Table 1.

The four criteria applied to Ebola Care and a virtual assistant.

	Ebola Care	Virtual assistant
Zero costs to users	Zero cost to user, when they are provided with a mobile phone if they do not have one already.	The costs to a patient will typically be zero. Further, a virtual assistant does not require a smartphone, but would work with voice or even SMS.
The health care provider can recoup costs	Yes, health care efforts can be directed much more efficiently.	If the system costs are borne at the national level or by an NGO, then costs to local health care providers are small, mainly consisting of investing in the ability to see the information that a patient has already supplied to the virtual assistant.
Increased demand can be handled	In this case a more accurate measure of Ebola spread allowed for more help from international organizations to handle demand.	This point is more complex. In particular, during the initial development a virtual assistant may refer patients to health care professionals quite often. In an area with poor access to health care, this increase in demand will probably not be offset by reducing the number of necessary visits. But, on the other hand, a virtual assistant can be controlled. Where health care access is rationed, the virtual assistant would be taught to suggest remedies that are feasible, such as self-treatment.​
The mHealth app does not duplicate other mHealth development efforts	Only to a minor extent, since Ebola was rare in the rest of the world, and the time frame was short.	While Babylon and some others are investing in ‘engines’ for virtual health care assistants, there will probably be few for-profit applications targeted to low-income countries and dealing with local language, dialects and awareness of local health issues.