Yellow fever vaccination coverage following massive emergency immunization campaigns in rural Uganda, May 2011: a community cluster survey

Yellow fever is an acute vaccine preventable viral haemorrhagic disease found mainly in tropical regions of Africa and America. It is caused by a Flavivirus which is transmitted by mosquito bites and has a natural reservoir in non-human primates. The virus is endemic in tropical areas of Africa and Latin America, with a combined population of over 900 million people [1]. The “yellow” in the name refers to the jaundice that affects some patients. Approximately 200,000 cases and 30,000 deaths occur annually especially in Africa [2] where the case-fatality rate ranges from 15–50% [3]. The yellow fever virus can cause devastating epidemics of potentially fatal, hemorrhagic disease. The disease can be confirmed through, serological testing by way of ELISA for yellow fever virus-specific IgM or isolation of the virus from blood samples. These are the recommended standard diagnostic tests for yellow fever although PCR tests which detect the viral genetic material have been used [1, 4]. The number of yellow fever cases has increased over the past two decades due to declining population immunity to infection, deforestation, urbanization, population movements and climate change. This disease has no cure and treatment is symptomatic, aiming at reducing symptoms for the comfort of the patient and mass vaccination is the most important preventive measure [5]. Immunization of people at risk interrupts human to human transmission of the yellow fever virus especially when adequate vaccination coverage levels are reached. The vaccine is safe, affordable and highly effective, and appears to provide protection for 30–35 years or more [1]. The vaccine provides effective immunity within one week for 95% of persons vaccinated [6, 7]. Despite mass vaccination campaigns to prevent and control these outbreaks, the risk of major yellow fever epidemics, especially in densely populated, poor urban settings, both in Africa and South America, has greatly increased [8]. Consequently, yellow fever is considered an emerging, or reemerging disease of considerable importance.

In December 2010, the Ugandan Ministry of Health (MOH) declared a Yellow fever outbreak based on nine laboratory confirmed samples from the five districts of Abim, Agago, Kitgum, Pader and Lamwo in northern Uganda. This part of Uganda, home to game reserves and national parks, borders South Sudan and Kenya where previous cases of yellow fever have been reported [9, 10]. This was the largest ever recorded yellow fever outbreak in the country with an overall attack rate of 13 cases per 100,000 population based on confirmed cases [4]. A cumulative total of 181 confirmed cases with 45 deaths (CFR = 24.8%) were recorded, with no new cases reported since February 2011 [4, 11]. The CFR among males (29.6%; 32/108) was nearly twice that of females (17.8%; 13/73) and the ages of suspected yellow fever cases varied from 3 months to 83 years with a mean age of 28.2 years and a standard deviation of 17.5 years [4]. Previous yellow fever outbreaks in Uganda had fewer or single cases [12, 13]. For about 40 years now, no yellow fever cases had been reported in Uganda [14, 15]. However, larger outbreaks had been reported in East Africa, with attack rates (cases per 100,000 persons) of 6,800 in Sudan [9] and 27.4 in Kenya [16]. In addition, the yellow fever vaccine is not part of the Expanded Program for Immunization (EPI) vaccines given routinely especially to children under five. It is mainly people travelling abroad from Uganda that are required to receive the yellow fever vaccine.

The Ugandan Ministry of Health developed a national response plan prioritizing surveillance and laboratory confirmation, case management, social mobilization and health education, and reactive vaccination to contain the outbreak. The national response team included Ministry of Health of Uganda, World Health Organization (WHO), Uganda Red Cross society (URCS), United Nations International Children Emergency Fund (UNICEF), and the U.S. Centers for Disease Control and Prevention (CDC). Others were African Field Epidemiology Network (AFENET), RESPOND (Responding to emerging pandemics and threats), MSF-Holland, Uganda National Expanded Program on Immunization (UNEPI) and World Vision. This campaign was conducted between the 22^nd and 26^th of January 2011 and targeted all persons aged 6 months and above in the affected districts. Approximately one million people were targeted in the affected region.

