Modeling the predictors of mobile health adoption by Rohingya Refugees in Bangladesh: An extension of UTAUT2 using combined SEM-Neural network approach.

While the healthcare facilities for the people is questionable in Bangladesh, Rohingya refugees is a burning issue for both Bangladesh and global community. Integrating Rohingya refugees into the framework of mHealth could be beneficial for both Bangladesh and Rohingya refugees in general, and in specific situation like COVID-19 outbreak . However, no research has been found on what motivates Rohingya refugees to accept mHealth in Bangladesh. Drawing on the UTAUT2 model, this study investigates the predictors of acceptance of mHealth services technologies among Rohingya refugees. The study also seeks to clarify the roles of mHealth developers, the Bangladesh government, and non-governmental organizations working with the 1.1 million Rohingya refugees in Bangladesh. Quantitative data were collected from refugee camps with the permission of the Refugee Relief and Repatriation Commissioner (RRRC). The data were analyzed in two stages using a mixed approach that combines PLS-SEM and Artificial Neural Network (ANN). This study revealed that Effort expectancy (EE, with t  = 5.629, β = 0.313) and facilitating conditions (FC with t  = 4.442, β = 0.269) in PLS-SEM, and FC (with 100 percent importance) and Health consciousness (HC, with 94.88 percent importance) in ANN analysis were found to be the most substantial predictors of mHealth adoption. The study also revealed that EE and FC are more important for low education group, while PE and Situational Constraint (SC) are more important for the high education group of refugees. In addition to providing insights for mHealth developers, this study particularly focuses on the role of government institutions and non-governmental social workers in working with the subjects to increase FC and HC among Rohingya refugees and bring them under mHealth services.