Tuberculosis in the aftermath of the 2010 earthquake in Haiti

Abstract Problem In 2010, Haiti sustained a devastating earthquake that crippled the health-care infrastructure in the capital city, Port-au-Prince, and left 1.5 million people homeless. Subsequently, there was an increase in reported tuberculosis in the affected population. Approach We conducted active tuberculosis case finding in a camp for internally displaced persons and a nearby slum. Community health workers screened for tuberculosis at the household level. People with persistent cough were referred to a physician. The National Tuberculosis Program continued its national tuberculosis reporting system. Local setting Even before the earthquake, Haiti had the highest tuberculosis incidence in the Americas. About half of the tuberculosis cases occur in the Port-au-Prince region. Relevant changes The number of reported tuberculosis cases in Haiti has increased after the earthquake, but data are too limited to determine if this is due to an increase in tuberculosis burden or to improved case detection. Compared to previous national estimates (230 per 100 000 population), undiagnosed tuberculosis was threefold higher in a camp for internally displaced persons (693 per 100 000) and fivefold higher in an urban slum (1165 per 100 000). With funding from the World Health Organization (WHO), active case finding is now being done systematically in slums and camps. Lessons learnt Household-level screening for prolonged cough was effective in identifying patients with active tuberculosis in this study. Without accurate data, early detection of rising tuberculosis rates is challenging; data collection should be incorporated into pragmatic disease response programmes.


Introduction
On January 12, 2010, Haiti sustained a devastating earthquake. Total damages are estimated at 7.8 billion United States dollars, more than 120% of the country's 2009 gross domestic product. 1 Over 1.5 million people lost their homes and about 279 000 remained internally displaced in tent camps nearly four years later. 2 Even before the earthquake, Haiti had the poorest economic and health indices in the region of the Americas. 3 Haiti also had the highest tuberculosis incidence in the Americas (230 per 100 000 population in 2010) -nearly 10-fold higher than the regional incidence of 30 per 100 000 -and higher than the overall incidence of the world's 22 high-burden countries (166 per 100 000). 4 About half of the tuberculosis cases in Haiti occur in the West Department that includes Port-au-Prince and that was most heavily affected by the earthquake. Government buildings and health-care centres were destroyed, including the building that housed the National Tuberculosis Program, the two largest tuberculosis sanatoria and many clinics. Though most clinics resumed services within months, thousands of patients were initially dispersed in camps without tuberculosis medication.
Typically, tuberculosis rates remain stable in the immediate aftermath of natural disasters. [5][6][7] In Haiti, however, the internally displaced persons camps were crowded, the sanitation was poor, the children were chronically malnourished and the duration of residence was often prolonged. 8 The physical, social and economic damage to an already limited health-care infrastructure worsened the risk for acquiring active tuberculosis, hampered surveillance and challenged the public health response. The National Tuberculosis Program worked to open tuberculosis clinics as quickly as possible and the Haitian Group for the Study of Kaposi's Sarcoma and Opportunistic Infections (GHESKIO) conducted active case finding in a camp and slum adjacent to the clinic using cough as an indicator for tuberculosis. Studies done over the prior decade have shown that about one-third of patients presenting for human immunodeficiency virus (HIV) testing with chronic cough had active tuberculosis. 9,10 At the largest treatment centre in Haiti, the annual number of tuberculosis cases has more than doubled since the earthquake. The increase was first noted in paediatric patients. In 2010, 242 children younger than 10 years were diagnosed with tuberculosis, compared to 72 in 2009, a 336% increase. Fiftytwo percent of cases were children younger than two years and 33% were two to five year-olds. Clinicians were concerned that this rise in paediatric cases indicated an increase in ongoing transmission from adults to children, as children are likely to develop active disease soon after exposure. Higher numbers of paediatric cases were followed by a progressive rise in adolescent and adult tuberculosis cases, from 1026 in 2009, to 2719 in 2014 (Fig. 1). Meanwhile, the total number of patient visits increased by only 5%, from 246 276 in 2009 to 258 089 in 2013.
Lessons from the field Tuberculosis in Haiti Serena P Koenig et al.
screened camp residents in their tents and referred those reporting a cough of more than two weeks' duration for physician evaluation with smear microscopy and a chest radiograph. From July 2010 to June 2011, 282 patients were evaluated and 34 diagnosed with pulmonary tuberculosis; 22 (65%) were sputum smear-positive (Table 1). Seven patients with tuberculosis were younger than eight years and five of these children had a parent sharing the same tent subsequently diagnosed with active tuberculosis. The estimated tuberculosis incidence in the camp was 693 cases per 100 000 person-years, about three times the 2010 World Health Organization estimate for Haiti. 3 In August 2011, using the same strategy, we expanded active tuberculosis case finding into a section of the Cité de Dieu slum, which has about 10 000 residents. From August 2011 to August 2013, 1420 patients were evaluated and 233 diagnosed with pulmonary tuberculosis; 183 (79%) were sputum smear-positive ( Table 1). The estimated tuberculosis incidence was 1165 cases per 100 000 person-years, more than five times WHO's 2010 estimate for Haiti. 3,4 Active case finding efforts are now being expanded to over 220 000 people in slums and camps in Port-au-Prince, with a grant from WHO, through a collaboration between the National Tuberculosis Program, the largest treat-ment centre and the Haitian National Laboratory. Community health workers are conducting systematic door-to-door surveillance to identify people with persistent coughs and refer them to a tuberculosis clinic for diagnosis and treatment. The targeted population meet WHO's criteria for systematic tuberculosis screening: poor socioeconomic conditions and a high prevalence of tuberculosis. 12 Data from this surveillance will be evaluated independently by WHO and will provide further evidence of the burden of tuberculosis in Haiti.

