AI-Driven Rapid Training of Health Emergency Response Workers

Using AI-driven patient simulations and learning analytics, Team Alelo’s solution rapidly trains workers for surge responses to public health emergencies. Data from trainees and practitioners are used to train, evaluate, and improve the simulation models, and to adapt training to keep pace with rapidly evolving health emergency conditions.

Dr. Lewis Johnson, President and CEO, Alelo Inc.

Public health emergencies such as COVID-19 require a surge capacity of trained public health workers. During the COVID-19 pandemic, it was suddenly necessary to hire 100,000 contact tracers in the US alone. Public health workers are now needed to administer vaccines, overcome vaccine hesitancy, and control new outbreaks of infection. The WHO projects a shortage of 12.9 million healthcare workers by 2035, making it difficult to combat emerging public health threats. Distrust of health workers has contributed to health crises such as Ebola. The key to effective healthcare emergency response is rapid training of sufficient numbers of public health workers who are trusted in the communities that they serve, e.g., community health workers (CHWs). Such training could build local capacity and reduce the strain on public health departments and international public health organizations. CHWs have communication skills that are critical for effective emergency response. Unfortunately, conventional training methods are not designed to train people quickly in critical public health competencies. Academic programs focus on academic knowledge and do not provide enough training in practical skills that can be applied immediately in a health emergency. Moreover, health emergency conditions can evolve rapidly, and so training must quickly adapt to keep pace.

We target workers who wish to be retrained as community health workers (CHWs), and CHWs who require updated training for new health emergencies. CHWs play an important role in the healthcare team due to their understanding of the local community, and the WHO has highlighted their role in augmenting health systems around the world. Local-community CHWs can gain trust in ways that government and foreign aid workers often cannot. Our solution provides rapid training so that trainees can quickly qualify to participate in health emergency response. This is particularly important in the post-pandemic economy, in which many jobs have been eliminated and workers have been dislocated. To refine our solution, we are interviewing both practicing community health professionals and displaced workers seeking employment in public health, to understand the barriers to employment that they face and their potential for transitioning into public health. Our pilot project addresses communities in Hampton Roads, Virginia that are underemployed and are especially at risk due to COVID-19. This includes minority populations as well as military veterans. We then plan to expand globally, in partnership with healthcare organizations operating in various countries.

We plan to publish additional peer-reviewed articles from this work. As we scale our solution to larger numbers of trainees, and we have received ample training data, we will be able to bring operational costs for the unsupervised self-study version of the training program down below £1 per trainee. Through a combination of grants, sponsorships, and paid subscriptions, we can then offer the self-study version free of charge. We also are interested in sharing anonymized data sets with other research organizations.

This solution can be applied in many locations to augment health system capacity, and so can have significant global impact on the ability of health systems to respond to health emergencies and promote population health generally, especially in settings with shortages of capable health professionals. Our competency-centred approach can be readily adapted to training requirements in various countries, as well as local languages. The community-based approach supports communities that are underserved by healthcare systems. Furthermore, it can be a driver of economic growth in such communities through the creation of qualified employment opportunities, in particular for women, since most community health workers are women.

In the next year, after we complete pilot trials, we plan to scale up the CHW solution nationally in the United States. This involves aligning the curriculum with CHW Core Consensus (C3) competency standards and then with the standards in local jurisdictions as appropriate. We will continue to collect training scenario data and use it to update the training in response to the evolving conditions of the COVID-19 pandemic. We will conduct pilot trials of CHW training in other countries, most likely starting in Canada and the UK. We will extend the training platform to rapid training in other occupations.

In three years’ time, we plan to expand broadly into countries with significant broadband Internet penetration. This includes many countries in Europe, Latin America, and East and Central Asia. From there we can expand into countries with less Internet connectivity.

A key part of our scale-up strategy is opening Enskill to expand data collection and authoring. We will have collaborators familiar with health systems in different countries annotate content and contribute data to reflect local conditions. Likely partners would be organizations that are already involved in building local health system capacity such as Médecins sans Frontières and Partners in Health.

As noted above, key metrics of success are training speed, course enrolments and completions, operational cost per trainee, and job placement rates. Previous Alelo training courses have demonstrated rapid training, e.g., an immersive training program for Arabic that trained functional communication skills in as little as 40 hours. Cumulative training enrolments for US Government unsupervised self-study cultural training courses is 535,000, with a completion rate of 39%. Operational costs per trainee for Enskill courses incorporating speech data collection and speech and language processing are typically £0.20 per month.

In the current CHW pilot we are targeting a training speed of twice that of the typical CHW course. Integrating team learning activities and progress monitoring by a program navigator and instructor will help us achieve completion rates that are much closer to 100%. We then will seek to scale up usage while continuing to keep completion rates high and operational costs low.

Expansion will require adapting course content to local certification requirements. However, since most US jurisdictions align with the C3 competency standards, we can reach many jurisdictions without much customization. We will focus on the most stringent local competency standards, to ensure the broadest possible adoption. We are members of the Community-Based Workforce Alliance, which can connect us with community health organizations across the United States as well as in other countries. Another potential barrier are national training standards that can make it difficult to certify training internationally. We plan to work first with countries such as Canada and the UK that have national health systems and nationwide training standards for CHWs. We also plan to work with ESTICOM, the EU health programme that is promoting Europe-wide CHW training.

Expansion will require reducing the time required to update content. To accomplish this, we will extend Enskill so that healthcare partners and even healthcare workers can annotate and contribute data to training scenarios. This will enable us to rapidly retrain the avatar dialogue models and update training content to reflect current conditions in the health emergency.

Additional funding will be required, as we discuss below in the financial plan.

Alelo Inc.

We are looking for help from the Trinity Challenge in achieving global scale. Although we have experience in marketing training solutions globally, we have not yet delivered public health solutions to a global audience. We recognize that strong partners are needed to facilitate entry into international markets, particularly in healthcare, where there are many national regulations to consider. We anticipate that you have contacts and experience in this area that we could benefit from.

We are interested in the financial support that The Trinity Challenge can provide, but we are pursuing other funding opportunities as well. We believe that recognition by the Trinity Challenge will give us added credibility with other investors and sponsors, which will aid our other fundraising efforts.

• The Bill and Melinda Gates Foundation. The Foundation works with organizations such as the Africa CDC to set up training programs for public health officials. Could they help connect us with similar organizations setting up training programs?
• Google. We are using Google Classroom as part of our solution. We provide Google Chromebooks to trainees in our pilot who lack adequate computing resources. Does Google recommend some sources for Chromebook equipment grants?
• Microsoft. We utilize Microsoft AI technology in our solution. Is Microsoft interested in supporting our efforts?
• Northeastern University. Northeastern is involved in community health education. Is there a possibility for us to collaborate, perhaps in curriculum development and certification?
• Optum. Does Optum engage in community health education? Could our solution help train their community health workers?
• Patrick J McGovern Foundation. How can we align our efforts with the Foundation’s goal to develop AI and data science for good?
• Reckitt. Can we align our efforts to train CHWs to promote health and wellness with Reckitt’s goal of promoting a cleaner, healthier world?
• Tencent. We would like to offer our AI-driven solutions in China. Could Tencent help facilitate this?