Abstract
While it is acknowledged that control of blood pressure (BP) in primary care is of major importance in reducing the burden of cardiovascular diseases, control is suboptimal in most populations. Previous work by Turner et al found that the presence of related comorbidities increased the likelihood of a patient’s physician addressing hypertension, while unrelated comorbidities competed for attention. We evaluated BP control over a 3 year interval according to JNC-VII guidelines in primary care patients managed in two Pennsylvania health care systems, one urban (Penn = P, n=16,690) and one rural (Geisinger = G, n=79,934). Factors assessed included number of comorbidities (diagnoses occurring >2 times that could compete for clinician attention), number of primary care visits, age, sex, race, insurance and health system. The reduced JNC-VII threshold of 130/80 was applied for diabetes and chronic kidney disease. The G population was more Caucasian (C)(98 vs 49 percent), less African American (AA)(1 vs 46%), marginally older (60.2 vs 59.2 years), and less female (56 vs 61%) than P. Mean months of follow-up and number of BP readings were 39 and 12.4 at G and 33 and 9.9 at P. The mean number of comorbidities was 3.4 at G and 3.7 at P and, by race, 3.5 for C, 3.9 for AA. Medicare coverage was slightly more common at P (32 vs 29%) and commercial insurance slightly more common at G (63 vs 58%). BP was controlled in a similar proportion of readings overall, 66% and 65% (G and P). Binary logit models were constructed using percent of visits with BP controlled (%vBPc) as the dependent variable as a function of the above factors. Overall, increasing comorbidity was associated with a slightly greater %vBPc (P<0.0001), with an effect size averaging 1–2% for each comorbidity. Increasing number of unrelated comorbidities was associated with better BP control, while the opposite was true for related comorbidities. Stratified analyses conducted because of significant interactions showed the effect of related comorbidities to be more dominant in African Americans than Caucasians, particularly in patients over age 60. Nevertheless, similar effects were seen independent of race, age, gender and insurance. After accounting for race, there were no differences between the urban and rural systems. These findings suggest that race may explain urban-rural contrasts in BP control, and management of comorbidities may complicate control in African Americans more so than in Caucasians.
