Leprosy records, regarding, firstly, the identification variables, which are used in the probabilistic linkage of the SINAN-Leprosy records with Cadastro Unico (CADU) database individual records. Errors in the patient's name and patient's mother's name were identified using Spark-Phyton. Non alpha-numeric characters and numbers were excluded. Other errors with identifiable patterns were corrected using a second identifier- a variable containing the patient's first and last name. We then proceeded to identify multiple registries of the same individual during the study period, which could affect the linkage procedure. Assuming that individuals carrying the same identification variables (name, mother's name, birth date, sex and place of residence) were one individual only, we found 15,176 patients with multiple registries. In short, using a decision tree based on date of registry and type of entry (new case, relapse or case transfer), we grouped in wide format multiple records of the same individual, or picked up randomly just one record when not enough information was available. The final database comprises 364,858 individuals, over the 9 years of study.We have developped two lingkage algorithyms: Buscador and altymo. We performed tests of linkage of SINAN_Leprosy 2010 records from two Brazilian states, Bahia (3260 registries) and Sergipe (490 registries), with CADU database from 2007 to 2015. Using Buscador, approximately 49% leprosy registries of Bahia were linked (ROC curve area [Sensitivity/Specificity]: 0.942[0.846/0.891]), and, for Sergipe, approximately 39% registries were linked (ROC curve area [Sensitivity/Specificity]: 0.963 [0.913/0.921]). Using AtyImo, 55.3% of leprosy registries from Sergipe were linked, with a similar accuracy as found with Buscados (ROC curve area [Sensitivity/Specificity]: 0.969 [0.908/0.914]). We have just linked the entire SINAN-Leprosy database using Buscador and we expect to find approximately 40% of linked pairs, with a similar accuracy. Formal validation tests on both linkage procedures will be performed using proportional samples for all Brazilian states. Machine learning techniques with AtyImo are also being developed, to provide the complete accuracy results. We have just linked the entire SINAN-Leprosy database using Buscador and we expect to find approximately 40% of linked pairs, with a similar accuracy. Formal validation tests on both linkage procedures will be performed using proportional samples for all Brazilian states. Machine learning techniques with AtyImo are also being developed, to provide the complete accuracy results. We do not expect all pairs to link as not all cases of leprosy applied to scial benefits, 
Type Of Material Database/Collection of data 
Provided To Others? No  
Impact The data is necessary for our research on social determinatsof leprosy and leprosy utcomes and the impact of social policies, but the next two years CIDACS will be open to provide anonymized datasets on requests form policymakers and researchers interested in exploring subgroups. 
 
Description Centro de Integração de Dados e Conhecimentos para Saúde (CIDACS), Fundacao Oswald Cruz, Bahia, Brazil 
Organisation Oswaldo Cruz Foundation (Fiocruz)
Country Brazil 
Sector Public 
PI Contribution Our research team has contributed expertise in infectious disease epidemiology, analysis of large-scale electronic health data, and arthropod-borne viruses.
Collaborator Contribution CIDACS is a world class data center based in Brazil with institutional expertise in data linkage and impact evaluation. Our collaborations utilize data from the CIDACS 100 Million Brazilian Cohort, which links electronic health and social records on more than 114 million individuals in Brazil. In addition, CIDACS will lead new data linkage efforts to wrangle data related to hospitalization records. This project is possible due to the cutting-edge infrastructure at CIDACS, which includes secure data handling facilities and high-performance computing equipment.
Impact The current project has greatly strengthened our collaboration, which began in 2017, and led to a successful spin-off grant focused on Zika virus that will allow us to continue collaborating over the next 7 years. To date, outputs of the current project include ethics preparation, manuscript preparation and publication (n=3 so far), scientific exchanges, and discussions of future funding applications. Through our collaboration, we have a total of four grants together, 15 publications as well as joint activities, icnluding presentations of the findings, stakeholder meetings, and scientist exchanges.
Start Year 2020
 
Description Computational Sciences Dept of the Faculty of Mathematics and Statistics of the universidade Federal da Bahia 
Organisation Federal University of Bahia
Country Brazil 
Sector Academic/University 
PI Contribution This is now a well established collaboration with the Computational Sciences Dept of the Faculty of Mathematics and Statistics of the universidade Federal da Bahia within the conetxt of CIDACS and the 100 000 000 cohort We defined the conditions for data cleaning, defined the specification for the linkage, worked with them in joint meetings in which progress was discussed, and participated in the planing and interpreting the validation of the linkage in collaboration with the computational sciences team and the collaborators in the Maths and Stats Faculty.
Collaborator Contribution They conducted computerized cleaning and data preparation of 378,215 SINAN-Leprosy records, usingSpark-Phyton, excluding duplicates ;they produced of two linkage algorithyms, Buscador and AtyImo, in the contxt of the 100 000 000 cohort programme; they linked a sample of the Leprosy and cadu data for validation. They have very recently completed the linkage of the whole Leprosy and CADU data using Buscador, and are conducting preliminary checks.
Impact Computerized cleaning and data preparation of 378,215 SINAN-Leprosy records, usingSpark-Phyton. Production of two linkage algorithyms, Buscador and AtyImo. Linkage of a sample of the Leprosy and cadu data for validation Linkage of the whole Leprosy and CADU data using buscador, preliminary checks being conducted.
Start Year 2016
 
