Beyond Agreement: a real-world study of the workflow gap between echocardiography and timely structural cardiac assessmentHow a Validation Study Exposed a Hidden Gap in Cardiac Care

While a prospective study at Cascais Hospital demonstrated that the Cardio-HART (CHART) tool provides clinically credible agreement with echocardiography for left ventricular ejection fraction estimation and heart failure classification, its most significant contribution was revealing a critical, previously overlooked gap between the nominal availability of echocardiography and the timely delivery of structural cardiac assessment in routine care.

The Gist

The Big Picture: A "Gold Standard" That Wasn't Always There

Imagine you are trying to fix a car, and the mechanic tells you, "Don't worry, we have a high-tech diagnostic computer in the garage. It's the gold standard for finding engine problems."

But here's the catch: Even though the computer exists in the garage, the mechanic often has to wait days to get a slot to use it. In the meantime, they have to guess what's wrong based on the sound of the engine or the smell of the exhaust. They know the computer is there, but they can't actually use it when they need to.

This study found that the same thing happens in heart care. Hospitals have the "gold standard" test for heart failure (an ultrasound called an echocardiogram), but there is often a hidden gap between having the machine and actually getting the results in time to make decisions.

The New Tool: The "Smart Stethoscope"

The researchers tested a new tool called CHART. Think of this as a "smart stethoscope." It doesn't take a picture of the heart; instead, it listens to the electrical signals (an ECG) and uses Artificial Intelligence to guess how well the heart is pumping.

The goal was to see if this "smart guess" was accurate enough to match the "gold standard" ultrasound.

What They Did

They took 47 patients who were already at a heart specialist's office. For almost every patient, they did two things on the same day:

1. The "Smart Guess" (CHART).
2. The "Gold Standard" Ultrasound (Echocardiogram).

They wanted to see if the two methods agreed on how strong the heart was pumping.

The Results: The Tool Worked, But It Revealed a Bigger Problem

1. The Tool Was Good (Especially for Weak Hearts)
The "Smart Guess" (CHART) was surprisingly accurate.

• The Analogy: If the ultrasound said the heart was pumping at 30%, the CHART tool guessed around 32%. That's very close.
• It was particularly good at spotting hearts that were pumping weakly (a condition called Heart Failure with Reduced Ejection Fraction).
• It was less accurate for hearts that were pumping normally but still had other issues (Heart Failure with Preserved Ejection Fraction), which is a common difficulty even for the best doctors.

2. The Real Discovery: The "Hidden Gap"
This is the most important part of the paper. The researchers expected to just see if the new tool was accurate. Instead, they found something else.

Because the new tool worked well right there in the doctor's office, it highlighted a problem they hadn't noticed before: Even in a specialist hospital, patients often wait too long to get the "Gold Standard" ultrasound.

• The Analogy: Imagine you are in a hospital waiting room. You know the MRI machine is in the building. But because of scheduling delays, you sit there for days without seeing it. You might get sicker while you wait.
• The study showed that even in a place where you think you have immediate access to the best heart test, you often don't. The new tool (CHART) could give a quick, reliable answer immediately, exposing the fact that the "gold standard" was actually unavailable at the moment it was needed.

The Conclusion

The study concludes two main things:

1. The Tool is Valid: The CHART tool is a credible way to estimate heart strength, especially for patients with weak hearts.
2. The System has a Flaw: The most surprising finding wasn't about the tool itself, but about the workflow. The study proved that there is a "hidden gap" between having a heart ultrasound machine and actually using it quickly enough to help patients.

What the paper does NOT say:

• It does not say this tool should replace the ultrasound forever. The ultrasound is still the boss.
• It does not say this tool should be used for every single patient immediately.
• It does not claim that this tool fixes the waiting list problem on its own.

Instead, it suggests that this tool acts like a "spotlight." By working well, it shines a light on the fact that patients are often waiting too long for the best tests, even in good hospitals. The authors say we need bigger studies to confirm this, especially for patients with normal pumping strength.

Technical Summary

Technical Summary: Beyond Agreement – A Real-World Study of the Workflow Gap Between Echocardiography and Timely Structural Cardiac Assessment

Problem Statement

Heart failure (HF) remains a significant global health burden, often under-recognized due to non-specific symptoms and the limitations of biomarkers and ECGs in providing structural or functional cardiac data. While echocardiography is the gold standard for assessing left ventricular ejection fraction (LVEF) and phenotyping HF, a critical limitation exists in real-world care: the gap between the nominal availability of echocardiography within a healthcare system and its timely access at the point of clinical decision-making.

