Setting
Borno State is located in North East of Nigeria with a land mass of 75,481 square kilometer and a population of about 6.4 million people (2019 population census projections) [11]. The state shares an international boundary with Niger Republic, Chad and Cameroon. It shares interstate borders with Yobe (to the West), Adamawa (to the South), and Gombe (to the South West). Borno has 27 LGAs comprising a total of 311 political wards, out of which 201 are accessible for immunization and surveillance activities, and were fully accessible at the time of the ORI. The 2018 DHS indicated that 46.5% of children aged 12-23 months were vaccinated with MCV in Borno [4].
Since 2009, an armed insurgency has been ongoing in Borno State. This has resulted in the displacement of communities and the disruption of health services due to destruction of health facilities and out-migration of health personnel. The last weeks of December 2018 witnessed increased attacks on civilian populations by armed groups and resulted in a massive population displacement from security-compromised areas, including inaccessible settlements (communities), into the metropolitan LGAs of the state. By January 2019, more than 700,000 people were living in overcrowded camp-like settings, significantly increasing the risk of epidemics and infectious diseases. As a result, approximately 20% of the population of Borno State are living in IDPs camps in hosted communities [12].
Case reporting and definition of outbreak and cases
In Nigeria, measles case-based surveillance is integrated with the polio acute flaccid paralysis surveillance structure and includes four national measles serological laboratories capable of testing on specimens from suspected measles cases for immunoglobulin (Ig) M antibodies as described elsewhere [13]. The surveillance performance is monitored regularly using performance monitoring indicators (i.e. proportion of LGAs with at least 1 suspected measles case reported with a blood specimen in a year, proportion of reported suspected cases from whom blood specimen is collected, non-measles febrile rash illness rate, and incidence of confirmed measles per million population). The measles surveillance system is sufficiently sensitive to identify and confirm measles cases. With one exception (i.e. incidence of confirmed measles per million population), the annual targets for the core measles surveillance performance indicators have been met since 2017 [14]. In brief, the surveillance network runs from the settlements and wards through the LGAs and States to the National level. The network consists of disease surveillance and notification officers (DSNOs) in all the LGAs in the country with State Epidemiologists at the State level. It is therefore institution-based and consists of formal and informal healthcare delivery points (reporting sites) and non-formal health care providers (informants). Additional to the installed surveillance network, a community-based surveillance makes use of community informants (i.e. traditional birth attendants, patent drug vendors, traditional healers) to report suspected cases.
A measles outbreak was defined as the occurrence of three or more confirmed measles cases in a unit (i.e. health facility/LGA/community with an approximate catchment population of ≥ 100,000) in a month [15][8]. We reviewed measles outbreak investigation reports and measles cases line-listed by the DSNO, WHO Borno State office and the rapid response team (RRT) deployed to Borno. Data collected included name, address (i.e. rural vs urban, LGA, ward, settlement/community), age, sex, date of rash onset, vaccination status, date of last vaccination, date of specimen collection and, final classification of cases as at epidemiologic week 9. Additional data was collected from the measles case based surveillance database regarding date of rash onset, age, vaccination status, lab-confirmed – positive serologic test for measles IgM antibody, epidemiologic linkage, and clinically compatible cases. WHO case definitions for measles were used for final classification of cases [16].
Measles Outbreak Response Immunization – ORI
To improve population immunity and interrupt the large measles outbreak, the ORI was conducted in 2 phases. An objective of the ORI was to reach the underserved children aged between 6 months and 71 months. Key elements considered by NPHCDA to determine the target population and scope of the vaccination activities included routine vaccination coverage, age-specific of lab-confirmed cases, absolute number of cases (suspected and lab-confirmed) and, results of the measles risk assessment. Phase 1 targeted children in 8 wards of the state capital of Maiduguri Metropolitan Council (MMC) from the 21st to the 25th of March 2019.
Phase 2 was implemented from the 14th to the 21st of May 2019, and included accessible wards of 13 LGAs out of the 27 LGAs of Borno States. In addition to conventional vaccination teams, special teams were used during the response. These teams, with the support of local vigilante and military, were deployed during the 2nd phase of the campaign to safely operate in security compromised areas, areas with migrant and other special populations (i.e. nomadic, IDPs in camps and those embedded host community). The processes of engagement of security personnel to access security-compromised areas was described elsewhere [17]. A total of 37 special teams were used to reach underserved children in partially accessible and/or security compromised areas of the LGA.
Estimation of measles vaccine effectiveness (–VE)
Measles VE was calculated using the screening method according to the following formula: VE = 1-((PCV/(1-PCV)) * (1-PPV)/PPV); where PCV is the proportion of cases vaccinated and PPV is the proportion of the population vaccinated. The screening method estimates VE by comparing vaccination coverage in cases of a disease (PCV) with the vaccination coverage in the population from which the cases are derived (PPV). PCV was calculated from the outbreak investigation reports and the measles surveillance database, and PPV was computed from the 2018 PCCS report. Evaluation of VE was assessed in children 9 months to 71 months of age for a more accurate estimate of the VE. Infants below the age of 9 months were not scheduled for measles vaccination and were excluded from the VE estimation. Likewise, cases reported from IDP camps were excluded from the VE estimation as the PCCS did not capture coverage from IDP camps.
Sensitivity Analysis
We conducted a sensitivity analysis where we compared the VE for lab-confirmed cases with the VE for epi-linked cases. Similarly, as the original VE calculation, we did not considered cases reported from IDP camps nor vaccination of under 9 months. The sensitivity analysis was performed to assess the effect of final classification of cases on the MCV VE estimate.
PCCS – Sample design
We designed a quantitative survey assessment of the ORI after completion of the 2nd phase of the ORI. The survey was a two-stage cluster household survey. A unique sample design was applied in the Borno PCCS. Due to absence of a recent sampling frame and limited funds to conduct household listing, the aggregate of ward micro-plans, which contained the number of settlements in a ward and the estimated population per settlement, was used as the sampling frame for this survey. The settlement was randomly selected as the primary sampling unit and individual households as secondary sampling units. The primary sampling frame consisted of 1,044 settlements from which 120 primary sampling units (settlements) were selected in the 12 accessible LGAs at the time of the survey. The secondary sampling involved the selection of households within each settlement by using systematic random sampling. All the households in a settlement were visited, and eligible households with children aged < 9 months to 71 months identified and line listed formed the settlement sampling frame. We calculated the sampling interval by dividing the total number of eligible households in the settlement with the desired cluster household size (15). Subsequently, we randomly selected a starting point on the line list and all subsequent households after addition of the sampling interval. Therefore, a sample of 15 households were randomly selected in each settlement providing a total of 1,800 households in 12 out of the 13 LGAs involved in the ORI (one LGA [Biu] was not surveyed due to insecurity).
PCCS – Data Collection
We collected data using Open Data Kit software running on android mobile phones. On completion of the household roster, only age-eligible respondents were present for interviewing. Information on the children receiving measles vaccination (i.e. children aged < 9 to 71 months with SIA cards and recall) including vaccination coverage during the measles campaign were collected. Design weights were calculated as the product of inverse probabilities of selection in the first and second stages. Data collection was conducted between 7th and 19th of July 2019.
PCCS – Data Cleaning and Analysis
We used the SIA module of Vaccination Coverage Quality Indicators software running on Stata version 15 (StataCorp. 2017) for data cleaning and analysis. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC) [18]. Results are based on the weighted data to account for the survey sampling design and non-response. Wilson’s 95% confidence intervals and upper and lower confidence bounds have been computed.