Addressing the global shortage of mathematical modellers

Mathematical modelling played a crucial role during the COVID-19 pandemic. Globally, however, there is a severe shortage of skilled mathematical modellers, especially in low- and middle-income countries (LMICs). Here, we explore some of the issues, discuss a successful example of how this problem can be addressed, and outline a new online training programme for the next generation of mathematical modellers.

A village in Cambodia - there is an urgent need for mathematical modellers in LMICs such as Cambodia, especially as these countries approach the ‘last mile’ of malaria elimination.

(Photo credit: Adam Bodley)

The profile of mathematical modelling of infectious diseases has increased dramatically in recent years. This was in large part due to the role modelling played during the COVID-19 pandemic. Models were used to describe and project the spread of SARS-CoV-2 (the virus that causes COVID-19), assess the potential impact of non-pharmaceutical interventions such as physical distancing and contact tracing, and inform policymaking (1). At the same time, there have been considerable technical advances in modelling, with new methods, improved computational tools, greater sharing of public data, improved code and code availability and better visualisation methods (2).

The COVID-19 pandemic increased the awareness of disease modelling among policymakers and the general public alike. However, it also shone a light on the substantial differences among countries in terms of their capacity to conduct disease modelling, particularly the apparent lack of such capacity in many LMICs (3). In response to this, considerable efforts were made by researchers in high-income countries (HICs) to provide modelling results relevant to LMICs. However, data limitations and the generalised nature of models produced for use globally weakened their utility and also led to concerns over their accuracy (3).

The lack of modelling capacity in LMICs is compounded by the lack of in‐country training in mathematical modelling (4). It is essential to have in-country modelling expertise so that local contexts and considerations are fully taken into account when developing models and interpreting their outputs. Furthermore, modelling-based evidence must be presented to local policymakers and other stakeholders in a context-dependent manner, to ensure appropriate decisions are made (5). A recent review of global research activity in relation to the mathematical modelling of infectious diseases provided further evidence of the need to prioritise training for modellers in LMICs (6). 

One such approach is collaborative modelling for policymaking that comprises an iterative process involving a range of stakeholders, including modellers, funders and policymakers (5). A good example of this type of participatory modelling approach was the COVID-19 Modelling Consortium (CoMo Consortium) (7). The CoMo Consortium was a collaboration involving more than 100 modellers from almost 50 countries across 5 continents (8). The participatory nature of the consortium was critical to its success in translating contextual factors associated with COVID-19 policy interventions into messages that were suitable for policymakers, thus gaining their trust.

Crucially, the CoMo Consortium included experienced modellers from HICs along with modellers from LMICs or individuals from LMICs who were interested in or had an aptitude for modelling. The more experienced modellers mentored and advised the less experienced modellers from LMICs. In turn, the LMIC modellers provided detailed information about their particular context. In many cases, the LMIC consortium members were able to gain access to relevant local datasets and analyse these data themselves. This meant they did not need to hand their data over to HIC modellers who lacked an in-depth understanding of the local context from which these data were derived. This also helped overcome the issue mentioned above regarding generalised models produced in HICs that proved to be unsuitable for use in LMICs.

As a result of the CoMo Consortium’s approach, there were multiple fruitful collaborations among modellers and policymakers in a range of LMICs around the world, including in Africa, Asia and South America (9-14). Although the peak of the COVID-19 pandemic has now passed, the CoMo Consortium members felt that the collaboration had been extremely useful and was worth continuing. It has therefore been rebranded as the International Consortium of Modellers (also referred to as CoMo). CoMo continues to have regular meetings, seminars and training sessions at which all members are welcome.

The CoMo Consortium was initiated and led by Model Health’s director, Prof Lisa White, who has now established an online training programme for modellers, the Strategic Modelling for Public Health Policy Training Programme. This is a brand-new, online programme of training in policy-facing public health modelling for a global audience. The programme is designed to widen global access and build in-country modelling capacity, especially in LMICs.

It is a modular programme, with a variety of modules suitable for modellers of differing levels of experience. There are modules covering the basics of modelling for those individuals who are interested in modelling but who have limited experience. There will also be modules in advanced techniques for modellers who have more experience. Furthermore, the programme is suitable for individuals working in related professions, across a range of sectors, who wish to learn more about the discipline.

