Validation Study of Adherence to Daily Multiple Micronutrient Supplementation by Palestine Refugee Women Attending United Nations Relief and Works Agency for Palestine Refugees in the Near East Health Clinics in Jordan

This is the authors' abstract. We don't add key points for this paper.

Background Multiple micronutrient supplementation (MMS) is a new standard-of-antenatal care emerging in lower income countries. Assessing adherence, a key implementation outcome, requires validation. Objective Validate antenatal adherence to daily MMS based on recalled tablet days missed during follow-up intervals in UNRWA clinics serving Palestine refugees in Jordan. Methods This study was conducted during an implementation trial of antenatal MMS versus standard-of-care iron-folic acid supplementation. In 13 clinics, at registration, pregnant women received daily MMS in 180-count bottles weighing 111 g (bottle with desiccant=26 grams, tablet=0.47 g). At seven follow-up visits, returning bottles were weighed to nearest gram on digital scales and recalled tablet days missed were recorded. Tablets removed from bottles, estimated from bottle weight decrements, were regressed on tablets taken, obtained by subtracting tablet days missed from interval lengths. Similar analyses were conducted for cumulative intervals since registration. Results Among 9754 registered women, 7608 (78%), 6,208 (63.6%), 4,833 (49.5%), 3,409 (34.9%), 2159 (22.1%), 1063 (10.9%) and 384 (3.9%) completed the 1 st -to-7 th follow-up visits, respectively, with decline mainly due to censoring at close-out. Approximately 25% of women failed to return bottles, whose tablet intakes were modestly lower than bottle-returnees. Notwithstanding, 75-84% in both groups recalled missing <2 days and <5.3% missing >30 days. Interval-specific regression slopes (β 1 ) were 0.88-0.78 [95% confidence intervals (CI) within +0.06 for 1 st 5 intervals], reflecting a ∼20% overreporting by recall, with r=0.77-0.58, and root mean square errors (RMSE) of 10-12 tablets, reflecting model error in prediction, normalized to 7-10% of interval ranges. Cumulative models yielded β 1 =0.83-0.77 with similarly narrow 95% CIs, r=0.79-0.64 and RMSE=12-24 tablets, normalized to 7-13% of range. Conclusions Histories of recalled tablet days missed provides a reliable and reasonably valid estimate of tablet disappearance from MMS bottles. Further research is warranted to correct a consistently reported ∼20% average overestimate of tablet intake. Trial registration Not required, ethically approved as public health practice.