- Email: [email protected]
Use of electronic medical records for clinical research in the management of type 2 diabetes
Research in Social and Administrative Pharmacy j (2014) j–j
Original Research
Use of electronic medical records for clinical research in the management of type 2 diabetes Khalid M. Kamal, Ph.D.*, Ishveen Chopra, M.S., Jennifer P. Elliott, Pharm.D., Thomas J. Mattei, B.S., Pharm.D. Division of Clinical, Social and Administrative Sciences, Duquesne University, Mylan School of Pharmacy, 600 Forbes Avenue, Pittsburgh, PA 15282, USA
Abstract Background: Essential to optimal diabetes care is the organization and management of complex clinical data. An EMR system can facilitate better management of clinical and clinical-related information by standardizing care and increasing the efficiency of delivering quality care to patients. However, studies have not described clinical characteristics of patients with type 2 diabetes in a primary practice setting that utilizes an EMR system. Objective: To describe the demographic characteristics, clinical measures, and resource utilization of patients with type 2 diabetes in a primary care setting that employs an EMR system. Methods: Patients R18 years of age, having two or more visits with their physicians (January 1, 2012 to December 31, 2012), and with a recorded diagnosis of diabetes (ICD-9-CM: 250.xx) were retrospectively identified from the GE CentricityÒ EMR database of a primary care physician group. Demographic characteristics, clinical measures, and resource utilization were evaluated. Descriptive statistics were conducted using frequencies and proportions for categorical data and means and standard deviations for continuous variables. Results: 5170 patients with type 2 diabetes were identified for year 2012. Majority of patients with type 2 diabetes were males (53.38%), whites (86.63%), and obese (62.19%); had HbA1c levels !7% (51.72%), LDL-C levels !100 mg/dL (59.09%), HDL-C levels R40 (56.25%); and had never smoked (54.89%). Compared to patients with HbA1c !7% and 7%–9%, those with HbA1c O9% were the youngest, had higher average office visits/patient, and had a higher prevalence of depression, obesity, elevated LDL-C and lower HDL-C levels. Conclusions: This study provides insight into the potential risk factors for diabetes such as the presence of obesity, dyslipidemia, and depression, specifically in patients with HbA1c levels above 9%. Physicians
Author contributions: All authors contributed equally to all stages of the study and to the preparation and revision of the manuscript. Funding source: This study was not funded by any source. Conflict of interest statement: The authors declare no conflicts of interest or financial interests in any product or service mentioned in the article, including grants, employment, gifts, stock holdings, and honoraria. * Corresponding author. Division of Clinical, Social and Administrative Sciences, Duquesne University, Mylan School of Pharmacy, 418 B Mellon, 600 Forbes Avenue, Pittsburgh, PA 15282, USA. Tel.: þ1 412 396 1926; fax: þ1 412 396 5130. E-mail address: [email protected] (K.M. Kamal). 1551-7411/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.sapharm.2014.01.001
2
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
should use evidence-based benchmarks in the development of EMR disease management programs to improve patient outcomes and quality of care. Ó 2014 Elsevier Inc. All rights reserved. Keywords: Type 2 diabetes; Electronic medical record; HbA1c; Retrospective study
Introduction The use of electronic medical records (EMR) by health care organizations promotes improved quality and efficiency of patient care and serves as a valuable resource for outcomes research.1 The term EMR is generally defined as a standardized electronic database for health care; with data recorded, developed, maintained, and/or provided by clinicians and providers in direct patient care.1,2 The EMR database allows for the capture of important demographic information, clinical data, and resource utilization from patient records. Since data are captured at each visit and stored in the EMR, measurement of clinical outcomes and resource utilization is possible for each patient. Several integrated health providers in the United States (US), such as Kaiser Permanente, Harvard Pilgrim Health System, and the Department of Veterans Affairs, have been leaders in EMR adoption.1,3 The appropriate use of EMR offers various potential benefits to clinical practice.4–6 These can include improved quality of patient care, decreased health care costs, and ease at which key clinical information is exchanged among providers.2,3 Several government initiatives have supported the development and utilization of EMR systems in the US. The American Recovery and Reinvestment Act of 2009 has provisions designed to advance the use of health information technology, including $19 billion to incentivize the adoption of certified EMR technology.1 Despite the various benefits and government incentives, adoption of such systems has been slow. According to the National Center for Health Statistics Physician Workflow study report, only 55% of physicians had adopted an EMR system in 2011; of which three quarters met federal “meaningful use” criteria.7 Some of the barriers to EMR adoption in practice settings include high implementation costs, concerns about privacy, lack of standardization, and a disagreement on who pays for and who profits from these systems.1,8 Continued adoption of EMR is important for health outcomes researchers, as it is a valuable tool in accessing community-based data. Retrospective
data (medical charts, administrative claims) and primary data (surveys, clinical trials) are contemporary data sources utilized in outcomes research. These data sources, however, have inherent limitations. Data is limited to patient claims and there are significant sample selection issues (e.g., sicker or possibly more motivated population).1,9 Clinical outcomes such as blood pressure and cholesterol levels can be extracted from a patient’s paper medical charts, however, the process is time and resource intensive and sometimes a large sample size may not be feasible.1 An EMR system can facilitate better management of clinical and clinical-related information by standardizing care and increasing the efficiency of delivering quality care to patients. From a health outcomes perspective, it can serve as a potential data source containing clinical and clinical-related information on access to community-based clinical measures, offers real time data retrieval from large sample sizes, and can be electronically linked to medical and pharmacy claims data, thus, enabling its use for a broad array of outcomes research.1,3 Diabetes is one chronic condition that has realized the benefit of an EMR enhanced disease management program.10 According to the Centers for Disease Control and Prevention report, 8.3% of the US population had diabetes in 2010, making it the sixth leading cause of death.11 Further, the estimated medical costs and reduced work productivity associated with diabetes accounted for $245 billion in 2012, a 41% increase from the previous estimate in 2007 ($174 billion).12 The management of diabetes requires coordinated medical care coupled with patient selfmanagement to decrease the risk of serious complications such as vascular, renal, and ophthalmologic morbidities.13 The need to provide optimal diabetes care at reduced costs has led several health care organizations to measure and improve the quality of care in diabetes through the use of EMRs.13 An EMR system can facilitate better management of patients by organizing complex clinical information, coordinating tasks among the health care team, and reducing inaccurate or incomplete information.
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
These benefits can support evidence-based clinical decision making and promote costs savings.2,7,8 The first step in providing optimal diabetes care is to identify individuals who are at increased risk for diabetes-related complications. The early identification and the subsequent careful monitoring of demographic and clinical information of at-risk individuals can assist physicians in tailoring comprehensive intervention programs to prevent disease progression and complications. EMR system provides physicians with a support structure to longitudinally monitor their patients, to track the impact of their recommended treatment and follow-up procedures, and to evaluate the impact of the intervention programs that have been implemented. A number of outcomes research studies have utilized EMR data to analyze different aspects of diabetes including quality of care, disease management, medication efficacy and side effects, patient identification, comorbidities and disease complications. However, none of the studies have described the demographic, clinical characteristics and resource utilization of patients with type 2 diabetes in a primary care setting that employs an EMR system. The specific study objectives are to (1) describe the demographic and clinical characteristics of patients with type 2 diabetes; (2) describe the demographic and clinical characteristics of patients based on different HbA1c levels; and (3) provide information on resource utilization in these patients.
