Clinical Therapeutics/Volume 33, Number 1, 2011

Clinical Therapeutics/Volume 33, Number 1, 2011 Concurrent Control of Blood Glucose, Body Mass, and Blood Pressure in Patients With Type 2 Diabetes: An Analysis of Data From Electronic Medical Records Carrie McAdam-Marx, PhD, RPh 1 ; Jonathan Bouchard, MS 2 ; Mark Aagren, MSc 2 ; Chris Conner, PhD 2 ; and Diana I. Brixner, PhD 1 1 Department of Pharmacotherapy, University of Utah, Salt Lake City, Utah; and 2 Novo Nordisk Inc., Princeton, New Jersey ABSTRACT Objective: This cross-sectional study describes glycemic, body mass, and blood pressure (BP) control in patients with type 2 diabetes mellitus (T2DM) in an ambulatory care-based database of electronic medical records (EMRs). Methods: Patients aged 18 years with T2DM documented in 2008 and with glycosylated hemoglobin (HbA 1c ) value, body mass index (BMI), and BP charted within 90 days before or after T2DM documentation were identified using the General Electric Centricity EMR research database. Control of glycemia (controlled, 7.0%; uncontrolled, 7.0% [intermediate, 7.0% 9.0%; poor, 9.0%]), body mass (nonobese, 30 kg/ m 2 ; obese, 30 kg/m 2 [intermediate, 30 35 kg/m 2 ; poor, 35 kg/m 2 ]), and BP (controlled, 130/ 80 mm Hg; uncontrolled, 130/ 80 mm Hg [intermediate, 130 160/80 100 mm Hg; poor, 160/ 100 mm Hg]) was identified, and patients were stratified by level of control individually and in combination. Comorbidities and antidiabetic and antihypertensive treatments prescribed in the year before the index date were identified in the EMR. Results: The mean (SD) age of the cohort (N 49,560) was 60.9 (12.4) years; 51.5% were female. A minority had controlled glycemia (36.4%), body mass (30.2%), and/or BP (36.4%). Of those with controlled glycemia and body mass, 44.8% also had controlled BP, representing 5.5% of the overall study population. Conclusions: Despite the potential selection bias, a minority of patients in this cross-sectional study were at HbA 1c, BP, or BMI goals, and even fewer gained control of all 3 of these risk factors. More intensive T2DM treatment with complementary efforts to manage body mass and BP is warranted in this population. (Clin Ther. 2011;33:110 120) 2011 Elsevier HS Journals, Inc. Key words: dyslipidemia, glycemic control, HbA 1c, hypertension, obesity, type 2 diabetes. INTRODUCTION Approximately 24 million people in the United States have diabetes, 95% of whom have type 2 diabetes mellitus (T2DM). 1,2 Diabetes has been associated with important patient and health system burdens, at an estimated direct and indirect cost of $174 billion per year in the United States. 3 This health care burden is due in part to cardiovascular (CV) complications. Patients with T2DM have a 2- to 4-fold higher risk for death related to CV disease compared with those without T2DM. 2 Contributing to this risk is the high prevalence of other CV risk factors in patients with T2DM. It is estimated that 84% of patients with diabetes are overweight or obese, and that 63% have blood pressure (BP) at or exceeding 130/ 80 mm Hg, 4 the goal level recommended for patients with diabetes. 5,6 Recognizing the high prevalence of these comorbidities, the American Diabetes Association guideline for diabetes care emphasizes the need to simultaneously address comorbidities to avoid CV complications and poor outcomes. 5 In planning for the care of patients with T2DM from a clinical practice or payer perspective, it is important to know how well glycemia, body mass, and BP are simultaneously controlled. Although the prevalence of obesity and hypertension has been documented in patients with diabetes, 5,6 the extent to which patients with T2DM have gained concurrent control of glycemia, body mass, and/or BP Accepted for publication November 22, 2010. doi:10.1016/j.clinthera.2011.01.018 0149-2918/$ - see front matter 2011 Elsevier HS Journals, Inc. All rights reserved. 110 Volume 33 Number 1