Changes in reported tuberculosis
Despite the loss of their offices, the National Tuberculosis Program officers continued their tuberculosis reporting system after the earthquake. Each tuberculosis clinic in Haiti provides quarterly data that the National Tuberculosis Program tallies and submits to WHO. In 2010, the National Tuberculosis Program reported a 4% (from 14 861 to 14 222) decrease in tuberculosis cases nationwide and 9.5% (from 6489 to 5871) decrease in the West Department. This apparent decline is probably due to incomplete reporting after the earthquake. Compared with 2009 (n = 6489), the number of cases in the West Department increased by 7% (n = 6944) in 2011 and 17% (n = 7596) in 2012. The number of cases nationwide increased by 19% from 2011 to 2013, from 14 315 to 17 040 ( Fig. 1); preliminary data indicate a slight decrease to 16 400 cases in 2014. 4 Though HIV infection is an important risk factor for the development of tuberculosis, the increase in tuberculosis cases is not explained by changes related to HIV. The proportion of tuberculosis patients who are HIV-infected has remained stable at 20% and the estimated  prevalence of HIV has been stable at 2.2%. 3,4,13 Haiti's response to HIV is likely to be responsible for the lack of increase in HIV-associated tuberculosis after the earthquake. Although many HIV clinics were damaged in the earthquake and testing declined in 2010, the annual number of HIV tests rose above pre-earthquake levels by 2011 and 86% of tuberculosis patients were tested for HIV in 2013, compared to the regional average of 69%. 4,14 Within months, 90% of the pre-earthquake patients had resumed antiretroviral therapy, reducing their risk of contracting tuberculosis. 15

Need for improved data
With only one source of tuberculosis data for Haiti, we cannot be certain whether there is a change in prevalence, case detection or a combination of the two. There were no structural changes during this time adequate to explain an increased case detection rate. Efforts to expand active case finding had been localized and intermittent, there were no major changes in the network of tuberculosis laboratories and the proportion of new pulmonary tuberculosis cases that are bacteriologically-confirmed has not changed substantially (66% in 2010, 64% in 2011, 65% in 2012 and 68% in 2013). 4 Better data on the epidemiology of tuberculosis in Haiti are needed to understand the true burden of disease. Tuberculosis prevalence has never been assessed from a population-based survey in Haiti and there are no recent tuberculosis mortality data from a national vital registration system. WHO is therefore left with incomplete data on which to base its tuberculosis estimates for Haiti and as a result, confidence intervals are wide; estimated tuberculosis prevalence in Haiti is 129 to 421 per 100 000, compared to 30 to 48 per 100 000 for the Americas. 3,4 This makes short-term changes in tuberculosis rates difficult to detect. The National Tuberculosis Program is planning a thorough review of tuberculosis surveillance data. The inclusion of tuberculosis mortality data would also be useful. Periodic surveys can measure trends and guide expansion of public health interventions, including active case finding.

Discussion
Available data point to a rise in the number of reported tuberculosis cases in Haiti after the earthquake, but data are too limited to determine if this is due to an increase in tuberculosis burden or to improved case detection. Although it is clear that active case finding efforts have identified additional patients with tuberculosis, the scale of these activities thus far does not account for this increase in cases. These findings demonstrate that tuberculosis should be monitored in post-disaster settings when the affected population lives in conditions conducive to tuberculosis transmission (Box 1). To improve tuberculosis control, better data are needed to detect changes, track trends and efficiently direct a response that focuses scarce resources on the populations with the highest tuberculosis burden.
In implementing active case finding activities, we learned that screening for chronic cough in slums and camps is effective in diagnosing patients with active tuberculosis in a post-disaster setting. 9,10 This approach, using trained community health workers to identify patients with chronic cough, is efficient, low-cost and feasible for scale-up. Future steps will include expansion of these activities in Haiti and training of community health workers to provide diagnostic and treatment services for other diseases, a strategy which is promoted by the Haitian Ministry of Health to improve the health of the population at large. ■

Acknowledgements
The project was supported in part by NIH grant number 3 U2R TW006896-04S1. SK is also affiliated with the Division of Global
Enfoque Se llevaron a cabo detecciones de casos activos de tuberculosis en un campamento para personas desplazadas de la ciudad y en un barrio pobre cercano. Los agentes de salud comunitarios investigaron los casos de tuberculosis a nivel doméstico. Las personas con tos persistente fueron derivadas a un médico. El programa nacional contra la tuberculosis continuó su sistema nacional de informes sobre la tuberculosis. Contexto local Incluso antes del terremoto, Haití tenía el mayor número de casos de tuberculosis de todo el continente americano. Alrededor de la mitad de los casos de tuberculosis se encontraba en la región de Puerto Príncipe.