- Email: [email protected]
Use of Technology to Track Program Outcomes in a Diabetes Self-Management Program
RESEARCH Perspectives in Practice
Use of Technology to Track Program Outcomes in a Diabetes Self-Management Program CINDA S. CHIMA, MS, RD; NANCY FARMER-DZIAK, MS, RD; PAT CARDWELL, MS, RD; SARA SNOW, MS, RD
ABSTRACT The Diabetes Self-Management Education Program at MetroHealth Medical Center in Cleveland, OH, uses widely available technology to facilitate outcomes tracking and market the diabetes program. Baseline assessment data are entered directly into an Access database form (Microsoft, Inc, Seattle, WA). Quarterly, updated weight and lab data are downloaded into the database from the Epicare electronic medical record (Epic Systems Corp, Madison, WI). This system has enabled staff to track outcomes of program participants on an ongoing basis. To date, 438 patients have been entered into the program database, though complete clinical data are not available for all patients. Mean (⫾standard deviation) baseline body mass index of program participants was 35.8⫾9.1 (range 18.0 to 70.0, n⫽261). Mean (⫾standard deviation) baseline hemoglobin A1c (HbA1c) for all patients was 9.5%⫾2.5%, range 4.5% to 18.3% (n⫽332). Median baseline HbA1c was 9.1%, and the median last available postprogram HbA1c was 7.5% (P⬍.001, n⫽216; patients ranged from 90 days to more than 3 years postprogram entry). Weight change was not significant. In patients 1-year postprogram (n⫽72), mean baseline HbA1c was 9.9%⫾2.9% and the mean 1-year HbA1c value was 7.4%⫾1.7%, P⬍.001. At 1 year, 75% of patients had HbA1c ⱕ8%. In response to these outcomes, an alert was implemented in the outpatient charting system triggered by an HbA1c ⬎8.5% and recommending referral to the Diabetes Self-Management Education Program. Since implementation of the prompt, referrals to the program have increased 40%. J Am Diet Assoc. 2005;105:1933-1938.
C. S. Chima is an assistant professor, Department of Family and Consumer Sciences, University of Akron, Akron, OH. N. Farmer-Dziak is director of Clinical Nutrition, MetroHealth System, Cleveland, OH. P. Cardwell is a clinical dietitian and S. Snow is a coordinator/dietitian in the Department of Pediatrics, both at MetroHealth Medical Center, Cleveland, OH. Address correspondence to: Cinda S. Chima, MS, RD, Department of Family and Consumer Sciences, University of Akron, Akron, OH 44325-6103. E-mail: csc19@ uakron.edu Copyright © 2005 by the American Dietetic Association. 0002-8223/05/10512-0009$30.00/0 doi: 10.1016/j.jada.2005.07.013
© 2005 by the American Dietetic Association
T
racking patient outcomes in a systematic manner is challenging for busy clinicians in outpatient settings. Yet outcomes are key in evaluating the effectiveness of medical nutrition therapy, meeting accreditation requirements (1), and marketing self-management programs to physicians, patients, and third-party payers (2). Although use of clinical outcomes analysis to improve care is appropriate (3), little has been published on this topic from a practice perspective. The MetroHealth System is the largest provider of indigent care in the state of Ohio, serving a racially diverse, socially complex population. The Diabetes Self-Management Education Program at MetroHealth Medical Center in Cleveland, OH, is based in the Department of Clinical Nutrition. The director of the department serves as the coordinator for the program, which is staffed by dietitian certified diabetes educators (CDEs) and a nurse CDE from the medical specialties department. The department’s manager of systems and ambulatory services provides systems and business management support. The program achieved recognition from the American Diabetes Association in 2001. American Diabetes Association–recognized programs are required to track demographics and outcomes of program participants in order to achieve and maintain that status (1). Good practice demands that clinicians use outcomes data to improve practice. Yet the logistics of tracking data consistently can overwhelm intent in an ambulatory nutrition program that logged nearly 6,000 visits in 2003 with minimal clerical support. Although outcomes data tracking in the Diabetes SelfManagement Education Program at MetroHealth began with a paper process, several technological advances have facilitated implementation of the outcomes management program. An electronic charting system (Epicare, Epic Systems Corp, Madison, WI) was implemented in ambulatory services, facilitating access to data and requiring installation of a personal computer in every patient-care service area. Establishment of a shared drive for clinical nutrition on the Intranet network allowed entry of data into a database from various sites and promoted standardized practice through use of online educational materials, class outlines, and protocols. Support from MetroHealth’s Information Systems Department helped managers overcome systems obstacles. METHODS An electronic order set for the Diabetes Self-Management Education Program was created in the Epic online order system (Figure 1). The order set was designed to meet regulatory requirements for patient referral and diagnosis. It includes explicit diagnostic criteria for diabetes,
Journal of the AMERICAN DIETETIC ASSOCIATION
1933
Figure 1. Diabetes Self-Management Education Program order set in Epic.