C. McAdam-Marx et al. is unknown. The purpose of this cross-sectional study was to describe the current state of glycemic, body mass, and BP control individually and concurrently in a population of patients with T2DM treated in an ambulatory care setting. MATERIALS AND METHODS This cross-sectional study was conducted at the University of Utah, Salt Lake City, Utah. The institutional review board at the university reviewed and approved the study protocol; patient informed consent was waived. Data Collection Data for this cross-sectional analysis were identified using the General Electric Centricity electronic medical records (EMR) research database (GE Healthcare, Waukesha, Wisconsin). 7 This EMR database contained longitudinal ambulatory electronic health data for 9 million patients from 1996 through 2008, with a mean duration of clinical activity of slightly over 2 years. The database contained patient information including demographic data, vital sign measurements, laboratory orders and results, medication list entries and prescriptions, and diagnoses or chief complaints. The EMR database included data from patients from 35 US states who were treated by a subset of physicians who used the Centricity EMR. Two thirds of the physicians were primary care practitioners. The index date for each patient was in 2008 and was based on the first indication of T2DM occurring in the EMR during 2008. The index date does not necessarily represent a patient s initial T2DM diagnosis. T2DM was defined as follows: a primary or secondary diagnostic code for T2DM (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD- 9-CM] 8 code 250.X0 or 250.X2); 2 consecutive fasting blood glucose values 126 mg/dl; a nonfasting blood glucose concentration 200 mg/dl; a medication order or medication list entry for an antidiabetic drug, including oral agents or insulin or other injectable agents; and/or a glycosylated hemoglobin (HbA 1c ) value 7.0%. Patients were included in this study if they were 18 years of age on the index date; had EMR activity for 395 days before and 90 days after the index date; and had values for HbA 1c, body mass index (BMI), and BP within 90 days of the index date. Eligible patients also had values in the EMR within 90 days from the index date for triglycerides (TG), HDL-C, and LDL-C. Patients without clinical values 90 days of the index date were excluded. Patients were also excluded if they had 2 diagnoses for type 1 diabetes (ICD-9-CM code 250.X1 or 250.X3). Demographic characteristics (age, sex, race, and region) were identified, and age was defined as that on the index date. In cases in which race was not specified, race was categorized as unknown. EMR data from the 395-day period before the index date were used to determine patient characteristics, including antidiabetic, antihypertensive, and antihyperlipidemic drug therapy. A preindex period of 395 days was chosen to ensure that medication use was fully characterized. Dispensing data were not included in the EMR; thus, analyses are based on prescription orders. For long-term treatments, a prescription order may have been written each year, with refills for a 1-year supply. One year plus 30 days provided a small lag from the time that a prescription expired until the new prescription order was written. Comorbidities reported in the preindex time frame (chronic kidney disease, CV disease, cerebrovascular disease, hypertension, dyslipidemia, myocardial infarction, stroke, retinopathy, and neuropathy) were also identified. Levels of glycemic, body mass, and BP control within 90 days of the index date were assessed. Lipid values were also identified during this time frame. A window of 90 days before and after the index date was used to help maximize the number of patients with clinical data near the study index date while still providing reasonable estimates of glycemic, body mass, and BP control on the index date. Data Analysis Both glycemic and BP control were defined based on the American Diabetes Association guidelines for the management of these factors, categorized as controlled (at goal) or uncontrolled (not at goal). 5 Controlled glycemia was defined as HbA 1c 7.0%; uncontrolled, 7.0%. Uncontrolled glycemia was subclassified as intermediate control (7.0% 9.0%) or poor control ( 9.0%) based on the American Association of Clinical Endocrinologists/American College of Endocrinology guideline for glycemic control. 9 Controlled BP was defined as systolic/diastolic BP 130/ 80 mm Hg; uncontrolled BP, 130/ 80 mm Hg. 5 Uncontrolled BP was subclassified as intermediate control (130 160/ 80 100 mm Hg) or poor control ( 160/ 100 mm January 2011 111

