Audio computer-assisted self-interviewing (ACASI) may avert socially desirable responses about infant feeding in the context of HIV

BACKGROUND Understanding infant feeding practices in the context of HIV and factors that put mothers at risk of HIV infection is an important step towards prevention of mother to child transmission of HIV (PMTCT). Face-to-face (FTF) interviewing may not be a suitable way of ascertaining this information because respondents may report what is socially desirable. Audio computer-assisted self-interviewing (ACASI) is thought to increase privacy, reporting of sensitive issues and to eliminate socially desirable responses. We compared ACASI with FTF interviewing and explored its feasibility, usability, and acceptability in a PMTCT program in Kenya. METHODS A graphic user interface (GUI) was developed using Macromedia Authorware and questions and instructions recorded in local languages Kikuyu and Kiswahili. Eighty mothers enrolled in the PMTCT program were interviewed with each of the interviewing mode (ACASI and FTF) and responses obtained in FTF interviews and ACASI compared using McNemar's chi2 for paired proportions. A paired Student's t-test was used to compare means of age, marital-time and parity when measuring interview mode effect and two-sample Student's t-test to compare means for samples stratified by education level - determined during the exit interview. A Chi-Square (chi2test) was used to compare ability to use ACASI by education level. RESULTS Mean ages for intended time for breastfeeding as reported by ACASI were 11 months by ACASI and 19 months by FTF interviewing (p < 0.001). Introduction of complementary foods at <or=3 months was reported more frequently by respondents in ACASI compared to FTF interviews for 7 of 13 complementary food items commonly utilized in the study area (p < 0.05). More respondents reported use of unsuitable utensils for infant feeding in ACASI than in FTF interviewing (p = 0.001). In other sensitive questions, 7% more respondents reported unstable relationships with ACASI than when interviewed FTF (p = 0.039). Regardless of education level, respondents used ACASI similarly and majority (65%) preferred it to FTF interviewing mainly due to enhanced usability and privacy. Most respondents (79%) preferred ACASI to FTF for future interviewing. CONCLUSION ACASI seems to improve quality of information by increasing response to sensitive questions, decreasing socially desirable responses, and by preventing null responses and was suitable for collecting data in a setting where formal education is low.