appropriate lab orders, and referrals for Medical Nutrition Therapy and Diabetes Self Management Training, the two components of the Diabetes Management Program. Referrals are transmitted electronically to the clinical nutrition department. An Excel spreadsheet program (Microsoft, Inc, Seattle, WA) is used to track the disposition of referrals (date of referral, contacts made, appointment dates, and appointment completion).
Good practice demands that clinicians use outcomes data to improve practice. An Access database (Microsoft, Inc) and online assessment form were created to mirror data collected in the initial program assessment and subsequent outcomes data (Figure 2). The online form is accessible through shortcuts from each program provider’s desktop. Database access is restricted to program staff. At the initial visit, assessment data are entered into
1934
December 2005 Volume 105 Number 12
the database form by the clinician as the patient is interviewed. The patient’s name is also entered into an electronic roster of program participants in Epic. Most referred patients receive the core curriculum of the program, based on American Diabetes Association content areas (1). The core curriculum is delivered in three 2-hour group classes taught by dietitians or nurse CDEs, and two to three individual medical nutrition therapy sessions. Program options include education on use of the blood glucose monitor, insulin teaching, hearthealthy classes, and preconception guidelines. The ambulatory/systems manager receives a quarterly download of clinical data from Epic for patients on the program roster. Thus, the database is populated with the most recent laboratory and other clinical data. The ambulatory manager queries the database and saves pertinent data to an Excel file. Sigma Stat statistical software version 2.03 (SPSS, 1997, Chicago, IL) is used to analyze and summarize program data. Data are reported to the Diabetes Self-Management Education Program Professional Advisory Board, the Utilization Management and Quality Management Committees, and provider groups
Figure 2. Diabetes Self-Management Education Program online assessment form. in medical departments. Use and publication of aggregate data from the program database has been determined to be exempt from institutional review board review by the MetroHealth Institutional Review Board as per Section 46.101(b) (4) of the Federal Regulations. The Risk Management Department has reviewed the concept and intended use of the program database and found it consistent with patient privacy regulations. RESULTS AND DISCUSSION To date, 438 patients have been entered onto the program roster, meaning they participated in at least one class session. Mean age (⫾standard deviation) at referral was 53⫾12.6 years (range 20 to 85 years, n⫽438). Demographically, patients in the program mirror MetroHealth’s diverse diabetic population, being 45% white, 43% African American, 9% Hispanic, and 3% other (n⫽435). Sixty-two percent are women, 38% men; 32% are married, while 68% are single, separated, widowed, or divorced. Mean (⫾standard deviation) body mass index of patients at referral is 35.8⫾9.1 (range 18.0 to 70.0). More than 75% of patients referred are classified as obese based on body mass index at baseline.
