Caregiver recall in childhood vaccination surveys: Systematic review of recall quality and use in low- and middle-income settings. [Review]

INTRODUCTION: High population coverage is key to the impact of vaccines. However, vaccine coverage estimates in low- and middle-income countries (LMICs) have repeatedly been shown to be of poor quality. LMICs often rely on 'caregiver recall' of vaccination, the validity and collection method of which remains uncertain. We aimed to critique the quality of caregiver recall and make recommendations for its collection and use. METHODS: We performed a systematic review for methods assessing childhood vaccination coverage in LMICs. We searched Medline using variations of the key terms: (child) AND (vaccinat*) AND (survey OR recall OR coverage) AND (reliab* OR valid*). We selected articles assessing the quality of recall in LMICs and extracted reported validity, reliability and completeness. We synthesised recommendations on collecting, analysing and presenting caregiver recall for varying resource availabilities. RESULTS: Of 1268 articles, 134 full texts were screened and eight were included for review. There was heterogeneity in study designs, ways of incorporating recall data and outcomes measured. Sensitivity of recall was 41-98%; specificity was 12-80%. There was a dearth of reliability measures and no consistent method for dealing with data incompleteness. CONCLUSION: There are quality concerns with caregiver recall and difficulty in assessing it given the lack of a 'gold standard' for vaccine status. To improve coverage estimates and the impact of vaccines, caregiver recall should be used. Other recommendations include: recall is included for those presenting vaccine records; missing data is imputed; recall and record quality are assessed in a sub-sample; and sensitivity analyses are performed.