Automated Referral Prompts Reveal Lack of Equitable Access to Care in Patients With Aortic Stenosis

Abstract

Background

Despite the high prevalence and significant mortality associated with aortic stenosis (AS), data demonstrate low rates of referral to specialized heart valve teams. The objectives of this study were therefore to highlight contemporary referral patterns in patients with significant AS and identify potential barriers to referral.

Methods

Patients undergoing transthoracic echocardiography (TTE) in a large public health catchment had automatically generated text inserted into their TTE report and electronic medical record if the TTE met the American Society of Echocardiography criteria for moderate or severe AS. Text recommended referral to a heart valve team for further assessment or treatment. Patients were prospectively identified and followed. Structured telehealth was performed to clarify symptom status and perceived barriers to referral.

Results

Over 6 months, 343 patients with severe (n = 142) or moderate (n = 201) AS were identified. Despite significant AS alongside a referral prompt, only 86 (61%) patients with severe and 44 (22%) with moderate AS were referred for assessment. Patient and echocardiographic characteristics were similar between referred and nonreferred. Mortality was significantly higher in nonreferred patients with severe AS (19.6 vs. 2.3% referred, p < 0.001). Of the nonreferred patients who died, 5/11 (45%) had no compelling reason for lack of referral. Most nonreferred patients with severe AS reported progressive symptoms (75% New York Heart Association class II/III). Over half (51.2%) reported being managed with a “watchful waiting” strategy despite meeting a class I indication for aortic valve replacement, and nearly one-third (27.5%) were unaware of their diagnosis.

Conclusions

Despite automatically generated referral prompts in patients with severe AS, many patients without a compelling reason for lack of referral were not appropriately assessed and died.

Graphical Abstract

Despite automated echocardiographic prompts encouraging referral of patients with severe AS to a Heart Valve Team, a significant portion of patients were not referred and had a significantly increased rate of mortality. AS, aortic stenosis; NYHA, New York Heart Association; LVEF, left ventricular ejection fraction; AVA, aortic valve area; Vmax, peak velocity. ∗Patients (n=40) with severe AS not referred and available for telehealth.

Highlights

• •Automated prompts were used to identify patients with moderate and severe aortic stenosis and alert referring providers of the need for further specialized assessment or valve intervention.
• •Despite the addition of an automated prompt, a significant proportion of these patients were still not appropriately referred to a member of the Heart Valve Team, and died.
• •Our findings raise concerns regarding the effectiveness of current processes of care within health care systems and the timely management of patients with severe aortic stenosis.

The burden of valvular heart disease increasingly poses a globally significant public health concern.¹ Aortic stenosis (AS) is now the most prevalent heart valve disorder in the Western world. AS effects over 12% of individuals aged 75 years and above, and its prevalence is projected to rise exponentially alongside a rapidly aging population.1, 2, 3, 4 Left untreated, severe AS is associated with significant loss of physical function and high rates of morbidity and mortality.^5,6 Even patients with moderate AS have been shown to have poor long-term rates of survival approaching that of severe AS.⁷ Recently, data suggest that asymptomatic patients also derive benefit from early referral and treatment of severe AS.⁸ Thus, timely intervention remains the cornerstone for the management of AS, with aortic valve (AV) replacement (AVR) now widely inclusive of both surgical and transcatheter options.^9,10

Current US and European guidelines outline management strategies according to the severity of valvular heart disease and emphasize the importance of consultation with a comprehensive heart valve team (HVT).^9,11 Despite clear recommendations and widespread availability of transcatheter valve replacement options, studies continue to suggest underutilization of valve replacement for patients with severe AS and low rates of referral to specialized HVTs.12, 13, 14 These findings have been replicated in different health care models across the United States of America and United Kingdom and are potentially attributable to multiple factors including age, sex, race, socioeconomic status, and geographic location.15, 16, 17 In the present study, we report on the efficacy of traditional referral patterns for patients with moderate and severe AS within a single, publicly funded, health catchment area of metropolitan Vancouver.