Mean (⫾standard deviation) baseline hemoglobin A1c (HbA1c) for all patients (within 6 months before or 1 month after entering program) was 9.5%⫾2.5% (range 4.5% to 18.3%, n⫽332). Of all patients who had pre- and postprogram HbA1c, median baseline HbA1c was 9.1%, and the median last available HbA1c was 7.5% (P⬍.001, n⫽216). Of these patients, 33% had most recent HbA1c ⬍7%; 29% had HbA1c ⬎7% and ⱕ8%, and 38% had HbA1c ⬎8%. Most recent postprogram HbA1c ranged from 90 days to more than 3 years out from program entry. There was no significant change in weight. In order to evaluate meaningful outcomes in a more homogeneous group, we examined data from patients who were at least 1 year out from program entry, using the follow-up HbA1c value closest to 1 year (n⫽72). In this group, mean baseline HbA1c was 9.8%⫾2.9%; mean 1-year value was 7.4%⫾1.7%, P⬍.001 by paired t test. At 1 year, 54% had HbA1c ⱕ7%; 19% had values ⬎7% and ⱕ8%; and 26% had values ⬎8%. Participants in the program appear to achieve better blood glucose control than the general diabetes population in the department of medicine, the program’s major referral source. In a random sample of 200 department of
December 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1935
Figure 3. Diabetes Self-Management Education Program prompt in Epic. medicine patients with diabetes, only 25% had HbA1c ⬍7%; 32% had HbA1c ⱖ7% and ⱕ9.5%; and 43% had HbA1c ⬎9.5% (C. Southwell, MD, 2001, unpublished data). The limitations of a clinical outcomes database are apparent. Participants in the program are self-selected, as they may elect not to participate in the program, or no-show or cancel when scheduled. These outcomes were the result of integrated therapy, including self-management training and medications as needed. The data are drawn from a clinical, not a research database; complete data are not available on all persons in the program. However, all available data are included in the analysis. Largely as a result of these outcomes, an interdisciplinary diabetes care improvement task force secured approval to implement an alert in the Epic system in March 2003. The alert is triggered by a diagnosis of diabetes and an HbA1c ⬎8.5%. Internal medicine physicians receive a message stating that the patient has poorly controlled blood glucose and recommending a referral to the Diabetes Self-Management Education Program (see Figure 3). Following implementation of the prompt, referrals
1936
December 2005 Volume 105 Number 12
increased 40%, mostly from internal medicine physicians. The program received 347 referrals in 2003 (246 from medicine). Referrals during the first quarter of 2004 were 2½ times the number of referrals during the same period in 2003 (186 vs 78) (see Figure 4). New patients seen in the program increased 84% from 2002 to 2003, and 90% from first quarter 2003 to first quarter 2004. The prompt was expanded to Family Practice patients in March 2004. Diabetes educators have established a framework and standards related to outcomes measurement in diabetes self-management training programs, focusing on behavioral change as the unique outcome of diabetes self-management training (5). A number of software programs are available that are specifically designed to meet the American Diabetes Association accreditation requirements (6). One disadvantage of their use in a mixed practice is that they are designed specifically for diabetes self-management training. The system developed at MetroHealth can be adapted to various patient populations. The American Dietetic Association has published Evidence-Based Guides for Practice that provide practice standards to facilitate outcomes-based practice for diabe-
Figure 4. Diabetes Self-Management Education Program referral volumes January to December 2003. Hemoglobin A1c prompt in medicine implemented March 10, 2003; revised September 2003.