Methods Data source This is a retrospective, cross-sectional study that utilizes GE CentricityÒ EMR database of a primary care physician group. At the time of the analysis, the database contained approximately 77,600 active patients receiving care from 35 primary care providers in Southwestern Pennsylvania. Study population Patients were included in the study if they were R18 years of age, had two or more visits with their primary care physicians between January 1, 2012 and December 31, 2012 (active status), and had a recorded diagnosis of diabetes (ICD-9-CM: 250.xx). The study examined two patient cohorts – patients with type 2 diabetes and all patients in the database for year 2012.
3
Study variables Measurement outcomes included demographic characteristics, clinical measures and resource utilization. The demographic characteristics examined included age, gender, race, marital status, and employment status. The clinical parameters analyzed were glycosylated hemoglobin A1c (HbA1c), low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), triglycerides, body mass index (BMI), smoking status, and depression. Resource utilization, such as average office visits per patient, medication use (all active medications, diabetes medications, and average medications/patient), total procedures/services claimed, and foot and eye exams were analyzed for the year 2012. Further, comparisons were made between selected demographic characteristics (e.g., age and gender), clinical measures (e.g., BMI, smoking status, LDL, HDL, triglycerides, and depression) and resource utilization (e.g., office visits, foot, and eye exams) across different HbA1c levels (! 7%, 7–9%, and O9%). Statistical analyses Descriptive statistics were conducted using frequencies and proportions for categorical data and means and standard deviations for continuous variables. Data from the EMR database were extracted using Microsoft SQLÒ and all statistical analyses were conducted using SPSSÒ version 18.0 (SPSS Inc., Chicago, IL). The study was approved by the Institutional Review Board at Duquesne University, Pittsburgh, PA. Results Initial review of the EMR database identified 5170 patients with type 2 diabetes in the year 2012. The mean age of patients with type 2 diabetes was 65.0 13.4 years. The demographic and clinical characteristics of patients with type 2 diabetes are reported in Table 1. The majority of patients with type 2 diabetes were male (53.38%) and white (86.63%). Both marital (20.39%) and employment status (83.54%) had considerable missing data. The majority of patients were obese with a BMI R30 (62.19%); had HbA1c levels ! 7% (51.72%), LDL-C levels !100 mg/dL (59.09%), HDL-C levels R40 (56.25%); and had never smoked (54.89%). Resource utilization in patients with type 2 diabetes compared with all patients in the EMR is
4
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
Table 1 Demographic and clinical characteristics of patients with type 2 diabetes
Table 2 Resource utilization of patients with type 2 diabetes compared to all the patients in the EMR system
Variables
N (5170)
Variables
Age in years (mean SD) Gender Male Female Missing Race Caucasian African-American Asian/Pacific Islander Other Missing Marital status Married Single Divorced/widowed/ separated Other Missing Employment status Employed Unemployed Retired Missing Body mass index Less than 18 18–25 25–30 More than 30 Missing Smoking status Smoker Previous Never Missing HbA1c levels !7% 7–9% O9% Missing LDL-C levels R100 mg/dL !100 mg/dL Missing HDL-C levels !40 mg/dL R40 mg/dL Missing
65 13.4
Percentage
2760 2410 0
53.38 46.62 0.00
4479 157 81 11 443
86.63 3.04 1.57 0.21 8.57
2501 521 618
48.38 10.08 11.95
472 1054
9.13 20.39
381 106 364 4319
7.37 2.05 7.04 83.54
13 471 1311 3215 160
0.25 9.11 25.35 62.19 3.09
752 1568 2838 12
14.55 30.33 54.89 0.23
2674 1874 511 111
51.72 36.25 9.88 2.15
1473 3055 642
28.49 59.09 12.42
2100 2908 162
40.62 56.25 3.13
HbA1c, glycosylated hemoglobin; HDL-C, high density lipoprotein cholesterol; LDL-C, low density lipoprotein cholesterol; SD, standard deviation.
reported in Table 2. Patients with type 2 diabetes had a total of 25,090 physician office visits (21.15% of total office visits for all patients) and
Age in years (mean SD) Total patients Total office visits Average office visits/ patient Average medication count Average DM medication count Total claims/ procedures Foot exam done Eye exam done Top 3 diabetes medications
N All patients
Patients with DM
56 19.6
65 13.4
50,676 118,634 2.34 2.0
5170 25,090 4.85 3.9
6.5 –
11 1.7
610,176
84,372
– –
4194 2218 Metformin, Lantus, Januvia
DM, type 2 diabetes; SD, standard deviation.
84,372 claims (13.83% of total claims for all patients). Patients with type 2 diabetes had 4.85 3.9 office visits per patient, compared to 2.34 2.0 for all patients. 81.12% and 42.90% of patients with type 2 diabetes had foot and eye exams conducted, respectively. Further, the average active medication count was greater for patients with type 2 diabetes when compared to other patients (11 vs. 6.5). On average, patients with type 2 diabetes in this practice were prescribed two diabetes medications, with the most commonly prescribed medications being Metformin, Lantus, and Januvia. The demographic characteristics, clinical measures, and resource utilization of the patients with type 2 diabetes categorized on the basis of HbA1c levels are reported in Table 3. Patients with HbA1c O9% were the youngest (58.4 13.2 years) compared to those with HbA1c between 7% and 9% (65.3 13.5 years) and HbA1c! 7% (66.5 13.1 years). Patients with HbA1c O9% had the highest average office visits/patient (5.32) followed by HbA1c 7%–9% (5.15) and HbA1c !7% (4.57); had a higher prevalence of depression (10.96%) followed by HbA1c !7% (9.54%) and HbA1c 7%–9% (7.84%); had a higher prevalence of obesity (BMIO30) (71.04%), LDL-C R100 mg/dL (35.61%), and HDL-C !40 mg/dL (51.47%) compared to the other two cohorts.