Clinical Therapeutics Hg) based on the cutoff for stage 2 hypertension defined in the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. 10 Patients were categorized as either nonobese (BMI 30 kg/m 2 )orobese ( 30 kg/m 2 ), with obese subclassified as BMI 30 to 35 or 35 kg/ m 2, all based on the National Heart, Lung, and Blood Institute guideline on the management of overweight and obesity in adults. 11 The proportion of patients at each level of glycemic control was calculated, and the characteristics of patients at each level were identified. The prevalence of comorbidities was identified and reported based on corresponding ICD-9-CM codes recorded during the preindex period. Mean TG, HDL-C, and LDL-C concentrations were identified, as was the proportion of patients at goal for these laboratory measures (TG 150 mg/dl, HDL-C 40 mg/dl for men and 50 mg/dl for women, and LDL-C 100 mg/dl). The prevalence of prescription orders for antidiabetic and antihypertensive therapy in the year before the index date was also identified by level of glycemic or BP control, respectively. Descriptive statistics were used to identify differences in patient demographic and clinical characteristics by level of glycemic control using independent t tests and 2 tests for continuous and discrete variables, respectively. Patients were further stratified by BMI ( 30, 30 35, or 35 kg/m 2 ) to assess the distribution of patients by glycemic and body mass control and to assess BP status by combined levels of glycemic and body mass control. Missing clinical data were not imputed due to the relatively high occurrence of missing values. Similarly, the last reported value was not carried forward due to a lack of confidence that it would accurately represent the patient s status on the index date. All statistical tests were conducted at an a priori significance level of 0.05 using Stata SE version 10.0 (StataCorp, College Station, Texas) and SAS version 9 (SAS Institute Inc., Cary, North Carolina). RESULTS The final study cohort included 49,560 patients (Figure). The mean (SD) age was 60.9 (12.4) years, and 51.5% were female (Table I). A total of 35.3% of the cohort were reported as being white; 7.3%, black; and 3.3%, Hispanic. Race was reported as unknown in 52.5% of patients. Mean (SD) HbA 1c was 7.6% (1.6%), and 36.4% (18,025) of patients had controlled glycemia. A total of 48.2% (23,905) of the study cohort had intermediate glycemic control, and 15.4% (7630) had poor glycemic control. Mean (SD) BMI was 34.0 (7.7) kg/m 2, and Figure. Population identification. T2DM type 2 diabetes mellitus; HbA 1c glycosylated hemoglobin; ICD- 9-CM International Classification of Diseases, Ninth Revision, Clinical Modification 8 ; FBG fasting blood glucose; BG blood glucose; BMI body mass index; BP blood pressure; EMR electronic medical record; T1DM type 1 diabetes mellitus. *Some patients had 1 criterion for T2DM. 112 Volume 33 Number 1

January 2011 113 Table I. Demographic and clinical characteristics of patients with type 2 diabetes mellitus in an ambulatory care setting in 2008. Data are number (%) of patients unless otherwise specified.* Variable (HbA 1c 7.0%) (n 18,025 [36.4%]) Intermediate (HbA 1c 7.0% 9.0%) (n 23,905 [48.2%]) Poor (HbA 1c 9.0%) (n 7630 [15.4%]) All Patients (N 49,560) Age Mean (SD), y 62.8 (12.2) 61.5 (12.0) 54.8 (12.2) 60.9 (12.4) Group 65 y 9266 (51.4) 13,678 (57.2) 5998 (78.6) 28,942 (58.4) 65 y 8759 (48.6) 10,227 (42.8) 1632 (21.4) 20,618 (41.6) Sex Female 9765 (54.2) 12,012 (50.2) 3755 (49.2) 25,532 (51.5) Male 8260 (45.8) 11,893 (49.8) 3875 (50.8) 24,028 (48.5) Race White 6435 (35.7) 8601 (36.0) 2461 (32.3) 17,497 (35.3) Black 1106 (6.1) 1624 (6.8) 873 (11.4) 3603 (7.3) Hispanic 591 (3.3) 718 (3.0) 340 (4.5) 1649 (3.3) Other 245 (1.4) 414 (1.7) 157 (2.1) 816 (1.6) Unknown 9648 (53.5) 12,548 (52.5) 3799 (49.8) 25,995 (52.5) Region Northeast 5704 (31.6) 8408 (35.2) 2451 (32.1) 16,563 (33.4) South 4646 (25.8) 6320 (26.4) 2180 (28.6) 13,146 (26.5) Midwest 4174 (23.2) 5195 (21.7) 1745 (22.9) 11,114 (22.4) West 3501 (19.4) 3982 (16.7) 1254 (16.4) 8737 (17.6) HbA 1c, mean (SD), % 6.2 (0.5) 7.6 (0.5) 10.6 (1.5) 7.6 (1.6) BMI Mean (SD), kg/m 2 33.4 (7.7) 34.2 (7.6) 34.9 (8.0) 34.0 (7.7) Controlled (nonobese; 30 kg/m 2 ) 6047 (33.5) 6909 (28.9) 2000 (26.2) 14,956 (30.2) Uncontrolled (obese; 30 kg/m 2 ) 11,978 (66.5) 16,996 (71.1) 5630 (73.8) 34,604 (69.8) Intermediate control (30 35 kg/m 2 ) 5118 (28.4) 6909 (28.9) 2099 (27.5) 14,126 (28.5) Poor control ( 35 kg/m 2 ) 6860 (38.1) 10,087 (42.2) 3531 (46.3) 20,478 (41.3) C. McAdam-Marx et al.