tes and several other diagnoses (7). These Guides for Practice include appropriate outcomes data sets for specific disease conditions and expectations for improvement through medical nutrition therapy. A database system such as this one could be used to automate ongoing data tracking using the guides. Although outcomes data have been published that support the efficacy of diabetes self-management training and medical nutrition therapy in the treatment of people with diabetes (8-11), the establishment of continuous outcomes-based practice requires a system that is easily usable by busy clinicians at the provider⫺patient interface and compatible with existing medical information systems. CONCLUSIONS Providers of medical nutrition therapy and diabetes selfmanagement training share the challenge of devising effective outcomes management systems that can be used in live clinical settings. Ideally, computerized systems should communicate with electronic medical records already present in the health care environment to minimize the manual work required to transfer data from one system to another. Because this system utilizes standard database and spreadsheet technology, user support is available within the organization. No purchase of specialized software was required and it is flexible enough to be adapted to a variety of patient populations. Little technical skill is required on the part of the front-line provider beyond keyboard entry into an electronic form. Database management expertise is required to design the database and extract the data needed. Continuous data collection allows quick retrieval of data to answer clinical questions. For example, in preparation for a possible disease management program, data regarding program participants covered by the employee health plan were reviewed. Results showed that HbA1c of MetroHealth System employees in the program declined
21%, from 8.6% to 7.1% (P⫽.02, n⫽16.) As the database grows, program data could be used to evaluate utilization and health care costs of plan participants. Moreover, this system can serve as a means of evaluating effectiveness of specific interventions and teaching methods. For example, we found that patients are significantly more likely to complete a concentrated diabetes self-management training group program delivered in 3 successive weeks than one-on-one teaching scheduled to coincide with physician visits over a longer time period. Presently, we are collecting data that will allow us to compare patient outcomes (HbA1c) in individual vs group program participants.
Providers of medical nutrition therapy and diabetes self-management training share the challenge of devising effective outcomes management systems that can be used in live clinical settings. Because MetroHealth is an integrated health system, system staff are well positioned to access clinical data that allow evaluation of program effectiveness. Our experience demonstrates how positive outcomes drive physician referrals. The aggregate data have been presented to primary care departments, endocrinology, utilization management staff, and quality management. Every patient that succeeds is an ambassador for program referrals. This program can be used as a template for expansion to other patient populations in a variety of health care settings.
December 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1937
References 1. American Diabetes Association. National Standards for Diabetes Self Management Education. Diabetes Care. 2004;27(suppl 1):S143-S150. 2. American Dietetic Association. Position of the American Dietetic Association: Nutrition services in managed care. J Am Diet Assoc. 2002;102:1474-1478. 3. American Association of Diabetes Educators. Standards for outcomes measurement of diabetes selfmanagement education. Diabetes Educ. 2003;29:804816. 4. US Department of Health and Human Services [Web site]. Protection of human subjects. 45 CFR §46.101(b). Available at: http://www.hhs.gov/ohrp/humansubjects/ guidance/45cfr46.htm#46.101. Accessed August 5, 2005. 5. Peeples M, Mulcahy K, Tomky D, Weaver R. The conceptual framework of the National Diabetes Education Outcomes System (NDEOS). Diabetes Educ. 2001;27:547-562. 6. Mulcahy K, Tomky D, Peeples M, Weaver T. An ed-
1938
December 2005 Volume 105 Number 12
7.
8.
9. 10. 11.
ucator’s guide to the diabetes outcomes measurement systems. Diabetes Educ. 2001;27:830-842. American Dietetic Association. Medical Nutrition Therapy Evidence-based Guides for Practice. Nutrition Practice Guidelines for Type 1 and Type 2 Diabetes Mellitus. Available at: http://www.eatright.org/Member/ ProductCatalog/SearchableProducts/90_8835.cfm. Accessed July 6, 2004. Banister NA, Jastrow ST, Hodges V, Loop R, Gillham MB. Diabetes self-management training program in a community clinic improves patient outcomes at modest cost. J Am Diet Assoc. 2004;104:807-810. Norris SL. Effectiveness of self-management training in type 2 diabetes: A systematic review of randomized controlled trials. Diabetes Care. 2001;24:561-587. Pastors JG, Franz M. The evidence for the effectiveness of medical nutrition therapy in diabetes management. Diabetes Care. 2002;26:608-613. Pastors JG, Franz M, Warshaw H, Daly A, Arnold MS. How effective is medical nutrition therapy in diabetes care? J Am Diet Assoc. 2003;103:827-831.
