*3.1.1 Improved/adapted mathematical projections/outbreak science to help guide decisions*

Mathematical projections/outbreak science has become the driving force behind the pandemic responses with growing calls to follow science including potential lessons from other risk science experiences [23]. However, we should bear in mind that mathematical models are useful exploration of questions are also dangerous way to assert answers. Various teams, primarily comprised of academic modelers, organized by, for example, the World Health Organization, the US Centers for Disease Control and Prevention have been involved [58, 59]. Obviously, research does not get much more policy-relevant since governments across the globe relying on these projections [59]. However, there is yet a lot to be learned about how the virus spreads that the models should be constantly updated with increasing knowledge and information; a formidable task even with the best surveillance systems [60]. Because any single data type is likely to yield under- or over- estimate of the extent and spread of the disease, it is important to consider multiple data types and be cautious in relying on estimates without considering sources of bias [41]. Models (equation-based, agent-based …) are, at best, simplified representations of reality based on assumptions on the behaviors of the virus (reproductive rate, incubation period, death rate …), environmental/climate and individuals/societies including demographic composition and mobility [61, 62]. Even the best systems need regular updating and improvement based more on real data derived from epidemiologic investigations rather than assumptions [41, 63, 64]. In short, models should be used with prudence and we should ensure that modelling should not be considered with certainty than the models deserve; and politicians must not be allowed to offload accountability to models of their choosing.

In the Ethiopian context, the situation is compounded by the weak health management information system, diverse population and limited experience at modeling [65]. A recent modeling, for example (**Figure 1**), seems to clearly underestimate deaths by their won assumptions, leaving out possible deaths among those not hospitalized [66]. However, evidences reveal that majority of confirmed cases and deaths are from Addis Ababa.

Countries are expected to develop their own estimates based on demographic and epidemiological characteristics and update them periodically as data/info improves while networking and learning from the various efforts elsewhere. Typically, repeated runs with varying inputs and assumptions are undertaken on

**Figure 1.**

*Attempt at Modeling the COVID Pandemic for Ethiopia. (Source: Adapted from [66], --- authors additions)*

several modes to avoid too much reliance on one mode [58]. The need, to develop the capacity to generate real-time, reliable, accessible and actionable data to empower leaders to act faster [42].

Based on experiences from the US, we should aim to change the health system by accelerating use of telemedicine; move away from traditional models of employerbased health insurance; move away from nursing homes; address health disparities and the social determinants of health; improve drugs affordability; increase local production of drugs; enhance epidemic preparedness with more task shifting and improved financial management [57].

Organizational structures vary from country to country [58] but, in the Ethiopian context, the Federal Ministry of Health (FMOH), the Ethiopian Public Health Institute (EPHI), Regional Health Bureaus (RHB) etc. are destined to play the major role. It seems advisable to create a multi/interdisciplinary (epidemiologists, clinicians, health managers, social scientists, mathematicians …) team in 2–3 universities and establish network or even a National Infectious Disease Forecasting Center or revamp EPHI to play this role [42]. The network/center should have a direct link with policy makers, adequate funding and access to data during outbreaks. Models are only as robust as the data used to build them. In many settings, the infrastructure for collecting, collating, and cleaning high-quality data is underdeveloped [58]. The network/center could attempt to create one data bank; the aim being to achieve precision public health which requires robust primary surveillance data, rapid application of sophisticated analytics to track the geographical distribution of disease, and the capacity to act on such information. It could also join the WHO Global Research and Innovation Forum [30, 62]. Its term of reference (TOR) could include other non-epidemic issues see, for example, BARDA, CDC Health Economics and Modeling Unit; The Research and Policy for Infectious Disease Dynamics (RAPIDD) group in the Fogarty International Center at the National Institutes of Health (NIH), The International Initiative on Spatial Life course Epidemiology (ISLE) which could serve as possible contacts. Collaboration

#### *Evidence-Based Preparedness for Post COVID-19 DOI: http://dx.doi.org/10.5772/intechopen.96931*

with other (neighboring) African countries should be promoted [42, 62]. It is also important to consolidate training in field epidemiology and outbreak science methodologies [58] and, remember that no public-health research is complete until the key findings are effectively communicated and, ideally, implemented [67].