114 Volume 33 Number 1 Table I (continued). Variable (HbA 1c 7.0%) (n 18,025 [36.4%]) Intermediate (HbA 1c 7.0% 9.0%) (n 23,905 [48.2%]) Poor (HbA 1c 9.0%) (n 7630 [15.4%]) All Patients (N 49,560) SBP Mean (SD), mm Hg 130 (16.8) 131 (17.0) 132 (18.1) 131 (17.1) Controlled ( 130 mm Hg) 9226 (51.2) 11,192 (46.8) 3396 (44.5) 23,814 (48.1) Uncontrolled ( 130 mm Hg) 8799 (48.8) 12,713 (53.2) 4234 (55.5) 25,746 (51.9) Intermediate control 7788 (43.2) 11,113 (46.5) 3612 (47.3) 22,513 (45.4) (130 160 mm Hg) Poor control ( 160 mm Hg) 1011 (5.6) 1600 (6.7) 622 (8.2) 3233 (6.5) DBP Mean (SD), mm Hg 75 (10.3) 76 (10.4) 78 (11) 76 (10.5) Controlled ( 80 mm Hg) 11,346 (62.9) 14,368 (60.1) 3763 (49.3) 29,477 (59.5) Uncontrolled ( 80 mm Hg) 6679 (37.1) 9537 (39.9) 3867 (50.7) 20,083 (40.5) Intermediate control 6408 (35.6) 9090 (38.0) 3564 (46.7) 19,062 (38.5) (80 100 mm Hg) Poor control ( 100 mm Hg) 271 (1.5) 447 (1.9) 303 (4.0) 1021 (2.1) Overall BP status Controlled ( 130/ 80 mm Hg) 7145 (39.6) 8560 (35.8) 2353 (30.8) 18,058 (36.4) Uncontrolled ( 130/ 80 mm Hg) 10,880 (60.4) 15,345 (64.2) 5277 (69.2) 31,502 (63.6) Intermediate control (130 160/ 9742 (54.1) 13,553 (56.7) 4520 (59.2) 27,815 (56.1) 80 100 mm Hg) Poor control ( 160/ 100 mm Hg) 1138 (6.3) 1792 (7.5) 757 (9.9) 3687 (7.4) Clinical Therapeutics

January 2011 115 Table I (continued). Variable (HbA 1c 7.0%) (n 18,025 [36.4%]) Intermediate (HbA 1c 7.0% 9.0%) (n 23,905 [48.2%]) Poor (HbA 1c 9.0%) (n 7630 [15.4%]) All Patients (N 49,560) TG Mean (SD), mg/dl 155.6 (91.5) 170.7 (106.0) 212.7 (200.9) 171.6 (123) Controlled ( 150 mg/dl) 10,331 (57.3) 12,107 (50.6) 3147 (41.2) 25,585 (51.6) Uncontrolled ( 150 mg/dl) 7694 (42.7) 11,798 (49.4) 4483 (58.8) 23,975 (48.4) HDL-C Mean (SD), mg/dl 45.3 (13.3) 44.1 (12.6) 42.8 (12.6) 44.3 (12.9) Controlled ( 40 mg/dl males; 8186 (45.4) 10,037 (42.0) 2880 (37.7) 21,103 (42.6) 50 mg/dl females) Uncontrolled ( 40 mg/dl males; 50 mg/dl females) 9839 (54.6) 13,868 (58.0) 4750 (62.3) 28,457 (57.4) LDL-C Mean (SD), mg/dl 94.1 (34.6) 94.4 (34.9) 105.7 (41.0) 96.0 (36.1) Controlled ( 100 mg/dl) 11,186 (62.1) 14,814 (62.0) 3723 (48.8) 29,723 (60.0) Uncontrolled ( 100 mg/dl) 6839 (37.9) 9091 (38.0) 3907 (51.2) 19,837 (40.0) Concurrent treatment Antihypertensive therapy 10,623 (58.9) 12,340 (51.6) 3884 (50.9) 26,847 (54.2) Antihyperlipidemic therapy 9300 (51.6) 10,918 (45.7) 3248 (42.6) 23,466 (47.3) No antidiabetics 5256 (29.2) 8613 (36.0) 2086 (27.3) 15,955 (32.2) 1 Oral antidiabetic 6822 (37.8) 5204 (21.8) 1314 (17.2) 13,340 (26.9) 2 Oral antidiabetics 2277 (12.6) 2760 (11.5) 878 (11.5) 5915 (11.9) 3 Oral antidiabetics 712 (4.0) 1113 (4.7) 349 (4.6) 2174 (4.4) 4 Oral antidiabetics 351 (1.9) 604 (2.5) 187 (2.5) 1142 (2.3) Insulin alone 1048 (5.8) 2424 (10.1) 1066 (14.0) 4538 (9.2) Insulin oral antidiabetic 876 (4.9) 2057 (8.6) 1292 (16.9) 4225 (8.5) Noninsulin injectable oral insulin 683 (3.8) 1130 (4.7) 458 (6.0) 2271 (4.6) C. McAdam-Marx et al.

