Lessons Learned From eHealth Tools in IMPACT AF and mWellcare

November 20, 2018

CHICAGO — Giving healthcare providers access to electronic decision-support tools failed to improve patient outcomes in two trials — one of outpatients with atrial fibrillation (AF) in Nova Scotia, Canada, and the other of outpatients with hypertension or diabetes in rural India.

But some of the findings from the trials, presented here at the American Heart Association Scientific Sessions 2018, are promising enough that some researchers proposed that the strategy "still has potential."

The Nova Scotia trial, called Integrated Management Program Advancing Community Treatment of Atrial Fibrillation (IMPACT-AF), fell short of showing a significant benefit in time to AF-related emergency-department (ED) visit or unplanned cardiovascular (CV) hospitalization when primary care physicians used a computerized decision-support system rather than usual care.

"I think we were underpowered," IMPACT-AF researcher Jafna Layth Cox, MD, Dalhousie University, Halifax, told theheart.org | Medscape Cardiology. However, electronic decision-support tools still "have the potential to markedly reduce events," he said.

In the study conducted in India, patients with diabetes or hypertension showed similar improvements in systolic blood pressure (BP) and in HbA1C levels whether they received "enhanced usual care" from physicians and trained nurses or the same care augmented by a computerized decision-support system.

Nonetheless, the Integrated mHealth System for the Prevention and Care of Chronic Disease (mWellcare) study "demonstrates the feasibility an ambitious multifunctional electronic health record decision-support system across multiple sites at a primary care level using available trained staff," said Dorairaj Prabhakaran, MD, Public Health Foundation of India, Gurgaon.

"The overall null result," he added, "is likely due to the benefits accrued by enhanced usual care and emphasizes the potential value of using nonphysician providers and improving access to much needed medications."

The mWellcare trial was published online in Circulation to coincide with Prabhakaran's presentation, along with an accompanying editorial by Anushka A. Patel, MBBS, PhD, The George Institute for Global Health, University of New South Wales, Sydney, Australia.

Patel called the findings from the mWellcare strategy "an important addition to our knowledge in this area contributing to a growing understanding around both design and evaluation aspects of such interventions."

Funders and researchers, she writes, "could pay greater attention to the need for mHealth approaches that support comprehensive primary care through disease integration, but that also contribute toward broader health system strengthening across multiple domains."

Commenting from the session panel, Gregory Piazza, MD, Brigham and Women's Hospital, Boston, said that going forward, "it's probably more important to learn from the [trials] that don't work and see what you can change."

During the same session, Piazza presented the AF-ALERT study of a computerized decision-support tool that helped clinicians in the United States prescribe more appropriate oral anticoagulation to hospitalized patients with AF who had not yet received therapy to reduce the risk for stroke.

Atrial Fib in Nova Scotia

Medical knowledge is expected to double every 73 days in 2020,[1] said Cox, and "a typical primary care doctor must now stay abreast of about 10,000 diseases and syndromes, 3000 medications, and 1100 laboratory tests."[2]

An electronic clinical decision-support system could help physicians, but "it needs to be rigorously tested in a randomized trial," he said.

IMPACT-AF enrolled and randomized 203 primary healthcare providers, 25% of all primary care providers in Nova Scotia, managing the care of 1146 patients with AF. They were assigned to provide care assisted by the decision tool or to provide usual care.

Physicians assigned to the electronic decision-support group could use a computer program to get information about a patient's stroke risk and anticoagulant dosing, for example, and they would also get alerts. Patients receiving care in that group could use an optional app to communicate with their provider.

At baseline, average age of the patients was 58 years, 60% were male, and mean CHA2DS2-VASc score was 3.7.

After 12 months, the primary efficacy outcome of a composite of unplanned CV hospitalizations and AF-related ED visits was similar in the intervention and control groups (hazard ratio [HR], 1.02; 95% CI, 0.73 - 1.41; P = .93).

About two-thirds of the events were AF-related ED visits and one-third were unplanned CV hospitalizations; these rates were similar in the two groups.

