Results
We identified 91 papers, of which 27 met the previously defined inclusion criteria. Of these 27, 5 were review articles,4,16-19 2 were reports,20,21 1 was an editorial,22 and 19 were clinical studies.10,12,14,15,23-37 From these clinical studies, we identified only 8 that specifically identified aspects of the relationship between SES and CHF.
Of the 8 relevant clinical studies, 2 were abstracts12,14 and 6 were papers.10,15,34-37Table 1 summarizes the demographic and design characteristics of these studies. Table 2 summarizes their outcome measures and key findings. Given the relative paucity of clinical trials, the results of the abstracts are included here. The first abstract, by Sharma and colleagues,12 used the US Survey (NHANES-11) database to determine the prevalence and mortality of CHF in a noninstitutionalized population. This involved 20,322 individuals in a population-based survey conducted between 1975 and 1980 and showed an increased relative risk (RR) for CHF in the low SES population (RR=2.33; P=.001).
The second abstract, by Philbin and associates,14 determined the relationship between SES and the risk of hospital readmission for CHF by conducting a retrospective review of 42,731 patients discharged alive with a primary diagnosis of CHF during 1995 from New York State Hospitals. They found that patients with readmissions had lower mean incomes than those who did not ($32,902 vs $33,757, P=.001). Although statistically significant, this difference is of questionable clinical or economic significance.
Some common themes emerged from the clinical studies. There was a focus on the frequency of admission and the relationship with SES. Philbin and colleagues’ findings14 were echoed by Struthers10 and MacIntyre35 and colleagues. The former found that the number of cardiac hospitalizations per patient varied from 0.71 in deprivation category 1 to 2 (most affluent) to 0.91 in category 5 to 6 (most deprived) (P =.007). In addition, the risk ratio for cardiac hospitalizations (for an increase of 1 category of social deprivation) was 1.11 (95% CI, 1.002-1.224). MacIntyre35 found that the admission rate was 56% higher in the most deprived quintile compared with the most affluent quintile (P <.001) and deprivation increased the short-term case fatality rate (by 26% in men and 11% in women).
Another common theme was barriers to care suffered by those with lower SES. Coughlin and colleagues34 showed that older age, lower income, and lack of private health insurance were inversely associated with having been placed on a waiting list for transplantation (P <.05). Factors significantly associated with not being put on the waiting list included old age, lower income, and a lack of private health insurance. Auerbach and associates36 also showed that patients with a lower income were less likely to receive care from a cardiologist (adjusted odds ratio [AOR] = 0.65; 95% CI, 0.45-0.93) and were more likely to receive a cardiologist’s care if they had a college education (AOR = 189; 95% CI, 1.02-3.51).
Severity of illness was another issue highlighted. Latour-Perez and colleagues15 found that those from a lower SES had higher simplified acute physiology scores (SAPS) on admission, indicating a higher severity of illness, SAPS 9 ± 5 compared with 7 ± 4 (P =.0052). However, their therapeutic intervention scores (TISS) were similar to the nondeprived population, 18±11 compared with 18 ± 9 (P =.666). The TISS assigns values ranging from 1 to 4 for 57 medical and surgical interventions to measure the intensity of care during a 24-hour period. Similarly, Romm and associates37 studied whether the CHF patient’s initial status is the most significant relationship to outcome, as measured by activity and symptomatology. He found that those from a lower SES had greater symptom and lower activity scores (correlation coefficients -0.181 and 0.185, respectively, P >.05).
The validity and reliability of the findings for the 6 clinical trials10,15,34-37 are subject to a number of conditions, including their definition of CHF, the study design, the representativeness of the study population, patients being at a similar stage of CHF, the dropout rate, and potential sources of bias and confounding. The limitations of each study in relation to these factors are summarized in Table 3. The applied gradings, based on the Oxford Centre for Evidence-Based Medicine Levels of Evidence,13 indicate that current evidence is suboptimal. There are 2 points that merit emphasis because of their relevance to future research work—the lack of a definition of CHF and the exclusive focus on SES and hospitalized CHF patients.
None of the studies explicitly defined the category of CHF patients included in the study, with only 1 study (MacIntyre and colleagues35) acknowledging their inability to define CHF because of a lack of information from discharge coding. Subsequently, as shown in some of the studies,15,35,37 it is difficult to establish the stage of CHF experienced by studied patients, therefore limiting the generalizability of findings.