Use of Technology to Track Program Outcomes in a Diabetes Self-Management Program CINDA S. CHIMA, MS, RD; NANCY FARMER-DZIAK, MS, RD; PAT CARDWELL, MS, RD; SARA SNOW, MS, RD
ABSTRACT The Diabetes Self-Management Education Program at MetroHealth Medical Center in Cleveland, OH, uses widely available technology to facilitate outcomes tracking and market the diabetes program. Baseline assessment data are entered directly into an Access database form (Microsoft, Inc, Seattle, WA). Quarterly, updated weight and lab data are downloaded into the database from the Epicare electronic medical record (Epic Systems Corp, Madison, WI). This system has enabled staff to track outcomes of program participants on an ongoing basis. To date, 438 patients have been entered into the program database, though complete clinical data are not available for all patients. Mean (⫾standard deviation) baseline body mass index of program participants was 35.8⫾9.1 (range 18.0 to 70.0, n⫽261). Mean (⫾standard deviation) baseline hemoglobin A1c (HbA1c) for all patients was 9.5%⫾2.5%, range 4.5% to 18.3% (n⫽332). Median baseline HbA1c was 9.1%, and the median last available postprogram HbA1c was 7.5% (P⬍.001, n⫽216; patients ranged from 90 days to more than 3 years postprogram entry). Weight change was not significant. In patients 1-year postprogram (n⫽72), mean baseline HbA1c was 9.9%⫾2.9% and the mean 1-year HbA1c value was 7.4%⫾1.7%, P⬍.001. At 1 year, 75% of patients had HbA1c ⱕ8%. In response to these outcomes, an alert was implemented in the outpatient charting system triggered by an HbA1c ⬎8.5% and recommending referral to the Diabetes Self-Management Education Program. Since implementation of the prompt, referrals to the program have increased 40%. J Am Diet Assoc. 2005;105:1933-1938.
C. S. Chima is an assistant professor, Department of Family and Consumer Sciences, University of Akron, Akron, OH. N. Farmer-Dziak is director of Clinical Nutrition, MetroHealth System, Cleveland, OH. P. Cardwell is a clinical dietitian and S. Snow is a coordinator/dietitian in the Department of Pediatrics, both at MetroHealth Medical Center, Cleveland, OH. Address correspondence to: Cinda S. Chima, MS, RD, Department of Family and Consumer Sciences, University of Akron, Akron, OH 44325-6103. E-mail: csc19@ uakron.edu Copyright © 2005 by the American Dietetic Association. 0002-8223/05/10512-0009$30.00/0 doi: 10.1016/j.jada.2005.07.013
© 2005 by the American Dietetic Association
T
racking patient outcomes in a systematic manner is challenging for busy clinicians in outpatient settings. Yet outcomes are key in evaluating the effectiveness of medical nutrition therapy, meeting accreditation requirements (1), and marketing self-management programs to physicians, patients, and third-party payers (2). Although use of clinical outcomes analysis to improve care is appropriate (3), little has been published on this topic from a practice perspective. The MetroHealth System is the largest provider of indigent care in the state of Ohio, serving a racially diverse, socially complex population. The Diabetes Self-Management Education Program at MetroHealth Medical Center in Cleveland, OH, is based in the Department of Clinical Nutrition. The director of the department serves as the coordinator for the program, which is staffed by dietitian certified diabetes educators (CDEs) and a nurse CDE from the medical specialties department. The department’s manager of systems and ambulatory services provides systems and business management support. The program achieved recognition from the American Diabetes Association in 2001. American Diabetes Association–recognized programs are required to track demographics and outcomes of program participants in order to achieve and maintain that status (1). Good practice demands that clinicians use outcomes data to improve practice. Yet the logistics of tracking data consistently can overwhelm intent in an ambulatory nutrition program that logged nearly 6,000 visits in 2003 with minimal clerical support. Although outcomes data tracking in the Diabetes SelfManagement Education Program at MetroHealth began with a paper process, several technological advances have facilitated implementation of the outcomes management program. An electronic charting system (Epicare, Epic Systems Corp, Madison, WI) was implemented in ambulatory services, facilitating access to data and requiring installation of a personal computer in every patient-care service area. Establishment of a shared drive for clinical nutrition on the Intranet network allowed entry of data into a database from various sites and promoted standardized practice through use of online educational materials, class outlines, and protocols. Support from MetroHealth’s Information Systems Department helped managers overcome systems obstacles. METHODS An electronic order set for the Diabetes Self-Management Education Program was created in the Epic online order system (Figure 1). The order set was designed to meet regulatory requirements for patient referral and diagnosis. It includes explicit diagnostic criteria for diabetes,
Journal of the AMERICAN DIETETIC ASSOCIATION
1933
Figure 1. Diabetes Self-Management Education Program order set in Epic.
