Researchers used artificial intelligence to create tools that can predict which postmenopausal women are most likely to break bones in the future. By studying two large groups of women over 8-10 years, scientists found that a computer program could accurately identify fracture risk by looking at previous breaks, bone density scans, vitamin D levels, and a hormone called PTH. This new approach could help doctors catch women at high risk before a fracture happens, allowing them to take preventive steps like medication or lifestyle changes.

The Quick Take

  • What they studied: Can a computer program predict which postmenopausal women will break bones by analyzing their medical information?
  • Who participated: Two separate groups of postmenopausal women (women past menopause) were followed for 8-10 years. Researchers developed their prediction tool using one group and tested it on another independent group to make sure it worked accurately.
  • Key finding: The AI tool correctly predicted fracture risk 92% of the time for women with osteoporosis and 88% of the time for all postmenopausal women. The most important clues were previous fractures, bone density measurements, vitamin D levels, and PTH hormone levels.
  • What it means for you: If you’re a postmenopausal woman, this research suggests doctors may soon have a better way to figure out your fracture risk before a break happens. This could lead to earlier treatment options. However, this is still new technology and should be used alongside your doctor’s regular care, not as a replacement.

The Research Details

Scientists used a method called machine learning, which is when computers learn patterns from data to make predictions. They studied two separate groups of postmenopausal women over many years (8-10 years of follow-up). The first group (HURH Cohort) was used to teach the computer program what patterns lead to fractures. The second group (Camargo Cohort) was used to test whether the program’s predictions actually worked in real women.

The researchers created two different prediction tools: one specifically for women already diagnosed with osteoporosis, and another for all postmenopausal women. For each tool, they tested different combinations of medical information to see which pieces of data were most helpful for making accurate predictions.

This approach is stronger than just looking at one group because it shows the tool works in different populations of women, not just the original group studied.

Using two separate groups of women (one to develop the tool, one to test it) is important because it shows the predictions actually work in real life, not just in the original study. This is called ‘validation’ and makes the results more trustworthy. Machine learning can spot patterns that doctors might miss, potentially catching high-risk women earlier.

This study has several strengths: it followed women for a long time (8-10 years), used two independent groups to test the tool, and was conducted at multiple medical centers. The high accuracy rates (92% and 88%) suggest the tool works well. However, the paper doesn’t specify exactly how many women were studied, which would help readers understand the study’s scope better. The tool should be considered a helpful addition to regular doctor visits, not a replacement for clinical judgment.

What the Results Show

The machine learning tool was highly accurate at predicting fracture risk. For postmenopausal women with osteoporosis, the tool correctly identified who would break bones 92% of the time. For all postmenopausal women (including those without osteoporosis), it was accurate 88% of the time. These are very high accuracy rates in medical prediction.

The study found that four pieces of information were most important for making these predictions: (1) whether a woman had broken a bone before, (2) bone density measurements from a DXA scan (a special X-ray that measures bone strength), (3) vitamin D levels in the blood, and (4) PTH hormone levels. When doctors had access to these four pieces of information, the computer program could make the best predictions.

Interestingly, the researchers found that using the most commonly available medical information (rather than rare or hard-to-get tests) gave the best results. This means the tool could be practical for regular doctor’s offices to use.

The study showed that previous fractures were one of the strongest warning signs for future breaks. Women who had already broken a bone were at much higher risk of breaking another one. The bone density scan (DXA) was also very important—women with lower bone density had higher fracture risk, which confirms what doctors already knew but shows the AI tool recognizes this pattern well.

This research builds on existing knowledge that previous fractures, low bone density, and low vitamin D increase fracture risk. What’s new is using artificial intelligence to combine all these pieces of information together to make more accurate predictions than doctors could make by looking at each factor separately. The high accuracy rates suggest this AI approach is better than traditional methods.

The study doesn’t clearly state how many women participated, which makes it harder to understand how large and representative the research was. The study focused only on postmenopausal women, so the results may not apply to younger women or men. The tool was developed and tested in specific populations, so it may work differently in other groups of women with different backgrounds or health characteristics. Additionally, this is a new tool that needs more testing in real-world doctor’s offices before it becomes standard practice.

The Bottom Line

If you’re a postmenopausal woman concerned about bone health, discuss fracture risk assessment with your doctor. Ask about bone density testing (DXA scan) and vitamin D level checks, as these are key to understanding your risk. This research suggests that combining these measurements with information about previous fractures gives doctors a better picture. However, this AI tool is still new and not yet standard in all doctor’s offices. Continue following your doctor’s current recommendations for bone health, which may include calcium and vitamin D supplements, exercise, and possibly bone-strengthening medications if you’re at high risk. Confidence level: Moderate—the research is promising but needs more real-world testing.

This research is most relevant for postmenopausal women, especially those with osteoporosis or a history of broken bones. Women with low vitamin D levels or family history of osteoporosis should also pay attention. Healthcare providers and researchers working on bone health will find this valuable. This doesn’t apply to premenopausal women or men, though similar tools might be developed for them in the future.

If your doctor uses this tool to identify high fracture risk, preventive treatments (like bone-strengthening medications) typically take 1-2 years to show measurable improvements in bone density. Lifestyle changes like exercise and better nutrition can help over similar timeframes. Fracture prevention is a long-term commitment, not something that shows quick results.

Want to Apply This Research?

  • Track your bone health markers monthly: record any new fractures or falls, note your vitamin D supplement intake, and log weight-bearing exercises (walking, dancing, strength training). If you’ve had a DXA scan, record the date and results to monitor changes over time.
  • Use the app to set reminders for vitamin D supplementation, schedule regular weight-bearing exercise sessions (aim for 30 minutes most days), and track calcium intake through meals. Set a goal to discuss fracture risk assessment with your doctor within the next 3 months if you haven’t had recent bone density testing.
  • Create a bone health dashboard showing: (1) exercise frequency and type, (2) supplement adherence (calcium and vitamin D), (3) any falls or injuries, and (4) dates of medical appointments and bone density tests. Review monthly to identify patterns and share with your healthcare provider at annual checkups.

This research describes a new artificial intelligence tool for predicting bone fracture risk in postmenopausal women. The tool is not yet standard medical practice and should not replace consultation with your healthcare provider. If you have concerns about bone health, osteoporosis, or fracture risk, please discuss them with your doctor who can order appropriate tests and recommend personalized treatment. This article is for educational purposes and does not constitute medical advice. Always consult with qualified healthcare professionals before making changes to your health regimen or treatment plan.