Description Departamento de estatistica, Instituto de Matematica e Estatistica-UFBa 
Organisation Federal University of Bahia
Country Brazil 
Sector Academic/University 
PI Contribution We work with the staistical team to defined the parameters for the validation study and conttibuted to the interpretation. We participated in the discussion of cut off points for defining two records as a "match". We We wrote the plan of anaysis to be discussed with them.
Collaborator Contribution Tere are two professors of statistics that work with us in the programme. So far, they lead workshops on statistical approaches for evaluation of interventionas for the team ( al the team working in the projects in the 100 000 0000 cohort) and conducted the validation study of the two approaches used by the IT team for linage as well as participating in the discussion of cut off points for defining two records as a "match".
Impact Plan of analysis Evaluation of linkage
Start Year 2016
 
Description Sao Paulo School of Economics, Fundação Getulio Vargas, Sao Paulo, Brazil 
Organisation Getulio Vargas Foundation
Country Brazil 
Sector Charity/Non Profit 
PI Contribution Our research team has contributed expertise in infectious disease epidemiology, analysis of large-scale electronic health data, and arthropod-borne viruses.
Collaborator Contribution Our partners are contributing expertise in health economics and impact evaluation.
Impact Outputs of this multidisciplinary collaboration include scientific exchange, completion of a systematic review, and manuscript preparation.
Start Year 2020
 
Description University of Glasgow & Centro de Integração de Dados e Conhecimentos para Saúde (CIDACS) 
Organisation University of Glasgow
Country United Kingdom 
Sector Academic/University 
PI Contribution Our team will be sharing some of our linked datasets produced by our MRC grant and best practices on data management and evaluation of social policies on leprosy and tuberculosis.
Collaborator Contribution Our collaborators at the University of Glasgow provide expertise related to the application of multilevel modelling to health data, particularly concerning its use as a tool for exploring inequalities, and expanding the uses of routinely collected and linked hospital discharge, mortality and register data.
Impact Planned future analyses include projects on the impact of alternative social protection programs (e.g., Minha Casa, Minha Vida housing program) on the incidence of leprosy in Brazil.
Start Year 2016
 
Description Conclusion of Project: Stakeholder Meeting (Salvador, Brazil) 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Policymakers/politicians
Results and Impact The presentations about the social determinants of Hansen's Disease and the impact of social welfare programs on the incidence and treatment of Hansen's Disease sparked important discussions with policymakers, patients, non-profit advocacy groups, and researchers about future efforts to reduce the burden of Hansen's Disease in Brazil.
Year(s) Of Engagement Activity 2019
URL https://www.yumpu.com/pt/document/read/62721980/seminario-de-entrega
 
Description Presentation at X Congresso Brasileiro de Epidemiologia 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Professional Practitioners
Results and Impact Dr. Julia Pescarini provided an update on leprosy in Brazil and its association with severe disabilities at the 10th biannual Congress of the Brazilian Society of Epidemiology in Florianopolis, Brazil, in October 2017.
Year(s) Of Engagement Activity 2017
URL https://proceedings.galoa.com.br/epi
 
Description Seminar at CIDACS - Centro de Integração de Dados e Conhecimentos para a Saúde,- Fiocruz BA at Tecnocentro, no Parque Tecnológico da Bahia. 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Other audiences
Results and Impact This was a two day event with the launch and visit to instalations of the Centro de Integração de Dados e Conhecimentos para a Saúde, in the first day and scientific presentations and discussions in the second day launch of scientific activities with a seminar aimed at many disciplines with international and national speakers. Themes ranged from ethics of BIG DATA research to methodoloical challenges of evauation of impact of policies.
Year(s) Of Engagement Activity 2016
URL http://rxrlnx01.eastus.cloudapp.azure.com/archive/201612
 
Description Talk and participation in Seminar for the Centre for Statistical Methodology "BIG DATA at LSHTM: who is doing what?" 14 March 
Form Of Engagement Activity A magazine, newsletter or online publication
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Other audiences
Results and Impact An open meeting of theLSHTM Centre for Statistical Methodology on "BIG DATA " to share knowledge and discuss methodological chalenges and solutions covering many disciplines using big data: -Omics, Pharmaco-epi, Environmental epi, evaluation of social policy and health care on health
Year(s) Of Engagement Activity 2017
 
Description UCL Grand Challenges: Global Health Day Conference 2017 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Postgraduate students
Results and Impact Dr. Elizabeth Brickley presented on "The Brazilian 100 Million Cohort: Innovation in Large-Scale Data Collection and Linkage" and served on a broader panel for Q & A on "The UK Contribution to Innovation in Global Health." She explained how the use of big data and linkage studies are giving us new opportunities to investigate relatively rare and persistently challenging diseases like leprosy.
Year(s) Of Engagement Activity 2017
URL http://www.ucl.ac.uk/grand-challenges/global-health/news/Global-Health-Day-2017-REVISED-programme
 
Description Workshop/Meeting with World Bank 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Third sector organisations
Results and Impact On 1 December 2017, Dr. Julia Pescarini presented her recent findings on the 'Effect of social determinants and impact of social programs on leprosy in the '100 million Brazilians' cohort' to an audience of World Bank representatives. This discussion informed global policy recommendations regarding the impact of conditional cash transfer programs on health outcomes.
Year(s) Of Engagement Activity 2017
 
Description Workshop/Meeting with the Bill & Melinda Gates Foundation 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Supporters
Results and Impact On 26 June 2017, in Salvador (Brazil), Dr. Julia Pescarini presented her recent work on "Evaluation of BF impact on leprosy in Bahia in the 100 million Brazilian cohort. A preliminary analysis using RDD approach" to program officers affiliated with the Bill and Melinda Gates Foundation.
Year(s) Of Engagement Activity 2017