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8 Table 3 Demographic characteristics, clinical measures, and resource utilization of patients with type 2 diabetes categorized on their HbA1c levels Variables
HbA1c
N Age in years (mean SD) Gender Male Female Body mass index Less than 18 18–25 25–30 More than 30 Smoking status Smoker Previous Never Office visits Average office visit/patient (mean SD) LDL-C (R100 mg/dL) HDL-C (! 40 mg/dL) Triglycerides (O150 mg/dL) Foot exam Eye exam Comorbidity – depression
2674 1874 511 66.5 13.1 65.3 13.5 58.4 13.2
!7%
7%–9%
O9%
1382 (51.68) 1037 (55.34) 280 (54.79) 1292 (48.32) 837 (44.66) 231 (45.21) 9 271 743 1577
(0.34) 3 (0.16) 1 (10.13) 155 (8.27) 38 (27.79) 456 (24.33) 90 (58.98) 1206 (64.35) 363
341 (12.75) 913 (34.14) 1603 (59.95) 12,209 4.57
(0.20) (7.44) (17.61) (71.04)
205 (10.94) 81 (15.85) 531 (28.34) 124 (24.27) 931 (49.68) 330 (64.58) 9642 2719 5.15 5.32
771 (28.83) 503 (26.84) 182 (35.61) 972 (36.35) 845 (45.09) 263 (51.47) 1526 (57.07) 885 (47.22) 176 (34.44) 2152 (80.48) 1581 (84.36) 401 (78.47) 1186 (44.35) 826 (44.08) 177 (34.64) 255 (9.54) 147 (7.84) 56 (10.96)
HbA1c, glycosylated hemoglobin; HDL-C, high density lipoprotein cholesterol; LDL-C, low density lipoprotein cholesterol; N, number; SD, standard deviation.
Discussions The implementation of EMR continues to grow nationwide and offers the potential to significantly enhance the process and outcomes of patient care. With continuous improvements in EMR quality and functionality, health care providers have the advantage of providing better and more efficient health care services across their practice settings. The use of EMR in outcomes research is increasing, and impacting health conditions such as diabetes, hypertension, and hyperlipidemia. The majority of studies in diabetes have utilized EMR data to analyze treatment efficacy and quality of care. This study was designed as a preliminary, observational analysis of
5
demographic characteristics, clinical measures and resource utilization of patients with type 2 diabetes in a primary care setting that employed the GE Centricity EMR database. The study results can be interpreted in the context of real-world medical practice and can be useful in the following aspects. First, the study has identified population at increased risk of developing diabetes-related complications, i.e., patients with HbA1c levels O7%. For example, our study identified patient population (such as those with younger age, presence of obesity, and depression) in whom HbA1c goals were not attained. Further, physicians using this EMR database can identify and follow-up these patients and implement appropriate strategies, tailored specifically for each patient based on their overall clinical profile. The physicians can longitudinally review the patient data and can specifically intervene on patients not attaining HbA1c goals by reviewing their monitoring parameters, treatment goals and follow-up procedures. The results can assist physicians in making clinical decisions regarding the intensity of preventive interventions (whether dietary advice should be strict and specific, when to provide suggestions for physical activity and when it should be intensified or individualized, and when and which medications should be prescribed) based on overall clinical profile. Finally, our study suggests that using EMR for outcomes research can improve the overall process of care, especially in a primary care setting by increasing the efficiency of complex data retrieval which is crucial in optimal diabetes care. For example, our study showed that higher percentage of patients with HbA1c O9% had LDL-C R100 mg/dL and HDL-C !40 mg/dL which can assist the physicians to ensure targeted management of patients. Demographic and clinical characteristics of patients with type 2 diabetes The demographic characteristics of our cohort of patients with type 2 diabetes is similar to the 2011 national estimates for the US population, which reported that type 2 diabetes affected 11.3% of the total adult population (R20 years), was common among non-Hispanic Whites, affects a higher proportion of men (11.8%) compared to women (10.8%), prevalence increases with age, and was highest among patients over 65 years of age.14 The prevalence of diabetes is reported to increase with increase in BMI.15 Similar trends were observed in our study where the proportion
6
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
of patients with type 2 diabetes increased with increase in BMI: %25 kg/m2 (9.36%), 25–30 kg/m2 (25.35%), and O30 kg/m2 (62.19%). It must be noted that the majority of patients with type 2 diabetes in this practice (51.72%) had HbA1c !7%, suggesting that most patients were well controlled with only 9.9% at increased risk for diabetesrelated complications (HbA1c O9%). Additionally, the majority of patients (56.25%) had optimal HDL-C levels (R40 mg/dL), and fewer than one-third (28.49%) had LDL-C levels R100 mg/dL. Tight control of cholesterol levels is important, as heart disease and stroke were noted on 68 and 16% of diabetes-related death certificates in 2004 among people aged 65 years or older, respectively. Improved control of LDL cholesterol can reduce cardiovascular complications by 20–50%.14 Resource utilization of patients with type 2 diabetes compared to all patients in the EMR system The total economic burden of diagnosed diabetes in the US was estimated to be $245 billion in 2012, with $176 billion attributed to direct medical costs. The average medical expenditures among people with diagnosed diabetes were 2.3 times higher than for those without diabetes.14 A similar trend was observed in our study, with the number of office visits/patient for those with type 2 diabetes more than doubling that of patients without type 2 diabetes. Patients with type 2 diabetes were also prescribed nearly double the number of medications than those without type 2 diabetes. Given that only 2 of the 11 medications prescribed for patients with type 2 diabetes were medications with an indication for treating diabetes, we suggest that this is a population who suffers from additional co-morbidities. The majority (81.12%) of patients with type 2 diabetes in this cohort reported receiving a foot exam in the past year. This is important, given that comprehensive foot care programs can reduce amputation rates by 45–85%.14 However, less than half (42.9%) of patients reported receiving an eye exam within the past year. Detecting and treating diabetic eye disease can reduce the development of severe vision loss by an estimated 50–60%.14 Based upon these findings, the patients in this practice could benefit from the development of a program to increase the rates of annual eye exams. This is just one example of how EMR based outcomes research can identify target areas to improve patient care.
