When we think of patient medical records, a lot of us think of billing and coding and maybe of health-care providers communicating with one another about how patients are doing. But increasingly medical records are becoming grist for the big-data mill.
According to Nigam Shah, MBBS, PhD, associate professor of biomedical informatics research at Stanford, it's now possible to artfully extract important biomedical information from pre-existing patient medical records. Such data can be anonymous for the patient, and it's virtually free for researchers, especially compared to the high cost of lengthy clinical trials that enroll thousands of people.
A just-published study by Shah and his colleagues used patient records to examine a suspected connection between a treatment for prostate cancer and the subsequent risk of developing Alzheimer's disease. Among a group of about 17,000 prostate cancer patients, those treated with a medication that suppresses testosterone -- so-called androgen blockers -- had nearly twice the overall rate of later developing Alzheimer's disease. In absolute numbers, more people are likely helped by the androgen blocking treatment than hurt, but the results are sobering.
The dilemma this finding raises -- to take the drug or not -- could be solved with a precision-health approach that would clarify which patients should take androgen blockers and which ones should pass. The trick will be to sort the prostate cancer patients who can benefit most from androgen blockers from those whose risk of developing Alzheimer's is most likely to be increased by the drug.
With any luck, patient medical records can help provide that answer, too.
Previously: Stanford-based Alzheimer's Disease Research Center to be launched, New technology enabling men to make more confident decisions about prostate cancer treatment and How efforts to mine electronic health records influence clinical care
Photo of Nigam Shah by Steven Fisch
