Vaccination coverage survey versus administrative data in the assessment of mass yellow fever immunization in internally displaced persons—Liberia, 2004
Introduction
Yellow fever (YF), a vaccine-preventable viral hemorrhagic disease, causes infection in approximately 200,000 persons annually and is responsible for an estimated 30,000 deaths per year [1]. The vast majority of disease and mortality occurs within the YF belt (latitude 15° north to 10° south) in Sub-Saharan Africa [1]. In West Africa, the virus is transmitted in three cycles – a sylvatic cycle in which transmission occurs between forest-dwelling mosquitoes and non-human primates, an intermediate cycle in which transmission occurs between mosquitoes and both non-human primates and humans in moist savanna areas of Africa, and an urban cycle where it can cause large epidemics [1]. Urban cycle epidemics develop from anthroponotic, also known as human-to-human, transmission in which humans serve as the sole host reservoir of the peridomestic Aedes aegypti mosquito vector. Urban epidemics occur when anicteric but viremic persons who are not yet severely ill travel from jungles and savannas to cities where they infect local A. aegypti mosquitoes, a species that is abundant in urban areas and in areas where humans store water. When YF is identified in any setting, the likelihood that it resulted from human-to-human transmission or its possible introduction into an urban setting must be rapidly assessed to determine the need for emergency vaccination.
Immunization coverage of the at-risk population is an important determinant for human-to-human transmission because the potential for an epidemic increases when there is low prevalence of neutralizing antibody to YF virus from previous vaccination or naturally acquired infection. The 17D YF vaccine is highly effective and safe; Over 400 million estimated doses have been administered worldwide, with rare reports of serious adverse events following immunization (AEFI) [2], [3], [4], [5], [6], [7], [8], [9], [10], [11]. Monitoring AEFI during mass vaccination campaigns are important to ensure not only the timely identification of possible events which may signal compromised safety of the vaccine, but to detect potential programmatic errors that may occur when a large number of doses are administered over a short time period.
In Liberia, 14 years of civil war during 1989–2003 devastated much of the country's healthcare infrastructure and severely disrupted public health disease surveillance and immunization programs. On February 13, 2004, the World Health Organization (WHO) declared an outbreak of YF after laboratory confirmation of four cases. All cases had illness onset January 1–9, 2004, of which three were fatal. Two cases occurred in men aged 19 and 26 years living in densely populated internally displaced person (IDP) camps, settlements of citizens displaced from their in-country homes because of civil strife, in Salala District, Bong County, in central Liberia. Roughly 365,000 Liberians, one-sixth of the country's estimated population, had lived in IDP camps throughout the country since the end of the civil war in August 2003 [12]. Because of the potential for human-to- human transmission in overcrowded IDP camps with extension into the surrounding area, YF mass vaccination campaigns were launched in Salala District IDP camps and neighboring communities during February 26 to March 6, 2004 and March 16 to March 20, 2004.
Administrative data, a simple formula used to estimate vaccination coverage that divides the number of persons vaccinated during the campaigns by the number of vaccine-eligible persons estimated residing in the camps, indicated that, of the estimated 87,000 persons living in these camps, only 49,395 (56.8%) were immunized against YF with the 17D YF vaccine (Institute Pasteur, Dakar, Senegal) during the campaigns [13], [14], [15], [16]. Because validation studies to determine IDP camp population size had not been performed, a coverage survey was needed to more precisely evaluate vaccination coverage after the campaigns. We report findings of a rapid vaccine coverage survey that underscores the importance of accurately estimating the population at risk for YF to assist decisions regarding future vaccination strategies.
Section snippets
Methods
The IDP camps in Salala District of Bong County were built in 2002 and administered by the Liberian government through the Liberia Refugee Repatriation and Resettlement Commission (LRRRC). Medical services were coordinated by Médecins Sans Frontières – France (MSF-F) and food was supplied by the World Food Programme (WFP). Camp shelters were uniformly designed, each measuring approximately 4 m× 5 m (Fig. 1). A census of the camps had not been performed. Instead, a WFP formula of five persons per
Results
One of the 260 shelters did not exist. Of 259 shelters that were visited, 22 were either unoccupied for three successive days during the survey (15 shelters) or were incomplete shelters that were not yet inhabited (seven shelters). Data were analyzed for 237 survey respondents (one person per shelter). There were no missing data from IDPs surveyed and no AEFI reported.
The median number of household members living in a shelter was four (range, 1–8). We estimated that 69,536 persons lived in
Discussion
Over the past two decades, YF epidemics have regularly occurred in West Africa [21]. In response to an outbreak in Liberia in 2004, the WHO, UNICEF, and MSF conducted an emergency vaccination campaign in large camps of IDPs after two of the initial four cases of yellow fever were confirmed among camp residents. In assessing the need for additional vaccination, administrative data suggested that less than 60% of potentially at-risk IDP camp inhabitants were vaccinated during a mass YF
Ethics requirement
The study was programmatic research in response to an emergency public health outbreak and therefore not subject to an ethical review board by CDC, WHO, UNICEF, or Medicins Sans Frontieres.
Acknowledgments
The authors wish to thank Angela Kearney of UNICEF and Lynn Thomas of USAID for their assistance and guidance in this study. We value the time and cooperation from all the survey team members and IDPs who participated in the study.