The top three diabetes medications prescribed in this cohort of patients were Metformin, LantusÒ and JanuviaÒ. According to the American Diabetes Association’s (ADA) 2013 Clinical Practice Recommendations, metformin is the preferred initial pharmacological agent for the treatment of type 2 diabetes. If non-insulin monotherapy at maximal tolerated dose does not achieve or maintain HbA1c targets over 3–6 months, the ADA recommends adding a second oral agent, a glucagon-like peptide-1 (GLP-1) receptor agonist, or insulin.16 LantusÒ, a long acting insulin, was the second most commonly prescribed medication among this cohort of patients, following the guideline recommendations. The HbA1c lowering efficacy of JanuviaÒ, a dipeptidyl peptidase IV (DPP-IV) inhibitor, is less than that of the guideline recommended GLP-1 receptor agonist class of medications. JanuviaÒ is also weight neutral, whereas the GLP-1 receptor agonists typically cause weight loss. Given that the majority of patients with type 2 diabetes in this cohort were obese, this property would be very beneficial. Using EMR to identify prescribing patterns can help practices assess how well their prescribing patterns adhere to practice guidelines, and initiate policies to increase compliance. Demographic characteristics, clinical measures, and resource utilization of patients with type 2 diabetes based on HbA1c levels Differences in demographic and clinical characteristics were also observed among patients based on their HbA1c levels. A patient is considered to be at goal or has ‘controlled’ diabetes if their HbA1c level is less than 7%. Patients who were not at goal were younger in age (HbA1c 7%–9%, mean age: 65.3 13.5 years and HbA1c O9%, mean age: 58.4 13.2 years) compared to those who were at goal (HbA1c !7%, mean age: 66.5 13.1 years). This trend could be due to increases in disease awareness as age increases, as well as a larger proportion of younger patients being newly diagnosed and in the process of optimizing pharmacologic and non-pharmacologic treatment regimens. Another reason could be increased monitoring in patients over 65 years of age since they likely have other co-morbid conditions requiring frequent follow-up.17 In this study, a higher proportion of patients with HbA1c levels O9% were obese, compared to those with levels !9%. Research suggests that obesity is associated with the development of
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
insulin resistance resulting in type 2 diabetes. It is also a risk factor for the cardiovascular complications affecting people with diabetes.15 Patients in this study with HbA1c O9% also had the highest proportion of LDL-C levels R100 mg/dL and HDL-C levels !40 mg/dL, when compared to patients with HbA1c levels !9%. This subset of patients warrants special attention, as lipid particles in diabetic dyslipidemia are more atherogenic than in non-diabetic dyslipidemia.18 The development of programs to aggressively lower cholesterol levels within this cohort of patients could potentially decrease cardiovascular burden. Another interesting finding was that a higher proportion of patients (10.96%) who were not at goal (HbA1c levels O9%) had a diagnosis of depression compared to those with HbA1c between 7–9% (7.84%) and !7% (9.54%). Previous research suggests that the presence of diabetes doubles the risk of depression, and the risk of developing depression further increases as diabetes complications worsen.19 Better disease recognition and treatment optimization for both diabetes and depression can improve health outcomes in patients with these co-existing conditions. Implications for clinical practice and future research The first step in primary, secondary and tertiary prevention of type 2 diabetes and diabetes-related morbidity/mortality is the identification of individuals at risk for diabetes and diabetes-related complications. EMR is an attractive tool for identifying and monitoring such patients. Appropriate treatment and monitoring is imperative for all patients with diabetes, and identifying the demographic characteristics, clinical measures and typical resource utilization of populations at increased risk for diabetes-related morbidity and mortality should enable physicians to develop comprehensive programs specifically targeting these patients. In addition, having clinical information via EMR can help physicians longitudinally assess prescribing patterns and treatment efficacy. Limitations The use of EMR data has inherent limitations. First, this study was designed with the intent of assessing preliminary results for identifying characteristics of patients with diabetes, thus, only descriptive statistics were conducted for the analysis. Thus, more sophisticated statistical analysis techniques such as multivariate analyses using
7
regression models were not utilized and confounding variables were not controlled for any of the reported analyses. Second, data were manually entered into the EMR which resulted in incomplete data sets for certain patient characteristics (race, marital status, and employment), therefore making it impossible to analyze such characteristics. Third, the EMR data did not contain information related to lifestyle measures such as exercise and diet. It also did not include information on prescription fills and medication adherence. Fourth, the study results may not be generalizable to other populations or other physician practices. Conclusions The finding of this study suggest that EMR data can be used to identify target populations as well as practice patterns in chronic conditions such as diabetes through the assessment of demographic characteristics, clinical measures and resource utilization. This retrospective, observational study also provides insight into the potential risk factors for diabetes such as the presence of obesity, dyslipidemia, and depression, specifically in patients with HbA1c levels above 9%. Physicians should use evidence-based benchmarks in the development of EMR disease management programs to improve patient outcomes and quality of care. Acknowledgments The authors would like to thank Dr. Lou Civitarese and Mitch Kwiatkowski from Preferred Primary Care Physician Group for providing the electronic medical records data. References 1. Belletti D, Zacker C, Mullins CD. Perspectives on electronic medical records adoption: electronic medical records (EMR) in outcomes research. Patient Relat Outcome Meas 2010;1:29–37. 2. Kanas G, Morimoto L, Mowat F, O’Malley C, Fryzek J, Nordyke R. Use of electronic medical records in oncology outcomes research. Clinicoecon Outcomes Res 2010;2:1–14. 3. www.fastercures.org [Internet]. Fastercures, Think Research, Using Electronic Medical Records to Bridge Patient Care and Research; White Paper. Available from: www.fastercures.org/documents/.../ white_papers/emr_whitepaper.pdf; Fall 2005 Cited 2012 Mar 20.
8
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
4. Gill JM, Ewen E, Nsereko M. Impact of an electronic medical record on quality of care in a primary care office. Del Med J 2001;73(5):187–194. 5. Pollard C, Bailey KA, Petitte T, Baus A, Swim M, Hendryx M. Electronic patient registries improve diabetes care and clinical outcomes in rural community health centers. J Rural Health 2009;25(1): 77–84. 6. Guzek J, Guzek S, Murphy K, Gallacher P, Lesneski C. Improving diabetes care using a multitiered quality improvement model. Am J Med Qual 2009;24(6):505–511. 7. DesRoches CM, Campbell EG, Rao SR, et al. Electronic health records in ambulatory care - a national survey of physicians. N Engl J Med 2008;359(1): 50–60. 8. Shortliffe EH. The evolution of electronic medical records. Acad Med 1999;74(4):414–419. 9. Cebul RD. Using electronic medical records to measure and improve performance. Trans Am Clin Climatol Assoc 2008;119:65–75. 10. O’Connor PJ, Sperl-Hillen JM, Rush WA, et al. Impact of electronic health record clinical decision support on diabetes care: a randomized trial. Ann Fam Med 2011;9(1):12–21. 11. Centers for Disease Control and Prevention [Internet]. NCHS Data Brief. Available from: http://www.cdc.gov/nchs/data/databriefs/db98.htm; Cited 2012 Dec 21.
12. American Diabetes Association. Economic costs of diabetes in the U.S. in 2012. Diabetes Care 2013; 36(4):1033–1046. 13. Love TE, Cebul RD, Einstadter D, et al. Electronic medical record-assisted design of a clusterrandomized trial to improve diabetes care and outcomes. J Gen Intern Med 2008;23(4):383–391. 14. American Diabetes Association [Internet]. National Diabetes Fact Sheet, 2011. Available from: http:// www.cdc.gov/diabetes/pubs/pdf/ndfs_2011.pdf; Cited 2013 May 15. 15. American Association of Diabetes Educators. Addressing obesity via diabetes self-management education and training. Diabetes Educ 2012;38(1): 151–154. 16. American Diabetes Association. Standards of medical care in diabetesd2013. Diabetes Care 2013; 36(suppl 1):S11–S66. 17. McDonald M, Hertz RP, Unger AN, Lustik MB. Prevalence, awareness, and management of hypertension, dyslipidemia, and diabetes among United States adults aged 65 and older. J Gerontol A Biol Sci Med Sci 2009;64(2):256–263. 18. Taskinan M. Diabetic dyslipidemia. Atheroscler Suppl 2002;3(1):47–51. 19. PsychCentral [Internet]. Diabetes and Depression [Internet]. Available from: http://psychcentral.com/ lib/2008/diabetes-and-depression/all/1/; Cited 2012 March 31.
