Curation of Mini Mental State Examination (MMSE) Scores in the VA Million Veteran Program (MVP): Applications for Cognitive Aging Research
This study demonstrates that Mini Mental State Examination (MMSE) scores systematically extracted from the VA Million Veteran Program's electronic health records are valid for cognitive aging research, showing significant associations with APOE ε4 status and dementia diagnoses across diverse ancestral groups.
Original paper dedicated to the public domain under CC0 1.0 (https://creativecommons.org/publicdomain/zero/1.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Imagine the human brain as a vast, bustling city. Sometimes, the traffic lights start to flicker, the roads get clogged, and the city's memory of where it's been begins to fade. This is the reality of Alzheimer's disease and related dementias, a growing challenge that affects millions of older adults. Scientists have long known that a specific genetic "key," called the APOE ε4 allele, can make it harder for the city to keep its lights on, increasing the risk of this decline. However, studying this on a massive scale has been like trying to map the city's traffic by asking every single driver to fill out a detailed, handwritten map. It's slow, messy, and often incomplete. To solve this, researchers needed a way to quickly and accurately pull out the "traffic reports" already hidden inside the massive digital logbooks of the Veterans Health Administration (the VA), where millions of veterans' health records are stored. The big question was: Could we dig through these digital archives, find the cognitive test scores hidden in the text, and trust that they tell the true story of how the brain is aging?
This paper is the story of a team of data detectives who decided to try exactly that. They focused on a specific test called the Mini Mental State Examination (MMSE), a quick check-up that doctors use to see how well a person is thinking and remembering. In the VA's electronic health records, these scores aren't neatly stored in a spreadsheet; they are scattered like treasure chests inside thousands of free-text doctor's notes. The team built a clever, two-step system to find them. First, they used computer rules to scan the notes and grab any numbers that looked like test scores. But because computers can get confused by dates or other numbers, they didn't stop there. They brought in a team of human reviewers to double-check the findings, acting like expert editors who read the context to make sure the number they found was actually a valid test score and not just a date or a vital sign.
After curating this massive dataset, the researchers looked at nearly 50,000 veterans to see if their new method worked. They asked: Do the scores they found match what we already know about genetics and dementia? The answer was a resounding yes. They found that among veterans of European ancestry, the more copies of the risky APOE ε4 gene a person had, the lower their test scores tended to be—a clear "dose-response" pattern where two copies meant lower scores than one, and one meant lower than none. Interestingly, this pattern was a bit different for veterans of African and Hispanic ancestry; while those with the gene still scored lower on average, the strict "two copies is worse than one" pattern wasn't as clear-cut. Furthermore, when they looked at veterans over 65, the test scores lined up perfectly with dementia diagnoses. Veterans with lower scores were much more likely to have a diagnosis of Alzheimer's or other dementias, with the odds of having the disease skyrocketing as the scores dropped. The study concludes that this "digital archaeology" approach works: we can successfully extract and trust these cognitive scores from messy medical records, opening the door to studying brain health in huge, diverse groups of people without needing to re-test everyone from scratch.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.