Thanks for sharing more ideas on how to go to UHC. In my response to your previous blog post "How to fill "quality" wine in the bottle?" I touched some of the issues you are discussing now.
I fully agree with you that CEA will not help us effectively move to UHC if we consider UHC is a process of "increasing access and financial protection without deteriorating equity". In addition to the absence of equity consideration in CEA, measuring cost-effectiveness of one intervention against another may be hardly possible, if not impossible, as the reality of the context is so complex with better health outcomes being contributed by many interventions (a combination of supply and demand-side interventions plus environmental factors), which are often necessary for low-income settings.
Your idea about computer-based randomization of entitlement to state subsidies for health care (or exemption from payment) at the point of use sounds interesting and innovative. It is clear that a computer-based "randomization" can be an alternative or implementation to "targeting or prioritization" which has been criticized for its exclusion and inclusion errors and associated cost. However, such randomization may undermine the principle of UHC: "pre-payment" and risk pooling. Giving entitlement at the point of use through such randomization is associated with a number of side-effects related to the so-called "uncertainty of entitlement", including inappropriate or cost-inefficient health seeking behaviors. Imagine what would be decision of an ill person who has no money and is not sure he will get free at an ideal health facility/provider. He would delay seeking care or seeking poorer quality care or forgo treatment. If he would seek care at that facility/provider and then fail to win the lottery, he then spends time and money to go there for nothing and also miss opportunity for get care early care. To be secure, this person may borrow money with high interests or sell assets before going to that facility/provider.
We have an explicit example of this lottery system in Cambodia applied by a charity hospital. They provide high quality free care to all people, and then they face over demand (e.g. everyday 200 people come for outpatient consultations while they can offer only 100). They then introduce a system of lottery. It works for the hospital, but causes a lot of consequences to patients. Everyday, over half of people, some coming from very faraway, fail to get a consultation, and some do so for many times. At the end, because of limited resources, the hospital leaders decide to open more clinics, but charging "fair" user fees from patients, which is to me not ideal, but a better alternative. I was also involved shortly in the discussion with the hospital leaders to find such solution.