Monitoring and management of cardiovascular risk factors by primary care physicians at an academic medical center
Cindy Ripsin, M.S., M.P.H., M.D.
Associate Residency Director
University of Texas Southwestern
Austin Family Medicine Program
One in 10 adults in the United States have diabetes and recent information suggests that will rise to one in three by 2050 if current trends continue.1 Although cardiovascular disease (CVD) is the leading cause of death for all Americans, it is accelerated in persons with diabetes.2 Targeted management of CVD risk factors—blood pressure, low-density lipoprotein cholesterol (LDL-C), and urine microalbumin in addition to hemoglobin A1C—for patients with type 2 diabetes mellitus (T2DM) has been shown to decrease total and CVD mortality,3,4 and detailed guidelines are available.5 However, published reports suggest that patients are not routinely receiving comprehensive management of these risk factors.6-9 A retrospective analysis using a system-wide electronic health record at an academic medical center was performed to determine to what extent patients with T2DM are receiving targeted management of CVD risk factors by primary care physicians (PCPs) according to the guidelines published by the American Diabetes Association.5
Study design: Retrospective cohort.
Subjects and setting: Electronic health records of patients who received management of their diabetes by PCPs (family medicine, internal medicine) at an academic medical center from January 2007 through July 2008 were analyzed after obtaining approval from the human subjects committee of the local institutional review board.
Inclusion/exclusion criteria: Subjects were initially included if they had any diagnosis of diabetes (ICD 250.xx), were at least 30 years old, non-pregnant, and had three or more visits to one of the primary care clinics during the study period. Once these records (n = 856) were received, patients were further excluded if at least three visits were not with the same physician during the 18-month study period (n = 217) or if diabetes care was managed at a specialty care clinic (n = 238). An additional 145 records were excluded for miscellaneous reasons that included a lack of laboratory data or a notice of death at the time of chart review; charts of deceased patients were not accessed.
Records representing 256 unique patients met all criteria and were included in the analysis. For the 18 months of the study period, mean frequencies for monitoring CVD risk factors were: hemoglobin A1C 2.7 (STD 1.48); blood pressure 6.83 (STD 2.96), LDL-C 1.82 (STD 1.16), serum creatinine 6.2 (STD 9.12), and urine microalbumin 0.60 (STD 0.75).
Average values for the study group as a whole for these same CVD risk factors were: A1C 7.6 percent (STD 1.74), blood pressure: systolic 134 mmHg (STD 22) and diastolic 76 mmHg (STD 12), and LDL-C 103 mg/dL (STD 31.3).
The values were significantly modified by age for A1C (< 65 years = 7.99 percent and > 65 years = 7.19 percent; p = 0.0004), and diastolic but not systolic blood pressure (< 65 years = 77 mmHg and > 65 years = 74 mmHg; p = 0.032). Only the value for LDL-C was significantly modified by gender (females = 110 mg/dL and males = 93 mg/dL; p = 0.0007).
Data were also analyzed to determine if there were significant differences in the surveillance of CVD risk factors between family medicine and internal medicine physicians. There was a significant difference in the frequency of monitoring A1C (p = 0.0001) and LDL-C (p = 0.04), but not in the actual values themselves. However, comparisons using physician type were found to be confounded by significant differences in referral patterns between family medicine and internal medicine physicians (p = 0.03). Although approximately half of the initial sample of 856 (56 percent) of patients were managed by family physicians, family medicine patients accounted for just over a third (38 percent) of excluded patients based upon this criterion. The discrepancy in referral pattern was most pronounced at the faculty level; in the final analysis, 15 of the 256 patients were managed by internal medicine faculty compared with 100 patients being managed by family medicine faculty, and 58 and 82 by family medicine and internal medicine residents, respectively.
The proportion of patients prescribed statins (57 percent), aspirin (34 percent) and ace-inhibitors (61 percent) was assessed. Controlling for gender approached but did not achieve statistical significance for statins (p = 0.069) and ace-inhibitors (p = 0.069), and adherence by age approached but did not achieve significance for statins (p = 0.068) and for ace-inhibitors (p = 0.068).
