External validation and clinical utility of NELA and P-POSSUM risk models in an ageing Singapore emergency laparotomy cohort
This study externally validates NELA and P-POSSUM risk models in an ageing Singapore emergency laparotomy cohort, demonstrating that NELA offers superior discrimination, calibration, and clinical utility compared to P-POSSUM for perioperative risk stratification in this specific setting.
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
Every year, thousands of older adults around the world face a sudden, life-threatening abdominal crisis that requires immediate surgery. These are not planned operations; they are emergency procedures where a doctor must open the abdomen to fix a perforated organ, a blockage, or a burst blood vessel before the patient's condition worsens. Because these patients are often elderly and carry other health problems, the risk of dying within a month of the operation is high, ranging from six to twenty percent depending on the country and the specific hospital. To help doctors make the right choices about how much care a patient needs, medical teams use mathematical tools called risk models. These tools take information about a patient's age, heart function, and the type of surgery required to calculate a percentage chance of survival. Two of the most famous tools are the Portsmouth Physiological and Operative Severity Score, known as P-POSSUM, and the National Emergency Laparotomy Audit model, or NELA. While these tools were built using data from the United Kingdom and other Western nations, it has remained unclear whether they work just as well for patients in Asia, where the population is aging rapidly and healthcare systems operate differently.
In a large hospital in Singapore, researchers set out to test whether these two models could accurately predict outcomes for their own local patients. They looked back at the records of 393 adults who had undergone emergency abdominal surgery between late 2020 and mid-2022. The team focused on a single, stark outcome: whether a patient died within thirty days of the operation. By comparing the predictions made by the P-POSSUM and NELA calculators against what actually happened in the hospital, the researchers could see which tool was telling the truth and which was guessing. The study included a diverse group of patients, with a median age of sixty-nine years, and found that eight percent of them did not survive the month following their surgery. This real-world result served as the yardstick against which both models were measured.
The investigation revealed that while both tools could separate high-risk patients from low-risk ones to a reasonable degree, one was clearly superior. The NELA model proved to be the more accurate predictor, correctly identifying the likelihood of death in a way that matched the hospital's actual outcomes much more closely than the P-POSSUM model. The older P-POSSUM tool tended to overestimate the danger, predicting that many patients would die when they actually survived. In contrast, the NELA model was better calibrated, meaning its predictions of risk were closer to the reality of the patients' conditions. When the researchers looked at the numbers, the NELA model showed a stronger ability to distinguish between those who would survive and those who would not, while the P-POSSUM model frequently flagged patients as high-risk when they were actually safe.
Beyond simple accuracy, the study examined how useful these tools would be for a surgeon making a split-second decision in a busy emergency room. The researchers used a method that simulates different scenarios to see which model would lead to better clinical choices, such as deciding whether a patient needs intensive care or a senior specialist. They found that the NELA model offered a greater net benefit across the range of risks that doctors actually care about. It helped identify the right patients for extra care without unnecessarily alarming doctors about patients who were stable. The P-POSSUM model, on the other hand, became less useful as the threshold for concern rose, often suggesting that too many patients needed intensive resources when they did not.
The authors suggest that these differences likely stem from the unique nature of the Singapore healthcare system and the specific characteristics of its patients. The hospital where the study took place operates with a dedicated team of senior surgeons who review emergency cases immediately, and the facility has rapid access to intensive care beds. These organizational strengths may improve survival rates beyond what the patient's age or disease severity alone would predict, a factor that the older models do not fully capture. Because the P-POSSUM tool was built on data from a different time and place, it struggled to account for these local advantages, leading it to predict higher death rates than actually occurred. The NELA model, being newer and derived from a larger, more recent dataset, adapted better to this environment.
Ultimately, the study concludes that while both tools have value, the NELA model is the better choice for guiding care in this specific setting. It provides a clearer, more honest picture of the risks facing an aging population in Singapore. The findings serve as a reminder that medical tools developed in one country cannot be assumed to work perfectly in another without testing. Before a hospital adopts a new risk calculator, it must verify that the tool fits its own patients and its own way of working. In this case, the evidence points to NELA as the more reliable compass for navigating the high-stakes world of emergency abdominal surgery.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.