A couple of interesting questions by @MsEyezee along with my responses and insights. I hope someone finds this useful and people take COVID-19 seriously! #TakeResponsibility #Covid_19 #CovidNGR #COVID19Nigeria #Nigeria #COVIDー19 #COVID19 #NCDC

Q1. Given the progression can we deduce/project transmission saturation?
Q2. How should this knowledge be used to shape the response?
Q3. - How (if any) are these used in planning our economic response to the pandemic.

Q4. Is it possible to anticipate how long the pandemic will be around locally?
Q5. All things being equal? What happens when the world begins to wake up? Would we be ready too?

(1) The accuracy is strong because apart from using relatively sophisticated models, the algorithms are online, it automatically tunes and updates its parameters every time NCDC releases a new result (a piece of code does this every 24hrs).

(2) Monthly forecasts as your rightly noted are tempting. But technically, it would be wrong to perform monthly forecasts. Why? because of error propagation. The most accurate prediction is 1-step ahead (which is one day ahead). But can we try? Yes, see (4) and (9-11)

(3) Another problem is that we do not have enough data. Cases are mainly between February 27 and April 25. Which is around 58 days. So, predicting 30 days using 58 days data is not recommended. I would suggest a week ahead as the best horizon (with the data we have so far).

(4) These models are based on time of events (time-series). Unfortunately, I am constrained by the data available from NCDC. We can apply Epidemiological (Epd) models as well and I'm doing this elsewhere.

(5) Nonetheless, I have observed that 1-day time-series forecasts meet the eager expectations of the common man. Epd. models can vary wildly and, in several places, predictions have been very different depending on analytic assumptions.

(6) Armed with an Epd. model and with the right data one can try to estimate the worst and best-case scenarios. They are excellent for evaluating how many people could get infected, recovered, die, curve flatten, and so on.

(7) However, countries that have done these (Epd. modeling) have used it to estimate and evaluate policies and worse case scenarios. If the model says we'll have 100k cases, and 10k could die, then this information helps in (8)

(8) How many hospitals do we need and where? how many ventilators should we get? what are the existing medical facilities/staff and how many PPEs we got? what about lockdowns? and so on!

(9) Since I don’t think we have an effective integrated national database, a more applicable way of estimating the real number of cases we have in Nigeria might be through Viral genome data.

(10) How does this work? Once the virus is transmitted from one person to another, it usually undergoes mutation. The virus from China may not be the same when it got to Nigeria. So, from one person to another, the virus mutates and changes.

(11) If we analyze these data, using genome sequencing, we can analyze how much it has changed to get indications of the number of people (mutation counts) it has infected before reaching the current host. You can look up the company known as "Nextstrain"

(12) Regarding the economy, I don't know what the plan is. First, lockdown is not easy for people. Second, there must be a strategy. For me, the whole idea of lockdown is to prevent further spread by forced isolation. During the lockdown, we are supposed to test aggressively.

(13) By testing aggressively, we can quickly identify the hot zones and cool zones. Cool zones can be reopened for activity while hot zones remain closed. If we do not do this, then that would be painful!

(14) Because it means we're not taking advantage of the lockdown to quickly restart economic activities. The entire efforts can be lost if we eventually reopen while there are still several infected people. We’ll have a second wave!!

(15) Somehow, we always manage to survive situations like this. Let's support the government and NCDC, I believe they're doing what they can. With the data available, we probably would not be ready. It seems Africa is generally taking backstage.

(16) Several countries have conducted hundreds of thousands of tests. With that, they know how many cases they have, the situation, and can address it. As a result, they have "peaked" in the number of cases and some have started flattening the curve.

(17) The problem is that we do not even know how many cases we have in Nigeria yet. How long will it take to know? Well, it depends on how long we test. How long till we reopen the economy? Tests! How prepared were we? All the same, we hope to come out stronger!