116 Volume 33 Number 1 Table I (continued). Variable (HbA 1c 7.0%) (n 18,025 [36.4%]) Intermediate (HbA 1c 7.0% 9.0%) (n 23,905 [48.2%]) Poor (HbA 1c 9.0%) (n 7630 [15.4%]) All Patients (N 49,560) Comorbidity Dyslipidemia 4305 (23.9) 4020 (16.8) 1299 (17.0) 9624 (19.4) Hypertension 3891 (21.6) 3711 (15.5) 1223 (16.0) 8825 (17.8) Cardiovascular disease 1695 (9.4) 1685 (7.0) 462 (6.1) 3842 (7.8) Chronic kidney disease 1294 (7.2) 1143 (4.8) 315 (4.1) 2752 (5.6) Cerebrovascular disease 487 (2.7) 459 (1.9) 116 (1.5) 1062 (2.1) Neuropathy 469 (2.6) 519 (2.2) 148 (1.9) 1136 (2.3) Stroke 187 (1.0) 162 (0.7) 54 (0.7) 403 (0.8) Myocardial infarction 100 (0.6) 97 (0.4) 34 (0.4) 231 (0.5) Retinopathy 94 (0.5) 135 (0.6) 43 (0.6) 272 (0.5) Peripheral vascular disease 35 (0.2) 28 (0.1) 7 (0.1) 70 (0.1) HBA 1c glycosylated hemoglobin; BMI body mass index; SBP systolic blood pressure; DBP diastolic blood pressure; TG triglycerides. *Percentages may not total 100 due to rounding. P 0.001 versus HbA 1c 7%. P 0.05 versus HbA 1c 7%. Clinical Therapeutics

C. McAdam-Marx et al. 30.2% (14,956) of the study cohort was nonobese. Mean systolic/diastolic BP was 131 (17.1)/76 (10.5) mm Hg, and 36.4% (18,058) had controlled BP. The proportions of patients with controlled TG, HDL-C, and LDL-C were 51.6% (25,585), 42.6% (21,103), and 60.0% (29,723), respectively. The most common comorbidity was dyslipidemia (19.4% [9624]). Of the overall study population, 67.8% (33,605) had received a prescription order for an antidiabetic agent in the year before the index date (70.8% [12,769] of patients with controlled glycemia, 64.0% [15,292] of those with intermediate glycemic control, and 72.7% [5544] of those with poor control). Before the index date, 54.2% (26,847) of patients had been prescribed an antihypertensive agent, which corresponds to 49.9% (9016) of those with controlled BP, 54.6% (15,182) of those with intermediate BP control, and 71.8% (2649) of those with poor BP control. In this study cohort, as glycemic control worsened, body mass increased (Table I). Although 30.2% (14,956) of the total cohort was nonobese, this group represented 33.5% (6047/18,025) of those with controlled glycemia, 28.9% (6909/23,905) of those with intermediate glycemic control and 26.2% (2000/7630) of those with poor glycemic control (both, P 0.001 vs controlled glycemia). Similarly, 41.3% (20,478) of the overall population had poorly controlled body mass; this group represented 38.1% (6860) of those with controlled glycemia, but 42.2% (10,087) of those with intermediate glycemic control and 46.3% (3531) of those with poorly controlled glycemia (both, P 0.001 vs controlled glycemia). Being at goal for glycemia was similarly associated with controlled BP. A greater proportion of patients with controlled glycemia (39.6% [7145]) had controlled BP compared with those with intermediate (35.8% [8560]) or poor glycemic control (30.8% [2353]) (both, P 0.001). On stratifying patients by both glycemic and body mass control, BP control was best in the subset of patients with controlled glycemia and BMI (Table II). A total of 44.8% (2712/6047) of this subset had controlled BP on the index date compared with 26.4% (931/3531) of patients with poor glycemic and body mass control. However, the number of patients who had simultaneous control of glycemia, body mass, and BP at the time of the study (2712) represented only 5.5% of the study cohort. DISCUSSION This study describes glycemic control in a population of 49,560 patients with T2DM treated in an ambulatory care setting in the United States during 2008 by physicians using an EMR system. Only 5.5% concurrently attained glycemic, body mass, and BP control at the time of the study compared with approximately one third who had control of one of these factors when considered individually. Therefore, the majority of patients with T2DM in this study may have been at risk for poor CV outcomes due to inadequate control of glycemia, body mass, and BP. Considering glycemic control with or without control of body mass or BP, the majority (63.6% [31,535]) of patients with T2DM in this study had inadequate glycemic control, with a mean (SD) HbA 1c of 7.6% (1.6%). This finding is comparable to national estimates (1999 2000 National Health and Nutrition Examination Survey [NHANES] data) of a mean HbA 1c of 7.8% and 63.0% not at goal. 4 Obesity was prevalent in 69.8% of patients in the present study. National obesity data (1999 2000 NHANES) estimate that 54.6% of US patients with any type of diabetes are obese. 