Original Research
Use of electronic medical records for clinical research in the management of type 2 diabetes Khalid M. Kamal, Ph.D.*, Ishveen Chopra, M.S., Jennifer P. Elliott, Pharm.D., Thomas J. Mattei, B.S., Pharm.D. Division of Clinical, Social and Administrative Sciences, Duquesne University, Mylan School of Pharmacy, 600 Forbes Avenue, Pittsburgh, PA 15282, USA
Abstract Background: Essential to optimal diabetes care is the organization and management of complex clinical data. An EMR system can facilitate better management of clinical and clinical-related information by standardizing care and increasing the efficiency of delivering quality care to patients. However, studies have not described clinical characteristics of patients with type 2 diabetes in a primary practice setting that utilizes an EMR system. Objective: To describe the demographic characteristics, clinical measures, and resource utilization of patients with type 2 diabetes in a primary care setting that employs an EMR system. Methods: Patients R18 years of age, having two or more visits with their physicians (January 1, 2012 to December 31, 2012), and with a recorded diagnosis of diabetes (ICD-9-CM: 250.xx) were retrospectively identified from the GE CentricityÒ EMR database of a primary care physician group. Demographic characteristics, clinical measures, and resource utilization were evaluated. Descriptive statistics were conducted using frequencies and proportions for categorical data and means and standard deviations for continuous variables. Results: 5170 patients with type 2 diabetes were identified for year 2012. Majority of patients with type 2 diabetes were males (53.38%), whites (86.63%), and obese (62.19%); had HbA1c levels !7% (51.72%), LDL-C levels !100 mg/dL (59.09%), HDL-C levels R40 (56.25%); and had never smoked (54.89%). Compared to patients with HbA1c !7% and 7%–9%, those with HbA1c O9% were the youngest, had higher average office visits/patient, and had a higher prevalence of depression, obesity, elevated LDL-C and lower HDL-C levels. Conclusions: This study provides insight into the potential risk factors for diabetes such as the presence of obesity, dyslipidemia, and depression, specifically in patients with HbA1c levels above 9%. Physicians
Author contributions: All authors contributed equally to all stages of the study and to the preparation and revision of the manuscript. Funding source: This study was not funded by any source. Conflict of interest statement: The authors declare no conflicts of interest or financial interests in any product or service mentioned in the article, including grants, employment, gifts, stock holdings, and honoraria. * Corresponding author. Division of Clinical, Social and Administrative Sciences, Duquesne University, Mylan School of Pharmacy, 418 B Mellon, 600 Forbes Avenue, Pittsburgh, PA 15282, USA. Tel.: þ1 412 396 1926; fax: þ1 412 396 5130. E-mail address: [email protected] (K.M. Kamal). 1551-7411/$ - see front matter Ó 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.sapharm.2014.01.001
2
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
should use evidence-based benchmarks in the development of EMR disease management programs to improve patient outcomes and quality of care. Ó 2014 Elsevier Inc. All rights reserved. Keywords: Type 2 diabetes; Electronic medical record; HbA1c; Retrospective study
Introduction The use of electronic medical records (EMR) by health care organizations promotes improved quality and efficiency of patient care and serves as a valuable resource for outcomes research.1 The term EMR is generally defined as a standardized electronic database for health care; with data recorded, developed, maintained, and/or provided by clinicians and providers in direct patient care.1,2 The EMR database allows for the capture of important demographic information, clinical data, and resource utilization from patient records. Since data are captured at each visit and stored in the EMR, measurement of clinical outcomes and resource utilization is possible for each patient. Several integrated health providers in the United States (US), such as Kaiser Permanente, Harvard Pilgrim Health System, and the Department of Veterans Affairs, have been leaders in EMR adoption.1,3 The appropriate use of EMR offers various potential benefits to clinical practice.4–6 These can include improved quality of patient care, decreased health care costs, and ease at which key clinical information is exchanged among providers.2,3 Several government initiatives have supported the development and utilization of EMR systems in the US. The American Recovery and Reinvestment Act of 2009 has provisions designed to advance the use of health information technology, including $19 billion to incentivize the adoption of certified EMR technology.1 Despite the various benefits and government incentives, adoption of such systems has been slow. According to the National Center for Health Statistics Physician Workflow study report, only 55% of physicians had adopted an EMR system in 2011; of which three quarters met federal “meaningful use” criteria.7 Some of the barriers to EMR adoption in practice settings include high implementation costs, concerns about privacy, lack of standardization, and a disagreement on who pays for and who profits from these systems.1,8 Continued adoption of EMR is important for health outcomes researchers, as it is a valuable tool in accessing community-based data. Retrospective
data (medical charts, administrative claims) and primary data (surveys, clinical trials) are contemporary data sources utilized in outcomes research. These data sources, however, have inherent limitations. Data is limited to patient claims and there are significant sample selection issues (e.g., sicker or possibly more motivated population).1,9 Clinical outcomes such as blood pressure and cholesterol levels can be extracted from a patient’s paper medical charts, however, the process is time and resource intensive and sometimes a large sample size may not be feasible.1 An EMR system can facilitate better management of clinical and clinical-related information by standardizing care and increasing the efficiency of delivering quality care to patients. From a health outcomes perspective, it can serve as a potential data source containing clinical and clinical-related information on access to community-based clinical measures, offers real time data retrieval from large sample sizes, and can be electronically linked to medical and pharmacy claims data, thus, enabling its use for a broad array of outcomes research.1,3 Diabetes is one chronic condition that has realized the benefit of an EMR enhanced disease management program.10 According to the Centers for Disease Control and Prevention report, 8.3% of the US population had diabetes in 2010, making it the sixth leading cause of death.11 Further, the estimated medical costs and reduced work productivity associated with diabetes accounted for $245 billion in 2012, a 41% increase from the previous estimate in 2007 ($174 billion).12 The management of diabetes requires coordinated medical care coupled with patient selfmanagement to decrease the risk of serious complications such as vascular, renal, and ophthalmologic morbidities.13 The need to provide optimal diabetes care at reduced costs has led several health care organizations to measure and improve the quality of care in diabetes through the use of EMRs.13 An EMR system can facilitate better management of patients by organizing complex clinical information, coordinating tasks among the health care team, and reducing inaccurate or incomplete information.
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
These benefits can support evidence-based clinical decision making and promote costs savings.2,7,8 The first step in providing optimal diabetes care is to identify individuals who are at increased risk for diabetes-related complications. The early identification and the subsequent careful monitoring of demographic and clinical information of at-risk individuals can assist physicians in tailoring comprehensive intervention programs to prevent disease progression and complications. EMR system provides physicians with a support structure to longitudinally monitor their patients, to track the impact of their recommended treatment and follow-up procedures, and to evaluate the impact of the intervention programs that have been implemented. A number of outcomes research studies have utilized EMR data to analyze different aspects of diabetes including quality of care, disease management, medication efficacy and side effects, patient identification, comorbidities and disease complications. However, none of the studies have described the demographic, clinical characteristics and resource utilization of patients with type 2 diabetes in a primary care setting that employs an EMR system. The specific study objectives are to (1) describe the demographic and clinical characteristics of patients with type 2 diabetes; (2) describe the demographic and clinical characteristics of patients based on different HbA1c levels; and (3) provide information on resource utilization in these patients.