Authorship and Contributorship: All authors – G. Huhn, J. Brown, W. Perea, A. Berthe, H. Otero, G. LiBeau, N. Maksha, M. Sankoh, S. Montgomery, A. Marfin, M. Admassu – can take responsibility for the content of the paper and the conception and design of the study.
References (40)
Yellow fever: an update
Lancet Infect Dis Aug
(2001)- et al.
Possible association of encephalitis and 17D yellow fever vaccination in a 29-year-old traveller
Vaccine
(1993) - et al.
Hepatitis and death following vaccination with 17D-204 yellow fever vaccine
Lancet
(2001) - et al.
Serious adverse events associated with yellow fever 17DD vaccine in Brazil: a report of two cases
Lancet
(2001) - et al.
Fever and multisystem organ failure associated with 17D-204 yellow fever vaccination: a report of four cases
Lancet
(2001) - et al.
Effects of yellow fever vaccination
Lancet
(2001) - et al.
Effects of yellow fever vaccination
Lancet
(2001) - et al.
Effects of yellow fever vaccination
Lancet
(2001) - et al.
Mass immunization campaigns and quality of services
Lancet
(1990) - Robertson S. Yellow fever: the immunological basis for immunization. Document...
Fatal myeloencephalitis following yellow fever vaccination in a case with HIV infection
J Med Assoc Thai
Impact of alternative approaches to accelerated measles control: experience in the African Region, 1996–2002
J Infect Dis
Use of administrative data to estimate mass vaccination campaign coverage, Burkina Faso, 1999
J Infect Dis
Benefit versus risk factors in immunization against yellow fever
Dev Biol Stand
Should yellow fever vaccine be included in the Expanded Program of Immunization in Africa? A cost-effectiveness analysis for Nigeria
Am J Trop Med Hyg
Surveys to measure programme coverage and impact: a review of the methodology used by the Expanded Programme on Immunization
World Health Stat Q
Cited by (35)
Understanding the health needs of internally displaced persons: A scoping review
2021, Journal of Migration and HealthCitation Excerpt :However, even in insecure conflict contexts where IDPs are periodically inaccessible, research shows that vaccination interventions among IDPs and local populations can work. Key factors include intensive social mobilization, collaboration with key community and agency partners, including grassroots organisations and mass media, well-trained health staff, additional human resources, appropriate vaccine delivery systems and accurate estimates of target populations, as well as the utility of negotiations with opposition armed forces to permit vaccination in remote areas (Lam et al., 2015; Huhn et al., 2006; Abubakar et al., 2015). Integrating vaccination services into other core interventions such as nutrition could also increase uptake among IDPs, as shown in South Sudan (Oladeji et al., 2019).
Lessons learned in the implementation of supplementary immunization activity (SIA) field guidelines for injectable vaccines – Experiences from Tanzania
2020, VaccineCitation Excerpt :The target population which formed the denominator, was projected by the National Bureau of Statistics (NBS) from the 2012 census [12]. Errors in the estimation of target population could be a result of complex computations that carry uncertainty because of incorrect assumptions that are sometimes made on population migration. [13,14,15] These assumptions are made on fertility and mortality which may have a greater impact at younger and older ages
Coverage and acceptability of cholera vaccine among high-risk population of urban Dhaka, Bangladesh
2014, VaccineCitation Excerpt :Evidence shows that a survey for assessing the vaccine coverage is a better option than review of the vaccination records [13] as it helps to ignore over-reporting and/or under-reporting [14]. Conducting a coverage survey with a subsample of population is useful to determine the consistency or accuracy of an administrative report [15]. On the other hand, vaccine campaigns may succeed or fail depending on the various behavioural and cultural issues, such as perceptions about disease, vaccines, local systems, acceptable practices, incentives, and considerations in vaccine delivery [16].
Routine childhood vaccination programme coverage, El Salvador, 2011-In search of timeliness
2014, VaccineCitation Excerpt :Administrative coverage is calculated dividing the number of vaccine doses administered by the number of children in the target age-group. These calculations are affected by data quality issues, both regarding the denominator (e.g., outdated census) and the numerator (e.g., incomplete/incorrect reporting) [9,10], and do not include information on exact age of vaccination, making it impossible to assess timeliness of administration. Especially in low- and middle-income countries, vaccination coverage surveys are used to assess coverage and compare it with administrative data, while answering specific questions to guide programme strategies [11,12].
Validity of vaccination cards and parental recall to estimate vaccination coverage: A systematic review of the literature
2013, VaccineCitation Excerpt :The two main methods used to obtain vaccination coverage data are service delivery administrative tallies and community-based household surveys [5]. While administrative tallies are routinely collected, they are often compromised by constraints in staffing, supply of vaccine cards or registers, consistency of supervision, and capacity to transmit data to higher levels, and are subject to both numerator bias (e.g., inaccurate counting, incomplete reporting, including doses given to children >12 months of age) as well as denominator bias (e.g., poor population estimates, out-dated census data, imprecise population growth correction, population movement) [6–10]. Lim et al. compared administrative data from 193 countries to household surveys conducted during the same time period and found administrative figures grossly over estimate coverage compared to survey estimates [11].