Sultan Buraena Duncan2011TDE

The Diabetes Educator http://tde.sagepub.com/

Assessing the Value of the Diabetes Educator Ian Duncan, Tamim Ahmed, Qijuan (Emily) Li, Barbara Stetson, Laurie Ruggiero, Kathryn Burton, Dawn Rosenthal and Karen Fitzner The Diabetes Educator 2011 37: 638 originally published online 30 August 2011 DOI: 10.1177/0145721711416256 The online version of this article can be found at: http://tde.sagepub.com/content/37/5/638

Published by: http://www.sagepublications.com

On behalf of:

American Association of Diabetes Educators

Additional services and information for The Diabetes Educator can be found at: Email Alerts: http://tde.sagepub.com/cgi/alerts Subscriptions: http://tde.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav

>> Version of Record - Sep 14, 2011 OnlineFirst Version of Record - Aug 30, 2011 What is This?

Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

The Diabetes EDUCATOR 638

Assessing the Value of the Diabetes Educator Ian Duncan, FSA, FIA, FCIA, MAAA

Purpose

Tamim Ahmed, PhD, MBA

The purpose of this study was to evaluate the effectiveness of diabetes self-management education or training provided by diabetes educators in reducing complications and improving quality of life.

Qijuan (Emily) Li, MPH Barbara Stetson, PhD Laurie Ruggiero, PhD Kathryn Burton, MS, RD, CDE Dawn Rosenthal, RD, CDE Karen Fitzner, PhD From Solucia Consulting, Farmington, Connecticut (Mr Duncan, Dr Ahmed, Ms Li), Department of Psychological and Brain Sciences, University of Louisville, Louisville, Kentucky (Dr Stetson), Institute for Health Research and Policy, School of Public Health, University of Illinois at Chicago, Chicago, Illinois (Dr Ruggiero), University of Cincinnati/VA Medical Center Research Service (151), Cincinnati, Ohio (Ms Burton), American Association of Diabetes Educators, Chicago, Illinois (Ms Rosenthal, Dr Fitzner). Correspondence to Ian Duncan, FSA, FIA, FCIA, MAAA, Solucia Consulting, 220 Farmington Avenue, Suite 4, Farmington, CT 06032 ([email protected]).

Methods Commercial and Medicare payer-derived claims data were used to assess the relationship between DSME/T and cost. Unlike the prior study that examined diabetes education provided by all professionals, the current study focused on the value of interventions performed as part of formal accredited/recognized diabetes education programs provided by diabetes educators only. Specifically, the current study focused on diabetes education delivered in diabetes self-management training programs based on 2 codes (G0108 and G0109).

Results

DOI: 10.1177/0145721711416256 © 2011 The Author(s)

Results of the study provide insights into the differences in trends between participants and nonparticipants in DSMT. People with diabetes who had DSMT encounters provided by diabetes educators in accredited/recognized programs are likely to show lower cost patterns when compared with a control group of people with diabetes without DSMT encounters. People with diabetes who have multiple episodes of DSMT are more likely to receive care in accordance with recommended guidelines and to comply with diabetes-related prescription regimens, resulting in lower costs and utilization trends.

Volume 37, Number 5, September/October 2011 Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

Assessing the Value of the Diabetes Educator 639

Conclusions and Policy Implications The collaboration between diabetes educators and patients continues to demonstrate positive clinical quality outcomes and cost savings. This analysis shows that repeated DSMT encounters over time result in a doseresponse effect on positive outcomes.

D

iabetes has been described as “one of the most challenging health problems of the 21st century.”1 The prevalence of diabetes in the United States is increasing rapidly and has been labeled as an “epidemic,”2 with recent estimates that 23.6 million people in the United States have diabetes (7.8% of the total population).3 Furthermore, the Centers for Disease Control and Prevention estimates that if current trends continue, “1 in 3 Americans will develop diabetes sometime in their lifetime.” In addition, diabetes is the leading cause of new cases of kidney failure, blindness among adults, and nontraumatic lower-extremity amputations.3 Diabetes is estimated to result in total costs of $174 billion ($116 billion direct and $58 billon indirect), and medical expenditures are approximately 2.3 times higher than for people without diabetes.3 Efforts are needed to reduce the burden of diabetes on the individual and society.

