PhD Studentship Opportunity (2026–27) – Opportunities for Youth

Facilitating Complex Phenotyping for Electronic Health Records Research Using Large Language Models

MRC London Intercollegiate Doctoral Training Partnership (MRC LID)
London School of Hygiene & Tropical Medicine (LSHTM)

The London School of Hygiene & Tropical Medicine invites purposes for a doctoral analysis studentship exploring how Large Language Models (LLMs) can remodel the best way advanced ailments are recognized in massive-scale Electronic Health Records (EHR). This progressive mission gives the chance to contribute to rising strategies on the intersection of epidemiology, synthetic intelligence, and well being knowledge science.

---

Supervisory Team

Supervisor
Dr Julian Matthewman
Faculty of Epidemiology & Population Health
Department of Non-communicable Disease Epidemiology
Email: julian.matthewman@lshtm.ac.uk

Co-Supervisor
Professor Sinéad Langan
Faculty of Epidemiology & Population Health
Department of Non-communicable Disease Epidemiology
Email: sinead.langan@lshtm.ac.uk

---

Project Summary

Accurate identification of affected person teams—often known as phenotyping—is crucial for excessive-high quality epidemiological analysis. Yet for advanced situations, normal coding techniques resembling SNOMED CT and ICD-10 are sometimes inadequate. This PhD mission will discover how LLMs can assist the creation of richer, extra exact and nuanced phenotypes inside EHR datasets.

The pupil will develop a framework that integrates medical experience with AI outputs, conduct a scientific literature evaluate, and apply the framework to construct a “phenotype atlas” utilizing the UK’s CPRD Aurum database. The mission goals to advance using LLMs in well being knowledge analysis and create a helpful methodological useful resource for the broader analysis neighborhood.

Project Keywords: Phenotyping; Large Language Models; Electronic Health Records; Health Data Science

---

MRC LID Themes

---

Skills You Will Gain

• Large Language Model (LLM) analysis, together with immediate engineering and nice-tuning

• Electronic Health Record (EHR) administration and evaluation

• Epidemiological design and interpretation

• Quantitative knowledge evaluation in R or Python

• Natural Language Processing (NLP) for medical purposes

---

Study Routes

Full-time study: Yes
Part-time study: Yes

---

Location & Working Arrangements

Students are anticipated to work totally on website at LSHTM – Bloomsbury, London, in keeping with institutional expectations for analysis diploma programmes. Funded college students might also attend as much as three conferences in the course of the studentship.

Travel necessities for this mission: None past normal coaching/convention attendance.

---

Eligibility Requirements

Applicants should meet LSHTM’s normal doctoral admissions standards and will display:

• A robust quantitative background and a well being-associated background, usually by means of a Master’s diploma in epidemiology, well being knowledge science, medical statistics or a associated discipline

• Proficiency in Python or R

• Interest in AI/ML strategies and well being knowledge science

Industrial CASE (iCASE) conversion: Not obtainable

---

Project in More Detail

Routinely collected EHR knowledge allow massive-scale well being analysis, however phenotyping advanced ailments stays labour-intensive and reliant on professional interpretation. LLMs supply the potential to course of medical information at scale, aiding with advanced resolution logic and lowering the time required for professional evaluate.

Objectives

1. Conduct and keep a residing scoping evaluate of LLM use in well being knowledge classification and phenotyping.

2. Develop a clear, reproducible framework for LLM-assisted phenotyping.

3. Apply the framework to create an Atlas of Complex Disease Phenotypes inside CPRD Aurum (seemingly specializing in pores and skin or inflammatory ailments).

4. Compare key epidemiological measures (e.g., incidence, prevalence) derived from LLM-assisted phenotypes with these from normal definitions.

Data Source

The mission makes use of CPRD Aurum, one of many largest UK major care datasets, which incorporates diagnoses, signs, prescriptions, referrals and checks, with linkages to further knowledge sources.

Risks

Access threat is low attributable to LSHTM’s present institutional licence for CPRD Aurum.

Learn extra and Apply here

Related