4 A greater proportion of patients with glycemic control in this study was nonobese (33.5%) compared with those with intermediate (28.9%) or poor glycemic control (26.2%). These findings are not surprising but may be less dramatic than expected. Although focusing on lifestyle changes, namely diet and exercise, can improve weight and glycemic control in patients with T2DM, 5 these data suggest that patients can attain glycemic control in the absence of body mass control. Most nonobese patients also failed to attain glycemic control. A notable characteristic of this study is the cross-stratification of patients by HbA 1c and BMI. The greatest proportion of patients in this study represented those who were obese with intermediate glycemic control (16,996 [34.3% overall]). By cross-stratifying patients by HbA 1c and BMI, this study identified that the highest proportion of patients with BP control at the time of the study occurred in nonobese patients who also had achieved the HbA 1c goal. The lowest proportion of patients with BP in the target range was observed in the subset of patients with poorly controlled body mass and uncontrolled glycemia. This finding indicates that attainment of the BP goal was more common in patients with better glycemic and body mass control. The clinical importance of these findings lies in the risks for diabetes complications and CV disease, which January 2011 117

Clinical Therapeutics Table II. Patients with type 2 diabetes mellitus in an ambulatory care setting in 2008 whose blood pressure (BP) was at goal ( 130/ 80 mm Hg) at the time of the study. BMI Glycemic Control (HbA 1c 7.0%) (n 18,025) Intermediate (HbA 1c 7.0% 9.0%) (n 23,905) Poor Glycemic Control (HbA 1c 9.0%) (n 7630) All Patients (N 49,560) Controlled BMI ( 30 kg/m 2 ) Controlled BMI by glycemic control 6047 6909 2000 14,956 BP control 2712 2984 785 6481 % of patients in BMI/HbA 1c control group 44.8 43.2 39.3 43.3 % Overall 5.5 6.0 1.6 13.1 Uncontrolled BMI Intermediate control (BMI 30 35 kg/m 2 ) Intermediate BMI by glycemic 5118 6909 2099 14,126 control BP control 2001 2439 637 5077 % of patients in BMI/HbA 1c control group 39.1 35.3 30.3 35.9 % Overall 4.0 4.9 1.3 10.2 Poor control (BMI 35 kg/m 2 ) Intermediate BMI by glycemic 6860 10,087 3531 20,487 control BP control 2432 3137 931 6500 % of Patients in BMI/HbA 1c control group 35.5 31.1 26.4 31.7 % Overall 4.9 6.3 1.9 13.1 HBA 1c glycosylated hemoglobin; BMI body mass index. may progress to serious CV events, including myocardial infarction and stroke. Studies have reported that effective management of diabetes to maintain HbA 1c 7% and effective hypertension management decrease the risk for macrovascular and CV complications. 5,6,12,13 Despite a lack of glycemia and BP control in this study cohort, 25% of patients with poor glycemic control and almost 50% of patients with poor BP control were not prescribed an antidiabetic or antihypertensive agent, respectively, by the EMR physician in the year before the index date. Although patients may have been receiving drug treatment prescribed by a non-emr physician, it is likely that some patients were not receiving antidiabetic or antihypertensive therapy as recommended by treatment guidelines. 5 This study did not assess specific or concurrent antidiabetic or antihypertensive therapies. However, studies have reported that a treatment approach favoring multiple risk-factor intervention and control can improve intermediate- to long-term CV event risk and mortality 14,15 ; therefore, multifactorial intervention and control in patients with T2DM is crucial. Limitations There were several limitations of this study. To assess the combined control of glycemia with body mass and BP and to describe comorbidities in the cohort, the inclusion criteria required that these variables and lipid data be documented on the study index date 90 days. This criterion led to important patient attrition and may have introduced a selection bias. To assess this 118 Volume 33 Number 1

