From Self-Reporting Accuracy to Therapy Access: AHA Posters Cover Issues in Disparities

November 18, 2019

Sunday’s poster session at the 2019 American Heart Association (AHA) Scientific Sessions in Philadelphia, Pennsylvania, included research that addressed disparities in clinical outcomes, healthcare delivery, and access to payer coverage.

In a healthcare system that increasingly asks consumers to shop for better prices and record patient-reported outcomes (PROs), are patients up to the task? Do patients have any idea what symptoms of health conditions look like? How do we reduce atherosclerotic cardiovascular disease (ASCVD) in high-risk patients if those most at risk lack access to insurance?

These are just some of the questions explored in Sunday’s poster session at the 2019 American Heart Association (AHA) Scientific Sessions, where many of the researchers address disparities in clinical outcomes, healthcare delivery, and access to payer coverage. The AHA Sessions are taking place through Monday in Philadelphia, Pennsylvania.

Self-reported CVD Risk Factors. Researchers often use self-reported questionnaires to assess risk factors of cardiovascular disease (CVD). However, the accuracy of self-reported risk factors—such as hypertension, diabetes, hyperlipidemia, and obesity—is not well-known, especially among African Americans, who face higher levels of risk than whites. Researchers from Emoy University assessed a group of 389 patients who were enrolled in the Morehouse-Emory Cardiovascular Center for Health Equity study; their average age was 53 years; 39% were male. Patients self-reported their history of hypertension, hyperlipidemia, and diabetes, along with their height and weight. Patients’ BMI was presented as “normal” or “overweight/obese.” For each condition, patients had to agree or disagree whether they had one of the risk factors, along with an estimate of the positive (PPV) or negative predictive value (NPV).

In this study, 47% of the participants inaccurately reported at least 1 CV risk factor:

  • 21% underreported hypertension; 1% over-reported it.
  • 25% underreported hyperlipidemia; 9% over-reported it.
  • 15% underreported diabetes; 1% over-reported it.
  • 3% underreported obesity; 0.5% over-reported it.

The positive predictive values for the self-reported conditions were: hypertension, 97.5% % (95% confidence interval-CI 94.2-98.9), 90.8% (84.6-94.7); hyperlipidemia, 90.8% (84.6-94.7); diabetes, 95.1% (88.0-98.1); and BMI, 99.7% (97.8-99.9). The NPVs were: hypertension, 57.1% (52.5-61.5); hyperlipidemia, 63.6% (60.1-66.9), diabetes, 81.4% (78.2-84.2); and BMI 82.7 % (73.7-89.1).

The authors concluded that among African Americans, “A lower proportion of subjects accurately identify the absence of a risk factor. Thus, a substantial number of subjects are not aware of, or under report, the presence of these risk factors.

“There are clear discrepancies between self-reported and objectively defined presence of cardiovascular risk factors. Caution should be utilized when using self-reported data to assess cardiovascular risk,” they wrote.

Using the EHR for Population Health. Treating hypertension is a fundamental step in avoiding the cascade of health problems that can result from uncontrolled blood pressure (BP). In 2017, the AHA and the American College of Cardiology lowered the threshold for what constitutes high BP; thus, some people may have untreated elevated BP and be unaware.

Researchers at Yale New Haven Hospital System leveraged an electronic health record (EHR) to identify patients with severe hypertension in an outpatient setting; of note, the poverty rate in New Haven has been reported at 26%. Researchers used data from 273,350 patients aged 18-85 years who have visited Yale New Haven Hospital System at least twice on an outpatient basis between January 1, 2013 and December 31, 2017. Those with severe hypertension had at least 1 measurement of systolic BP ≥160 mmHg or diastolic BP ≥100 mmHg. The researchers validated an algorithm for use with chart review with random subsets of patients to identify patients.

