skip to Main Content

Adherence interventions and artificial intelligence: an overview

Adherence Interventions And Artificial Intelligence: An Overview

Medication adherence is the biggest problem that healthcare keeps beating its collective head against. We have heard the numbers. US Pharmacist, in January 2018, cited statistics that nonadherence can account for up to 50% of treatment failures, around 125,000 deaths, and up to 25% of hospitalizations each year in the United States. Also according to US Pharmacist, while adherence rates of 80% or more are needed for optimal therapeutic efficacy, it is estimated that adherence to chronic medications is around 50%.

Tackling low medication adherence rates via the pharmacy

Some of the more effective—and cost-efficient—ways of promoting medication adherence have been through pharmacy-based programs. One of the most-cited examples of the success of pharmacy-based interventions is the Asheville Project that covered 12 community and hospital pharmacy clinics in Asheville, N.C., over a six-year period from 2000 through 2005.

More recently, in 2019, a study examined whether a pharmacist-led nonadherence intervention reduced the burden on the Australian healthcare system. The retrospective observational study examined a de-identified database of dispensing data from 20,335 patients (n=11,257 on rosuvastatin, n=6,797 on Irbesartan, and n=2,281 on desvenlafaxine). Each patient received a pharmacist-led medication intervention, and the data six months before was compared with six months after. Across the three disease states, the cost of medication nonadherence was estimated to be $517 per adult, or $10.4 billion. After the intervention, costs were estimated to decrease by $95 per adult and saving the Australian healthcare system $1.9 billion.

Electronic vs. in-person interventions

What kinds of interventions are effective? American Pharmacist cited one study of randomized controlled trials that evaluated adherence-intervention models for patients with diabetes and cardiovascular disease. Overall, the study found that in-person interventions had similar efficacy to indirect ones, with success rates of 56% vs. 42% respectively. In-person tactics included face-to-face interviews, hospital discharge instructions, clinic-based interventions, and phone calls. Indirect tactics include traditional paper-based mailed and faxed information, as well as electronic—automated phone calls, electronic pill boxes, and computer-generated targeted interventions. Of the indirect tactics, the electronic ones were found more effective than paper-based, at 67% versus 33% .

One area of electronic intervention that is perceived as having great potential is text messaging, but according to the Agency for Healthcare Research and Quality (AHRQ), many mobile health services are not agile enough to meet patients’ complex and changing needs, with one-size-fits-all messages sent on a rigid schedule. It’s difficult for a standard text messaging strategy to overcome the complex challenges to adherence, which include patient beliefs about a disease and its treatment, organizational barriers, and cost.

During a substudy in the Phase II clinical trial for Abbott and Gilead’s ABT-126, an α7 nicotinic receptor agonist in development for the treatment of schizophrenia, researchers used an artificial intelligence platform, AiCure on mobile devices to measure medication adherence. The AiCure app, an interactive medical assistant, watches patients take their medication and aggregates data for clinical insights. The study found that mean cumulative pharmacokinetic adherence over 24 weeks was 89.7% for subjects monitored using the AI platform compared with 71.9% for subjects who were monitored by modified directly observed therapy.

Artificial Intelligence takes on adherence

According to TrialBulletin.com, as this was being written, there are 204 clinical trials in process or pending on the use of artificial intelligence in medication adherence, monitoring, diagnosis and other areas.

AI has the potential to enhance all kinds of medication adherence interventions, not just electronic ones. In 2016, Blue Cross Blue Shield of North Carolina engaged AllazoHealth to improve the medication adherence rates for their Medicare Advantage Part D population of 104,392 patients. The authors believe the study is the first randomized controlled trial to isolate the effect of machine learning technology on enhancing the effectiveness of medication adherence interventions.

The study examined adherence rates across the three groupings of medication classes that are directly tied to end-of-year Star Ratings and bonus payments: renin angiotensin system antagonists (RASAs), oral anti-diabetics (OADs), and statins. The study measured the impact of live calls delivered with Allazo targeting versus the same live calls delivered with only traditional targeting versus a control group without any interventions. Ultimately, all patients in the AllazoHealth group showed a 5.5 times greater uplift in adherence, with 23% less spent on interventions for the AllazoHealth-targeted group vs. the traditionally targeted group.

There is no doubt about it: AI has an essential role to play in healthcare and medication adherence. As methods and platforms are tested, we expect solutions that are predictive, rather than reactive, will win out.

About AllazoHealth

AllazoHealth uses artificial intelligence to make a positive impact on individual patient behavior. We optimize medication adherence outcomes for pharmaceuticals, payers, and pharmacies. Our AI engine targets the right individual patient with the right intervention at the right time.

Find out more about how AI benefits patient adherence. Contact us today.   

Learn more about the impact of AllazoHealth's technology

OUR IMPACT

Improving the Effectiveness of Adherence Interventions by

5.45x

We worked alongside Blue Cross Blue Shield of North Carolina and their call center vendor to launch one of our biggest programs, working to improve adherence rates across their population of 104,392 Medicare Advantage Part D (MAPD) patients. We found that AllazoHealth targeted interventions accounted for 5.45 times the uplift in adherence compared to traditionally targeted interventions.