C. McAdam-Marx et al. bias potential, a post hoc analysis of patients with lipid data versus those without lipid data was conducted. A small age difference was identified in patients with available lipid data compared with those without lipid data (60.9 vs 60.7 years, respectively; P 0.007), but the proportions of males and females did not differ significantly. Small differences were identified between those with and without lipid data in baseline HbA 1c (7.59% vs 7.54%, respectively), BMI (34.0 vs 33.6 kg/m 2 ), systolic BP (131 vs 130 mm Hg), and diastolic BP (76 vs 75 mm Hg) (all, P 0.001). The findings from this post hoc analysis suggest that differences in clinical values between patients who had lipid data at the index date versus those who did not were statistically significant but not clinically meaningful. Nonetheless, this limitation should be considered when interpreting the study findings. Other limitations were related to the use of an EMR research database, including incomplete patient data. Patient care delivered by physicians or practices other than those contributing data to the research database, including laboratory values and prescription drug orders, were captured only if reported to the EMR physician and recorded in a patient s chart. Data on patients behaviors (adherence to diet and exercise regimens, medication compliance) that may influence disease control were not comprehensively collected in the EMR and could not be assessed or controlled in this study. There were limits to the generalizability of this study s findings because the EMR database was not nationally representative from either a patient or provider perspective. There was greater distribution of patients aged 45 and fewer aged 45 years than in the general US population, and the GE EMR had a higher representation of women than the US population, based on 2005 US Census estimates. 12 These patients resided in 35 of 50 US states, and all had access to a source of medical care. Physicians who use EMRs tend to have urban rather than rural practices and see fewer low-income and Medicaid patients than do non-emr physicians. 16 Provider s specialties were not documented at the patient-encounter level at the time of the study. Therefore, it was not possible to stratify patients by physician specialty, which may have influenced treatment and disease management. This study described the management of diabetes and related CV risk factors in patients with T2DM in an ambulatory care-based setting based on EMR data. Because this was accomplished using a cross-sectional design, inferences cannot be made about the causative nature of the relationships between glycemic control, body mass control, and the management of other CV risk factors. Therefore, additional, longitudinal research is necessary to elucidate the associations between glycemic control and simultaneous control of body mass, BP, and lipids, and to more thoroughly understand how achieving control of any one of these CV risk factors individually influences a patient s likelihood of gaining and maintaining control of other risk factors. It would be valuable to repeat this analysis in a nationally representative database that would provide greater external validity. CONCLUSIONS Despite the potential selection bias in this cross-sectional study, a majority of patients with T2DM were not at the recommended goal for glycemic control, and still fewer were concurrently at goal for BMI and BP. Although greater proportions of patients were at goal for BP when both glycemia and BMI were controlled, attainment of goal for all 3 risk factors was uncommon (5.5%). More intensive T2DM treatment with complementary efforts to manage body mass and BP is warranted in this population. ACKNOWLEDGMENTS This study was funded in part through an unrestricted research grant made to the Pharmacotherapy Outcomes Research Center at the University of Utah by Novo