Patient feedback surveys among pregnant women in Eswatini to improve antenatal care retention

BACKGROUND: Uptake and retention in antenatal care (ANC) is critical for preventing adverse pregnancy outcomes for both mothers and infants. METHODS: We implemented a rapid quality improvement project to improve ANC retention at seven health facilities in Eswatini (October-December 2017). All pregnant women attending ANC visits were eligible to participate in anonymous tablet-based audio assisted computer self-interview (ACASI) surveys. The 24-question survey asked about women’s interactions with health facility staff (HFS) (nurses, mentor mothers, receptionists and lab workers) with a three-level symbolic response options (agree/happy, neutral, disagree/sad). Women were asked to self-report HIV status. Survey results were shared with HFS at monthly quality improvement sessions. Chi-square tests were used to assess differences in responses between months one and three, and between HIV-positive and negative women. Routine medical record data were used to compare retention among pregnant women newly enrolled in ANC two periods, January-February 2017 (‘pre-period’) and January-February 2018 (‘post-period’) at two of the participating health facilities. Proportions of women retained at 3 and 6 months were compared using Cochran-Mantel-Haenszel and Wilcoxon tests. RESULTS: A total of 1,483 surveys were completed by pregnant women attending ANC, of whom 508 (34.3%) self-reported to be HIV-positive. The only significant change in responses from month one to three was whether nurses listened with agreement increasing from 88.3% to 94.8% (p<0.01). Overall, WLHIV had significantly higher proportions of reported satisfaction with HFS interactions compared to HIV-negative women. A total of 680 pregnant women were included in the retention analysis; 454 (66.8%) HIV-negative and 226 (33.2%) WLHIV. In the pre- and post-periods, 59.4% and 64.6%, respectively, attended at least four ANC visits (p = 0.16). The proportion of women retained at six months increased from 60.9% in the pre-period to 72.7% in the post-period (p = 0.03). For HIV-negative women, pre- and post-period six-month retention significantly increased from 56.6% to 71.6% (p = 0.02); however, the increase in WLHIV retained at six months from 70.7% (pre-period) to 75.0% (post-period) was not statistically significant (p = 0.64). CONCLUSION: The type of rapid quality improvement intervention we implemented may be useful in improving patient-provider relationships although whether it can improve retention remains unclear.