Designing effective training programs for applied mathematical modeling

By guest contributors Jaline Gerardin and Sheetal P. Silal

The direct application of mathematical models of infectious disease to inform policy decisions has greatly expanded in recent years. Decision-makers are increasingly interested in considering evidence from modeled scenarios as they devise disease management and response strategies. In the cross-disciplinary field of applied mathematical modeling, models are closely integrated with local data, and modelers closely collaborate with decision-makers.

In low- and middle-income countries (LMICs) with high burden of disease, capacity for training applied modelers is often limited. Much of the local modeling talent received their training in departments of mathematics and thus have strong skills in theoretical modeling but usually have little background in public health. Similarly, local epidemiologists have excellent knowledge about disease systems but limited or no modeling skills.

To meet the needs of LMIC health programs by upskilling local modelers, existing applied modeling teams are developing training programs with the support of global donors, who are increasingly recognizing the need for dedicated initiatives for training in applied modeling. Borne of the experience of running these capacity strengthening initiatives in the last 10 years, the authors have identified eight technical (hard) and non-technical (soft) skills essential to trainees in applied modeling (Figure 1). Training programs in applied modeling should consider how they are strengthening the weakest of these skills amongst their participants, and individual modelers can use this list to identify areas for personal growth.

1. Mathematical modeling and simulation: design, develop, implement, calibrate, analyze, and interpret mathematical models; be comfortable in programming and in using computing clusters.

2. Data analysis and contextualisation: understand different types of data and data collection methods, their strengths and limitations; be familiar with data analysis and visualization methods.

3. Domain knowledge of disease and health systems: including but not limited to disease biology, epidemiology, entomology, health system dynamics, population dynamics, health economics and financing, and public health.

4. Research, methods, and evidence synthesis: build on others’ previous work, select and apply the appropriate methods, innovate, and assemble multiple sources of evidence to inform the design and interpretation of models.

5. Knowledge translation: work with consumers of modeling to consider modeled evidence with other relevant evidence, contextualize the evidence to local considerations, and thereby advise on policies and support operational decisions.

6. Leadership and engagement: engender trust and confidence, behave professionally, and build strong relationships.

7. Scientific communication: communicate complex concepts effectively across different audiences in oral presentations and policy briefs, convey both strengths and limitations of approaches and outputs, and persuasively organize ideas into proposals.

8. Project management: work flexibly and adaptively with short timelines, be responsive to requests from decision-makers, and stay organized across multiple requests and through the course of a work program.

Figure 1. Eight essential skills for applied modeling.

While striving to equip aspiring modelers with these eight essential skills, training programmes should consider that applied modeling is best learned via practical experience through a total immersion process of co-designing a project with policy-makers, executing the analysis, communicating results back to policy-makers, and receiving feedback and coaching at every step. This approach additionally supports the career development of their participants, building relationships between participants and policy-makers that participants can continue independently post-program. We find that training programs are effective when they are intensive, immersive experiences for sufficient duration that participants can grow their capabilities across all eight skills. In-person, full time availability for multiple months are preferred to virtual or short-term sessions where participants may be distracted by other responsibilities or hampered by local constraints.

Participants build valuable bonds with others in their cohort, the instructional staff, and other members of the scientific community of the hosting institution. We have found it helpful to purposefully build community by enrolling multiple participants together and by deliberately connecting participants with relevant professional associations such as the Applied Malaria Modeling Network (AMMnet). The bonds built during the program help sustain momentum after program completion. Targeting training to university faculty who can pass on their skills to their own trainees can help scale training efforts.

Whether training programmes are hosted by LMIC or northern institutions, trainers should be mindful of trainees’ different backgrounds, lived experiences, and modes of learning to which they are accustomed, in order to develop training material that can be delivered in a manner to maximize everyone’s ability to learn together.

Applied modeling is a highly cross-disciplinary field, requiring a unique mixture of hard skills, soft skills, and domain knowledge. Successful training programs are those that include hands-on simultaneous learning of multiple skills, delivered in a considered format to maximise the sustainability of training efforts.

About the authors:

Links:

https://www.globalhealth.northwestern.edu/

https://www.appliedhealthanalytics.org/

https://science.uct.ac.za/masha

Handles:

@FSMGlobalHealth

@ah4di

@sheetalsilal

@MASHA_UCT

Disclaimer: Views expressed by contributors are solely those of individual contributors, and not necessarily those of PLOS.

The post Designing effective training programs for applied mathematical modeling appeared first on Speaking of Medicine and Health.