There were numerically fewer AF-related ED visits in the intervention group than in the usual care group (12.6% vs 17.9%), an absolute difference of 5%, and a relative difference of 30%, although the differences weren't significant (P = .92). But, Cox said, "that's fewer emergency department beds that are being taken up, and it's a heck of a lot less cost to the system."

There were seven ischemic strokes, four intracranial hemorrhages, and 14 major bleeding events, but none of those end points differed significantly between the two groups.

Although the decision-support system suggested approaches to care, physicians could choose otherwise, Cox noted. He and his colleagues plan to study that aspect of the trial in further analysis.

Diabetes, Hypertension in India

mWellcare assessed the management of hypertension, diabetes, current tobacco and alcohol use, and depression using an electronic decision-support system plus enhanced usual care and also using enhanced usual care alone in patients with hypertension and/or diabetes.

The researchers enrolled 3698 adults 30 years or older with hypertension (49.7%), diabetes (35.4%), or both (14.9%) who were receiving treatment at one of 20 community health centers in Haryana in Northern India and 20 in Karnataka in Southern India.

Mean age of the patients was 55 years, 55% were men, and, at baseline, mean HbA1C was 9.4% and mean BP was 154/90 mm Hg.

Physicians in the enhanced usual care group received training on clinical practice guidelines, and nurses gave patients pamphlets on lifestyle advice and kept patient records. Physicians treated patients based on their clinical judgment.

In the intervention group, nurses received training on clinical practice guidelines, and both physicians and nurses were trained to use the decision-support system. Nurses received prompts and recorded patient parameters in the electronic record system. Patients received text reminders about medication adherence. Clinicians were sent a recommended treatment plan.

At 12 months, in the control and intervention groups, systolic BP dropped by 12.7 mm Hg and 13.7 mm Hg, respectively, and HbA1C dropped by 0.58 and 0.48 absolute percentage points, respectively. None of the changes were significant.

Systolic BP was adjusted for education, lipid-lowering drugs, aspirin use, peripheral vascular disease, and smoking status. Changes in HbA1c were adjusted for age, employment status, antihyperglycemic use, peripheral vascular disease, and alcohol use.

How should we be learning from these studies, asked session comoderator Clara K. Chow, MBBS, PhD, The George Institute for Global Health. That is, "if you find no statistically significant effects in your interventions, do you think you should not implement them, or do you think you should implement them?"

"If I had gotten a null result with no change in the two groups, then I would not implement it. But here, there is a clear case of change in both groups," Prabhakaran replied.

It's unfortunate that the trial didn't produce the desired result, "but that doesn't mean we need to give up," he continued. The decision-support system "can provide a lot of surveillance data, especially through the electronic health record. I believe that it needs to be implemented."

"India is a country that's rapidly changing, and we saw an improvement" in blood pressure and HbA1C in mWellcare that was "pretty substantial," invited discussant Mark  D. Huffman, MD, MPH, Northwestern University Feinberg School of Medicine, Chicago, told theheart.org | Medscape Cardiology.

"Having a surveillance of what's happening," he said, "will allow not only researchers but patients and policy makers to identify the really big gaps" in care.

IMPACT AF was funded by an unrestricted educational grant from Bayer. Cox discloses that he received a research grant from Bayer and is on the speaker's bureau for Bayer, Pfizer, Boehringer Ingelheim, and Servier. mWellcare was supported by grants from the WellcomeTrust in the United Kingdom and from the National Institutes of Health in the United States . Prabhakaran holds part of a joint copyright for the mPowerHeart mHealthSystem, which has features on electronic storage of health records and electronic clinical decision-support computation. Patel discloses that The George Institute for Global Health's wholly-owned commercial enterprise, George Health, has commercial relationships involving digital health initiatives, and she has received multiple grants for digital health and mhealth research.

American Heart Association (AHA) Scientific Sessions 2018: Abstract 19507 and 19526. Presented November 10, 2018.

Circulation. Published online November 10, 2018. Full text, Editorial

For more from theheart.org, follow us on Twitter and Facebook


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.