Findings were as follows:

  • The review found 50,264 patients (18.64%) with severe hypertension.
  • Of these, 83.3% had systolic BP ≥160 mmHg, 40.6% had diastolic BP ≥100 mmHg, and 20.1% of patients met both measures.
  • The mean age of patients with severe hypertension was 61.2 years (SD 14.1 years); 46.6% were male.
  • 70.9% were white, 17.5% were black, 8.3% were Hispanic.

Of those with severe hypertension, 66.6% had not had a follow-up visit within 1 month, per current guidelines, and 64.3% did not have a later BP measure to meet the pre-2017 BP targets of 140/100 mmHg within 3 months.

The researchers also found 12,302 (24.5%) had 3 readings of severe BP, 7469 (14.9%) had 4 readings, and 4859 (9.7%) had 5 such readings. Researchers said their work shows that the clinically relevant data needed to at-risk patients exists within the EHR. “This tool can be used to improve the diagnosis and treatment of severe hypertension in health systems,” they wrote.

Lipid therapy in underserved populations. A poster presented by Xue-Qiao Zha, MD, of the University of Washington, showed that underserved groups may have higher measures of metabolic syndrome, but they can improve on therapy to treat lipids if given access.

Researchers from the University of Washington measured carotid plaque composition with MRI and enrolled 217 patients who had ApoB ≥120 mg/dL, carotid atherosclerosis (≥15% stenosis by ultrasound), and had been treated for lipids for less than 1 year. The patients all received atorvastatin and dietary counseling for 2 years. Two-thirds also received niacin.

Patients who were “adequately served” (184 patients) were defined as living in urban areas and covered by health insurance. Those who were “underserved” (33 patients) were those living in rural areas and were uninsured or underinsured. They had a mean age of 56 years; 60% were male.

The research team compared metabolic and inflammatory risk between the 2 groups at baseline and 2 years into the intervention. Results were as follows:

  • At baseline, 58% of the underserved group had metabolic syndrome, compared with 33% of those adequately served. The underserved group had higher high-sensitivity C-reactive protein levels [median (IQR)] than the adequately served; 3.5 (3.4) vs 1.9 (2.8) mg/L, indicating higher levels of inflammation.
  • Over 2 years, both the underserved and the adequately served showed similar reductions in low-density lipoprotein (LDL) cholesterol, 5.3 mg/dL, hsCRP (log 0.05 mg/dL, and triglycerides (log 0.04 mg/dl). However, the underserved group had a non-significantly higher rate of metabolic syndrome at 1 year (24% vs 16%) and at 2 years (18% vs 12%). Levels of hsCRP remained significantly higher in the underserved group [2.6 (3.3) vs 1.0 (1.6) mg/L] at 1 year (P <.001) and 2.4 (4.1) vs 0.9 (1.5) mg/L at 2 years (P <.001).
  • Underserved patients were more likely to stop taking therapy compared with adequately served patients at 2 years (24% vs 18%; P = .4).

In the discussion that followed her presentation, Zha said challenges with obesity and metabolic syndrome in underserved populations are increasing, showing the need for better solutions to reduce the risk of ASCVD among underserved populations. The authors of the study call for better implementation of proven therapy interventions and methods of retention on therapy.


1. Khan AR, Kim JH, Ejaz K. Validity of self-reported cardiovascular disease risk factors in African American adults. Presented at the American Heart Association 2019 Scientific Sessions; Philadelphia, Pennsylvania; November 16-18, 2019. Abstract Su3012.

2. Lu Y, Huang C, Mahajan S, et al. Leveraging the electronic health records for population health: a case study for patients with severe hypertension. Presented at the American Heart Association 2019 Scientific Sessions; Philadelphia, Pennsylvania; November 16-18, 2019. Abstract Su3052.

3. Chu MP, Isquith DA, McKeeth S, et al. Metabolic and inflammatory risk during lipid-lowering and lifestyle therapy in underserved populations. Presented at the American Heart Association 2019 Scientific Sessions; Philadelphia, Pennsylvania; November 16-18, 2019. Abstract MDP259.