Patient Awareness of Their Heart Failure Diagnosis and Its Implications for Epidemiologic Studies and Clinical Care

Self-reported heart failure (HF) has been used in myriad epidemiologic studies, including the National Health and Nutrition Examination Survey.¹ Clinicians also often rely on patient self-report to obtain a medical history. However, studies to quantify concordance between self-reported HF and adjudicated HF events are lacking. We therefore examined the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study, a well-characterized biracial cohort with many adjudicated HF hospitalizations^2,3 and data on self-reported HF.

REGARDS is a prospective longitudinal cohort study with ongoing follow-up that originally recruited 30,239 adults dwelling in the community, aged at least 45 years, who self-reported non-Hispanic White or Black race, from 2003 to 2007 from the 48 contiguous states.⁴ Participants received telephone calls every 6 months to ascertain key study end points including death and hospitalization for a cardiovascular cause. This was followed by retrieval and adjudication of medical records for the event. Adjudication of HF hospitalizations was performed by 2 independent clinician investigators on the basis of signs and symptoms and diagnostic studies including biomarkers and echocardiographic parameters documented in the medical records.⁵ Approximately 10 years after initial recruitment, a follow-up assessment was performed to collect more data including self-reported HF. REGARDS participants provided written informed consent to participate, and the protocol was approved by institutional review boards of the participating institutions. For this study, we examined REGARDS participants who completed the 10-year follow-up assessment and experienced a previous adjudicated HF hospitalization.

Among those with adjudicated HF hospitalizations, self-reported HF was defined according to the following question at the 10-year follow-up assessment: “Has a doctor or health professional ever told you that you have heart failure or congestive heart failure?” We calculated the proportion who self-reported HF at the 10-year follow-up assessment and examined the probability of death after the 10-year follow-up assessment using the Kaplan-Meier method. We also constructed a multivariable-adjusted logistic regression model to identify correlates of self-reporting HF—covariates were based on the Sorensen integrated model of health literacy.⁶ For missing covariates, we conducted multiple imputation using chained equations. We used 2-sided hypothesis testing and p value <0.05 to determine statistical significance.

Among 362 participants with adjudicated HF hospitalization, 114 participants (31%) did not self-report HF (Table 1). Over an average of 4.5 years after the 10-year follow-up assessment, mortality was 33.3% among participants who did not self-report HF and 26.2% among those who did self-report HF. Previous HF hospitalization was significantly associated with self-report of HF (Figure 1, Table 2). In addition, patients with HF with reduced ejection fraction (HFrEF)/HF with mildly reduced ejection fraction (HFmrEF) trended toward an association with self-report of HF. Black race (compared with White race) was also associated with self-report of HF, and social determinants of health including low education and ZIP-code poverty trended toward an inverse association, although these associations did not reach statistical significance.

Table 1.

Characteristics of study participants

Variables	All with Adjudicated HF Hospitalization (n=362)	Aware, with self-reported HF history (n=248)	Unaware, without self-reported HF history (n=114)
HF-characteristics
Number of HF hospitalizations prior to 2nd CATI, mean (SD)	1.4 (0.9)	1.5 (1.0)	1.2 (0.5)
Duration between first HF hospitalization and 2nd CATI in years, mean (SD)	4.4 (2.7)	4.3 (2.8)	4.6 (2.5)
HFrEF/HFmrEF, n (%)	183 (50.6%)	132 (53.2%)	51 (44.7%)
Age	75.9 (8.5)	75.2 (8.6)	77.4 (7.9)
Female, n (%)	187 (51.7%)	127 (51.2%)	60 (52.6%)
Black, n (%)	175 (48.3%)	130 (52.4%)	45 (39.5%)
Region, n (%)
Southeast (Stroke belt or stroke buckle)	206 (56.9%)	137 (55.2%)	69 (60.5%)
Non-stroke belt	152 (42.0%)	108 (43.6%)	44 (38.6%)
Low education (less than high school), n (%)	53 (14.6%)	33 (13.3%)	20 (17.5%)
Low income (<$35,000)	180 (49.7%)	126 (50.8%)	54 (47.4%)
Lives in high-poverty zip code	76 (21.0%)	48 (19.4%)	28 (24.6%)
Poor public health infrastructure, n (%)	130 (35.9%)	85 (34.3%)	45 (39.5%)
Lack of health insurance	6 (1.7%)	3 (1.2%)	3 (2.6%)
DSST score (cognition)	28.9 (12.0)	29.5 (12.4)	27.8 (11.0)
Confidence in filling out medical forms (health literacy), mean (SD)	3.8 (1.2)	3.9 (1.2)	3.7 (1.2)
History of CHD	202 (55.8%)	146 (58.9%)	56 (49.1%)
Number of medications	11.9 (4.8)	12.3 (4.8)	11.1 (4.6)

CATI = computer-assisted telephone interview; CHD = coronary heart disease; DSST = digit symbol substitution test (score based on the count of numbers matched to symbols; higher scores indicate better cognition); HF = heart failure characteristics; HFrEF = heart failure with reduced ejection fraction.

Frequency (%) and mean (SD) are presented as appropriate. Confidence in filling out medical forms is scored from 1 to 5, wherein 5 is most confident and 1 is least confident.

Figure 1.

