Using A.I. for Preventative Care

jordangiali

One of the most promising use cases for A.I. in frontline health is preventative care, including behavior changes around diet and exercise.

• What are some other ways that A.I. can be used to increase preventative care in low- and middle-income countries?
• What are the challenges in using this technology for preventative care? For example, there could be cultural barriers to adopting this technology.

Please link to any relevant research reports.

Shashi

Hi @Shabbir, @synhodo, @jonc101 and @kenjisuzuki - As you all have background in AI/ML, we feel you may have thoughts on how AI can be used to increase preventive care in LMICs. Please join the discussion to share your thoughts. Thanks.

Shashi

Hi @msrjoy, @ajchenx, @elekaja, @jda - We would love to hear your thoughts on using AI for preventive care. Thanks.

NickOttens

@henry_bro, I know you have expertise in preventative care as well. I wonder if you have any suggestion?

This discussion is related to one of the other prizes we're designing: Frontline Health.

NickOttens

@nastyahaut, you may be interested in this discussion as well.

@Shashi, @jordangiali, Anastasia is the Founder of Haut.AI, an AI-powered platform for skincare and pharma companies.

nastyahaut

Hi @jordangiali, @NickOttens thanks for thinking of me.
While some biological data types like "omics" can be quite expensive and difficult to obtain, other data types are more affordable. For example, our research group works on nin-invasive imaging ageing biomarkers (PhotoAgeClok algorithms - published by our group https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6286834/) and biomarkers of skin health.

Skin diseases rate might be underestimated since not all affected people address a doctor (https://doi.org/10.1111/jdv.15494 ). At the same time, they significantly affect the quality of life, may cause social distancing, and become life-threatening over time (Actinic Keratosis lesions may evolve in malignant skin lesions).

Extracting skin ageing and skin health biomarkers from images technology-wise is possible. What is more important, one can easily track progress over time in response to anti-ageing interventions, lifestyle changes. The limitation is image quality and validation in longitudinal studies.

Shashi

Thanks @nastyahaut for sharing your thoughts.
Hi @ajchenx, @joshnesbit and @owen - Do you have any thoughts on using AI for preventive care. Thanks.

ajchenx

@Shashi Yes, we have been building information technologies to increase adoption of clinical preventive care, including AI. The problems we have identified and rough approaches we are taking:
1. How to enable less trained village health workers to know a patient needs a preventive measure now? We are designing automated tool or AI-powered chatbot to assist health workers to check symptoms and learn what to do on the spot. If needed, health workers can also use the tool to easily refer patient to the right clinic and doctor.
2. How to disseminate clinical prevention guidelines and best practices to people who need it in villages and communities? Precision dissemination is a key and AI should be part of the solution.
3. How to educate and engage people to take preventive measures. Again, AI could play a key role in figuring out when, what and how to deliver patient education contents and prevention reminders to an individual, which can drive the individual to take action.

Shashi

Thanks @ajchenx for sharing your experience. Great Ideas.
Hi @addy_kulkarni, @Mellie64, @ClaireM and @dollendorf - Do you have any ideas on how we can use AI to improve preventive care?

jordangiali

@nastyahaut @ajchenx Thank you both for your feedback!

Nvargas2

Excited to join this community to contribute to a very important topic!

Nvargas2

I have come across conversational artificial intelligence (AI) approaches to promote chronic disease prevention. Previous studies have demonstrated value in the implementation of a fully automated text-based mobile coaching service for prevention purposes. One such intervention, specifically, promotes health behaviors related to diabetes prevention. I would like to share a study that measured user acceptability of AI coaches as alternatives to live health care professionals. These are good interventions to consider which may inform the variety and acceptability of these technologies among users, and these are particularly feasible and low tech for MLICs. Here is the link to the article that I referenced: https://diabetes.jmir.org/2017/2/e28/?utm_source=TrendMD&utm_medium=cpc&utm_campaign=JMIR_Diabetes_TrendMD_0

Nvargas2

@Shashi your idea to train less skilled health village workers to identify patients in need of health prevention efforts now is something that has been successfully done. Anecdotally through my experience in identifying children who are malnourished in the remote mountains of the Ecuadorian Andes is one example—the village workers were paid (small incentives) to work in partnership with the hospital. The two ladies from the Quichuan village knew the land very well and brought me along places where I would never be able to go on my own to measure children and derive z-scores. The effort was slow, and could have been boosted by AI in many ways. Particularly if I was able to have an automated tool that created a patient registry with spatial capabilities because the remoteness of the land made it difficult for my medical colleagues to follow up with the families. The village workers were key to identifying these remote villages and getting me access to the community. However, it took a month of training them, and building trust to get there. Once I got their trust and buy-in, they were extremely powerful stakeholders. I was able to update the Hospital with a new approach to identify and detect families at higher risk (as defined by one or more children with low Z score) and we identified a process, bought new equipment, and overall the small intervention improved the quality of nutritional services. I would have loved to have AI powered tools to be able to build on this intervention and scale up. So if it helps at all, through my experience, I know with good planning and the right partnerships, it can be done!

Shashi

@Nvargas2 - Welcome Natalia! We're excited to have you join our online community of experts. Thank you for sharing conversational AI approaches, other resource and your experience on skilling village health workers and partnering with them in your health prevention efforts. Your comments have been very insightful.

I also wanted to understand from you did you personally train the local health workers?
if yes; What problems you encountered while approaching the local health workers? What were the other challenges in providing training to the local health workers. What were the incentives paid?
If not; Can you connect us to people who provide training to local health workers. We would like to invite them to the community, as it would be interesting to know the challenges they face while approaching / training local health workers in low and middle income countries.
Thanks.

Shashi

Hi @Lauren, @andwhite and @gajewski - You have ideas on this topic. Please share it with us. Thanks!

Nvargas2

Thank you! In terms of What problems we encountered while approaching the local health workers, there were a few that we learned from designing the program with the two PIs that collaborated on this project (both University/academic PIs). First, that we underestimated the time it takes to build trust within the community in order for the training to be perceived as important and helpful. We had to battle the cultural beliefs related to being foreigners in that village. Many people began to talk about our intent to sell their blood to the United States. As a result, I pulled in two anthropology doctoral students to assist me with the implementation of a community-wide health fair and offered free, cultural appropriate food which was a huge hit! We had local health workers participate as mentors in a train-the-trainer model. In terms of Participation incentives, we provided $5 to thank participants for their time and guaranteed treatment of the children were found to be at risk of malnutrition or malnourished. The $5 token of appreciation per family went a long way as these folks lived on less than $2 per day. I hope this helps. This was a project funded by the University of South Florida, and I have worked closely with an expert on this type of work who has much experience in engaging healthcare workers in the field. He operates sentinel labs for parasite testing in several parts of South America. His name is Dr. Ricardo Izurieta—I learned a great deal from this fantastic colleague. I can find his email and post it as a follow up message.

Shashi

@Nvargas2 - Thanks Natalia for sharing this experience with us. This helps us to understand the cultural barriers one faces. Thank you for sharing Dr. Ricardo details.