Pader district is found in northern Uganda and this region borders South Sudan which is prone to sporadic yellow fever outbreaks [17]. In this district, residents lead a primarily agricultural lifestyle, growing crops and raising cattle, sheep, and poultry. Community settlements near bushes and forests are common since most people have just returned from internally displaced peoples camps after the war in northern Uganda. They venture into the forests mainly for hunting and harvesting bamboo for sale. Potential mosquito-breeding sites include baobab trees with tree holes, rock holes, and a water reservoir on the outskirts of villages. The district has a projected population of 205,898 [18] and the yellow fever attack rate was 2.9 cases per 100 000 population with females being more affected than males.

During the response, 177 vaccination posts were created, 354 health workers and 590 volunteers from 295 villages were identified and trained. Vaccination posts were mainly health facilities (both public and private not for profit), outreach sites for routine EPI vaccines and a few schools. During the campaigns, Uganda Red Cross society (URCS) supported community-based activities such as social mobilization, transportation of vaccines and provision of logistics. The URCS volunteers also supported organization of beneficiaries, issuing of vaccination cards and other support activities at the various vaccination posts in this district. The target population for vaccination in this campaign was all persons aged 6 months and above with a target vaccination coverage of at least 80%. After the campaign, the reported or administrative yellow fever vaccination coverage was 75.9% and this was below the targeted 80% [11]. Administrative yellow fever vaccination coverages in the other districts were: Abim 120.5%, Agago 81.6%, Kitgum 77.6% and Lamwo 73.9% [11]. When the target coverage is not reached, the potential for an epidemic increases since there is a low prevalence of neutralizing antibodies to the yellow fever virus from previous vaccination or naturally acquired infection.

Underreporting of vaccine doses administered usually results in underestimation of the coverage. It is important to conduct coverage surveys to validate reported administrative coverages so as to inform imminent (mop-up activities) and future immunizations strategies. Probability-based surveys such as cluster surveys can estimate vaccination coverage even if the size of the target population is not known [19, 20]. The aim of this survey was to estimate yellow fever vaccination coverage in Pader district and determine the reasons for non vaccination to recommend possible mop-up actions, guide future vaccination efforts and contribute to the control of yellow fever in the country.

This post campaign evaluation survey for yellow fever vaccination coverage was conducted between the 6^th and 12^th of May 2011. Of the 680 study participants, only 679 (99.8%) were included in our analysis. One respondent (0.2%) was excluded from further analysis on account of having an incompletely filled survey form. More than half of the respondents (54.6%) were females.

This post campaign survey, which was part of the outbreak response activities by the Ugandan Ministry of Health and WHO, was done to evaluate the vaccination coverage following an outbreak of yellow fever in northern Uganda. In Pader district, the overall estimate of yellow fever vaccination coverage was 96.1% with a card retention rate of 51.6%. The main reasons for not receiving vaccination during the mass campaign were: having travelled out of the district, lack of transport to vaccination posts and being sick during the campaigns. Others reasons were: being sick during the campaigns, being at school and rudeness of health workers at vaccination posts.

This overall vaccination coverage (by card and recall) in Pader district exceeded 90% and was above the desired minimum threshold coverage of 80% according to WHO [24]. This theoretically implies that a need to conduct an emergency follow up revaccination campaign was avoided since almost everybody (96.3%) was protected against yellow fever. When the desired immunization coverage is exceeded, the potential for human to human transmission reduces since a higher proportion of people at risk have circulating yellow fever antibodies. The other affected districts also reported high vaccination coverage estimates: Abim 95.4%, Agago 97.5%, Kitgum 97.8% and Lamwo 97.1%. A cluster survey done in internally displaced people’s camps in Liberia following mass yellow fever vaccination campaigns revealed comparably high coverages [25]. However, a high vaccination coverage doesn’t necessarily mean that the country is not at risk since Uganda borders South Sudan which experiences sporadic yellow fever outbreaks [17]. Secondly, yellow fever vaccination is not routinely given along with the routine EPI vaccines thus only people who travel out of the country might be protected against the disease. Indeed one case of yellow fever was reported in October 2012 in Agago district, northern Uganda and this was confirmed by the Epidemiology and Surveillance unit of Ministry of Health [26].