The frequency of monitoring CVD risk factors in this group of patients did not uniformly comply with recommended guidelines. On average, monitoring hemoglobin A1C and urine microalbumin occurred less often than recommended by the guidelines, while blood pressure and serum creatinine were monitored at least as often as the guidelines recommend. One would expect that the frequency of monitoring hemoglobin A1C should approach six (every three months for an 18-month study period) since the average A1C was less than ideal at 7.6 percent, and urine microalbumin should be measured annually. Although the frequency of monitoring for serum creatinine and blood pressure was on target with stated guidelines, this doesn’t necessarily reflect a conscious effort to monitor CVD risk factors since blood pressure is routinely measured at every office visit regardless of the reason for the visit, and serum creatinine is a part of virtually every chemistry panel. One of the touted benefits of an electronic health record is the ability to track and monitor health parameters in a systematic fashion. This analysis shows that simply using an electronic health record did not lead to improved monitoring of CVD risk factors.
The disparity in referral patterns between family physicians and internal medicine physicians was a significant and unexpected finding. Determining the reason for this is beyond the scope of this study, but one could speculate that perhaps internists see patients with more advanced disease and therefore need to refer proportionately more patients, or that family physicians are more comfortable managing patients with more advanced disease. Perhaps some combination of the two created this discrepancy. Regardless of the reason, it seems prudent to exercise caution when interpreting these as well as other study results that compare or contrast the management of primary care patients according to medical specialty.
Since statins, aspirin, and ace-inhibitors are standard medications for most patients with diabetes, it is disappointing to see the relatively low proportion of patients on these medications in this study, but these values are in line with those from the published literature.7 Aspirin use is almost certainly under-reported in this group of patients; electronically-generated prescriptions formed the data set for this part of the analysis with a secondary assessment of medication lists. Since aspirin is not a prescription medication it will not appear on an electronic medication list unless the provider enters it separately.
This study was designed to reliably represent management of diabetes and CVD risk factors by PCPs. Because of the strict inclusion criteria the final sample was quite small, and these results need to be validated with a much larger sample of patients. In addition, this study was conducted at an academic medical center so the ability to apply the results to other primary care settings is limited.
Monitoring and management of CVD risk factors did not uniformly occur according to recommended guidelines in this sample of patients in spite of the use of an electronic health record. Comparisons between the management practices of family medicine and internal medicine physicians were confounded by significantly disparate referral patterns to subspecialty care, so comparing management practices across primary care specialties should be done cautiously if at all.
- Centers for Disease Control and Prevention. 1600 Clifton Road, Atlanta, GA 30333. Number of Americans with diabetes expected to double or triple by 2050. Oct. 22, 2010.
- Rao SV and McGuire DK. Epidemiology of diabetes mellitus and cardiovascular disease. In Diabetes and Cardiovascular Disease: Integrating Science and Clinical Medicine. Editors: Steven P Marso and David M. Stern. Lippincott Williams and Wilkins, 2004. Philadelphia Penn.
- Gaede P, Lund-Anderson H, Parving H, Pedersen O. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Eng J Med. 2008;358:580-591.
- Gaede P, Vedel P, Larsen N, et.al. Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes. N Eng J Med. 2003;348:383-393.
- American Diabetes Association. Clinical practice recommendations 2008. Diabetes Care. 2008;31(Supp 1).
- Beaton S, Nag S, Gunter M, Gleeson J, Saijan S, Alexander C. Adequacy of glycemic, lipid, and blood pressure management for patients with diabetes mellitus in a managed care setting. Diabetes Care. 2004;27(3):694-698.
- George P, Tobin K, Corpus R, Devlin W, O’Neill W. Treatment of cardiac risk factors in diabetic patients. How well do we follow guidelines? Am Heart J. 2001;142(5):857-863.
- Massing M, Henley N, Carter-Edwards L, Schenck A, Simpson Jr. R. Lipid testing among patients with diabetes who receive care from primary care physicians. Diabetes Care. 2003;26(5):1369-1373.
- Putzer G, Roetzheim R, Rameriz A, Sneed K, Brownlee Jr. H, Campbell R. Compliance with recommendations for lipid management among patients with type 2 diabetes in an academic family medicine practice. J Am Board Fam Pract. 2004;17(2):101-107.
This study was funded in part by a research grant from the TAFP Foundation. Support for TAFP Foundation Research is made possible by the Family Medicine Research Champions.
Richard Garrison, M.D.
David Katerndahl, M.D.
Carol and Dale Moquist, M.D.
Thank you to all who have donated to an endowment.
For information on donating or creating a new endowment, contact Kathy McCarthy at email@example.com.