When Nudges Aren’t Enough: Study Ponders AS Referral System Changes
•February 13, 2026
0
Why It Matters
Low referral rates translate into preventable deaths, underscoring the need for system‑level interventions to ensure timely valve assessment. Implementing automated referrals could standardize care and narrow existing treatment disparities.
Key Takeaways
- •Only 60% severe AS patients received specialist referral
- •Unreferred severe AS patients faced 19.6% mortality
- •Passive EMR alerts modestly increase referral rates
- •Automatic referral with opt‑out proposed to close care gaps
- •Study highlights disparity in AS management across demographics
Pulse Analysis
Aortic stenosis remains a leading cause of morbidity in older adults, and current guidelines emphasize prompt evaluation by a heart‑team once severe disease is identified. Yet real‑world practice often stalls at the echo report, leaving patients in a diagnostic limbo. The Canadian cohort, drawn from a relatively homogeneous, affluent population, illustrates how even well‑designed electronic nudges can fall short when clinician workflow and patient autonomy intersect. By quantifying referral gaps—60% for severe disease and a mere 20% for moderate cases—the study spotlights a systemic blind spot that directly impacts survival.
When compared with the DETECT AS trial, which reported higher referral upticks and modest mortality benefits, the new data suggest that passive alerts alone are insufficient. The Canadian investigators observed that most unreferred patients lacked documented frailty or comorbidities that would justify non‑referral, and mortality among the unreferred severe cohort surged to nearly 20%. These outcomes reinforce the argument that alerts must be coupled with actionable pathways, such as automated referral orders, to move patients from identification to definitive treatment. Moreover, the opt‑out model championed by UCSF offers a pragmatic compromise, preserving clinician discretion while establishing a safety net for patients who might otherwise fall through the cracks.
For health systems grappling with similar gaps, the study provides a blueprint: integrate automatic referral triggers for Class I indications, embed clear language in EMR reports, and allow a brief opt‑out window before the referral is finalized. While implementation challenges—provider resistance, workflow redesign, and data privacy concerns—remain, the potential to reduce preventable deaths justifies the investment. As value‑based care models increasingly tie reimbursement to outcomes, automated, evidence‑backed referral pathways could become a standard component of cardiovascular quality improvement programs.
When Nudges Aren’t Enough: Study Ponders AS Referral System Changes
Moving to an automatic referral for moderate or severe AS may be an option in some situations, with opt‑out caveats.
Adding automated prompts to transthoracic echocardiography (TTE) reports and electronic medical records (EMRs) in patients with moderate to severe aortic stenosis (AS) does not dramatically improve referral for specialized assessment or valve intervention, a Canadian study suggests.
Only about 60 % of patients with severe AS and a little over 20 % with moderate AS were referred by the ordering physician to a heart team for assessment, with the majority (80 %) of unreferred patients who remained alive having no clinical documentation of conditions that would exclude them from consideration for AVR, such as significant frailty or a life‑limiting comorbidity.
Unreferred patients with severe AS were far more likely to die than those who did get a referral (19.6 % vs 2.3 %; P < 0.001). Among unreferred patients who died, 45 % had no compelling reason for not being referred.
“In a homogeneous, affluent, predominantly white population, our detection and treatment of aortic stenosis is abysmal. That's what this study says to me,” senior author David A. Wood, MD (Vancouver General Hospital/Dilawri Cardiovascular Institute, Canada), told TCTMD. “If you were to replicate this in a more heterogeneous population with marginalized populations, I would imagine it would be significantly worse.”
The study was published online February 10, 2026, ahead of print in Structural Heart, with lead author Sophie Offen, MBBS, PhD (Dilawri Cardiovascular Institute).
Other studies that have used similar prompts, like last year’s DETECT AS, have shown an increase in AVR and reductions in treatment disparities—including among elderly patients and women. Despite the gain in referrals, about 40 % of symptomatic AS patients in that study didn’t get timely treatment.
Sammy Elmariah, MD (University of California, San Francisco), senior investigator of DETECT AS, said he believes the Canadian experience confirms his study’s findings that these passive alerts, when acted upon, can make a difference.
“There's a robust difference in mortality, for example, in the patients that are referred versus those that are not referred,” he noted. “In DETECT AS, we very clearly saw significant improvement in rates of referral that had beneficial impact on patient outcomes. I think what this study is pointing to that was also highlighted in DETECT AS is simply the fact that we need to do even better. This gets us partway there to the ideal of having all of these patients evaluated in a timely fashion, but unfortunately it doesn't get us fully over the finish line.”
