Lloyd Price
- Jan 23

The intersection of LLMs, EHRs and SDoH

Exec Summary:

The intersection of large language models (LLMs), electronic health records (EHRs), and social determinants of health (SDoH) holds immense potential for revolutionising healthcare delivery and improving population health outcomes. Here's how:

LLMs Analysing EHRs:

Extracting SDoH data: LLMs can analyse the vast amount of text data in EHRs to identify and extract information about SDoH factors, such as housing insecurity, food insecurity, and lack of social support, which are often hidden or unstructured.
Predicting health risks: LLMs can analyze patient data and SDoH factors to predict their risk of developing chronic diseases, allowing for early intervention and preventive measures.
Personalising care: LLMs can help healthcare providers tailor treatment plans and interventions based on a patient's individual SDoH needs and circumstances.

Challenges and Considerations:

Data privacy: Ensuring patient data privacy and security is crucial when using LLMs to analyze EHRs. Robust data governance and anonymisation techniques are essential.
Bias and fairness: LLMs trained on biased data can perpetuate health inequalities. Careful data selection and bias mitigation strategies are necessary to ensure fair and equitable outcomes.
Interpretability and explainability: LLMs' decision-making processes can be complex and opaque. Making their predictions and recommendations understandable for healthcare providers and patients is crucial for building trust and acceptance.

Opportunities for Collaboration:

LLMs and public health researchers: Collaborations can leverage LLMs to analyze large datasets and identify patterns in SDoH factors that contribute to health disparities.
LLMs and healthcare providers: LLMs can assist healthcare providers in identifying patients at risk, tailoring interventions, and managing chronic conditions more effectively.
LLMs and policymakers: LLMs can inform policy decisions by analyzing the impact of SDoH factors on health outcomes and identifying effective interventions to address health inequalities.

Overall, the integration of LLMs, EHRs, and SDoH data analysis holds promise for a more personalised, equitable, and data-driven approach to healthcare. However, addressing ethical considerations, ensuring data privacy, and fostering collaboration are crucial for realizing the full potential of this powerful combination.

Corporate Development for Healthcare Technology companies in EMEA

Healthcare Technology Thought Leadership from Nelson Advisors – Market Insights, Analysis & Predictions. Visit https://www.healthcare.digital

HealthTech Corporate Development - Buy Side, Sell Side, Growth & Strategy services for Founders, Owners and Investors. Email lloyd@nelsonadvisors.co.uk

HealthTech M&A Newsletter from Nelson Advisors - HealthTech, Health IT, Digital Health Insights and Analysis. Subscribe Today! https://lnkd.in/e5hTp_xb

HealthTech Corporate Development and M&A - Buy Side, Sell Side, Growth & Strategy services for companies in Europe, Middle East and Africa. Visit www.nelsonadvisors.co.uk

Large language models to identify social determinants of health in electronic health records

A recent paper published in 'npj Digital Medicine' discussed how 'SDoH are rarely documented comprehensively in structured data in the electronic health records (EHRs), creating an obstacle to research and clinical care.

Instead, issues related to SDoH are most frequently described in the free text of clinic notes, which creates a bottleneck for incorporating these critical factors into databases to research the full impact and drivers of SDoH, and for proactively identifying patients who may benefit from additional social work and resource support.'

Launched in 2014, the npj journals, which are part of the prestigious Nature Portfolio, collaborate with preeminent scientists and global partners to publish high-quality open access research. Since then, the series of journals has expanded to span the full gamut of research — from the physical and applied sciences, to the life and health sciences, and society & the environment.

The journals operate under a collaborative editorial model, with active researchers engaged as manuscript-handling editors, supported by in-house, PhD-trained Nature Portfolio managing editors. As online-only, open-access titles published by Springer Nature, the npj series brings together strong editorial leadership with a world-class publication infrastructure to deliver high-quality, peer-reviewed original research to the global scientific community.

Social determinants of health (SDoH) play a critical role in patient outcomes, yet their documentation is often missing or incomplete in the structured data of electronic health records (EHRs). Large language models (LLMs) could enable high-throughput extraction of SDoH from the EHR to support research and clinical care. However, class imbalance and data limitations present challenges for this sparsely documented yet critical information. Here, we investigated the optimal methods for using LLMs to extract six SDoH categories from narrative text in the EHR: employment, housing, transportation, parental status, relationship, and social support.

Source: https://www.nature.com/articles/s41746-023-00970-0

Source : https://www.nature.com/nature-portfolio/about/npj-series

The future of large language models, electronic health records and social determinants of health

The future potential for the intersection of LLMs, EHRs, and SDoH is brimming with possibilities to revolutionise healthcare delivery and address long-standing challenges. Here are some exciting prospects:

Personalised Medicine:

LLMs analyzing EHRs and SDoH data can create detailed patient profiles encompassing medical history, social factors, and genetic information. This can lead to tailored treatment plans,preventative measures, and interventions addressing the root causes of disease beyond just the biological aspects.
Predictive models powered by LLMs can flag individuals at high risk for specific diseases or complications based on their unique SDoH profile and medical data. This allows for early intervention and proactive steps to improve health outcomes.

Enhancing Healthcare Access and Equity:

LLMs can analyze large datasets to identify communities with high SDoH burdens and healthcare disparities. This information can inform policy decisions and resource allocation to address inequities and improve access to healthcare for underserved populations.
By analyzing language in EHRs and patient surveys, LLMs can detect implicit biases in healthcare practices and communication. This awareness can pave the way for more inclusive and culturally sensitive healthcare experiences for diverse groups.

Addressing Social Determinants of Health:

LLMs can help healthcare providers and communities develop targeted interventions addressing specific SDoH factors contributing to health disparities. This could involve connecting patients with social services, job training programs, or housing assistance.
LLMs can analyze communication patterns and sentiment in social media to understand broader SDoH trends and public health concerns. This information can inform public health campaigns and social policy changes to address systemic issues impacting health.

Research and Innovation:

LLMs can analyze vast amounts of EHR and SDoH data to identify previously unknown correlations between social factors, lifestyle choices, and disease patterns. This can lead to breakthroughs in our understanding of disease origins and potential new treatment strategies.
LLMs can accelerate clinical research by analyzing data from large patient cohorts to identify promising drug candidates and treatment pathways. This can streamline the research process and bring new therapies to patients faster.