Predicting Osteoporosis using Machine learning

Posted by Akash Gutha on July 23, 2016

What is Osteoporosis? Osteoporosis is a disease where increased bone weakness increases the risk of a broken bone. It is the most common reason for a broken bone among the elderly. Bones that commonly break include the vertebrae in the spine, the bones of the forearm, and the hip.

Is there any middle stage? Yes. it’s called Osteopenia. Osteopenia is a condition in which bone mineral density is lower than normal. It is considered by many doctors to be a precursor to osteoporosis. However, not every person diagnosed with Osteopenia will develop osteoporosis.

Who are the common victims? Osteoporosis becomes more common with age.According to Wikipedia, About 15% of white people in their 50s and 70% of those over 80 are affected. It is more common in women than men. In the developed world, depending on the method of diagnosis, 2% to 8% of males and 9% to 38% of females are affected. In my personal experience women especially elderly mothers (my own mother) are the main victims.

Is this curable? At the time of writing, No.

Is it controllable? At the time of writing. Some techniques might work, but not for everyone.

How do we tackle this issue? Premature care has shown positive effects. So, this research’s the aim is to collect data from existing patients and use it to predict the probability that a person might develop Osteoporosis.

How do we collect the data? There are many ways we can get meaningful data regarding Osteoporosis. Many of them are invasive or require heavy equipment. But for this experiment, i teamed up with a friend of mine. We gather frequency of vibration from the Tibial bones using impulses from a medical hammer and an accelerometer.

Link to the project: Predicting Osteoporosis