Objective: the percentage of births attended by health professionals is widely used to measure skilled attendance. This indicator is based on women's reports of their birth attendant. This study explores how women identify health professionals attending their births.
Design: exit interviews, focus groups, in-depth interviews and a community survey. Qualitative data were analysed by topic. Frequency of women's responses on how they identify the birth attendant and other characteristics of birth care were generated through the community survey.
Setting and participants: women in Ghana who had birthed with a health professional in the last 5 years.
Measurements and findings: role, prior knowledge and uniform are the most common means by which women identify their attendant. These means of identification do not distinguish accurately between different types of health professional. Delivery events are more complex than is suggested through use of the indicator 'percentage of deliveries with health professionals'. Fifty-five per cent of births were attended by more than one person. In 11.6% of births, women were attended only after the partial birth of their baby.
Key conclusions and implications for practice: there is potential for incorrect identification of birth attendant by the use of women's reports. None of the methods used could verify women's reporting. Methodological developments in this area are necessary along with improved recording systems. Opportunities for women to identify health professionals should be enhanced.