Determinants of Routine Immunisation Data Quality in Malawi: A PRISM-Framed Mixed-Methods Assessment
Year: 2026
Type: preprint
Study design: Observational Study
Abstract: Background/Objectives: Efforts to improve routine health data often focus on tools and forms, yet data quality depends as much on the organization around the data and the behavior of the people who produce it. The Performance of Routine Information System Management framework separates these technical, organizational, and behavioral determinants. This study applied that framing to routine immunization data in five districts of Malawi, in order to ask not only how good the data were but why. Materials and Methods: A mixed-methods facility assessment measured data accuracy by recounting the third dose of pentavalent vaccine against monthly reports in 50 facilities, of which 44 being the subset that had both reported and verified values. Technical, organisational, and behavioural determinants were assessed through a structured checklist, a survey of data staff, and key-informant interviews, and were interpreted through the framework. The relationship between the determinants and accuracy was examined using a composite readiness score and descriptive comparison. Results: The technical determinants were largely in place, because registers, reporting forms, and the electronic system were available in about 98% of facilities. The organizational and behavioral determinants were weaker. Governance structures were widely present, yet none of ten governance or data-use functions was associated with accuracy. A composite score of the organizational and behavioral determinants showed no association with accuracy (Spearman ρ = 0.22, p = 0.153), so facilities that met almost every condition were no more accurate than those that met few. Staff described audits, supervision, and training as dependent on external funding, and as lapsing under the 2025 withdrawal of that funding. Confidence in the data was high at 91 to 100%, and review before submission was near universal, although accuracy was uneven. The median accuracy ratio was 100%, but only 34 of 44 facilities (77%) fell within the acceptable range, while over-reporting and under-reporting nearly cancelled, leaving the aggregate accurate to within 1%. Conclusions: The technical foundations of the routine data systems were largely in place, but organizational and behavioral mechanisms did not consistently translate these foundations into accurate facility level data. Strengthening routine immunization data quality will require greater attention to how data verification, feedback and corrective actions are routinely implemented and sustained.
Source: Preprints.org
Authors: Esther Malata, Beverly Shailla Laher, Jacqueline R. Chinkonde, Blessings H. Likoswe, Beatrice Mkomadzinja, Susan Barnaba, Jimmy Banda, Hannah Kachule, Lonjezo Dzekedzeke, Andrew Matchado, Humphrey Chatenga, Ajib Phiri, Temwa Mzengeza, Patrick N. Omar, Abiba L. P. Longwe-Ngwira, Chiara Pierotti, Charles Nwosisi, Jennifer Barak, Mike Chisema, Brenda Mhone, Hastings Chiumia, John Phuka, A. S. M. Shahabuddin
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W7214799505