Improving risk prediction models for high-risk patients using electronic health records
Official title: Addressing Algorithmic Unreliability and Dataset Shift in EHR-based Risk Prediction Models
Funded at University of Pennsylvania · NIH-11092330 · Data updated 2026-11-12
This study is working on improving computer programs that use health records to find patients with advanced cancers who might be at higher risk, especially by looking at how changes in healthcare during the COVID-19 pandemic affect these predictions, so that patients can get better care tailored to their needs.
Quick facts
| Grant type | Fellowship grant |
|---|---|
| Study type | NIH-funded research |
| Funding institution | University of Pennsylvania NIH-funded |
| Lab location | 1 site (Philadelphia, United States) |
| Project start | June 1, 2023 |
| Project end | May 31, 2026 |
| Funding year | 2025 |
| Award (this year) | $37,360 |
| NIH project number | 5F31LM014282-03 |
| Project ID | NIH-11092330 on NIH RePORTER |
Who could benefit from this research
✓Could help
Ideal candidates for this research are patients with advanced cancers who are being monitored through electronic health records.
✕Less likely to help
Patients with stable, non-complex conditions or those not utilizing electronic health records may not benefit from this research.
What this research studies
This research focuses on enhancing predictive algorithms that utilize electronic health record (EHR) data to identify high-risk patients, particularly those with advanced cancers. It addresses the issue of dataset shift, where changes in patient data over time can lead to inaccurate predictions. By analyzing how shifts in healthcare practices, especially during the COVID-19 pandemic, affect these algorithms, the research aims to develop more reliable models for clinical decision-making. Patients may benefit from improved risk assessments that lead to better-targeted care and resource allocation.
Why it matters
Potential benefit: If successful, this research could lead to more accurate risk predictions for patients, ensuring they receive timely and appropriate care.
How similar studies have gone: Previous research has shown that addressing dataset shift can significantly improve the performance of predictive models in healthcare settings.
Where this research is happening
-
University of Pennsylvania
Philadelphia, United States Active
Researchers
- Principal investigator: Kolla, Likhitha — University of Pennsylvania
- Study coordinator: Kolla, Likhitha
Related research
Other NIH-funded projects on the same topics.
-
Comparing universal and guideline-directed genetic testing for young adults with cancer
This study is looking at whether a broader genetic test can help young adults with cancer find inherited risks that regular tests might miss, so they can get better treatment and care tailored just for them.
-
Better, searchable data for human pain biology
A data integration center will standardize, combine, and share molecular and cellular data from human pain research to help scientists understand how pain works and to accelerate discoveries.
-
Understanding how sweat glands form and work to help burn and skin injury patients
A study at the University of Pennsylvania looks at how sweat glands develop and function, with the goal of informing treatments for people who suffer traumatic skin injuries.
-
Cardiovascular risk planning and reduction for men with prostate cancer
A program to better predict and reduce heart disease risk in men treated for prostate cancer, especially older patients.
-
Understanding how a special airway cell helps protect and repair breathing passages
A researcher is studying a newly found airway cell called the hillock to see how it protects the airways and helps repair them after injury, with the aim of helping people with asthma, COPD, or bronchiectasis.
-
Defining TMJ cartilage progenitors and their regulatory niche for TMJ regeneration
Researchers are studying the cells that could help repair the jaw’s cartilage in TMJ osteoarthritis to lay the groundwork for regenerative therapies.
About this research
- This is an active NIH-funded research project — typically early-stage science, not a clinical trial accepting patient enrollment.
- Some NIH-funded labs run parallel clinical studies or seek volunteers for related work. To check, contact the principal investigator or institution listed above.
- For full project details, budget, and progress reports, visit the official NIH RePORTER page.