Methods

Echocardiographic Prompts

The provincial echocardiogram reporting system was modified to identify all transthoracic echocardiography (TTE) studies that met the American Society of Echocardiography (ASE) guideline criteria for the diagnosis of moderate or severe AS.¹⁸ An automatically generated text prompt was inserted into the TTE report. The automatic prompt for moderate or severe AS could be confirmed or removed by the level 3 ASE echocardiographers reviewing the images. If confirmed, the prompt was also uploaded and appeared in the associated hospital electronic medical record (EMR). The prompt alerted the ordering provider to the presence of significant valvular heart disease and recommended referral to the local HVT for either further assessment or consideration of valve replacement if indicated. There is no private echocardiography available in British Columbia; thus, all inpatients and elective patients receiving TTEs in downtown Vancouver were prospectively included during the enrollment period.

Patient Population

All consecutive adult patients meeting the echocardiographic criteria for the diagnosis of moderate or severe AS on TTE within a 6-month period (May to October 2023) were included in the study. Duplicate studies, patients who had already received a referral or definitive treatment for their condition, and those with previous AVR or severe multivalve disease were excluded from further study. All clinical and echocardiographic data were prospectively collected and stored in a secure institutional database. Neither patients nor the public were involved in the design, conduct, reporting, or dissemination plans of our research. The study was approved by a local ethics board who waived the need for individual written consent.

Clinical Data

Baseline clinical data, TTE indication, and ordering provider specialty were collected from the echocardiographic reports. Patient comorbidities, previous referrals, and survival data were collected from a thorough review of the health authority hospital EMR system (Cerner, Oracle Health Care, Austin, Texas). All referrals to a heart valve specialist—either a structural cardiologist or cardiothoracic surgeon—for further assessment or possible surgical or transcatheter AVR were recorded.

Echocardiographic Data

Echocardiographic measurements were taken by experienced sonographers in accordance with the guidelines of the ASE.¹⁸ Left ventricular (LV) ejection fraction (LVEF) was calculated using the Simpson biplane method or by visual estimation when the biplane method was not possible. LV stroke volume (SV) was calculated using the LV outflow tract diameter (measured at the insertion of the AV leaflets) and velocity time integral. SV was indexed to body surface area to obtain the indexed SV. LV outflow tract diameter and Doppler tracings were made by a sonographer and confirmed or remeasured by the echocardiographer interpreting the study. The mean gradient (MG) was calculated using the simplified Bernoulli equation and the AV area (AVA) using the continuity equation.

The hemodynamic classification of severe AS was in line with the current ASE guidelines. Any one of three criteria—MG ≥ 40 mmHg, Vmax ≥4.0 m/s, or AVA <1.0 cm²—was considered to suggest severe AS. In cases where there were discordant criteria, the reporting physician adjudicated the severity of AS based on the combination of hemodynamic criteria, integrated with additional imaging findings and clinical data. Low-flow, low-gradient AS was considered in patients with a small valve area (AVA <1.0 cm²) and low-volume flow across the AV (i.e. indexed SV <35 mL/m² and/or LVEF <50%) with a mean AV gradient <40 mmHg.

Symptom Assessment

All patients alive at the time of study census were contacted for a structured telehealth appointment, conducted by a single senior cardiologist as part of the patient’s clinical evaluation. Patients were asked to report their current symptom status (according to the New York Heart Association [NYHA] classification) as well as any significant progression within the past 6 months. They were also asked about their knowledge of the echocardiographic results and, for those who were not referred, any reasons they may perceive to have contributed to why they had not yet received treatment. Primary language spoken, racial background, and marital status were also recorded. Through these structured patient telehealth calls, we tried to understand any potential barriers to referral and thereby identify opportunities to mitigate potential care gaps in the timely referral and management of patients with significant AS.

Study Follow-Up

All individuals were followed from the date of the index echocardiographic prompt until the point of death or study census. Study census occurred at least 4 months following the last included TTE to ensure timely referral of patients by their usual clinical providers was captured.