Methods Data source This is a retrospective, cross-sectional study that utilizes GE CentricityÒ EMR database of a primary care physician group. At the time of the analysis, the database contained approximately 77,600 active patients receiving care from 35 primary care providers in Southwestern Pennsylvania. Study population Patients were included in the study if they were R18 years of age, had two or more visits with their primary care physicians between January 1, 2012 and December 31, 2012 (active status), and had a recorded diagnosis of diabetes (ICD-9-CM: 250.xx). The study examined two patient cohorts – patients with type 2 diabetes and all patients in the database for year 2012.
3
Study variables Measurement outcomes included demographic characteristics, clinical measures and resource utilization. The demographic characteristics examined included age, gender, race, marital status, and employment status. The clinical parameters analyzed were glycosylated hemoglobin A1c (HbA1c), low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), triglycerides, body mass index (BMI), smoking status, and depression. Resource utilization, such as average office visits per patient, medication use (all active medications, diabetes medications, and average medications/patient), total procedures/services claimed, and foot and eye exams were analyzed for the year 2012. Further, comparisons were made between selected demographic characteristics (e.g., age and gender), clinical measures (e.g., BMI, smoking status, LDL, HDL, triglycerides, and depression) and resource utilization (e.g., office visits, foot, and eye exams) across different HbA1c levels (! 7%, 7–9%, and O9%). Statistical analyses Descriptive statistics were conducted using frequencies and proportions for categorical data and means and standard deviations for continuous variables. Data from the EMR database were extracted using Microsoft SQLÒ and all statistical analyses were conducted using SPSSÒ version 18.0 (SPSS Inc., Chicago, IL). The study was approved by the Institutional Review Board at Duquesne University, Pittsburgh, PA. Results Initial review of the EMR database identified 5170 patients with type 2 diabetes in the year 2012. The mean age of patients with type 2 diabetes was 65.0 13.4 years. The demographic and clinical characteristics of patients with type 2 diabetes are reported in Table 1. The majority of patients with type 2 diabetes were male (53.38%) and white (86.63%). Both marital (20.39%) and employment status (83.54%) had considerable missing data. The majority of patients were obese with a BMI R30 (62.19%); had HbA1c levels ! 7% (51.72%), LDL-C levels !100 mg/dL (59.09%), HDL-C levels R40 (56.25%); and had never smoked (54.89%). Resource utilization in patients with type 2 diabetes compared with all patients in the EMR is
4
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
Table 1 Demographic and clinical characteristics of patients with type 2 diabetes
Table 2 Resource utilization of patients with type 2 diabetes compared to all the patients in the EMR system
Variables
N (5170)
Variables
Age in years (mean SD) Gender Male Female Missing Race Caucasian African-American Asian/Pacific Islander Other Missing Marital status Married Single Divorced/widowed/ separated Other Missing Employment status Employed Unemployed Retired Missing Body mass index Less than 18 18–25 25–30 More than 30 Missing Smoking status Smoker Previous Never Missing HbA1c levels !7% 7–9% O9% Missing LDL-C levels R100 mg/dL !100 mg/dL Missing HDL-C levels !40 mg/dL R40 mg/dL Missing
65 13.4
Percentage
2760 2410 0
53.38 46.62 0.00
4479 157 81 11 443
86.63 3.04 1.57 0.21 8.57
2501 521 618
48.38 10.08 11.95
472 1054
9.13 20.39
381 106 364 4319
7.37 2.05 7.04 83.54
13 471 1311 3215 160
0.25 9.11 25.35 62.19 3.09
752 1568 2838 12
14.55 30.33 54.89 0.23
2674 1874 511 111
51.72 36.25 9.88 2.15
1473 3055 642
28.49 59.09 12.42
2100 2908 162
40.62 56.25 3.13
HbA1c, glycosylated hemoglobin; HDL-C, high density lipoprotein cholesterol; LDL-C, low density lipoprotein cholesterol; SD, standard deviation.
reported in Table 2. Patients with type 2 diabetes had a total of 25,090 physician office visits (21.15% of total office visits for all patients) and
Age in years (mean SD) Total patients Total office visits Average office visits/ patient Average medication count Average DM medication count Total claims/ procedures Foot exam done Eye exam done Top 3 diabetes medications
N All patients
Patients with DM
56 19.6
65 13.4
50,676 118,634 2.34 2.0
5170 25,090 4.85 3.9
6.5 –
11 1.7
610,176
84,372
– –
4194 2218 Metformin, Lantus, Januvia
DM, type 2 diabetes; SD, standard deviation.
84,372 claims (13.83% of total claims for all patients). Patients with type 2 diabetes had 4.85 3.9 office visits per patient, compared to 2.34 2.0 for all patients. 81.12% and 42.90% of patients with type 2 diabetes had foot and eye exams conducted, respectively. Further, the average active medication count was greater for patients with type 2 diabetes when compared to other patients (11 vs. 6.5). On average, patients with type 2 diabetes in this practice were prescribed two diabetes medications, with the most commonly prescribed medications being Metformin, Lantus, and Januvia. The demographic characteristics, clinical measures, and resource utilization of the patients with type 2 diabetes categorized on the basis of HbA1c levels are reported in Table 3. Patients with HbA1c O9% were the youngest (58.4 13.2 years) compared to those with HbA1c between 7% and 9% (65.3 13.5 years) and HbA1c! 7% (66.5 13.1 years). Patients with HbA1c O9% had the highest average office visits/patient (5.32) followed by HbA1c 7%–9% (5.15) and HbA1c !7% (4.57); had a higher prevalence of depression (10.96%) followed by HbA1c !7% (9.54%) and HbA1c 7%–9% (7.84%); had a higher prevalence of obesity (BMIO30) (71.04%), LDL-C R100 mg/dL (35.61%), and HDL-C !40 mg/dL (51.47%) compared to the other two cohorts.
