The first author and research assistants transcribed the interview recordings verbatim. Handwritten notes taken during interviews were used to enhance data during transcriptions. Interviews conducted in the local language were translated during transcriptions. The first author participated in all the transcriptions and translations of the interview recordings. Interview transcripts were shared between the investigators who read through them independently to gain insights into the data before formulation of codes for data analysis. Each investigator developed codes. The three investigators discussed, and agreed on the final codes for analyses through consensus. The final set of codes constituted the codebook used in data analysis. Manual data analysis using the thematic content analysis approach was used [30, 31]. The first author applied the codes on every meaning segments of each transcript, reading through the transcript line by line. The coded segments with similar meanings were aggregated to form subthemes and themes that reflected their central meaning. The analysis framework was shared between the investigators who discussed and agreed on the themes and subthemes. Further analysis was informed by the data and involved iterative coding and aggregating meaning segments from each transcript. Recurrent themes formed the main framework for data interpretation based on the main outcome variables of health-seeking for symptoms of esophageal cancer and adherence to recommended esophageal cancer treatment.

Participants’ socio-demographic variables including sex, age, marital status, education attainment, and duration of work as healthcare professional, and patients’ experiences right from first symptoms realization and attributions, help-seeking, and diagnosis as well as perceived and manifest stigma, stress, and threat to or actual loss of employment guided contextual data interpretation under the themes and subthemes. We did not aim to compare and contrast views from the patients and healthcare professionals, but rather identify common issues that potentially influenced health-seeking and adherence from the perspectives of the two categories of participants.