Statistical Analysis

All analyses in this study were conducted in a descriptive manner. Patient baseline characteristics, echocardiographic parameters, phone call questionnaire data, and patient survival status for severe and moderate AS cohorts were reported as frequency and percentage for categorical variables, mean and SD for normally distributed variables, and median and interquartile range (IQR) for non-normally distributed variables. Comparative analyses of referred and nonreferred patients in each cohort were performed using Chi-square test or Fisher exact test for categorical variables, depending on the expected cell counts, two-sample t-test for normally distributed continuous variables, and Wilcoxon rank-sum test for nonnormally distributed continuous variables. A two-sided alpha level of 0.05 was considered statistically significant with no formal adjustment of multiplicity given the exploratory nature of the study. All statistical analyses were performed using SAS (version 9.4; SAS Institute Inc., Cary, North Carolina).

Results

Baseline Clinical Characteristics

During the study period, a total of 343 individuals with moderate or severe AS were identified. Of these, 201 patients (59%) had moderate AS and 142 patients (41%) had severe AS. At the time of study census, over one-third of patients (39%) with severe AS had not been referred to a HVT for further assessment or valve replacement. In those with moderate AS, only 45 (22%) had been referred.

Baseline demographic and clinical data for patients with severe AS are outlined in Table 1. The mean age was similar among nonreferred and referred patients with severe AS (80.4 ± 11.1 years vs. 77.4 ± 11.1 years, p = 0.116). Numerically, more women were nonreferred (51.8 vs. 37.2%, p = 0.086); however, this difference was not statistically significant. Renal function (estimated glomerular filtration rate 62.2 ± 25.8 vs. 62.3 ± 25.2 mL/min, p = 0.973) and N-terminal pro–B-type natriuretic peptide (median 3444 ng/L vs. 1498 ng/L, p = 0.123) were also similar between severe AS nonreferred and referred groups, respectively. Likewise, comorbidities were similar between groups with respect to the prevalence of coronary artery disease, prior stroke, previous (nonaortic) valve replacement, and presence of a pacemaker or defibrillator. Only the presence of chronic lung disease was statistically higher in the nonreferred group (9/48 [18.8%] vs. 5/84 [6.0%], p = 0.022), although the difference was numerically small.

Table 1.

Severe AS: baseline demographic, clinical, and echocardiographic characteristics

Variable	Overall (n = 142)	Referred (n = 86)	Nonreferred (n = 56)	p value
Demographics
Age, y	78.6 ± 11.2	77.4 ± 11.1	80.4 ± 11.1	0.116
Female sex	61 (43.0)	32 (37.2)	29 (51.8)	0.086
BMI, kg/m2	26.3 ± 5.0	26.3 ± 5	26.2 ± 5.1	0.9
White	72/102 (70.6)	43/62 (69.4)	29/40 (72.5)	0.734
Comorbidities
Diabetes	24/132 (18.2)	13/84 (15.5)	11/48 (22.9)	0.286
CAD (PCI or CABG)	11/132 (8.3)	8/84 (9.5)	3/48 (6.3)	0.745
Prior stroke	12/132 (9.1)	5/84 (6.0)	7/48 (14.6)	0.12
PPM or AICD	10 (7.0)	4 (4.7)	6 (10.7)	0.192
Chronic lung disease	14/132 (10.6)	5/84 (6.0)	9/48 (18.8)	0.022
Serum laboratory data
eGFR, ml/min/1.73 m2	62.3 ± 25.3	62.3 ± 25.2	62.2 ± 25.8	0.973
NT-proBNP, ng/L	2276 (739-5658)	1498 (259-4590)	3444 (803-8167)	0.123
Echocardiographic findings
LVEF, %	60 (55-65)	60 (50-65)	60 (52-65)	0.101
AVA, cm2	0.8 ± 0.2	0.8 ± 0.2	0.8 ± 0.2	0.519
AVA (indexed), cm/m2	0.4 ± 0.1	0.4 ± 0.1	0.4 ± 0.1	0.384
Mean AV gradient, mmHg	39.9 ± 13.8	42.2 ± 13.2	36.4 ± 14.0	0.014
Peak AV velocity, m/s	4.1 ± 07	4.2 ± 0.6	4.0 ± 0.7	0.031
Stroke volume index, mL/m2	40.1 ± 10.9	42.4 ± 9.9	36.4 ± 11.5	0.004
Bicuspid morphology	12 (8.5)	9 (10.5)	3 (5.4)	0.365
RVSP, mmHg	37.9 ± 13.2	36.6 ± 12.6	39.8 ± 13.9	0.231

Values are mean ± SD, n (%), or median (IQR).