Correlates of HF diagnosis awareness.

CI = confidence interval; OR = odds ratio.

Table 2.

Odds ratios for awareness of HF diagnosis among participants with an adjudicated HF hospitalization

Variables	Unadjusted		Adjusted
	Unaware OR (CI)	Aware OR (CI)	Unaware OR (CI)	Aware OR (CI)
HF-characteristics
Number of HF hospitalizations prior to 2nd CATI (per hospitalization)	1 (ref)	1.90 (1.28, 2.82)	1 (ref)	1.98 (1.29, 3.05)
Duration between first HF hospitalization and 2nd CATI in years	1 (ref)	0.96 (0.88, 1.04)	1 (ref)	0.91 (0.83, 1.00)
HFrEF/HFmrEF	1 (ref)	1.45 (0.93, 2.27)	1 (ref)	1.30 (0.78, 2.16)
Age (per year)	1 (ref)	0.97 (0.94, 1.00)	1 (ref)	0.98 (0.95, 1.01)
Female	1 (ref)	0.95 (0.61, 1.47)	1 (ref)	1.01 (0.60, 1.68)
Black	1 (ref)	1.69 (1.08, 2.65)	1 (ref)	1.87 (1.06, 3.30)
Low education (less than high school)	1 (ref)	0.72 (0.39, 1.32)	1 (ref)	0.81 (0.40, 1.66)
Low income (<$35,000)	1 (ref)	1.15 (0.74, 1.79)	1 (ref)	1.02 (0.63, 1.66)
Lives in high-poverty zip code	1 (ref)	0.72 (0.42, 1.23)	1 (ref)	0.55 (0.30, 1.02)
Cognition via DSST score (per correct answer)	1 (ref)	1.01 (0.99, 1.03)	1 (ref)	1.00 (0.98, 1.03)
Health literacy via confidence in filling out medical forms (per 1 point)	1 (ref)	1.17 (0.97, 1.40)	1 (ref)	1.15 (0.92, 1.43)
History of CHD	1 (ref)	1.44 (0.90, 2.28)	1 (ref)	1.37 (0.81, 2.30)
Number of medications	1 (ref)	1.06 (1.00, 1.11)	1 (ref)	1.06 (1.00, 1.12)

CATI = computer-assisted telephone interview; CHD = coronary heart disease; DSST = digit symbol substitution test; HF = heart failure; HFmrEF = heart failure with mildly reduced ejection fraction; HFrEF = heart failure with reduced ejection fraction.

N = 362 with multiple imputations.

From this geographically diverse cohort study, we found that that nearly 1 of 3 adults (31%) with a history of adjudicated HF hospitalization did not report a history of HF, suggesting that many patients with HF may not be aware of this diagnosis.

Self-reported HF is the primary source for identifying HF populations in several key research studies, including NHANES,⁷ which is used to estimate HF prevalence in the United States.¹ Our study extends the previous finding that self-reported HF does not correlate well with Medicare claims⁸ by indicating limited concordance between self-reported HF and expert-adjudicated HF hospitalizations. Our findings indicate that epidemiologic studies that rely on self-report may be more likely to miss subjects with HF with preserved ejection fraction (HFpEF) and those with low income and education. Moreover, because patients with HF who did not report a history of HF may have higher mortality than those who do report a history of HF, studies examining HF based on self-report may underestimate mortality. Future work that relies on self-report will need to recognize and acknowledge these important limitations.

Our findings also suggest a gap in patient knowledge that could have substantial clinical consequences. Awareness of HF diagnosis is necessary for engagement in meaningful self-care, an essential component of HF management.⁹ Given our finding that lack of awareness of HF is common, our study identifies a potentially important but overlooked barrier to self-care and supports the need to develop strategies to improve awareness and knowledge. Interestingly, neither the measures of health literacy nor cognition were associated with self-report, suggesting that other factors lead to lack of awareness or that the metrics used to assess these concepts did not capture facets important for accurate self-report of HF.

Awareness of an HF diagnosis and accurate self-reporting of HF is also important to ensure appropriate medical care from clinicians. Given the well-documented fragmentation of the healthcare system and prevalent use of electronic medical records that do not always communicate with one another, patient self-report is frequently a primary source of information for clinicians obtaining a medical history. If patients do not report a known history of HF, they may be at risk of receiving suboptimal or even inappropriate medical care. This further highlights the importance of future interoperability of electronic medical records.

Although our sample size precluded more definitive inferences, our data suggest that number of HF hospitalizations, history of coronary heart disease, and HFrEF/HFmrEF may be associated with HF awareness. HFrEF and HFmrEF are subtypes of HF in which abnormalities can easily be seen (and understood) on an echocardiogram. In contradistinction, HFpEF remains a confusing entity to many clinicians;¹⁰ the notion that patients can have an HF syndrome with a preserved ejection fraction is not intuitive. As a potential consequence, our study shows that patients with HFpEF may be less likely to be aware of their HF diagnosis, supporting the need to develop interventions that can improve knowledge and awareness of HFpEF among clinicians and patients.

In conclusion, a significant proportion of patients with an adjudicated HF hospitalization do not self-report a diagnosis of HF. This finding should be considered when interpreting HF research studies based on self-report and when caring for adults with HF.