Risk Prediction Model for Urinary Retention After Lumbar Interbody Fusion Using Logistic Regression and Artificial Neural Network

Imagine your body as a bustling city where nerves are the telephone lines and muscles are the workers. Usually, when you need to use the bathroom, your brain sends a clear message down the line: "Relax the gate, squeeze the muscle, and let the water flow!" But sometimes, after a major construction project on the city's main highway—like surgery on the lower back—those phone lines get tangled, or the workers get too scared to move. This is called urinary retention, a common hiccup where a person simply can't pee even though their bladder is full. It's not just uncomfortable; it can lead to infections and make recovery take much longer. Doctors have long known that things like getting older, having certain health conditions, or feeling super nervous can make this "traffic jam" more likely, but they didn't have a crystal ball to predict exactly who would get stuck. That's where this study steps in, trying to build a digital crystal ball to spot the trouble before it happens.

This paper is like a team of detectives from Linfen Central Hospital who decided to solve the mystery of why some patients get stuck in the bathroom after a specific type of back surgery called "lumbar interbody fusion." Think of this surgery as swapping out a broken, wobbly floorboard in your spine and bolting new ones in place to make the whole structure solid. The researchers gathered data from 437 people who had this surgery, acting like detectives collecting clues. They asked questions about everything: How old are you? Are you male or female? Do you have diabetes? Are you worried about the surgery? Do you have an enlarged prostate? They even checked for things like constipation and depression.

Using two different "super-brains" to crunch the numbers, the team built a prediction model. The first was a Logistic Regression, which is like a very organized checklist that weighs each clue to see how much it adds to the risk. The second was an Artificial Neural Network, which is more like a video game AI that learns by looking at thousands of patterns to find hidden connections that a simple checklist might miss.

The investigation revealed that out of the 437 patients, 51 of them (about 11.67%) ended up having trouble peeing after their surgery. The detectives found five specific "super-suspects" that were the main culprits: being male, having diabetes, having benign prostatic hyperplasia (an enlarged prostate), feeling anxious, and having the surgery involve the lumbosacral area (the very bottom of the spine where the nerves for the bladder hang out).

The team then tested their new "crystal ball" on a group of patients they hadn't seen before. The results were incredibly sharp. Both the checklist model and the AI model achieved an Area Under the Curve (AUC) score of 0.974, indicating excellent predictive power. In the validation group, the specific accuracy was 96.2%. It's as if they built a weather forecast that could tell you with very high reliability whether it was going to rain on your specific street.

The paper doesn't just stop at the numbers; it explains why these suspects are dangerous. For men with an enlarged prostate, the "gate" is already narrow, and surgery makes it even harder to open. For people with diabetes, the "telephone lines" (nerves) might be damaged, so the message to pee never gets through. Anxiety acts like a panic button that tightens the muscles, making it impossible to relax and let go. And surgery right at the bottom of the spine is like construction right next to the main power switch for the bladder.

The authors are careful to say that while their model is excellent at predicting risk, it was built using data from just one hospital, so it needs to be tested in other places to make sure it works everywhere. But the takeaway is clear: by looking at these five specific factors, doctors can now identify high-risk patients early. Instead of waiting for a problem to happen, they can give extra help—like calming a nervous patient, adjusting medications, or keeping a closer eye on the bladder—to keep the "traffic" flowing smoothly and help everyone get back to their lives faster.