Abbreviations: AICD, automated implantable cardiac defibrillator; AS, aortic stenosis; AV, aortic valve; AVA, aortic valve area; BMI, body mass index; CABG, coronary artery bypass grafting; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; IQR, interquartile range; LVEF, left ventricular ejection fraction; NT-proBNP, N-terminal pro–B-type natriuretic peptide; PCI, percutaneous coronary intervention; PPM, permanent pacemaker; RVSP, right ventricular systolic pressure.

For those with moderate AS, baseline demographic and clinical data were also similar between referred and nonreferred groups. Only renal function was statistically poorer in the nonreferred cohort with moderate AS (estimated glomerular filtration rate 59.1 ± 26.1 mL/min vs. 70.3 ± 21.4 mL/min, p = 0.008) (Table 2).

Table 2.

Moderate AS: baseline demographic, clinical, and echocardiographic characteristics

Variable	Overall (n = 201)	Referred (n = 44)	Nonreferred (n = 157)	p value
Demographics
Age, y	76.4 ± 12.2	76.4 ± 12.9	76.5 ± 12.1	0.993
Female sex	89 (44.3)	22 (50.0)	67 (42.7)	0.387
BMI, kg/m2	27.0 ± 5.6	26.4 ± 6.2	27.1 ± 5.5	0.488
White	86 (74.1)	19 (82.6)	67 (72.0)	0.300
Comorbidities
Diabetes	34/167 (20.4)	6/43 (14.0)	28/124 (22.6)	0.226
CAD (PCI or CABG)	30/167 (18.0)	7/43 (16.3)	23/124 (18.5)	0.738
Prior stroke	20/166 (12.0)	4/43 (9.3)	16/123 (13.0)	0.520
PPM or AICD	12 (6.0)	0 (0.0)	12 (7.6)	0.073
Chronic lung disease	32/167 (19.2)	8/43 (18.6)	24/124 (19.4)	0.914
Serum laboratory data
eGFR, ml/min/1.73 m2	61.7 ± 25.5	70.3 ± 21.4	59.1 ± 26.1	0.008
NT-proBNP, ng/L	747 (285-3717)	712 (524-3564)	789 (280-3717)	0.960
Echocardiographic findings
LVEF, %	64 (60-65)	61 (58-65)	65 (60-65)	0.684
AVA, cm2	1.1 ± 0.2	1.1 ± 0.2	1.1 ± 0.2	0.493
AVA (indexed), cm/m2	0.6 ± 0.1	0.6 ± 0.1	0.6 ± 0.1	0.962
Mean AV gradient, mmHg	24.8 ± 5.8	25.5 ± 6.3	24.6 ± 5.6	0.404
Peak AV velocity, m/s	3.3 ± 0.3	3.4 ± 0.3	3.3 ± 0.3	0.167
Stroke volume index, mL/m2	42.9 ± 6.6	44.1 ± 7.4	42.4 ± 6.3	0.410
Bicuspid morphology	24 (11.9)	7 (15.9)	17 (10.8)	0.358
RVSP, mmHg	33.9 ± 10.2	33.7 ± 9.4	34.0 ± 10.4	0.884

Values are mean ± SD, n (%), or median (IQR).

Abbreviations: AICD, automated implantable cardiac defibrillator; AS, aortic stenosis; AV, aortic valve; AVA, aortic valve area; BMI, body mass index; CABG, coronary artery bypass grafting; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; IQR, interquartile range; LVEF, left ventricular ejection fraction; NT-proBNP, N-terminal pro–B-type natriuretic peptide; PCI, percutaneous coronary intervention; PPM, permanent pacemaker; RVSP, right ventricular systolic pressure.

