Voice technology + AI: A powerful partnership in patient recovery

Marie is a patient who has been discharged from the hospital after being admitted with Congestive heart failure (CHF). Once home, Marie recognizes that she must make some fundamental lifestyle changes, taking an active role in managing her heart condition and monitoring signs and symptoms that could point to another trip to the emergency room.

What is Marie’s likelihood of readmission? Heart failure consistently ranks as one of the top five principle diagnoses causing readmissions within 30 days.¹ And it’s expensive for hospitals. CMS readmissions data for acute myocardial infarction and heart failure suggests that patients spend nearly a week in the hospital and these patients cost nearly $10,000 per readmission.²

But Marie isn’t on her own in her health journey. She will have access to Voice User Interface (VUI) technologies – personalized, conversational, even empathetic voice-prompted care that guides people through the challenges of recovery and enables them to become an active participant in their own health. VUI is proving to be incredibly versatile and helpful in reaching patients: from educating them to ensuring compliance and adherence to medication routines and motivating patients to carefully self-manage their care once they are home.

VUI can also capture data that is rarely tracked as a form of outcomes: patient-reported data. Patient-reported outcomes data is the “holy grail” among providers and life science companies looking to tap this rich data set as a way to better understand patients’ overall health, their adherence to medications and behavior change.

Since Marie will be answering questions in response to prompts about her vital signs, pain or discomfort, physical signs like swelling and her adherence to medications, her providers will now have the opportunity to provide personalized, consistent, follow-up care based on her feedback during recovery.

Responses from patients about their own health status could then enable them to be identified as high-risk and prioritized for rapid intervention. This patient-reported data has historically been a major “blind spot” for clinicians that typically have sporadic interactions with patients once they have been discharged.

But does this data also present the possibility to actually predict Marie’s likelihood of being readmitted based on her responses and other attributes and data?

Leveraging AI to predict risk scores

Through advanced artificial intelligence (AI) modeling, data scientists and caregivers now have the ability to draw connections between patient answers and patient attributes to predict readmissions, revealing patterns of relative risk and combinations of patient responses or other data that warrant attention and can improve care management efforts.

AI helps learn the relationships between patient characteristics and subsequent outcomes by analyzing the complex interaction effects between many variables that are hard to detect without these innovative technologies. For example, AI can analyze the time patients spend on a call, the number of questions they answered, and their admissions history; AI can even leverage data deep within the EHR, as well as a broader set of data sources such as zip code. These can provide a deeper understanding of patient engagement, motivations, compliance and barriers to care a patient faces that might be less obvious and influenced by other factors.

By analyzing the combination of patient responses, AI can help identify that portion of patients showing signs of danger or need based on their individual readmission score, which predicts probability. And, this gives care teams the opportunity to intervene and help with greater efficiency than they would otherwise.

AI and VUI: A voice for good

Together, AI and VUI are changing the way patients like Marie and her care teams monitor how things are progressing after discharge, helping patients take charge of their health. Research shows that patients who engage with VUI technology are less likely to be readmitted to the hospital than those who were disengaged.³

Adding interactive empathetic conversations enhanced by AI to the equation helps hospitals scale their impact into the community and helps reduce readmissions by aggregating patient-reported data to identify those who are at risk and need more attention. These technological advances can help patients like Marie stay more in control of their recovery and, ultimately, out of the hospital.

1. Healthcare Cost and Utilization Project, sponsored by the Agency for Healthcare Research and Quality (AHRQ)
2. Cardiovascular Business
3. Wolters Kluwer data: 30-day readmission rates based on study from Sept. 1, 2015 – June 29, 2018. 80% of patients answered and responded to calls; those who engaged, had 10.7% readmission rate, lower than those who did not engage.