A new computer tool developed by a University of California-Davis doctor can help predict - and may help prevent - potentially deadly hip fractures in women.
Internist John Robbins hopes the computer program can guide physicians in their efforts to protect vulnerable patients from fractures that too often lead to disability or death. Robbins' research appears in the latest edition of the Journal of the American Medical Association.
Robbins cautioned that his computerized algorithm is not foolproof, and that it needs to be tested further. But Robbins said he believes the modeling can help women get a more personalized indication of their risk.
Analyzing data collected from the nationwide Women's Health Initiative, Robbins and other researchers identified 11 risk factors for hip fractures among postmenopausal women. They go beyond age and weight and include things like steroid drug use, smoking and ethnic background.
Researchers then developed a model that can quickly evaluate an individual's risk of suffering a hip fracture within five years. The model has been adapted for public use via the Internet.
"Because of the hundreds of thousands of women involved in the Women's Health Initiative, we were able to develop a more powerful tool for the prediction of hip fracture than has been possible before," Robbins said.
Currently in the United States, there are an estimated 329,000 hip fractures per year, many of them resulting in long-term hospitalization and disability or causing complications that lead to death.
Robbins' tool relies on factors other than bone-mineral density associated with a higher risk of hip fracture. Taken together, they can further refine the predicted risk of future fractures.
Race, for example, can play a role in fracture risk, but must be weighed against other risk factors, said Robbins.
For example, black women generally have a much lower risk of hip fracture than white women. But a 60-year-old black woman in good health has just as high a risk of breaking a hip as a 60-year-old white woman in good health if the black woman has broken her wrist and has a mother who suffered a broken hip.
These variables can and should play a role in a doctor's decision to prescribe drugs called bisphosphonates, which help prevent osteoporosis, Robbins said. Osteoporosis is a bone-weakening disease that typically afflicts women.
"I don't like the choice to use medication to be based on one simple piece of information (such as a bone-density scan)," he said. "If a woman has a very low predictive score but has already broken a hip, she ought to be treated."
In addition to race, prior fractures after age 54 and family history of fractures, other variables used in Robbins' model are weight, long-term use of corticosteroid drugs (which can weaken bone), level of physical activity, cigarette smoking and diabetes.
Sherry Sherman, who directs clinical aging and reproductive hormone research for the National Institute on Aging, welcomed the predictive tool.
"Certainly, age is one of the most powerful effects, as is weight loss, and use of corticosteroids," she said. "We have all known these are factors, but what the authors have done in this paper is to actually put these together and convert them into a score so you can understand your risk. I think it's very helpful."
Sherman added that the tool, developed based on data from a largely healthy population, should now be validated using a population of less healthy people.
Robbins agreed, adding that more research is needed to determine whether medical treatment or other lifestyle changes would be beneficial for those deemed at high risk by the new model.