Diabetes Self-management Education One central approach to reducing the burden of diabetes is diabetes self-management education (DSME), which has been described as “a critical element of care for all people with diabetes and is necessary in order to improve patient outcomes.”4 The professional society representing diabetes educators in the United States is the American Association of Diabetes Educators (AADE). Credentialed diabetes educators are health care professionals who have specialized training in diabetes care and represent multiple disciplines, such as nursing, dietetics, and pharmacy.5 The AADE defines diabetes self-management training (DSMT) as “an interactive, ongoing process involving the person with diabetes

(or the caregiver or family) and a diabetes educator(s).” DSME and DSMT are used interchangeably to refer to “a collaborative process through which people with or at risk for diabetes gain the knowledge and skills needed to modify their behavior and successfully self-manage the disease and its related conditions.”5,6). The overall objectives of DSME/DSMT are to support informed decision making and problem solving, facilitate optimal self-care behaviors, and promote active collaboration with the health care team to improve clinical outcomes, health status, and quality of life.4 DSME/DSMT “interventions aim to achieve optimal health status, better quality of life and reduce the need for costly health care.”7 Despite the expanding epidemic of diabetes in the United States, a recent state-focused analysis of a diabetes educator workforce by the New York State Health Foundation’s Diabetes Policy Center and Center for Health Workforce Studies found unmet need in rural areas and in lowincome and minority populations.8 The Centers for Medicare and Medicaid Services has designated 2 accrediting bodies for diabetes education programs: the AADE’s Diabetes Education Accreditation Program and the American Diabetes Association’s Education Recognition Program. Education provided within these programs is eligible for Medicare reimbursement. A number of studies have now documented improvements in outcomes in people living with diabetes following participation in DSME/DSMT.9 Dissemination of this information to key stakeholders in diabetes care and advocacy may greatly assist efforts to provide quality diabetes education to promote optimal health care in the midst of the expanding diabetes public health crisis. Prior studies examining the impact of DSME are described below to summarize what is known about the association between diabetes education participation and outcomes and to highlight what has yet to be demonstrated regarding the value of the diabetes educator.

Prior Studies Since the initial study of the value of diabetes education10 was conducted, new guidelines for the practice of DSME (published for the practice of DSME/DSMT) identify diabetes education as a distinct health care specialty and an integral part of diabetes care.5 Recent studies11 and reviews12-15 suggest that more information is needed about the impact of diabetes education provided

Duncan et al Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

The Diabetes EDUCATOR 640

by diabetes educators. For example, a study by Peyrot et al of a large national survey of patients, educators, and physicians found broad support for DSME/DSMT among physicians and patients and that people with diabetes may not receive all the DSME that is needed.11 Patients expressed a wish to receive more self-management support (personal planning, behavior change, coping) than they report receiving. Boren et al recently reviewed the costs and benefits associated with diabetes education.12 Of 26 articles reviewed, 18 reported findings that associated diabetes education (and disease management) with decreased cost, cost savings, cost-effectiveness, or positive return on investment. A recent review article by Urbanski et al13 indicated that there are few controlled trials that include full cost analyses of diabetes education. The review demonstrated that published studies involved differing amounts of diabetes education provided by a variety of health care professionals. Despite the limitations in the existing research, the review summary indicates that diabetes education appears to be costeffective, particularly when provided to individuals with the poorest glycemic control.13 Norris et al14 performed a meta-analysis of the effect of diabetes education on glycemic control. The authors identified 31 articles from 1980 to 1999 that presented results of randomized controlled trials evaluating the efficacy of self-management education on glycohemoglobin in adults with type 2 diabetes. Given the meta-analysis, the authors suggested, “On average, the intervention decreased glycohemoglobin by 0.76% (95% confidence interval, 0.341.18) more than the control group at immediate follow-up; by 0.26% (0.21% increase-0.73% decrease) at 1 to 3 months of follow-up; and by 0.26% (0.05-0.48) at 4 months of follow-up.” An additional finding of this study was that glycohemoglobin “decreased more with additional contact time between participant and educator; a decrease of 1% was noted for every additional 23.6 hours (13.3-105.4) of contact.” Overall, there is limited information regarding dose-response relationships between diabetes education and clinical outcomes. Examination of this association is an aim of the present study. A recent study examined the cost-effectiveness of diabetes self-management programs in community primary care settings.15 The programs examined were all demonstration projects that were a part of the Diabetes Initiative of the Robert Wood Johnson Foundation and that focused on evaluating the benefits of various diabetes self-management programs implemented in economically diverse

and disadvantaged locales with health disparities. Based on costs incurred over a 30-month implementation period, a simulation model estimated that the intervention reduces discounted lifetime treatment and complication costs by $3385 but that the costs were more than offset by the $15 031 projected cost of implementing the intervention and maintaining its effects in subsequent years. Study authors estimated that the DSME interventions would reduce long-term complications, leading to an increase in remaining life years and quality-adjusted life years. Sensitivity analyses tested the robustness of the model’s estimates under various alternative assumptions and revealed that the model generally predicted acceptable cost-effectiveness ratios. It was concluded that the diverse, “real world” DSME programs were cost-effective from a health systems perspective when the cost savings due to reductions in long-term complications were recognized. A study by Duncan et al focused on the value of diabetes education and examined the outcomes of a large population of people with diabetes covered by commercial and Medicare insurance.10 This study included diabetes education delivered by diabetes educators and other health care professionals. The study did not distinguish whether the DSME/DSMT was delivered in an accredited/recognized diabetes education program or another health care setting. To date, no published studies have examined the impact on key outcomes and costs or examined doseresponse effects of DSMT from diabetes education provided by diabetes educators in an accredited/recognized DSME/DSMT program. The Centers for Medicare and Medicaid Services cover and reimburse for DSMT for beneficiaries who meet specified criteria. The present study aims to extend prior research to examine the effect of delivery of DSMT (codes G0108 and G0109) on alignment with HEDIS (Healthcare Effectiveness Data and Information Set) measures, national standards, clinical outcomes, and costs. This study also aims to examine dose-response effects of multiple DSME encounters relative to single episodes.

