Artificial Intelligence Reviews Patient Perceptions of Cholesterol Drugs

Researchers at Stanford Medicine developed an artificial intelligence (AI) algorithm that can analyze Reddit data containing patient perceptions of statins, assessing positive and negative feedback along with discussion topics to determine how providers can better connect with patients and counter the prevalence of misinformation about these drugs.

A report from the World Health Organization indicated that among the top causes of death worldwide, cardiovascular disease ranked No. 1 in 2019, with ischaemic heart disease accounting for 16 percent of total deaths worldwide.

Despite the prevalence of this condition, treatment in the form of cholesterol-lowering drugs known as stations is effective and affordable. But many patients do not use them, according to the press release.

To gather insight into why patients might avoid statins, a group of researchers from Stanford Medicine created an AI algorithm to gather and analyze Reddit posts referencing statins. After using the algorithm to analyze more than 10,000 posts that relate to statins, they found that 67 percent were neutral, 30 percent expressed negative views on statins, and 3 percent expressed positivity.

Although researchers acknowledged a potential negativity bias associated with Reddit, as people with negative reviews are more likely to post online, they noted that the findings align with those of previous studies. Prior research has indicated that 30 to 40 percent of patients who are prescribed statins do not use them as they are supposed to.

Researchers also intended to gather specific information regarding the perception of statins, leading them to direct the AI algorithm to divide Reddit group discussions into topics such as ketogenic diets, diabetes, and statin hesitancy. 

They found that a common reason for avoiding statins was the ketogenic diet. The AI algorithm showed that many patients who attempted the ketogenic diet reportedly felt very healthy, leading them to mistrust the correlation between low-density lipoproteins and heart disease.

Internal medicine resident Sulaiman Somani, MD, a member of the team of Stanford Medicine researchers, noted that AI is beneficial as it helps researchers to collect unknown information.

"It makes a case for why AI is a good tool for this kind of research, because it helped find these new topics that we weren't expecting," he said in a press release.

Through their research, the team also noted a high level of misinformation in the Reddit posts overall.

For instance, many patients believe red yeast rice supplements may serve as valid alternatives to statins. These, however, would not have a significant impact on cholesterol levels.

"There's this idea that supplements are natural and better for the body, but we've actually done randomized control trials where we give individuals these over-the-counter supplements and see that they do not lower LDL nearly as effectively as statins," said Fatima Rodriguez, MD, associate professor of cardiovascular medicine and a member of the team of Stanford Medicine researchers, in the press release.  

Despite the findings that allude to high levels of ‘fake news’ surrounding statins, researchers acknowledged that knowing patient perceptions of medications can help clinicians battle these false narratives.

The use of AI to gather information on healthcare misinformation is not new.

A study published in April 2021 in JMIR revealed that a machine-learning algorithm could assist public health officials in exposing COVID-19-related conspiracy theories on social media, serving as a potential tool to limit the spread of misinformation online.

Researchers used publicly available data from Twitter that allowed for the characterization of four themes surrounding COVID-19 conspiracies and provided context for each of them through the initial five months of the public health emergency.