Referred vs Nonreferred Patients
The study included 343 patients (mean age 77 years; 43 % women) who underwent TTE and met echocardiographic criteria for the diagnosis of moderate (59 %) or severe (41 %) AS. No significant differences by age, sex, or other patient or echocardiographic characteristics were seen between referred and nonreferred patients with severe AS. Comorbidities also were similar between groups with the exception of chronic lung disease, which had a slightly higher rate of nonreferral. Characteristics in those with moderate AS also were mostly similar between those who were referred and not referred, with the only comorbidity difference being lower renal function in the nonreferred group.
In the nonreferred group with severe AS, 27.5 % had NYHA class II symptoms and 47.5 % (19/40) had class III symptoms. More than two‑thirds of nonreferred patients reported worsening symptoms over the past 6 months and about 50 % were being managed with a watchful‑waiting strategy.
A look at referral patterns by provider found that a cardiologist or cardiothoracic surgeon was the ordering physician for the TTE in 55.4 % of nonreferred severe AS patients and in 66.2 % of moderate AS patients.
“We will save lives and prevent morbidity by giving up some autonomy to a system that allows automatic prompts and ultimately automatic referrals.” – David A. Wood
Based on the study’s results, Wood said his healthcare system has now decided on an automatic referral strategy based on echo results, “because we really don't see any other solution for preventing otherwise healthy patients in their seventies and eighties from being missed and then coming in when they're too sick to be effectively treated.”
To TCTMD, he stressed that this decision pathway is not meant to take away physician or patient autonomy.
“There are absolutely scenarios [where] due to comorbidities, projected lifespan, or personal preference, patients will not want to proceed with either transcatheter or surgical valve replacement,” Wood said. “We fully acknowledge that, but they should be given the opportunity to make an informed decision.
“The current system of doing ultrasounds and then following up with your referring clinician leads to a problem where 40 to 50 % of people are not referred and have a much higher mortality than they need to have,” he continued. “We interviewed the clinicians in our health catchment area and heard loud and clear that anything we can do to automate and simplify the referral process for a Class I indication, like severe aortic stenosis, would be welcomed in 2026.”
Ultimately, Wood said, he believes individual physicians will need to accept the automated referrals to fix the broken system that is preventing so many AS patients from getting the care they need in a timely manner.
“We will save lives and prevent morbidity by giving up some autonomy to a system that allows automatic prompts and ultimately automatic referrals,” he added. “We need to rethink what the standard of care is, and if that means a loss of autonomy to more effectively treat and triage patients, that's what's needed.”
Elmariah noted that it may be easier to implement an automatic referral process in a healthcare system like Canada than in the US.
“Having said that, we too are in the midst of testing a system that is a little bit more automated [and] culturally a little bit more in tune with the US healthcare system and some of the concerns that providers have about automatic referral, but does seek to move the needle further beyond just these passive nudges,” he added.
“Specifically what we're doing here at UCSF is we have a pathway by which alerts are sent to providers and there is an automated or facilitated referral to the heart valve team, but there's an opt‑out feature for providers so that if a provider specifically doesn't want a patient referred, they can decline the referral,” Elmariah continued. “If there is not action taken, either to accept or decline the referral, then we send that referral anyway after a 2‑week period. So, what we're doing is setting a safety net to ensure that patients don't fall through the cracks.”
Elmariah said the opt‑out feature is important because the clinician sending their patient for the echocardiography is the one who knows them best. At the same time, there are opportunities for patient education via the prompts.
He noted the recent guidelines from the American Society of Echocardiography for standardized reporting of TTE require the addition of very clear language in the EMR that states that a patient has significant aortic stenosis and suggests that, if medically appropriate, the patient should be referred for further evaluation.
“The reason that's important is that more and more patients are in possession of results of their clinical testing,” Elmariah said. “So, if they see clear language like that as opposed to a bunch of numbers, quantitating the severity of aortic stenosis, they at least will become more aware that there actually is a problem that warrants further evaluation.”
For Wood and colleagues, the problem identified in the paper may just be the tip of the iceberg since they believe that the number of potential patients identified in the study would have been much higher if the echocardiograms had been reviewed by a core lab.
0
Comments
Want to join the conversation?
Loading comments...