appropriate lab orders, and referrals for Medical Nutrition Therapy and Diabetes Self Management Training, the two components of the Diabetes Management Program. Referrals are transmitted electronically to the clinical nutrition department. An Excel spreadsheet program (Microsoft, Inc, Seattle, WA) is used to track the disposition of referrals (date of referral, contacts made, appointment dates, and appointment completion).
Good practice demands that clinicians use outcomes data to improve practice. An Access database (Microsoft, Inc) and online assessment form were created to mirror data collected in the initial program assessment and subsequent outcomes data (Figure 2). The online form is accessible through shortcuts from each program provider’s desktop. Database access is restricted to program staff. At the initial visit, assessment data are entered into
1934
December 2005 Volume 105 Number 12
the database form by the clinician as the patient is interviewed. The patient’s name is also entered into an electronic roster of program participants in Epic. Most referred patients receive the core curriculum of the program, based on American Diabetes Association content areas (1). The core curriculum is delivered in three 2-hour group classes taught by dietitians or nurse CDEs, and two to three individual medical nutrition therapy sessions. Program options include education on use of the blood glucose monitor, insulin teaching, hearthealthy classes, and preconception guidelines. The ambulatory/systems manager receives a quarterly download of clinical data from Epic for patients on the program roster. Thus, the database is populated with the most recent laboratory and other clinical data. The ambulatory manager queries the database and saves pertinent data to an Excel file. Sigma Stat statistical software version 2.03 (SPSS, 1997, Chicago, IL) is used to analyze and summarize program data. Data are reported to the Diabetes Self-Management Education Program Professional Advisory Board, the Utilization Management and Quality Management Committees, and provider groups
Figure 2. Diabetes Self-Management Education Program online assessment form. in medical departments. Use and publication of aggregate data from the program database has been determined to be exempt from institutional review board review by the MetroHealth Institutional Review Board as per Section 46.101(b) (4) of the Federal Regulations. The Risk Management Department has reviewed the concept and intended use of the program database and found it consistent with patient privacy regulations. RESULTS AND DISCUSSION To date, 438 patients have been entered onto the program roster, meaning they participated in at least one class session. Mean age (⫾standard deviation) at referral was 53⫾12.6 years (range 20 to 85 years, n⫽438). Demographically, patients in the program mirror MetroHealth’s diverse diabetic population, being 45% white, 43% African American, 9% Hispanic, and 3% other (n⫽435). Sixty-two percent are women, 38% men; 32% are married, while 68% are single, separated, widowed, or divorced. Mean (⫾standard deviation) body mass index of patients at referral is 35.8⫾9.1 (range 18.0 to 70.0). More than 75% of patients referred are classified as obese based on body mass index at baseline.
