Assessing courtesy reporting bias in facility-based surveys on person-centred maternity care: evidence from urban informal settlements in Nairobi and Lusaka.

Background: Experience of care is typically measured through client exit surveys administered in the facility. Evidence suggests that such measures suffer from courtesy reporting bias whereby respondents do not accurately report on their experiences while in the facility. We explored the presence of courtesy bias by comparing women's reported experience of person-centred maternity care (PCMC) from facility-based client exit surveys to mobile phone-based surveys out of the facility in Nairobi and Lusaka's urban informal settlements. Methods: We randomly and independently sampled women in the facilities for either a facility-based survey (n = 233 in Lusaka and n = 112 in Nairobi) or a mobile phone-based survey (n = 203 in Lusaka and n = 300 in Nairobi) within one to two weeks of facility discharge. The questionnaire included a validated PCMC scale. After adjusting for differences in women's characteristics across groups, we compared PCMC scores between facility and phone-based samples. We ran multilevel linear regression models to assess PCMC by survey modality in each city. Results: In both cities, over 70.0% of women were aged 20-34 years and were married, at least two thirds had secondary education, and over 95.0% were unaccompanied during labour/delivery. The overall PCMC score was 69.3% among women surveyed on the phone compared to 70.2% among those surveyed in the facility in Nairobi. In Lusaka, it was 57.5% on the phone compared to 56.8% in-facility. We found no statistically significant differences in PCMC scores between survey modalities in both cities, after adjusting for differences in women's characteristics. Conclusions: We did not detect significant courtesy reporting bias in PCMC in facility-based client exit surveys in the context of urban informal settlements in Nairobi and Lusaka. Experience of PCMC can be measured through in-facility client exit surveys or mobile phone surveys. However, it is critical to address challenges related to a mobile phone-based approach.