Even in this relatively adherent cohort of diabetic adults enrolled in managed care, with ample time for clinical intervention, optimal control of risk factors was often not achieved. Over the 24-month interval, control of glycemia and lipids showed modest improvement, but not to target. Blood pressure control showed the least progress. Only those patients with SBP
≥160 mmHg at baseline showed any improvement over the 24 months. Control of sub-optimal risk factors is associated with significant declines in mortality and morbidity. For example, each 10-mmHg decline is associated with a 15–32% decline in risk of death.17–19
Sub-optimal control was strongly related to three modifiable aspects of primary care: missed visits (lipids showed a non-significant effect), failure to monitor even when visits are kept (lipids especially), and failure to intensify therapy, even when visits are kept and monitoring is completed. These relationships were graded and were independent of socio-demographic factors and co-morbidities. Strengths of the study are: (1) a longitudinal design with detailed information on clinical activity from visit to visit; (2) use of a HEDIS-based sample, the traditional reference population for judging quality of care; (3) data from 14 health center locations throughout the state.
Several studies have evaluated changes in A1c, blood pressure, and lipid control in patients with diabetes.20–22
All found improvement in risk factors during follow-up, but also identified a substantial percentage of patients who failed to reach targets. While there are many studies evaluating the processes of care,23–26
few studies relate these processes of care to outcomes, i.e., relating lack of testing, monitoring, or intensification with ultimate control of risk factors.20,27–29
We were unable to find any studies that evaluated visit adherence in relation to lipid or blood pressure control in diabetic adults. Two studies have evaluated the effects of appointment adherence on glycemic control.28,29
Karter et al.29
found 13% of patients missed >30% of appointments over 1 year, which was associated with a 0.70%–0.79% higher A1c compared to their visit-compliant counterparts. Rhee et al. also showed an increase of 0.12% in A1c for each missed visit.28
These studies are consistent with our findings that 10% of patients were non-adherent to visits, and visit non-adherence was associated with worse glycemic control.
Since 1980, several studies have evaluated the effects of failure to intensify medications on risk factor control, with generally congruent results.20,27,30
Berlowitz et al. found that intensification of therapy was related to better blood pressure control27
and that intensification of diabetes medications improved glycemic control over 2 years.30
Additionally, Brown et al.’s study20
found a peak in glycemic excursion prior to the addition of the metformin and that patients spent numerous months above target A1c before a therapeutic change was made. Straka et al.31
found that 70% of subjects who were above the LDL target of 100 mg/dl were not on lipid medications. The Hypertension Optimal Treatment (HOT) trial and United Kingdom Prospective Diabetes Study (UKPDS) trials demonstrated that a 4- and 10-mmHg difference, respectively, in blood pressure control in patients with diabetes resulted in a 66% and 32% decline in mortality, respectively.5,32
However, our study provides several novel findings. First, while many studies have evaluated the effects of failing to monitor appropriately,33–36
no longitudinal study has directly shown the effects of lack of monitoring on cardiovascular risk factor control after removing patients non-adherent to their visits. As we hypothesized, we found that failure to monitor is associated with poorer glycemic, lipid, and blood pressure control, and further, that lack of adequate monitoring of LDL cholesterol is a significant problem.
Second, our study identified blood pressure control as a special problem in diabetes care. Several explanations are likely. First, blood pressure variability may reduce physician confidence in the reading at any particular visit, encouraging “watchful waiting.” Such variability, whether from biologic variation, measurement error, or transient stimuli like nicotine and anxiety/white coat hypertension, provides the physician with a ‘soft’ rationale to delay treatment.37–42
In contrast, A1c and LDL measurements are much less variable. Second, physicians have underestimated the danger posed by mildly elevated blood pressure. In previous studies, physicians have expressed satisfaction with a target BP of 150 mmHg43,44
and report reluctance to treat diastolic blood pressures between 90–100 mmHg or systolic blood pressures between 140–160 mmHg.45
Several limitations of the study also deserve comment. First, because we used the lowest A1c, SBP, or LDL values in the quarter to represent the entire quarter, we may have underestimated the degree of sub-optimal control. Second, the HEDIS-based sample selection process was not designed for comparing performance between doctors or between facilities.46
However, analyses with clinic site and provider did not affect the results (Table ). Medication side effects, medication costs, complexity of the regimen, polypharmacy, co-morbidity, patients’ resistance to intensify therapy, visits for acute care, and time constraints in typical primary care office setting may all play a role in failure to intensify therapy. However, the main finding of this study, the selective failure to intensity SBP therapy, is notable.
Table 6 Failures to Visit, Monitor, and Intensify Treatment in Relation to A1c, SBP, and LDL Control after 24 Months, Adjusted for Age, Sex, Race, Case Mix, and Baseline Control (All Models Adjusted for Age, Race, Sex, Co-morbidity, Physician Effects, (more ...)
Third, we cannot exclude the possibility of missing data on visits, monitoring, and medication from outside of the Managed Care Organization network, but this is likely to be small in a staff model Health Maintenance Organization,47
and in the visit and monitoring compliant population used for the intensification analysis, drug refill compliance was high.
Fourth, our data were collected from 1999–2001 and may not be fully generalizable to the present day. Systems changes, including electronic medical record availability and increased awareness of clinical inertia, may have caused an increase in intensification rates by providers since the time of this study. However, a few more recent studies had similar rates of intensification.42–46
Further, there has been no change in ADA guidelines up to 2007 with regards to “actionable” values of risk factors. Finally, because we limited our sample to generally adherent patients under the care of academically affiliated physicians, our findings may not be fully generalizable.
In summary, failure to appropriately intensify therapy for glucose, lipids, and especially blood pressure is common and leads to poor risk factor control in diabetic adults, even with adherent patients.
Future research evaluating barriers and promoters to intensification, and evaluation of interventions to improve treatment intensification, especially of blood pressure, patient visit adherence, and LDL-cholesterol testing, are urgently required.