Mean (⫾standard deviation) baseline hemoglobin A1c (HbA1c) for all patients (within 6 months before or 1 month after entering program) was 9.5%⫾2.5% (range 4.5% to 18.3%, n⫽332). Of all patients who had pre- and postprogram HbA1c, median baseline HbA1c was 9.1%, and the median last available HbA1c was 7.5% (P⬍.001, n⫽216). Of these patients, 33% had most recent HbA1c ⬍7%; 29% had HbA1c ⬎7% and ⱕ8%, and 38% had HbA1c ⬎8%. Most recent postprogram HbA1c ranged from 90 days to more than 3 years out from program entry. There was no significant change in weight. In order to evaluate meaningful outcomes in a more homogeneous group, we examined data from patients who were at least 1 year out from program entry, using the follow-up HbA1c value closest to 1 year (n⫽72). In this group, mean baseline HbA1c was 9.8%⫾2.9%; mean 1-year value was 7.4%⫾1.7%, P⬍.001 by paired t test. At 1 year, 54% had HbA1c ⱕ7%; 19% had values ⬎7% and ⱕ8%; and 26% had values ⬎8%. Participants in the program appear to achieve better blood glucose control than the general diabetes population in the department of medicine, the program’s major referral source. In a random sample of 200 department of
December 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1935
Figure 3. Diabetes Self-Management Education Program prompt in Epic. medicine patients with diabetes, only 25% had HbA1c ⬍7%; 32% had HbA1c ⱖ7% and ⱕ9.5%; and 43% had HbA1c ⬎9.5% (C. Southwell, MD, 2001, unpublished data). The limitations of a clinical outcomes database are apparent. Participants in the program are self-selected, as they may elect not to participate in the program, or no-show or cancel when scheduled. These outcomes were the result of integrated therapy, including self-management training and medications as needed. The data are drawn from a clinical, not a research database; complete data are not available on all persons in the program. However, all available data are included in the analysis. Largely as a result of these outcomes, an interdisciplinary diabetes care improvement task force secured approval to implement an alert in the Epic system in March 2003. The alert is triggered by a diagnosis of diabetes and an HbA1c ⬎8.5%. Internal medicine physicians receive a message stating that the patient has poorly controlled blood glucose and recommending a referral to the Diabetes Self-Management Education Program (see Figure 3). Following implementation of the prompt, referrals
1936
December 2005 Volume 105 Number 12
increased 40%, mostly from internal medicine physicians. The program received 347 referrals in 2003 (246 from medicine). Referrals during the first quarter of 2004 were 2½ times the number of referrals during the same period in 2003 (186 vs 78) (see Figure 4). New patients seen in the program increased 84% from 2002 to 2003, and 90% from first quarter 2003 to first quarter 2004. The prompt was expanded to Family Practice patients in March 2004. Diabetes educators have established a framework and standards related to outcomes measurement in diabetes self-management training programs, focusing on behavioral change as the unique outcome of diabetes self-management training (5). A number of software programs are available that are specifically designed to meet the American Diabetes Association accreditation requirements (6). One disadvantage of their use in a mixed practice is that they are designed specifically for diabetes self-management training. The system developed at MetroHealth can be adapted to various patient populations. The American Dietetic Association has published Evidence-Based Guides for Practice that provide practice standards to facilitate outcomes-based practice for diabe-
Figure 4. Diabetes Self-Management Education Program referral volumes January to December 2003. Hemoglobin A1c prompt in medicine implemented March 10, 2003; revised September 2003.
