First-day heart rate improves post-landmark mortality risk stratification in critically ill adults with type 2 diabetes and acute myocardial infarction: development in MIMIC-IV and multicentre validation in eICU
This study demonstrates that incorporating mean heart rate from the first 24 hours of ICU admission modestly improves post-landmark in-hospital mortality risk stratification for critically ill adults with type 2 diabetes and acute myocardial infarction, as validated across MIMIC-IV and eICU databases, though the resulting model did not outperform the established APACHE IVa score.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
When a person suffers a heart attack, the body enters a state of intense crisis, and doctors must act quickly to decide who needs the most urgent care. For patients who also have type 2 diabetes, this situation is even more complex. Diabetes can silently damage blood vessels and nerves, often masking the usual warning signs of a heart attack and making recovery harder. In the intensive care unit, where these patients are monitored closely, doctors rely on a constant stream of data—blood pressure, oxygen levels, and heart rate—to guess who might survive and who might not. The challenge is knowing exactly when to look at this data. If a doctor tries to predict a patient's fate using information gathered over the first day, they cannot use that information to predict deaths that happen before the day is over. To solve this, researchers have begun using a method called a "landmark" approach. This means they wait until a specific time, such as twenty-four hours after a patient arrives, and then ask a new question: among the patients who are still alive and in the hospital at that moment, what does their first day of data tell us about their chances of surviving the rest of their stay?
A team of researchers set out to answer this question for a specific group of high-risk patients: adults with type 2 diabetes who were admitted to the intensive care unit after a heart attack. They wanted to know if the average heart rate during that first twenty-four hours could improve the accuracy of survival predictions, beyond what doctors could already guess from basic facts like age, gender, kidney function, and oxygen levels. To find the answer, they turned to two massive, anonymized databases of real patient records from hospitals in the United States. They first built their prediction tool using data from one large hospital system, and then they tested it strictly on data from dozens of other hospitals to see if it would hold up in the real world.
The researchers focused on patients who had survived the first twenty-four hours in the intensive care unit. They started with a simple model that used only four pieces of information: the patient's age, their sex, a measure of how well their kidneys were working, and their average oxygen level during that first day. This basic model gave a reasonable estimate of risk. However, the team suspected that the heart rate, which is recorded constantly by monitors, might hold a hidden clue. They created a second, slightly more complex model that added the average heart rate from that first day to the original four factors. When they tested this new model on the second set of hospitals, the results showed a clear improvement. The ability of the model to correctly distinguish between patients who would survive and those who would not increased noticeably when the heart rate was included. Specifically, the measure of this predictive accuracy rose by a small but meaningful amount, moving from a score of roughly 0.69 to 0.74. This suggests that a faster average heart rate during the first day is a genuine warning sign of higher risk for these specific patients, even after accounting for their age, kidney health, and oxygen levels.
The study also looked at other potential clues to see if they offered similar help. The researchers tested whether changes in blood pressure over the first day could improve the prediction, but they found that this information did not add any stable value. They also tried adding liver function tests, but because those records were often missing from the hospital databases, the model became unreliable. Furthermore, when they compared their new, heart-rate-based model against a well-known, complex scoring system called APACHE IVa, their simpler model did not perform better. The established system still predicted outcomes more accurately in this specific test. This is an important finding because it shows that while heart rate is a useful addition, it is not a magic bullet that replaces the need for comprehensive, detailed scoring systems used in critical care.
The researchers were careful to explain what their findings mean and, just as importantly, what they do not mean. They emphasized that this model is not a tool for deciding who gets admitted to the intensive care unit in the first place. Instead, it is designed as a reassessment tool for the second day of a hospital stay. It helps doctors look at a patient who has already survived the first twenty-four hours and understand if their heart rate suggests they are still in danger. The study also noted that the model tended to predict slightly higher risks than what actually happened in the external hospitals, meaning that if a doctor were to use this tool in a real hospital, they would need to adjust the numbers locally to make them fit their specific patient population.
Ultimately, this work provides a clearer picture of how to use the data already available in hospital monitors. It confirms that for patients with diabetes and heart attacks, the average heart rate over the first day is a valuable piece of the puzzle for predicting survival. However, the researchers concluded that this tool is not yet ready for daily use in a hospital ward. Before it can be trusted to guide life-or-death decisions, it needs to be tested again in a forward-looking study where doctors use the model in real time and see how it affects patient care. For now, it stands as a scientifically sound step toward better risk assessment, proving that even a simple number like the average heart rate can add important information when the right questions are asked at the right time.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.