Echocardiographic Variables

Most patients had preserved LVEF, irrespective of the severity of AS or referral status (median LVEF 60% vs. 60%, p = 0.101 for severe AS and 65% vs. 61%, p = 0.684 for moderate AS; nonreferred vs. referred, respectively). In patients with severe AS, however, we observed that SV, peak velocity (Vmax), and AV MG were all lower in the nonreferred group, likely reflecting the presence of patients with severe low-flow, low-gradient AS. However, peak velocity (Vmax 4.0 ± 0.7 m/s) and mean aortic valve area (AVA 0.8 ± 0.2 cm²), including those indexed for body surface area (AVA indexed [AVAi] 0.4 ± 0.1 cm/m²), were still reflective of severe AS in the nonreferred cohort (Table 1). All echocardiographic variables for patients with moderate AS were similar between referred and nonreferred groups (Table 2).

Referral Patterns

Most echocardiographic studies reporting severe AS were performed on outpatients (76%). Echocardiography performed on inpatients was significantly more common in nonreferred patients with severe AS (21/56 [37.5%]) in comparison to referred patients (12/86 [14%]); (p = 0.001) (Table 3). To better characterize any referral bias between cohorts, we separated the provider who ordered the TTE into those in primary care, those who were internists or other non-cardiac specialists, and those who were either cardiologists or cardiac surgeons. Contrary to expectation, the specialty of the ordering provider was a cardiologist or cardiothoracic surgeon in most nonreferred patients with severe (55.4%) or moderate (66.2%) AS.

Table 3.

Hospitalization, symptom status, and mortality stratified by referral status

	Referred	Nonreferred	p value
Severe AS
Inpatient status	12/86 (14.0)	21/56 (37.5)	0.001
NYHA class II-IV	46/62 (74.2)	30/40 (75.0)	0.927
Death	2/86 (2.3)	11/56 (19.6)	<0.001
Moderate AS
Inpatient status	12/44 (27.3)	37/157 (23.6)	0.613
NYHA class II-IV	16/23 (69.6)	44/87 (50.6)	0.104
Death	4/44 (9.1)	15/157 (9.6)	0.999

Abbreviations: AS, aortic stenosis; NYHA, New York Heart Association.

Follow-Up and Mortality

The median duration of follow-up for the entire cohort with severe AS was 226 days (IQR 176, 273). For the 56 individuals who had not been referred, the median follow-up was 211 days (IQR 162, 272). For those referred, the median follow-up was 239 days (IQR 191, 275). There was no significant difference in the duration of follow-up between referred and nonreferred groups (p = 0.18).

During the follow-up period, mortality was significantly higher in patients who were not referred for valve intervention with severe AS compared to those referred (19.6 vs. 2.3%, respectively; p < 0.001). Of the 11 nonreferred patients who died, 5/11 (45%) had no compelling reason for lack of referral to a HVT for further assessment or AVR. Of the nonreferred patients with severe AS who remained alive at time of study census, 36 (80%) had no clinical documentation of either significant frailty or a life-limiting comorbidity that may have excluded them from consideration of AVR. In patients with moderate AS, mortality was similar between nonreferred and referred groups (9.6% vs. 9.1%, respectively, p = 0.999) (Table 3).

Symptom Assessment

A total of 103 patients with severe AS, who were alive and contactable, completed the telehealth follow-up. The majority reported being symptomatic irrespective of referral status. Of the 40 nonreferred patients, 27.5% (11/41) had NYHA class II symptoms and 47.5% (19/40) had class III symptoms. They reported similar symptoms compared to the 63 patients who had been referred, of whom 46.8% had NYHA class II, 24.2% class III, and 3.2% class IV (p = 0.052).

More than two-thirds of the nonreferred patients (30/41 or 73.2%) also reported worsening symptoms over the past 6 months. Moreover, despite their symptom status, 51.2% (21/41) of these patients were being managed with a “watchful waiting” strategy. Of particular concern, nearly one-third of nonreferred patients (12/41 or 29.3%) reported they were unaware of a diagnosis of severe AS.

Language and Ethnicity

Consistent with the demographic profile of the catchment area in which this study was performed, patients predominantly identified as ethnically White, and there was no important difference when referral status was accounted for in patients with severe AS (69.4% referred vs. 72.5% nonreferred, p = 0.734). Patients whose primary language was not English were also equally represented in the cohorts who were nonreferred vs. referred (45% vs. 40.3%, p = 0.64).