The authors argue that this gap is frequently normalized, even in specialist settings, because the gold standard is assumed to exist somewhere in the system. Consequently, clinicians often rely on incomplete data (symptoms, biomarkers, ECG) while awaiting imaging, potentially delaying the recognition of structural deterioration, particularly in high-risk populations like those with ischemic heart disease. The study posits that this "workflow gap" is a distinct operational failure separate from simple resource scarcity.

Methodology

• Study Design: A prospective, single-centre cohort study conducted at Cascais Hospital, Portugal, between January 2022 and July 2023.
• Population: 47 adults referred for cardiology evaluation with suspected HF or followed in an HF clinic. The cohort was predominantly ischemic (43.2% of HF cases), with a mean age of 65.6 years and 70% male.
• Intervention: Patients underwent a Cardio-HART (CHART) assessment, an AI-powered 12-lead ECG platform, on the same day as a standard transthoracic echocardiography (TTE) in all but one case.
• Reference Standard: LVEF derived via Simpson's biplane method by cardiologists, adhering to European Association of Cardiovascular Imaging (EACVI) guidelines.
• Blinding: Cardiologists reviewing echocardiograms and technicians performing CHART were blinded to each other's results.
• Analysis:
  • Agreement: Assessed via Bland-Altman analysis (mean bias, 95% limits of agreement) and mean absolute error.
  • Performance: Evaluated using sensitivity, specificity, AUC, F1-score, and Cohen's kappa for HF classification (All HF, HFrEF, HFmrEF, HFpEF).
  • Comparison: CHART results were compared against both the Simpson's biplane reference and the conventional Teichholz method.

Key Results

• Diagnostic Agreement: CHART demonstrated a mean LVEF bias of +1.92% compared to Simpson's biplane echocardiography, with 95% limits of agreement ranging from -14.6% to +18.4%. The mean absolute error was 6.09%.
• Phenotype Performance:
  • Reduced/Mildly Reduced EF: Agreement was strongest in HFrEF and HFmrEF. CHART showed a higher bias (+5.06%) but lower variability (LoA -7.5% to +17.6%) in these groups.
  • Preserved EF (HFpEF): Agreement was lower, with higher variability (LoA -19.3% to +16.5%) and a mean absolute error of 6.82%.
• Classification Metrics:
  • For distinguishing HF from unlikely HF, CHART achieved 86.5% sensitivity and 80.0% specificity (F1-score: 90.1%).
  • For HFrEF classification, the AUC was 0.89.
• Comparison with Teichholz: CHART outperformed the Teichholz method, showing a lower Root Mean Square Error (RMSE) relative to the Simpson's biplane reference (8.80% for CHART vs. 11.80% for Teichholz).

Key Contributions and Claims

The paper makes two primary contributions, distinguishing between technical performance and operational insight:

1. Technical Validation: CHART demonstrated clinically credible performance for LVEF estimation and HF stratification, particularly in reduced-EF phenotypes. It showed superior alignment with the Simpson's biplane reference standard compared to the Teichholz method.
2. Exposure of a Workflow Gap: The study's most significant finding, according to the authors, is not the agreement data itself, but the revelation of a hidden workflow deficit. By performing credibly in a cardiology-led environment where echocardiography was theoretically available, CHART exposed the "distance between nominal echocardiographic availability and timely structural cardiac assessment."
   • The authors claim that the study revealed a clinically meaningful gap where structural assessment is deferred even in specialist settings, a deficit that becomes visible only when a rapid, point-of-care tool is introduced.
   • The value of CHART is framed not as a replacement for echocardiography, but as a mechanism to make this invisible delay visible and potentially narrow the gap between gold-standard capability and real-world access.

Limitations and Scope

The authors maintain a modest stance regarding the study's scope:

• Sample Size: The small cohort (n=47) limits precision, particularly for subgroup analyses.
• HFpEF Findings: Results for HFpEF are explicitly described as exploratory and require confirmation in larger, multicentre cohorts.
• Generalizability: The cohort was enriched for ischemic disease, which strengthens relevance for that population but limits generalizability to unselected populations.
• Role of CHART: The paper explicitly states that CHART is not a replacement for echocardiography and does not claim interchangeability across all clinical circumstances.

In conclusion, the study suggests that the primary significance of CHART lies in its ability to reveal and potentially mitigate a hidden, consequential gap in cardiac assessment pathways, rather than solely in its diagnostic metrics.