The training programme is a practical, skills-based programme of study. The aim is to train modellers so that they can carry out modelling work themselves. This work will be based on their specific context, using their own data, to generate modelling outputs relevant to their local context. These modellers will also be trained to communicate the results of their model outputs and the underlying assumptions to policymakers and other relevant stakeholders.

Further details about the online training programme for modellers can be found on the CoMo website.

References

1.           Jit M, Ainslie K, Althaus C, Caetano C, Colizza V, Paolotti D et al. Reflections on epidemiological modeling to inform policy during the COVID-19 pandemic In Western Europe, 2020-23. Health Aff (Millwood). 2023;42:1630-6. doi: 10.1377/hlthaff.2023.00688.

2.           McBryde ES, Meehan MT, Adegboye OA, Adekunle AI, Caldwell JM, Pak A et al. Role of modelling in COVID-19 policy development. Paediatr Respir Rev. 2020;35:57-60. doi: 10.1016/j.prrv.2020.06.013.

3.           Silal S, Bardsley C, Menon R, Abdullahi L, White LJ. Epidemiological modelling for public health decision making in sub-Saharan Africa. 2022 (https://www.opml.co.uk/files/Projects/a4964-epidem-modelling-capacity-strengthening.pdf, accessed 8 February 2024).

4.           Kupfer LE, Beecroft B, Viboud C, Wang X, Brouwers P. A call to action: strengthening the capacity for data capture and computational modelling of HIV integrated care in low- and middle-income countries. J Int AIDS Soc. 2020;23 Suppl 1:e25475. doi: 10.1002/jia2.25475.

5.           Teerawattananon Y, Kc S, Chi YL, Dabak S, Kazibwe J, Clapham H et al. Recalibrating the notion of modelling for policymaking during pandemics. Epidemics. 2022;38:100552. doi: 10.1016/j.epidem.2022.100552.

6.           Sweileh WM. Global research activity on mathematical modeling of transmission and control of 23 selected infectious disease outbreak. Glob Health. 2022;18:4. doi: 10.1186/s12992-022-00803-x.

7.           CoMo International. nd (https://como-international.github.io/, accessed 5 March 2024).

8.           Aguas R, White L, Hupert N, Shretta R, Pan-Ngum W, Celhay O et al. Modelling the COVID-19 pandemic in context: an international participatory approach. BMJ Glob Health. 2020;5. doi: 10.1136/bmjgh-2020-003126.

9.           Adib K, Hancock PA, Rahimli A, Mugisa B, Abdulrazeq F, Aguas R et al. A participatory modelling approach for investigating the spread of COVID-19 in countries of the Eastern Mediterranean Region to support public health decision-making. BMJ Glob Health. 2021;6. doi: 10.1136/bmjgh-2021-005207.

10.         Bellizzi S, Letchford N, Adib K, Probert WJM, Hancock P, Alsawalha L et al. Participatory mathematical modeling approach for policymaking during the first year of the COVID-19 crisis, Jordan. Emerg Infect Dis. 2023;29:1738-46. doi: 10.3201/eid2909.221493.

11.         Franco C, Ferreira LS, Sudbrack V, Borges ME, Poloni S, Prado PI et al. Percolation across households in mechanistic models of non-pharmaceutical interventions in SARS-CoV-2 disease dynamics. Epidemics. 2022;39:100551. doi: 10.1016/j.epidem.2022.100551.

12.         Diarra M, Kebir A, Talla C, Barry A, Faye J, Louati D et al. Non-pharmaceutical interventions and COVID-19 vaccination strategies in Senegal: a modelling study. BMJ Glob Health. 2022;7. doi: 10.1136/bmjgh-2021-007236.

13.         Marzouk M, Alhiraki OA, Aguas R, Gao B, Clapham H, Obaid W et al. SARS-CoV-2 transmission in opposition-controlled Northwest Syria: modeling pandemic responses during political conflict. Int J Infect Dis. 2022;117:103-15. doi: 10.1016/j.ijid.2022.01.062.

14.         Saeedzai SA, Sahak MN, Arifi F, Abdelkreem Aly E, Gurp MV, White LJ et al. COVID-19 morbidity in Afghanistan: a nationwide, population-based seroepidemiological study. BMJ Open. 2022;12:e060739. doi: 10.1136/bmjopen-2021-060739.