Our study found a difference between the survey and administrative coverage (75.9%) and this is in agreement with previous studies which suggest that the quality of administrative coverage data is unreliable [27, 28]. Such data, often times over or under estimates coverage due to wrong population estimates and incomplete tallying or reporting of vaccination doses. This is compounded by poorly documented shifts in population and reliance on imprecise or outdated census data [23, 25]. To our knowledge, this was the first reported cluster survey to evaluate yellow vaccination coverage in Uganda and no yellow fever outbreaks had been reported in Uganda in the last 40 years [14]. One possible reason for the discrepancy between administrative and our survey coverage is that the Ministry of Health during the immunization campaigns relied on old projected population estimates from the district population office. Secondly, the estimated target population used to calculate the administrative coverage might not have taken into account new residents, migrant workers and nomadic population.

Although, the yellow fever vaccination coverage exceeded 90% by self report (card or recall), only half of the respondents had cards as evidence for vaccination. A similar vaccination card retention rates was reported in a cluster survey for routine childhood EPI and yellow fever vaccination campaigns in West Africa [29]. In our study, we considered both card evidence and recall to estimate vaccination status. A previous survey showed that caregiver recall is highly sensitive (>90%) and fairly accurate in determining vaccination status [30]. Because of the high sensitivity of caregiver recall in estimating vaccination status, it is now acceptable practice both in developed and resource constrained countries to use caregiver recall in estimating immunization coverage [30, 31]. In our survey, we validated vaccination status by recall by also asking the site of vaccination during the campaigns. The low card retention rate after a short period of mass immunization (3 months) may suggests programmatic or logistical challenges in organizing vaccination activities. It is also possible that not all people vaccinated actually received their vaccination cards. Although there were unconfirmed reports of fake yellow fever vaccination cards being sold in clinics in Kampala, the capital city of Uganda, to Ugandans travelling abroad [32], no fake vaccination cards were found during our survey.

Our study revealed that the 5–15 years age category had the highest vaccination coverage. This is probably due to the fact that this age category is easy to mobilize since most of them are in schools. Nearly fifty percent of those not vaccinated had travelled out of the district possibly to engage in other gainful activities during the vaccination exercise. Furthermore, it is possible that adults are more likely to question the rationale of vaccination than children and this may influence their decision regarding vaccination. For children below 5 years, chances of being vaccinated largely depend on adults (caregivers). Others reasons for non vaccination from our study were lack of transport, sickness, long waiting time and rudeness of health workers. This is in agreement with studies done to assess the determinants of vaccination coverage in Nairobi Kenya and Niger [33, 34]. Such factors are major barriers to access of immunization and other health services by communities and thus may reinforce poor health seeking behaviors. This also reinforces the need for continuous staff motivation and regular supervision to overcome health system challenges to vaccination campaigns.

The mass vaccination campaigns against yellow fever in Pader district were successful since the estimated coverage exceeded the desired threshold of 80%. Consequently, mop up revaccination activities were avoided. However, active surveillance activities for early case detection of yellow fever are necessary since this area borders South Sudan which is prone to sporadic yellow fever outbreaks. Mass campaign evaluations should be done immediately or shortly after response activities in order to estimate vaccination coverage more accurately. In addition, more efforts should be put to ensure immunization card retention. In future, more vaccination posts should be established to increase access to immunization services by the population. Yellow fever survey tools could be modified to capture information on: reasons for poor card retention, religion, household size and distance from vaccination posts since such variables might affect vaccination status.