Given the nature of the study, a holistic inductive paradigm was used with a grounded theory analysis. Although the study took a grounded theory approach, it differed from the grounded theory as originally defined by Glaser & Strauss5 in that it did not adopt a theoretical sampling strategy. The sample was determined by the PHC set up6, and followed a purposeful sampling strategy as described by Patton (1990). The study, therefore, focussed on two categories of people who form the base of PHC service delivery, namely the women and health workers. The total sample was eighty-two (48 women leaders and 34 health workers).
Face-to-face interviews were conducted using an interview schedule that consisted of open questions and one relating to health information critical incidents. Two sets were used: one for the women and the other for health workers. As a semi-structured (open-ended) interview method was used to collect data, it was preferred to do a cross-case analysis for each question in the interview schedule. This involved grouping together and comparing answers from different interviewees to common questions. The analyst then abstracted from the data and generated concepts and categories inductively. This was preceded by open and selective coding, based on the original Glaser and Strauss (1967) approach.