Discussion

In a contemporary cohort of patients with significant AS, identified by automated prompts at the time of TTE that alerted the clinical provider to the presence of either moderate or severe AS, we present several key findings (Graphical Abstract). First, over one-third of patients with severe AS were not referred to a HVT for further assessment or AVR. This failure of necessary follow-up care occurred despite patient reports of significant symptoms and biochemical evidence of heart failure, comparable to their contemporaries who had been referred. Second, we found 19.6% of nonreferred patients with severe AS had died within the follow-up period from index TTE compared to only 2.3% of those who had been referred. Notably, of the nonreferred patients with severe AS who died, 5/11 (45%) had no compelling reason for lack of referral to a HVT on a detailed review of the EMR. Third, although most nonreferred patients reported progression of symptoms, two-thirds were still being managed with a “watchful waiting” strategy. Fourth, nearly one-third of nonreferred patients were unaware of their diagnosis of severe AS. Collectively, these findings raise significant concerns regarding the effectiveness of processes of care in the timely management of patients with severe AS within our health care system.

In 1986, Eugene Braunwald¹⁰ noted that the most important decision in the management of patients with AS was when to refer them for a timely intervention. Three decades later, and given recent data demonstrating that a strategy of early transcatheter AVR (TAVR) in asymptomatic AS patients was superior to clinical surveillance in reducing the incidence of death, stroke, or unplanned hospitalization for cardiovascular causes, Braunwald’s statement has never been more relevant. The natural history of AS is both unpredictable and poor, with mean survival rates of less than 2 years following the onset of symptoms and less than 1 year following the onset of LV dysfunction.¹⁹ Two contemporary real-world registries recently reported a 4–5-year all-cause mortality of 45%-67% in patients with severe AS.^7,20 There is also robust evidence that surgical AVR or TAVR improves long-term survival, symptoms, and LV systolic function in patients with severe symptomatic high-gradient AS.^5,9,21,22 Despite this, we found an alarming number of patients who were not referred for AVR; the majority of whom had NYHA class II-III symptoms and elevated levels of N-terminal pro–B-type natriuretic peptide (3444 ng/L), both independent predictors of cardiovascular mortality.²³

Our findings demonstrate that greater scientific knowledge, availability of less invasive treatment options, and well-defined treatment guidelines do not necessarily translate directly into improvements in patient outcomes. The reasons for this substantial gap in clinical care is likely multifactorial and may include deficits in knowledge base for referring clinicians as well as conscious and unconscious bias. Although there are conflicting results regarding the impact of advanced age on the outcomes of TAVR,^24,25 there is evidence to suggest that older individuals can derive significant benefit from TAVR even into their 10th decade.²³ In the present study, we found no significant difference with regard to age and referral status. However, numerically, the patients who were not referred were older. Whether age played a role in providers' decision to refer to the HVT is unknown, yet assessment of frailty has been shown to be a more robust tool to evaluate the likelihood of a significant improvement in quality of life measures post TAVR.²⁶

The presence of complicating medical issues and active hospitalization are other factors that may influence clinicians’ decision for referral. Only the presence of comorbid chronic lung disease met statistical significance between nonreferred and referred patients with severe AS, though the absolute numerical difference was small (9 vs. 5 patients, respectively). Furthermore, of the nonreferred patients with severe AS who remained alive at time of study census, 36 (80%) had no clinical documentation on EMR review of either significant frailty or a life-limiting comorbidity that may have excluded them from consideration of AVR. Inpatient status itself should also not have been a barrier for referral, and there is evidence that prioritizing urgent assessment and treatment of patients with acute decompensated AS not only reduces hospital length of stay but also reduces rates of acute kidney injury and death.²⁷

Despite significant advancements in therapeutic options for the treatment of AS, its uptake remains low. The largest contemporary cohort study to date recently reported that only 60% of patients with severe AS received treatment within a 4-year time period.²⁰ Multiple previous studies have demonstrated significant variability in rates of AS diagnosis and time-to-referral for AVR based on race, ethnicity, and socioeconomic influences.^16,17,28 Although racial disparity did not appear to be a significant factor in the current analysis, notably, the area of metropolitan Vancouver included in the current study was primarily Caucasian and affluent. We anticipate that the rates of nonreferral and mortality reported in the current study will be significantly higher when we expand the automatic echo prompts to the entire population of British Columbia with many remote and indigenous populations.

