JMIR Publications

ABSTRACT

Background:

Patient satisfaction surveys can offer crucial information on quality of care but are rarely conducted in low-income settings. In this study, we tested the feasibility of using an interactive voice response (IVR) system to assess patient satisfaction with primary care services in Addis Ababa, Ethiopia.

Objective:

We described the methodology, response rates, survey costs and identify factors associated with survey participation, completion and duration.

Methods:

Patients were recruited in-person from 18 public and private health facilities in Addis Ababa. Patients’ gender, age, education, reasons for seeking care and mobile phone numbers were collected. The survey included 15 questions that respondents answered using their phone keypad. We used a Heckman probit regression model to identify factors influencing the likelihood of IVR survey participation (picking up and answering at least one question) and completion (answering all survey questions) and a Weibull regression model to identify factors influencing the survey completion time.

Results:

Nearly all patients approached (94%) had a functioning mobile phone and 90% of those eligible agreed to be enrolled in the study. 93% of those enrolled picked up the call but only 66% answered at least one survey question, and 43% (N=1,037) completed the full survey. Average survey completion time was 8.1 minutes We found that those aged 40 and above were substantially less likely to participate in and complete the IVR survey compared to those aged 18-30. Higher education levels were also strongly associated with survey participation and completion. In adjusted models, those enrolled in private facilities were less likely to participate and complete the survey compared to those in public health centers. Being male, younger, speaking Amharic, using a private hospital, and being called after 8pm were associated with a shorter survey duration. Average survey costs were $7.90 per completed survey. During recruitment, individuals expressed concerns with privacy and with sharing their mobile phone number.

Conclusions:

Our findings reveal that an IVR survey is a feasible, low-cost, and rapid solution to to assess patient satisfaction in an urban context in Ethiopia. However, survey implementation must be carefully planned and tailored to local challenges. Governments and health facilities should consider IVR to routinely collect patient satisfaction data to inform quality improvement strategies. Clinical Trial: N/A