Clinical guidelines are available for the diagnosis and management of CHF (4,5), and evidence exists that physicians could better adhere to these guidelines (6). Other ways to reduce the likelihood of hospitalization are disease management programs (eg, increased follow-up) and self-management programs including symptom monitoring, weight monitoring, or medication dosage adjustment (7,8).

Prevention of future hospitalizations can occur even before symptoms of CHF occur. Evidence-based strategies in the outpatient setting for these patients include control of hypertension and low-density lipoprotein cholesterol, use of angiotensin-converting enzyme inhibitors if appropriate, smoking cessation, physical activity, weight management, and diabetes management (4,9).

Approximately three-fourths of preventable hospitalizations for CHF were among people aged 65 or older (Table 1). In 2007 through 2009, as in every other period, a higher percentage of these hospitalizations occurred for women (53%) than for men (47%). The South had the highest percentage of hospitalizations (37%-39%). The West had less than 15% of hospitalizations. The most frequent payer was Medicare, ranging from 73% to 77%, depending on the period. No other payer came close to this percentage; for example, in 2007 through 2009, Medicaid paid for 8%, private insurance paid for 13%, and all other types of coverage accounted for 5%.

For women, the crude rates for preventable CHF hospitalizations increased with age for both racial groups in each period (Table 2). In all subgroups of women, the crude estimates for black women were higher than those for white women during all 7 periods (Table 2). For white women, we found a significant decreasing linear trend for those aged 45 through 64 and for those aged 65 or older. For black women, we found a decreasing linear trend for those aged 65 or older (Table 2). For white and black women, the highest rate occurred among those aged 65 or older; it peaked in 1997 through 1998 for whites (1,903.8/100,000) and in 1999 through 2000 for blacks (3,303.6/100,000).

Crude rates for preventable CHF hospitalizations among men increased with age during every period (Table 3). In all subgroups of men, the point estimates for black men were significantly larger than for white men (Table 3). We found an increasing linear trend in rates for black men aged 18 to 44. For white men aged 65 or older we found a decreasing linear time trend (Table 3). For white and black men, the highest rate occurred among those aged 65 or older; it peaked in 1997 through 1998 for whites (1,932.6/100,000) and in 1995 through 1996 for blacks (2,706.5/100,000). For both women and men, the ratio of the rates for blacks compared to the rates for whites was highest in the youngest age group (~6.5) and lowest in the oldest age groups (~1.5).

Age- and sex-standardized rates also show that blacks had higher rates of hospitalization for CHF than did whites (Figure). The negative linear trend for whites was significant and the negative trend for blacks was of borderline significance. The slopes for whites and blacks were not significantly different from each other.

Other investigators have reported racial differences in preventable CHF hospitalizations. Using hospital data from 1991 through 1998 in California for people aged 20 to 64, Davis and colleagues found that non-Hispanic black men and women had unadjusted rates that were 4.1 and 5.5 times higher, respectively, than those for their non-Hispanic white counterparts (18). Using 1997 data from 22 states, Laditka and others studied a similarly aged population of non-Hispanics and found that black men were 3.4 times as likely to be hospitalized for CHF as white men, and black women were 6.5 times as likely to be hospitalized as white women (19). In 2003, Russo et al studied hospitalizations for people aged 18 or older from 23 states and found that age- and sex-adjusted rates for non-Hispanic black men and women combined were 2.5 times higher than those for non-Hispanic white men and women (20). Using Medicare data from Maryland in 2006, O'Neill and colleagues (21) found that adjusted rates for blacks were 60% higher than for whites. Using the 1995 through 2004 NHDS but with a different case definition than in our study, Zhang and Watanabe-Galloway found that blacks aged 65 or older had higher CHF hospitalization rates than did whites (12).

Among adults aged 45 to 84, blacks have approximately twice the incidence of CHF as whites (22), and blacks with CHF may have more comorbidities such as uncontrolled high blood pressure (23,24). This higher incidence and comorbidities may be reflected in the higher hospitalization rates among blacks (24). Furthermore, blacks, when affected by heart failure, experience a unique epidemiology and natural history (ie, disease occurring at an earlier age resulting in more substantial left ventricular dysfunction), which also may contribute to these increased rates (25). The cost of medications and type of insurance may create barriers to appropriate care, for example, by leading to the underuse of dietary and medication therapies (26-28). Blacks are less likely than whites to receive medical care such as appropriate diagnostic procedures, thrombolytic therapy, and revascularization procedures for acute coronary syndrome (29), and these delays in receiving quality care could help explain the higher hospitalization rates for blacks.

Our study contributes to the literature in the following ways: 1) we confirm a declining linear trend in CHF hospitalizations in much of the older population using data from hospitals sampled throughout the 50 states and the District of Columbia, 2) we provide national age- and sex-specific rates for multiple periods, thus establishing a baseline for future monitoring of health disparities, 3) we provide confidence intervals (calculated appropriately from a confidential database that includes design variables) showing the degree of certainty or uncertainty of point estimates and allowing judgments about the significance of differences between subpopulations, and 4) our results cover all adults and extend through 2009.

Our study also has several limitations. First, race was not reported for 17% to 21% of the CHF hospitalizations. These missing values likely are disproportionately occurring in the white population because hospitals that did not report race had a higher proportion of white discharges than hospitals that did; thus, our results may overestimate the differences in rates between blacks and whites (16). We believe that because the differences between racial groups are already large, it is unlikely that the distribution would be so skewed as to eliminate the differences between the 2 groups (16). For example, using data from 2005 through 2006, and assigning all missing hospitalizations to whites, we observed an increase in the rate for whites (from 353/100,000 to 461/100,000). Even after such an extreme assumption, blacks still have a higher rate (601/100,000) than whites. Also, because our findings are supported by many other researchers (12,19-22), we believe that these differences are likely real. Second, starting in 2000, NCHS allowed abstractors to record multiple races and to record them as such in the database, likely leading to some people who would have been classified before 2000 as either black or white now being classified as multiracial. Because only 3% of the records have race classified as multiracial, the effect is likely minimal. The difference between blacks and whites is smaller in our study than in previously published studies, most likely because we did not exclude all Hispanics from our groups of blacks and whites (other studies used the classification of non-Hispanic white compared to non-Hispanic black). Finally, as required by AHRQ's definition of a PQI, we were unable to exclude CHF hospitalizations that occurred because of transfers from another facility (ie, hospital, skilled nursing facility, or intermediate health facility) for the years 1995 through 2000 because transfer data were not collected during those years. However, on the basis of the years for which transfer data were available, we believe that excluding them from 2001 through 2009 created only a small error, given that transfers accounted for approximately 3% of CHF hospitalizations.

Julie C. Will, Centers for Disease Control and Prevention, Division for Heart Disease and Stroke Prevention. 4770 Buford Hwy NE, Mailstop F-72, Atlanta, GA 30341, Phone: 770-488-6024, Email: jxw6/at/cdc.gov.