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8 Table 3 Demographic characteristics, clinical measures, and resource utilization of patients with type 2 diabetes categorized on their HbA1c levels Variables
HbA1c
N Age in years (mean SD) Gender Male Female Body mass index Less than 18 18–25 25–30 More than 30 Smoking status Smoker Previous Never Office visits Average office visit/patient (mean SD) LDL-C (R100 mg/dL) HDL-C (! 40 mg/dL) Triglycerides (O150 mg/dL) Foot exam Eye exam Comorbidity – depression
2674 1874 511 66.5 13.1 65.3 13.5 58.4 13.2
!7%
7%–9%
O9%
1382 (51.68) 1037 (55.34) 280 (54.79) 1292 (48.32) 837 (44.66) 231 (45.21) 9 271 743 1577
(0.34) 3 (0.16) 1 (10.13) 155 (8.27) 38 (27.79) 456 (24.33) 90 (58.98) 1206 (64.35) 363
341 (12.75) 913 (34.14) 1603 (59.95) 12,209 4.57
(0.20) (7.44) (17.61) (71.04)
205 (10.94) 81 (15.85) 531 (28.34) 124 (24.27) 931 (49.68) 330 (64.58) 9642 2719 5.15 5.32
771 (28.83) 503 (26.84) 182 (35.61) 972 (36.35) 845 (45.09) 263 (51.47) 1526 (57.07) 885 (47.22) 176 (34.44) 2152 (80.48) 1581 (84.36) 401 (78.47) 1186 (44.35) 826 (44.08) 177 (34.64) 255 (9.54) 147 (7.84) 56 (10.96)
HbA1c, glycosylated hemoglobin; HDL-C, high density lipoprotein cholesterol; LDL-C, low density lipoprotein cholesterol; N, number; SD, standard deviation.
Discussions The implementation of EMR continues to grow nationwide and offers the potential to significantly enhance the process and outcomes of patient care. With continuous improvements in EMR quality and functionality, health care providers have the advantage of providing better and more efficient health care services across their practice settings. The use of EMR in outcomes research is increasing, and impacting health conditions such as diabetes, hypertension, and hyperlipidemia. The majority of studies in diabetes have utilized EMR data to analyze treatment efficacy and quality of care. This study was designed as a preliminary, observational analysis of
5
demographic characteristics, clinical measures and resource utilization of patients with type 2 diabetes in a primary care setting that employed the GE Centricity EMR database. The study results can be interpreted in the context of real-world medical practice and can be useful in the following aspects. First, the study has identified population at increased risk of developing diabetes-related complications, i.e., patients with HbA1c levels O7%. For example, our study identified patient population (such as those with younger age, presence of obesity, and depression) in whom HbA1c goals were not attained. Further, physicians using this EMR database can identify and follow-up these patients and implement appropriate strategies, tailored specifically for each patient based on their overall clinical profile. The physicians can longitudinally review the patient data and can specifically intervene on patients not attaining HbA1c goals by reviewing their monitoring parameters, treatment goals and follow-up procedures. The results can assist physicians in making clinical decisions regarding the intensity of preventive interventions (whether dietary advice should be strict and specific, when to provide suggestions for physical activity and when it should be intensified or individualized, and when and which medications should be prescribed) based on overall clinical profile. Finally, our study suggests that using EMR for outcomes research can improve the overall process of care, especially in a primary care setting by increasing the efficiency of complex data retrieval which is crucial in optimal diabetes care. For example, our study showed that higher percentage of patients with HbA1c O9% had LDL-C R100 mg/dL and HDL-C !40 mg/dL which can assist the physicians to ensure targeted management of patients. Demographic and clinical characteristics of patients with type 2 diabetes The demographic characteristics of our cohort of patients with type 2 diabetes is similar to the 2011 national estimates for the US population, which reported that type 2 diabetes affected 11.3% of the total adult population (R20 years), was common among non-Hispanic Whites, affects a higher proportion of men (11.8%) compared to women (10.8%), prevalence increases with age, and was highest among patients over 65 years of age.14 The prevalence of diabetes is reported to increase with increase in BMI.15 Similar trends were observed in our study where the proportion
6
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
of patients with type 2 diabetes increased with increase in BMI: %25 kg/m2 (9.36%), 25–30 kg/m2 (25.35%), and O30 kg/m2 (62.19%). It must be noted that the majority of patients with type 2 diabetes in this practice (51.72%) had HbA1c !7%, suggesting that most patients were well controlled with only 9.9% at increased risk for diabetesrelated complications (HbA1c O9%). Additionally, the majority of patients (56.25%) had optimal HDL-C levels (R40 mg/dL), and fewer than one-third (28.49%) had LDL-C levels R100 mg/dL. Tight control of cholesterol levels is important, as heart disease and stroke were noted on 68 and 16% of diabetes-related death certificates in 2004 among people aged 65 years or older, respectively. Improved control of LDL cholesterol can reduce cardiovascular complications by 20–50%.14 Resource utilization of patients with type 2 diabetes compared to all patients in the EMR system The total economic burden of diagnosed diabetes in the US was estimated to be $245 billion in 2012, with $176 billion attributed to direct medical costs. The average medical expenditures among people with diagnosed diabetes were 2.3 times higher than for those without diabetes.14 A similar trend was observed in our study, with the number of office visits/patient for those with type 2 diabetes more than doubling that of patients without type 2 diabetes. Patients with type 2 diabetes were also prescribed nearly double the number of medications than those without type 2 diabetes. Given that only 2 of the 11 medications prescribed for patients with type 2 diabetes were medications with an indication for treating diabetes, we suggest that this is a population who suffers from additional co-morbidities. The majority (81.12%) of patients with type 2 diabetes in this cohort reported receiving a foot exam in the past year. This is important, given that comprehensive foot care programs can reduce amputation rates by 45–85%.14 However, less than half (42.9%) of patients reported receiving an eye exam within the past year. Detecting and treating diabetic eye disease can reduce the development of severe vision loss by an estimated 50–60%.14 Based upon these findings, the patients in this practice could benefit from the development of a program to increase the rates of annual eye exams. This is just one example of how EMR based outcomes research can identify target areas to improve patient care.
