This paper provides new information on what might improve the quality and use of routine immunization data for decision making and accountability in a sub-national immunisation programme in Nigeria. As found elsewhere, the evidence shows that context matters in the implementation of quality improvement initiatives in immunisation programmes[12, 27], as will be discussed to highlight the contextual factors that shape acceptability, adoption, and feasibility the interventions and self-selected changes in immunisation data management practices and their implications for accountability within the immunization programme.
The findings show that the two components of the intervention, which were data review meetings and supervision, were acceptable and implementable. Nevertheless, the feasibility of both was low. The main driving factor for the feasibility is lack of funding. Our finding is consistent with existing evidence that inadequate funding undermines interventions to improve immunization data quality including data review meetings[20, 28], and data quality-focused supervision[20, 29–33]. The critical challenges to funding are lack domestic financing from government budget and dependence on external funding[34, 35]. Funds are required for transport and incentives for members of the supervision team as well as participants at the monthly data review meetings. From an accountability perspective, lack of funding limits the ability of immunization service providers to implement data review meetings and supervision, which support change in data management practices[36]. Also, lack of funding limits availability of incentives (transport refunds and refreshments) to service providers to participate in the data review meetings. Therefore, improving feasibility of data review meetings and supportive supervision requires predictable funding from government.
Despite being perceived as a satisfactory recording practice that can be implemented to ensure consistency of routine immunization data, resistance among health workers constrained the feasibility of ‘tally as you vaccinate’. Tallying appropriately is a rate-limiting step to recording of routine immunization data. The resistance among health workers was attributed to shortage of trained primary health care staff, long waiting time consequent on the practice, and unwillingness of service providers to use electronic data tool (open data kit). Similar resistance among health workers to interventions to improve recording of immunization data were reported in a previous study[32]. Even though a meaningful change to sustain ‘tally as you vaccinate’ is to increase staffing, this was not possible during the intervention because recruitment requires action beyond the health sector. Nonetheless, when staff strength improves, accountability can be demanded from resisting health workers. Nigeria’s RI accountability framework prescribes reward and sanction strategies for outstanding and poor performance[23].
The study found that ‘adherence to reporting timeline’ was an acceptable and considerably adopted practice. Nevertheless, the feasibility depended on resource availability notably funding, staffing, and information and communication technology. These findings align with results of a previous study that associate predictable funding, adequate staffing, and access to computers with timely reporting of routine health data[37]. In our study, lack of funds meant that heath workers use their personal money to pay transport to the headquarters or purchase internet access. Also, as previously stated, inadequate funding hindered data review meetings, which is an opportunity for peer review and submission of reports[20, 28]. Staffing improved timely reporting through teamwork between immunization focal persons and monitoring and evaluation focal persons at the facility and LGA levels. Although computers were identified as important input to facilitate reporting at LGA levels, training is needed to translate use of technology into gains in data quality[23, 38].
The study revealed that size of a target population used as a denominator to calculate coverage limited the feasibility of use of demographic information to set the target population for immunization despite its acceptability and adoption. Our finding is consistent with evidence from an Ethiopian study where inappropriate denominators were used to calculate service coverages[37]. In our study, health workers can estimate immunization target population within their wards using reference populations and formula with minimal external support. Although the estimates provided basis for monitoring performance of immunization programs, they were based on outdated population projections. Prior studies noted that outdated census was a major obstacle to target setting within local health systems[38, 39]. A more reliable data would be the actual total number of children in the targeted cohort in each ward but the system to collect these vital registration data is weak[21, 40]. To improve estimation of immunisation service coverage, the vital registration and health information systems need to be strengthened to track all childbirths in Nigeria.
The study brought to light the use of local alternatives in archiving of routine immunization records and reports. Similarly, health workers in other low-resource settings have used their discretions to fill service deliver gaps. For instance, in Ethiopia, health workers prepared local formats on their own when standard tools for recording and reporting immunisation data were not available in their facilities[41]. In our study, health workers used baskets in place of cupboards and shelves to store registers and reporting booklets. Adoption of this local alternatives not only ensured that records are easily retrieved, but it also overcame the cost of providing cupboards and shelves, and the endless waiting for the government to provide them.
Consistent with previous evidence that use of monitoring chart were increased tracking of immunisation tracking and ability of health workers to detect and respond to problems[6, 42, 43], our study indicated that updating monitoring chart was acceptable, adopted, and feasible. Three factors affecting effective use of monitoring chart were identified in this study. First, availability of monitoring chart. There should be regular and uninterrupted supply of monitoring charts. Secondly, training of health workers. Health workers in this study received training on the core outputs and how to conduct analysis and interpret the chart. Elsewhere, inadequate staff capacity, unclear roles and responsibilities related to data analysis, and staff attrition constrained capacity to update monitoring chart in health facilities[42, 44]. Thirdly, supervision enabled the supervisors to monitor revision of the chart and to offer needed technical support to update the chart similar to earlier evidence[43]. In sum, strategies to ensure effective use of monitoring chart would include these three factors.
Our findings highlight the acceptability and adoption of data use for decision making evident in the self-reported use of data to track defaulters, dropout, missed children, and vaccine usage. These findings are consistent with evidence from Nigeria, Tanzania, and Zambia where health workers self-reported using data to guide their actions[20, 45]. In contrast, other studies did not find evidence of data use. Three contextual factors influenced data use for action in our study. First, shortage of staff reduced the quality of data as well as the capacity of health facilities to conduct immunization outreaches. The second factor is demand for evidence by health facility committees (HFCs). HFCs, who understood the monitoring chart, used the evidence to mobilize their communities to increase immunization uptake. HFCs’ response to data communication confirms the need for a value proposition for immunisation data supported by multiway communication[21]. The third factor is funding. Lack of funds not only limited the capacity of health facilities to undertake immunization outreaches, but also the regularity of meetings of HFCs.
However, the study has some limitations. The study is limited to six of 17 LGAs in Enugu State and the quantitative findings have limited generalizability to the entire state nor Nigeria. Similar research should be conducted in other parts of Nigeria to validate our findings. Secondly, the intervention took relatively a short period of three months, which limited our exploration of other dimensions of implementation outcomes such as penetration, integration, and sustainability. Thirdly, the views expressed by the participants in the qualitative component of the study may be prone to social desirability bias, but the use of experienced researcher who understand the local health system context and constantly reminding the participants the purpose and implications of the study reduced such bias.