Most concerningly, nearly one-third (29.3%) of patients with severe symptomatic AS reported they were unaware of their diagnosis. Without significant disparities in sex, ethnicity, or language between nonreferred and referred groups, one has to consider non–patient-related factors for this seemingly significant failure in open disclosure regarding important aspects of a patient’s clinical care. In a prior study, patients with severe AS and noncardiac specialty providers were shown to be less likely to receive timely care and AVR.²⁹ However, within our cohort, there was also a high rate of nonreferral to the HVT by cardiologists and cardiac surgeons.

Even cardiologists or cardiac surgeons may have been unaware of their patient’s symptom status or believed they were not appropriate candidates for intervention. In EARLY-TAVR, for example, 15% of patients were only found to be symptomatic by treadmill testing, although this does not explain the significant discrepancy between patient-reported symptoms and likelihood for referral in our cohort.⁸ Time-poor providers with high clinical loads may also have overlooked EMR reporting of the findings, suggesting that direct alerts requiring provider signoff could improve clinician engagement and consideration of onward referral. The use of electronic provider notifications has in fact been shown to be associated with improvements in AVR rates and increased survival for patients with severe AS in a large multicenter trial conducted within the United States.³⁰

Certainly, the high rate of short-term mortality within our cohort suggests that a specialist opinion from a heart valve physician or team may have been useful. With the veritable armamentarium of transcatheter treatment options now available for valvular heart disease both within clinical trials and commercially, the expectation for up-to-date continuing education of primary care clinicians, and even general cardiologists, is becoming increasingly difficult. The introduction of an automatic referral to a local HVT may lessen the already intense clinical load on clinical providers, ensure timely assessment, and potentially improve patient outcomes. Of course, with the potential for expanding indications for treatment of asymptomatic severe AS, the onus will then fall on specialized HVTs to navigate the complexities of comprehensive symptom assessments and patient triage to avoid waitlist morbidity and mortality. Further randomized studies will be required to evaluate this strategy before widespread adoption.

Study Limitations

Our study must be interpreted in the context of several potential limitations. Firstly, the follow-up period was brief with a maximum of 10 months and minimum of 4 months following the last collected TTE. Therefore, additional factors—including providers' or patients' availability—may plausibly have motivated the lag to referral during that period. Still, given the unpredictable natural history of severe symptomatic AS, even this short delay in referral could negatively affect patient outcomes. Second, although a limited list of patient comorbidities were obtained from hospital EMRs, additional life-limiting conditions may not have been accounted for in the analysis. Third, our inability to formally assess patient frailty or functional status could partly explain the differential referral pattern. However, after a thorough evaluation of all available clinical documentation, only a small proportion of patients had documented evidence of either significant frailty or life-limiting comorbidities that would be expected to limit any benefit from TAVR. As mentioned previously, the lack of any significant disparity in referral for AVR based on race, ethnicity, or socioeconomic factors is likely a factor of the relatively insular study population and should be interpreted with caution. Replication of the study in a more generalized population would be of significant value. Finally, the generalizability of these results to other health care models in different countries must be performed with caution. However, the predominant message that more attention to referral processes is required to improve patient outcomes with severe AS is reflected in similar recent studies across alternate health care models.^29,30

Conclusion

Despite the use of automatically generated referral prompts in patients with severe AS, many patients without a compelling reason for lack of referral were not appropriately assessed and died. Further studies are needed to determine if automatic referral via an EMR to a HVT, regardless of symptom status, would provide improved access to care and decrease morbidity and mortality in this patient population.

Ethics Statement

All research was carried out in accordance with the guidelines from Ethical Conduct for Research Involving Humans. The study was approved by a local ethics board who waived the need for individual written consent, as the project was determined to meet criteria for Quality Improvement.

Funding

The authors have no funding to report.

Disclosure Statement

The authors report no conflict of interest.