MDNews - Long Island

Mining Hidden EHR Data TO CONDUCT A retrospective study examining androgen deprivation therapy as a risk factor for dementia, researchers at Stanford University School of Medicine and the University of Pennsylvania Perelman School of Medicine needed more information than the billing and diagnostic codes in patient EHRs could provide. "Most observational studies rely on [such] discrete, coded variables," says lead author Kevin T. Nead, MD, MPhil, a resident in the Department of Radiation Oncology at the Perelman School of Medicine and a fellow at the Leonard Davis Institute of Health Economics at the University of Pennsylvania. "However, most of the EHR is non-discrete variables — diagnoses and medications recorded in clinical notes, for example. The group at Stanford has worked to develop very robust and accurate methods to extract data from clinical notes so that we can capture the data hiding in the EHR that we currently don't have access to." To illustrate the concept, Dr. Nead uses the example of a clinician who sees a patient with a new diagnosis of prostate cancer and creates a clinical note about the encounter in the EHR without entering a diagnostic or billing code related to the disease. Previously, the lack of a code meant an algorithm searching the EHR for eligible individuals to include in a study related to prostate cancer would have missed the patient, but no longer — the Stanford methods allow for identifi cation and inclusion of such patients. MEN IN THE United States have about a one in seven chance of being diagnosed with prostate cancer during their lifetimes, according to the American Cancer Society. For decades, clinicians have used ADT to treat the disease. The treatment reduces the levels of androgens — primarily testos- terone and dihydrotestosterone — in the body or prevents them from reaching the prostate, where they stimulate the growth of both normal and cancer-causing cells. Approximately half a million prostate ca ncer patients in the United States receive ADT. The past several years, studies have found an association between ADT and reduced cog nitive function. In 2015, researchers led by Kevin T. Nead, MD, MPhil, a resident in the Department of Radiation Oncolog y at the University of Pennsylva nia Perelma n School of Medicine and a fellow at the Leonard Davis Institute of Health Economics at the University of Pennsylvania, and Nigam Shah, MBBS, PhD, Associate Professor of Biomedical Informatics Research at Stanford University School of Medicine, found that prostate cancer patients who underwent ADT were nearly twice as likely to develop Alzheimer's disease as patients who did not have the treatment. Risk correlated with treatment duration: Men who received ADT the longest had the highest risk of Alzheimer's. "When we published [that study], a letter to the editor pointed out that Alzheimer's is often confused with vascular dementia," Dr. Shah says in a Stanford press release. "So instead of looking for Alzheimer's and dementia separately, we decided to aggre- gate them into a higher-level category — all dementias and cognitive decline." DEPRIVING THE BODY OF TESTOSTERONE'S PROTECTION To investigate ADT's e ect on all demen- tias, Drs. Nead and Shah and colleagues conducted a retrospective study using the EHRs of nearly 10,000 Stanford Medicine prostate cancer patients. Slightly more than 1,800 of the patients had undergone ADT. Individuals who had the treatment were more than twice as likely to develop dementia within fi ve years as those who did not have it; nearly 8 percent of ADT patient s went on to have dementia , whereas only 3.5 percent of non-ADT patients did. "No matter how we looked at the data, we saw a strong association between ADT and dementia," says Dr. Nead, the lead author. ADT-associated cognitive sequelae can happen quickly. A 2015 Mo tt Cancer Center study found that patients who had ADT were more likely to experience mental impairment six months after treat- ment than non-ADT patients. Dr. Nead theorizes the treatment, by lowering testosterone levels, may rob the body of some of the androgen's critical benefi ts. "Testosterone and testosterone ana- logues have been shown to be directly neuroprotective," he says. "Low testos- terone and ADT have also been shown to increase cardiometabolic disease, which is an independent risk factor for dementia, by impacting neurovascular function. Through these mechanisms, ADT may globally decrease neurovascular function, thereby increasing the risk of dementia." The nex t step in ex ploring A DT 's association with dementia is to conduct a prospective study to confi rm causality and stratify risk, Dr. Nead says. CATALYST FOR CLINICAL CONVERSATION Many questions about the link between ADT and dementia remain. "We need to learn more about which specifi c types of ADT raise risk of cogni- tive impairment the most," says Brian D. Gonzalez, PhD, Assistant Member at the Mo tt Cancer Center. Gonzalez was the lead author of the 2015 Mo tt study but was not involved with the 2016 Penn and Stanford study. "We also need to fi nd out which participants are at greatest risk. I'm working to design studies that would take a personalized medicine approach to fi nding which men are at risk under which conditions based on their genotype, age, previous history, ADT regimen, behavior patterns and other factors." Clinicians should thoroughly discuss ADT's risks and benefi ts with patients, not stop prescribing it, Dr. Gonzalez says. Dr. Nead agrees. "Based on the body of literature that now exists, it would be reasonable to coun- sel patients regarding potential cognitive risks," he says. "We would, however, not recommend specifi c changes to clinical practice based on this study alone, given that ADT is a life-extending treatment in some men with prostate cancer." ■ SPECIAL CLINICAL SECTION: MEN'S HEALTH 2 3 M D N E W S .C O M /// M D N E W S L O N G I S L A N D ■ 2 017

• Cover