(HealthDay)—Providing patients with online (computer and mobile) portal access is associated with significantly improved diabetes medication adherence and glycemic control, according to a study recently published in JAMA Network Open.

Ilana Graetz, Ph.D., from Emory University in Atlanta, and colleagues evaluated the association between adding mobile patient portal access and diabetes medication adherence and glycemic levels among adults with diabetes. Patient portal access status was classified monthly (April 1, 2015, to Dec. 31, 2017) as never used, used from a computer only, used from a mobile device only, or used from both computer and mobile device for 111,463 patients.

The researchers found that the number of patients using the portal from both a computer and mobile device increased over time from 34.42 percent in April 2015 to 61.71 percent in December 2017. Adding computer-only portal access for patients with no prior portal access was associated with an increase in medication adherence (measured by percentage of days covered [PDC]) of 1.16 percentage points and a change of −0.06 percentage points in hemoglobin A1c (HbA1c) level. Adding both mobile and computer portal access was associated with greater change (1.67 percentage point increase in PDC and −0.13 percentage points in HbA1c level). Changes were greatest for adding both computer and mobile access for patients with higher baseline HbA1c level (>8.0 percent) and no previous portal access (increase in PDC of 5.09 percentage points and a change of −0.19 percentage points in HbA1c level).

“This is an example of how the health care system, by offering patients access to their own information and the ability to manage their health care online, can improve their health,” a coauthor said in a statement. “Offering this in a mobile-friendly way can give even more patients the ability to engage with their health care.”

A large team of researchers affiliated with multiple institutions across the U.S. has found that it might be possible to use machine learning to detect early-stage lung cancer in human patients. In their paper published in the journal Nature, the group describes their work, which involved testing machine learning systems and their ability to find circulating tumor DNA (ctDNA) in blood samples.

Lung cancer is one of the deadliest of the cancers, and like many others, the earlier it is found, the better chance of survival a patient has. Unfortunately, the only way to detect it currently is via CT scans, which are not only costly, but have high false positive rates. For that reason, medical researchers have been looking into the possibility of developing a blood test that can detect lung cancer in its early stages. Such blood tests would involve scanning samples of blood for bits of ctDNA—a daunting prospect for human lab technicians. For that reason, the researchers with this new effort looked to machine learning—prior studies have shown it to be useful in identifying early stage breast and other cancers.

In their work, the researchers trained a machine learning model on data typically associated with a type of lung cancer known as non-small-cell lung cancer. Once trained, the model was programmed to provide an estimate of lung cancer in a given patient.

During testing, the system found 63 percent of tumors for stage 1 lung cancer patients—not as good as CT scans, but possibly good enough to serve as an initial screen for people who are considered to be at high risk of developing lung cancer. The researchers note that currently, many such patients are not being screened at all. If a patient received a positive result, they would then be