Electronic diagnostic algorithms to assist mid-level health care workers in Nepal: a mixed-method exploratory study

Abstract

Introduction: The use of mobile health applications for data collection and disease management by rural health care workers in developing countries has been shown to be accepted by patients and health care workers. However, the acceptances of diagnostic decision applications have not been studied. Moreover, verbal acceptance of these tools has not been shown to equate with actual usage by the health care workers when use is not compulsory.

Objective: To measure the acceptance, usage and reasons for use or non-use of electronic diagnostic applications by health care workers to aid in clinical diagnosis.

Methods: Eleven health care workers (HCW) from rural facilities were asked to use the e-algo application on an electronic tablet with patients over the age of five presenting with acute complaints. Use was compulsory for the first 30 working days and after that optional. Patients were asked by questionnaire about their preference and confidence between the traditional approach and that of the e-algo. HCW acceptance was measured by focus group discussions after the compulsory period. The HCW was then told to use the application as they desired. After two months of non-compulsory usage, reasons were explored for use or non-use through a focus group discussion and interviews.

Results: A total of 1410 out-patient encounters occurred in the first phase. Of this, the e-algo was used with 1177 encounters (83%). 496 patients were asked about their preference and confidence in the use of the e-algo. 325 preferred the e-algo over the traditional visit 65.8-25.1%. Patient confidence was higher in the e-algo 72.2-17.4%. In the second phase, three of the nine HCWs did not use the e-algo at all, the remaining six HCWs reported e-algo use dropped to approximately 15% of total OPD visits. E-algos were reported to be used primarily with more complicated or confusing cases. Reasons for non-use was primarily time related.

Conclusions: We concluded that patients had confidence in and preferred the HCW using the e-algo in their patient care. The HCW users were also positive about the e-algo application, seeing its primary benefit as assisting them in more difficult cases through the use of a differential diagnosis and focused questions. HCWs also reported that the e-algo functioned as a learning tool as well as a diagnostic tool. However, actual usage of the application dropped off significantly when its use was not mandatory. The primary reason was that they did not feel the time required to use the application was warranted in the vast majority of their cases which they perceived as being simple and easily diagnose without the assistance of the application. Unless the HCW perceives the decision-support application to be valid, time-saving and easy to use, they will not use them.

Keywords: Algorithm; Community health worker; Decision support; Diagnosis; Mobile apps; PDA; Rural health.