Nordisk, Inc. Mr. Bouchard, Mr. Aagren, and Dr. Conner are employees of Novo Nordisk, Inc. Dr. Brixner has received consultancy funds from Novo Nordisk, Inc. The authors acknowledge Brandon Bellows, PharmD, for his assistance in drafting and revising the manuscript, Brian Oberg, MBA, for database support, and Xiangyang Ye, MS, for assistance with statistical analysis. The authors have indicated that they have no other conflicts of interest regarding the content of this article. REFERENCES 1. National Diabetes Information Clearing House. National Diabetes Statistics, 2007. Bethesda, Md: National Institute of Diabetes, Digestive and Kidney Diseases; 2008. 2. Centers for Disease Control and Prevention. National Diabetes Fact Sheet: General Information and National Estimates on Diabetes in the United States, 2007. Atlanta, January 2011 119

Clinical Therapeutics Ga: US Dept of Health and Human Services, Centers for Disease Control and Prevention; 2008. 3. American Diabetes Association. Economic Costs of Diabetes in the US in 2007 [published correction appears in Diabetes Care. 2008;31:1271]. Diabetes Care. 2008;31:596 615. 4. Saydah SH, Fradkin J, Cowie CC. Poor control of risk factors for vascular disease among adults with previously diagnosed diabetes. JAMA. 2004;291:335 342. 5. American Diabetes Association. Standards of medical care in diabetes 2010 [published correction appears in Diabetes Care. 2010;33: 692]. Diabetes Care. 2010;33(Suppl 1):S11 S61. 6. US Dept of Health and Human Services. The Seventh Report of the Joint National Committee on Prevention, Detection, and Treatment of High Blood Pressure. NIH publication 04-5230. 7. GE Healthcare. General Electric Clinical Data Services. 2010. https://www2. gehealthcare.com/portal/site/usen/ Consulting/. Accessed November 12, 2010. 8. Centers for Disease Control. International Classification of Diseases, Ninth Revision, Clinical Modification (ICD- 9-CM). http://www.cdc.gov/nchs/ icd/icd9cm.htm. Accessed December 7, 2010. 9. Rodbard HW, Jellinger PS, Davidson JA, et al. Statement by an American Association of Clinical Endocrinologists/American College of Endocrinology consensus panel on type 2 diabetes mellitus: An algorithm for glycemic control [published correction appears in Endocr Pract. 2009;15: 768 770]. Endocr Pract. 2009;15: 540 559. 10. Chobanian AV, Bakris GL, Black HR, et al, for the National Heart, Lung, and Blood Institute Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure; National High Blood Pressure Education Program Coordinating Committee. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: The JNC 7 Report [published correction appears in JAMA. 2003;290:197]. JAMA. 2003;289:2560 2572. 11. National Institutes of Health. Clinical Guidelines on the Identification, Evaluation and Treatment of Overweight and Obesity in Adults: The Evidence Report. Rockville, Md: National Institutes of Health, National Heart, Lung, and Blood Institute; 1998. Report no. 98-4083. 12. Duckworth W, Abraira C, Moritz T, et al, for the VADT Investigators. Glucose control and vascular complications in veterans with type 2 diabetes [published corrections appear in N Engl J Med. 2009;361:1028 and N Engl J Med. 2009;361:1024 1025]. N Engl J Med. 2009;360:129 139. 13. Gerstein HC, Miller ME, Byington RP, et al, for the Action to Control Cardiovascular Risk in Diabetes Study Group. Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med. 2008;358:2545 2559. 14. Gaede P, Lund-Andersen H, Parving HH, Pedersen O. Effect of a multifactorial intervention on mortality in type 2 diabetes. N Engl J Med. 2008; 358:580 591. 15. Gaede P, Vedel P, Larsen N, et al. Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes. N Engl J Med. 2003; 348:383 393. 16. National Center for Health Statistics. Electronic Medical Record Use by Office-Based Physicians: United States, 2005. Hyattsville, Md: Centers for Disease Control and Prevention; 2006. Address correspondence to: Carrie McAdam-Marx, PhD, RPh, Department of Pharmacotherapy, University of Utah College of Pharmacy, 421 Wakara Way, Room 208, Salt Lake City, UT 84108. E-mail: carrie.mcadammarx@pharm.utah.edu 120 Volume 33 Number 1