Researchers looked at 50 recent studies testing special devices that can automatically detect when people eat and drink. These devices use sensors and artificial intelligence to track food intake without people having to write everything down. While no device is perfect yet, the good news is that newer versions are becoming more comfortable to wear and faster at detecting meals. The main challenges remaining are battery life and the ability to work in real-time. This technology could eventually help doctors and nutritionists give personalized eating advice automatically, making it easier for people to get help with their diet.

The Quick Take

  • What they studied: Whether new sensor-based devices (like smartwatches and special glasses) can accurately detect when people eat and drink, and if they’re ready to be used by doctors and nutritionists in real life.
  • Who participated: This review analyzed 50 published studies from 2022-2025 that tested 52 different sensor devices designed to detect eating and drinking behaviors.
  • Key finding: No device currently meets all the requirements needed for real-world use, but recent improvements show promise. The biggest remaining problems are battery life (43 devices failed this test) and real-time processing speed (37 devices).
  • What it means for you: Smart eating-detection devices aren’t quite ready for everyday use yet, but they’re getting closer. Within a few years, you might be able to wear a device that automatically tracks your meals and gives you instant feedback to help you eat healthier—without having to manually log everything.

The Research Details

Researchers conducted a scoping review, which is like a comprehensive survey of recent scientific literature. They searched for all studies published between January 2022 and September 2025 that tested sensor-based devices for detecting eating and drinking. A scoping review is broader than a traditional systematic review—it’s designed to map out what research exists on a topic and identify gaps in knowledge.

The researchers evaluated each device against six practical requirements: achieving at least 80% accuracy, allowing people to eat any foods they want, being socially acceptable and comfortable to wear, having long battery life, providing real-time detection (instant feedback), and being able to detect both eating and drinking. These criteria were chosen because they represent what would be needed for these devices to actually work in a doctor’s office or at home.

The review included 50 studies testing 52 different devices. The devices came in different forms: some worn on the wrist (like smartwatches), some around the neck, some on the ear, some as special glasses, and some using multiple positions on the body.

This research approach is important because it shows the real-world readiness of technology, not just whether it works in a laboratory. By checking devices against practical criteria, researchers can tell us which ones are closest to being useful for actual patients and doctors. This helps guide future development and tells us how soon this technology might become available.

This is a well-structured review using established guidelines (PRISMA-ScR framework) for conducting scoping reviews. The researchers clearly defined what they were looking for and used consistent criteria to evaluate all devices. The main strength is that it focuses on recent research (2022-2025), capturing the latest advancements. One limitation is that the review depends on how well the original studies reported their results—if a study didn’t mention battery life, for example, it might have been marked as not meeting that criterion even if the device actually had good battery life.

What the Results Show

The review found that while sensor technology is advancing, no single device currently meets all six requirements for real-world use. The most common problem was battery life—43 out of 52 devices couldn’t run long enough to be practical for daily use. The second biggest issue was real-time processing, where 37 devices couldn’t provide instant feedback about what was eaten.

However, there’s encouraging news: recent devices are becoming more socially acceptable and comfortable to wear. This is a significant improvement from earlier versions that were bulky or obvious. The researchers also noted that computational efficiency (how fast the device can process information) has improved, suggesting that real-time detection may soon become feasible.

Wrist-worn devices (like smartwatches) were the most common type tested (19 devices), followed by neck-worn devices (8) and ear-worn devices (7). Interestingly, different device types had different strengths and weaknesses—for example, some were more accurate but less comfortable, while others were comfortable but less accurate.

Only 10 devices achieved the 80% accuracy target, meaning most devices still make mistakes when identifying what people are eating. About 31 devices were tested on a limited variety of foods, which means we don’t know how well they’d work with all the different foods people actually eat in real life. Similarly, 31 devices couldn’t detect both eating and drinking—they were good at one but not the other. Fifteen devices had issues with social acceptability and comfort, meaning people might not want to wear them regularly.

The researchers had conducted a similar review in 2021 that concluded these devices weren’t ready for practical use. This new review shows meaningful progress in some areas, particularly in making devices more comfortable and improving how fast they can process information. However, the fundamental challenges of battery life and real-time detection persist, suggesting that while progress is real, it’s been slower than some might have hoped.

The review’s findings depend entirely on how well the original studies reported their results. If a study didn’t mention battery life or comfort, the device was marked as not meeting those criteria, even though it might have actually performed well. Additionally, the review only included published studies, so unpublished research or devices still in development weren’t included. The criteria used (like 80% accuracy) are somewhat arbitrary—a device with 75% accuracy might still be useful in some situations. Finally, how well a device works in a research study might differ from how it works in real life, where people eat different foods and in different environments.

The Bottom Line

At this time, sensor-based eating detection devices are not ready to replace traditional dietary tracking methods in medical practice (low confidence). However, they show promise for future use, and development should continue focusing on battery life and real-time processing improvements (moderate confidence). If you’re interested in trying these devices for personal use, understand that they may not be perfectly accurate and should be used as a general guide rather than precise measurement (low to moderate confidence).

This research matters most to doctors, nutritionists, and health researchers who want to help patients track their eating more easily. It’s also relevant to people with chronic diseases like diabetes or obesity who struggle with manual food tracking. Technology companies developing these devices should pay attention to the specific barriers identified (battery life and real-time processing). The general public should be aware that while these devices are improving, they’re not yet ready for widespread medical use.

Based on current progress, meaningful improvements in battery life and real-time processing could take 2-5 years. A device that meets most or all of the feasibility criteria might be available for clinical use within 3-7 years, though this depends on continued research funding and technological breakthroughs. Early adopters might see useful versions within 1-2 years, but they should expect some limitations.

Want to Apply This Research?

  • If using a sensor-based eating detection device, track the accuracy by comparing the device’s detected meals to your own food diary for one week each month. Record: number of meals detected, number of meals missed, and any false alarms (times the device thought you were eating when you weren’t).
  • Start by wearing the device during one meal per day to get comfortable with it, then gradually increase usage. Use any real-time feedback the device provides to pause and reflect on your eating choices in the moment. For example, if the device alerts you that you’re eating, use that as a reminder to eat slowly and notice how full you feel.
  • Track battery life by noting when you charge the device and how long it lasts. Also monitor comfort by rating how natural it feels to wear on a scale of 1-10 weekly. Keep notes on which foods or eating situations the device struggles to detect, as this helps identify its limitations and when you might need to manually log meals.

This review summarizes research on emerging technology and does not constitute medical advice. Sensor-based eating detection devices are not yet approved or recommended as replacements for traditional dietary assessment methods in medical practice. If you have questions about tracking your diet for health reasons, consult with your doctor or a registered dietitian. Any device you use should be considered a tool to support, not replace, professional nutritional guidance. The accuracy and reliability of these devices vary significantly, and results from research studies may not reflect real-world performance in your specific situation.