Real-time Monitoring of a Circulating Vaccine-Derived Poliovirus Outbreak Immunization Campaign using Digital Health Technologies in South Sudan

"...the battle [to eradicate polio] is not over, as there are outbreaks of circulating vaccine-derived poliovirus (cVDPV), occurring mainly in places with low immunization coverage, most especially in conflict areas and remote communities experiencing huge migration..."

Although there is documented use of mobile and geographical information systems (GIS) technologies to support, guide, and monitor polio immunisation campaigns across different geographical settings, limited studies exist that cover the entire component of the campaign (pre-, intra-, and post-implementation) and that take place in complex and humanitarian settings like South Sudan. Such settings require more robust and flexible solutions due to factors such as insecurity, lack of communication and internet in many areas, and high numbers of displaced persons (e.g., 1.6 million internally displaced persons (IDPs) in South Sudan and 2.2 million South Sudanese refugees). This study details development of a tool to monitor and evaluate vaccination campaign processes in real time using open data kits (ODKs) that are linked to an interactive dashboard. It also evaluates the impact of the tool on the timeliness and completeness of data reporting and on its ability to provide real-time information to stakeholders at all levels so they can take informed action while increasing transparency and accountability.

In response to the circulating vaccine-derived poliovirus (cVDPV) type 2 outbreak declared in South Sudan on September 18 2020, the country conducted the first round of polio vaccination campaigns in six states and 44 counties, with monovalent type 2 oral polio vaccine (mOPV2) administered to children under the age of five. The ministry of health (MOH) alongside other partners organised the campaign. For example, the United Nations Children's Fund (UNICEF) supported vaccine management, logistics, communication, and social mobilisation.

Pre-campaign activities started a month ahead of the campaign, with each state updating and sharing the status of implementation of activities (e.g., community mobilisation and sensitisation). A Google spreadsheet was used to update and monitor the status of implementation by states, which could be viewed at the national level to decide if a state or county is ready to implement. With the support of the World Health Organization (WHO), UNICEF, and MOH, technology training was conducted at all levels (national, state, county, and payam). A total of 285 supervisors and monitor teams were trained, using mobile smartphones and tablets, on ODK Collect configuration, installation, and downloading of the e-checklists and were assigned to areas (counties) where they would supervise and monitor the campaign.

Intra-campaign monitoring took place during the four days of the campaign, with supervisors being allocated to payams. Each supervisor completed a team supervision checklist electronically using the ODK checklist for every team he/she supervised, conducted inside and outside monitoring, and filled in the appropriate checklist using a simple random sampling methodology to select ten households in each village supervised. The data collected were sent in real time to a central server hosted at the WHO Regional Office for Africa (AFRO) in Brazzaville. All checklists for the campaign were embedded with coordinates that gave the location of all places supervised and houses visited during the campaign. The post-campaign survey took place over two days independently by the CORE group immediately after the immunisation campaign, and the supervisors selected twenty households from within each cluster for which to apply the monitoring tool using ODK. ODK Collect was the application of choice due to its ability to cater to large data collection on mobile phones, more tenable to low-resource settings like Africa.

The article also details how data collection and management, as well as data analysis, work in this system. Some of the notable features include the fact that ODK Collect can work in both real-time and offline; hence, in places where the internet and communication posed challenges, the users saved the forms and later sent them to the server when internet connectivity became available. However, greater emphasis was placed on data being transmitted in real time to ensure that programme officers could monitor the progress of the campaign and take decisions where necessary. Power BI was employed to build reports and a dashboard, which was connected to the ODK server to fetch the data in real time, with a visualisation that was customised and filtered to highlight the status of the campaign and showcase coverage and the quality of the outbreak response. The dashboard report was then published as a web link, with all partners given access. For example, the report depicts the proportion of children with finger marked and children not finger marked (missed), with the reasons why they were missed. During the conduct of this campaign, there was a daily feedback meeting during which use of data from the dashboard allowed meeting participants to make decisions informed by real-time data.

It was found that, thanks to this system, the duration (time taken to receive data) of data transmission from the field to the national level improved for the different phases of the campaign (pre, intra, and post), with a P<0.001 for all the data sets compared between the paper-based and the electronic-based using ODK. The post-campaign data for lot quality assurance (LQA) surveyors also improved, with the final status of county acceptance level visible in real time, as compared to previous campaigns, where the data had to be downloaded, analysed before a presentation is made, and shared with all stakeholders. The real-time data has helped monitor places with missed children (including coordinates/location) to allow the team to plan for revisit of these missed children; these data can also be used to plan strategies that may ensure good coverage in these locations in subsequent rounds. The country is now able to meet the expected deadline for submitting LQAs data to the AFRO rapid response team within 24-48 hours after the campaign, as stipulated in the cVDPV outbreaks response guideline.

Furthermore, the tool facilitated real-time analysis of disaggregated data up to the payam and village level, especially in the case of the team supervision checklist, where details of each team supervised and the location where they worked are visible to all stakeholders. The tool also highlighted processes and challenges that the teams faced in the field and allowed for critical challenges to be communicated to the next level for prompt discussion and for corrective action to be taken on the spot. "One critical component that the tool addressed is the issue of transparency across all stakeholders and partners supporting the campaign due to improved access to data and information in real-time which foster communication, collaboration and follow up between different program managers and partners at the national and states level where decisions to guide the campaign are taken."

In conclusion: "In South Sudan, the implementation of this tool had helped the global polio eradication initiative (GPEI) partners and the ministry of health to identify achievements, gaps, and challenges in reaching children in the field, and support in identifying locations with missed children promptly to appropriately plan for means of reaching these children, in ensuring all children are reached with the polio vaccines." The researchers encourage further studies that compare the effects of this tool on the outcome of polio campaigns (coverage).