The top three diabetes medications prescribed in this cohort of patients were Metformin, LantusÒ and JanuviaÒ. According to the American Diabetes Association’s (ADA) 2013 Clinical Practice Recommendations, metformin is the preferred initial pharmacological agent for the treatment of type 2 diabetes. If non-insulin monotherapy at maximal tolerated dose does not achieve or maintain HbA1c targets over 3–6 months, the ADA recommends adding a second oral agent, a glucagon-like peptide-1 (GLP-1) receptor agonist, or insulin.16 LantusÒ, a long acting insulin, was the second most commonly prescribed medication among this cohort of patients, following the guideline recommendations. The HbA1c lowering efficacy of JanuviaÒ, a dipeptidyl peptidase IV (DPP-IV) inhibitor, is less than that of the guideline recommended GLP-1 receptor agonist class of medications. JanuviaÒ is also weight neutral, whereas the GLP-1 receptor agonists typically cause weight loss. Given that the majority of patients with type 2 diabetes in this cohort were obese, this property would be very beneficial. Using EMR to identify prescribing patterns can help practices assess how well their prescribing patterns adhere to practice guidelines, and initiate policies to increase compliance. Demographic characteristics, clinical measures, and resource utilization of patients with type 2 diabetes based on HbA1c levels Differences in demographic and clinical characteristics were also observed among patients based on their HbA1c levels. A patient is considered to be at goal or has ‘controlled’ diabetes if their HbA1c level is less than 7%. Patients who were not at goal were younger in age (HbA1c 7%–9%, mean age: 65.3 13.5 years and HbA1c O9%, mean age: 58.4 13.2 years) compared to those who were at goal (HbA1c !7%, mean age: 66.5 13.1 years). This trend could be due to increases in disease awareness as age increases, as well as a larger proportion of younger patients being newly diagnosed and in the process of optimizing pharmacologic and non-pharmacologic treatment regimens. Another reason could be increased monitoring in patients over 65 years of age since they likely have other co-morbid conditions requiring frequent follow-up.17 In this study, a higher proportion of patients with HbA1c levels O9% were obese, compared to those with levels !9%. Research suggests that obesity is associated with the development of
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
insulin resistance resulting in type 2 diabetes. It is also a risk factor for the cardiovascular complications affecting people with diabetes.15 Patients in this study with HbA1c O9% also had the highest proportion of LDL-C levels R100 mg/dL and HDL-C levels !40 mg/dL, when compared to patients with HbA1c levels !9%. This subset of patients warrants special attention, as lipid particles in diabetic dyslipidemia are more atherogenic than in non-diabetic dyslipidemia.18 The development of programs to aggressively lower cholesterol levels within this cohort of patients could potentially decrease cardiovascular burden. Another interesting finding was that a higher proportion of patients (10.96%) who were not at goal (HbA1c levels O9%) had a diagnosis of depression compared to those with HbA1c between 7–9% (7.84%) and !7% (9.54%). Previous research suggests that the presence of diabetes doubles the risk of depression, and the risk of developing depression further increases as diabetes complications worsen.19 Better disease recognition and treatment optimization for both diabetes and depression can improve health outcomes in patients with these co-existing conditions. Implications for clinical practice and future research The first step in primary, secondary and tertiary prevention of type 2 diabetes and diabetes-related morbidity/mortality is the identification of individuals at risk for diabetes and diabetes-related complications. EMR is an attractive tool for identifying and monitoring such patients. Appropriate treatment and monitoring is imperative for all patients with diabetes, and identifying the demographic characteristics, clinical measures and typical resource utilization of populations at increased risk for diabetes-related morbidity and mortality should enable physicians to develop comprehensive programs specifically targeting these patients. In addition, having clinical information via EMR can help physicians longitudinally assess prescribing patterns and treatment efficacy. Limitations The use of EMR data has inherent limitations. First, this study was designed with the intent of assessing preliminary results for identifying characteristics of patients with diabetes, thus, only descriptive statistics were conducted for the analysis. Thus, more sophisticated statistical analysis techniques such as multivariate analyses using
7
regression models were not utilized and confounding variables were not controlled for any of the reported analyses. Second, data were manually entered into the EMR which resulted in incomplete data sets for certain patient characteristics (race, marital status, and employment), therefore making it impossible to analyze such characteristics. Third, the EMR data did not contain information related to lifestyle measures such as exercise and diet. It also did not include information on prescription fills and medication adherence. Fourth, the study results may not be generalizable to other populations or other physician practices. Conclusions The finding of this study suggest that EMR data can be used to identify target populations as well as practice patterns in chronic conditions such as diabetes through the assessment of demographic characteristics, clinical measures and resource utilization. This retrospective, observational study also provides insight into the potential risk factors for diabetes such as the presence of obesity, dyslipidemia, and depression, specifically in patients with HbA1c levels above 9%. Physicians should use evidence-based benchmarks in the development of EMR disease management programs to improve patient outcomes and quality of care. Acknowledgments The authors would like to thank Dr. Lou Civitarese and Mitch Kwiatkowski from Preferred Primary Care Physician Group for providing the electronic medical records data. References 1. Belletti D, Zacker C, Mullins CD. Perspectives on electronic medical records adoption: electronic medical records (EMR) in outcomes research. Patient Relat Outcome Meas 2010;1:29–37. 2. Kanas G, Morimoto L, Mowat F, O’Malley C, Fryzek J, Nordyke R. Use of electronic medical records in oncology outcomes research. Clinicoecon Outcomes Res 2010;2:1–14. 3. www.fastercures.org [Internet]. Fastercures, Think Research, Using Electronic Medical Records to Bridge Patient Care and Research; White Paper. Available from: www.fastercures.org/documents/.../ white_papers/emr_whitepaper.pdf; Fall 2005 Cited 2012 Mar 20.
8
Kamal et al. / Research in Social and Administrative Pharmacy j (2014) 1–8
4. Gill JM, Ewen E, Nsereko M. Impact of an electronic medical record on quality of care in a primary care office. Del Med J 2001;73(5):187–194. 5. Pollard C, Bailey KA, Petitte T, Baus A, Swim M, Hendryx M. Electronic patient registries improve diabetes care and clinical outcomes in rural community health centers. J Rural Health 2009;25(1): 77–84. 6. Guzek J, Guzek S, Murphy K, Gallacher P, Lesneski C. Improving diabetes care using a multitiered quality improvement model. Am J Med Qual 2009;24(6):505–511. 7. DesRoches CM, Campbell EG, Rao SR, et al. Electronic health records in ambulatory care - a national survey of physicians. N Engl J Med 2008;359(1): 50–60. 8. Shortliffe EH. The evolution of electronic medical records. Acad Med 1999;74(4):414–419. 9. Cebul RD. Using electronic medical records to measure and improve performance. Trans Am Clin Climatol Assoc 2008;119:65–75. 10. O’Connor PJ, Sperl-Hillen JM, Rush WA, et al. Impact of electronic health record clinical decision support on diabetes care: a randomized trial. Ann Fam Med 2011;9(1):12–21. 11. Centers for Disease Control and Prevention [Internet]. NCHS Data Brief. Available from: http://www.cdc.gov/nchs/data/databriefs/db98.htm; Cited 2012 Dec 21.
12. American Diabetes Association. Economic costs of diabetes in the U.S. in 2012. Diabetes Care 2013; 36(4):1033–1046. 13. Love TE, Cebul RD, Einstadter D, et al. Electronic medical record-assisted design of a clusterrandomized trial to improve diabetes care and outcomes. J Gen Intern Med 2008;23(4):383–391. 14. American Diabetes Association [Internet]. National Diabetes Fact Sheet, 2011. Available from: http:// www.cdc.gov/diabetes/pubs/pdf/ndfs_2011.pdf; Cited 2013 May 15. 15. American Association of Diabetes Educators. Addressing obesity via diabetes self-management education and training. Diabetes Educ 2012;38(1): 151–154. 16. American Diabetes Association. Standards of medical care in diabetesd2013. Diabetes Care 2013; 36(suppl 1):S11–S66. 17. McDonald M, Hertz RP, Unger AN, Lustik MB. Prevalence, awareness, and management of hypertension, dyslipidemia, and diabetes among United States adults aged 65 and older. J Gerontol A Biol Sci Med Sci 2009;64(2):256–263. 18. Taskinan M. Diabetic dyslipidemia. Atheroscler Suppl 2002;3(1):47–51. 19. PsychCentral [Internet]. Diabetes and Depression [Internet]. Available from: http://psychcentral.com/ lib/2008/diabetes-and-depression/all/1/; Cited 2012 March 31.