Assessing the validity of maternal report on breastfeeding counselling in Kosovo’s primary health facilities

Background Improving the quality of breastfeeding counselling delivered by primary care providers can improve breastfeeding outcomes and ultimately reduce mortality and morbidity of children and mothers. Accurate data on coverage and quality of primary care breastfeeding counselling is essential for monitoring progress; however, global and national indicators are limited. To help address this gap, this study validated indicators of receipt and quality of breastfeeding counselling during routine consultations for infant care at seven primary health facilities across Kosovo. Methods Mothers’ reports of breastfeeding counselling received during routine consultations for their infants (0–12 months of age) were collected by exit interview in 2019 and 2021 (n = 609). Responses were compared against direct observation of their consultation using a structured checklist (reference standard) by a trained third-party observer at the primary care facility. We assessed 13 indicators; ten were related to the receipt and content of breastfeeding counselling, and three were specific to the provider’s interpersonal skills. We calculated sensitivity, specificity, and area under the receiver operating curve (AUC) to determine individual-level reporting accuracy. Results Ten indicators had an agreement rate above 70% and seven indicators had high overall individual-level validity (AUC ≥ 0.7). High prevalence indicators recorded high sensitivity and low specificity, and the inverse for low prevalence indicators. More subjective indicators were less reliable, e.g., mothers over-reported the prevalence of all three indicators related to providers’ interpersonal skills. Conclusions This study offers evidence on breastfeeding counselling quality by validating maternal reports of whether a provider discussed breastfeeding, the clinical content of that counselling, and how it was delivered. It is also situated in a primary care setting within a fragile state of which there is limited evidence. We observed that mothers reported accurately when asked directly to recall breastfeeding counselling services received. However, there is a need to further validate subjective questions about interpersonal skills and other measures for the ‘experience of care’ quality dimension.