tes and several other diagnoses (7). These Guides for Practice include appropriate outcomes data sets for specific disease conditions and expectations for improvement through medical nutrition therapy. A database system such as this one could be used to automate ongoing data tracking using the guides. Although outcomes data have been published that support the efficacy of diabetes self-management training and medical nutrition therapy in the treatment of people with diabetes (8-11), the establishment of continuous outcomes-based practice requires a system that is easily usable by busy clinicians at the provider⫺patient interface and compatible with existing medical information systems. CONCLUSIONS Providers of medical nutrition therapy and diabetes selfmanagement training share the challenge of devising effective outcomes management systems that can be used in live clinical settings. Ideally, computerized systems should communicate with electronic medical records already present in the health care environment to minimize the manual work required to transfer data from one system to another. Because this system utilizes standard database and spreadsheet technology, user support is available within the organization. No purchase of specialized software was required and it is flexible enough to be adapted to a variety of patient populations. Little technical skill is required on the part of the front-line provider beyond keyboard entry into an electronic form. Database management expertise is required to design the database and extract the data needed. Continuous data collection allows quick retrieval of data to answer clinical questions. For example, in preparation for a possible disease management program, data regarding program participants covered by the employee health plan were reviewed. Results showed that HbA1c of MetroHealth System employees in the program declined
21%, from 8.6% to 7.1% (P⫽.02, n⫽16.) As the database grows, program data could be used to evaluate utilization and health care costs of plan participants. Moreover, this system can serve as a means of evaluating effectiveness of specific interventions and teaching methods. For example, we found that patients are significantly more likely to complete a concentrated diabetes self-management training group program delivered in 3 successive weeks than one-on-one teaching scheduled to coincide with physician visits over a longer time period. Presently, we are collecting data that will allow us to compare patient outcomes (HbA1c) in individual vs group program participants.
Providers of medical nutrition therapy and diabetes self-management training share the challenge of devising effective outcomes management systems that can be used in live clinical settings. Because MetroHealth is an integrated health system, system staff are well positioned to access clinical data that allow evaluation of program effectiveness. Our experience demonstrates how positive outcomes drive physician referrals. The aggregate data have been presented to primary care departments, endocrinology, utilization management staff, and quality management. Every patient that succeeds is an ambassador for program referrals. This program can be used as a template for expansion to other patient populations in a variety of health care settings.
December 2005 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1937
References 1. American Diabetes Association. National Standards for Diabetes Self Management Education. Diabetes Care. 2004;27(suppl 1):S143-S150. 2. American Dietetic Association. Position of the American Dietetic Association: Nutrition services in managed care. J Am Diet Assoc. 2002;102:1474-1478. 3. American Association of Diabetes Educators. Standards for outcomes measurement of diabetes selfmanagement education. Diabetes Educ. 2003;29:804816. 4. US Department of Health and Human Services [Web site]. Protection of human subjects. 45 CFR §46.101(b). Available at: http://www.hhs.gov/ohrp/humansubjects/ guidance/45cfr46.htm#46.101. Accessed August 5, 2005. 5. Peeples M, Mulcahy K, Tomky D, Weaver R. The conceptual framework of the National Diabetes Education Outcomes System (NDEOS). Diabetes Educ. 2001;27:547-562. 6. Mulcahy K, Tomky D, Peeples M, Weaver T. An ed-
1938
December 2005 Volume 105 Number 12
7.
8.
9. 10. 11.
ucator’s guide to the diabetes outcomes measurement systems. Diabetes Educ. 2001;27:830-842. American Dietetic Association. Medical Nutrition Therapy Evidence-based Guides for Practice. Nutrition Practice Guidelines for Type 1 and Type 2 Diabetes Mellitus. Available at: http://www.eatright.org/Member/ ProductCatalog/SearchableProducts/90_8835.cfm. Accessed July 6, 2004. Banister NA, Jastrow ST, Hodges V, Loop R, Gillham MB. Diabetes self-management training program in a community clinic improves patient outcomes at modest cost. J Am Diet Assoc. 2004;104:807-810. Norris SL. Effectiveness of self-management training in type 2 diabetes: A systematic review of randomized controlled trials. Diabetes Care. 2001;24:561-587. Pastors JG, Franz M. The evidence for the effectiveness of medical nutrition therapy in diabetes management. Diabetes Care. 2002;26:608-613. Pastors JG, Franz M, Warshaw H, Daly A, Arnold MS. How effective is medical nutrition therapy in diabetes care? J Am Diet Assoc. 2003;103:827-831.