Hypotheses In each hypothesis, “diabetes education” was defined as education provided, either individually or in groups, within a diabetes education program recognized by the AADE’s Diabetes Education Accreditation Program

Volume 37, Number 5, September/October 2011 Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

Assessing the Value of the Diabetes Educator 641

or the American Diabetes Association’s Education Recognition Program or an AADE-accredited program supported by a standardized curriculum. The Centers for Medicare and Medicaid Services defines such education as DSMT. The value of the diabetes educator was tested using the following 4 hypotheses: 1. Diabetes education delivered in the Diabetes Education Accreditation Program or the Education Recognition Program by diabetes educators (eg, DSMT) is at least as effective as diabetes education delivered by other professionals. 2. Alignment with HEDIS measures and diabetes care guidelines is greater for patients who participate in DSMT within an accredited/recognized program, and such patients are more likely to follow diabetes care guidelines than are similar patients who do not participate in diabetes education. 3. The cost of care (claims) of patients who participate in DSMT programs is lower than that of similar patients who do not participate in diabetes education. 4. Participation in more educational contacts within a DSMT program and more regular diabetes education is associated with better outcomes.

These hypotheses were tested using administrative claims data from the Solucia Consulting database of millions of lives of claims experience (nationally) over several years. In total, a cohort comprising 4 years of continuous enrollment was identified yielding 58 409 members with diabetes. Description of the data may be found below.

Methods Study Design

Two longitudinal analyses were performed in this study, focused on diabetes education delivered via DSMT programs. The first longitudinal analysis replicated our prior analysis by comparing the patients receiving DSMT identified by the 2 codes used by diabetes educators in accredited/recognized programs—G0108 (individual delivery) and G0109 (group delivery)—with the prior group of patients receiving the broader range of diabetes education over the same 3-year study period used in our prior study. The broader range of diabetes education codes that identify services are provided later in the article.

Table 1

Study Group Categoriesa Study Group Category No diabetes education DSMT = 1 DSMT > 1

2005 No diabetes  education DSMT DSMT

2006-2008 No diabetes  education No DSMT DSMT

No diabetes education for a specific calendar year is defined as not having any claims with diabetes education procedure codes G0108 and G0109 or any of the other 6 diabetes education procedure codes in that calendar year. DSMT (diabetes self-management training) for a specific calendar year is defined as having claims with diabetes education procedure codes G0108 or G0109 in that calendar year period.

a

The cohort of patients for the first longitudinal study consisted of members of commercial and Medicare health plans identified with diabetes in 2005 who were continuously enrolled in the health plan throughout the entire 2005-2007 calendar year study period. To enhance the analysis, a second longitudinal study was performed of diabetes patients continuously enrolled in the health plan between 2005 and 2008. Patients in the 4-year longitudinal analysis were assigned to one of the categories noted in Table 1. The challenge in many studies is to achieve equivalence between intervention and comparison groups. There are many methods used to achieve equivalence, including randomization, matching of specific population members, and statistical techniques. Tracking members of the DSMT and no diabetes education cohorts over time has the advantage of allowing us to observe member progress over time as well as test the dose-response effect (if any) of repeated episodes of DSMT. Measures tracked include alignment with best practice and HEDIS clinical process measures over time (eg, A1c scores, lipids testing, microalbumin, and eye exams). Study Population

The study population for both longitudinal studies consisted of members of commercial and Medicare Advantage health plans from Solucia’s national database

Duncan et al Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

The Diabetes EDUCATOR 642

Table 2

Population Cohortsa Study Group Category

Commercial

Medicare

Total

Three-year cohort   No diabetes education

59 619

27 183

86 802

4506

989

5495

31 075

23 342

54 417

  DSMT = 1

2044

762

2806

  DSMT > 1

1050

136

1186

 DSMT Four-year cohort   No diabetes education

DSMT, diabetes self-management training.

a

of payer data. Details of member composition may be found in appendices 1 to 3, which are available on the AADE website (http://www.diabeteseducator.org/export/ sites/aade/_resources/pdf/Value_of_the_Diabetes_ Educator_Appendices.pdf). Data available to us consist of claims of health plan members who are employees and dependents of health plan purchasers (often employers), referred to as commercial, and members who are eligible for Medicare benefits as enrollees in Medicare Advantage plans, referred to as Medicare. In addition to the clinical (service and diagnosis) information included in claims records, claims include financial information. Aggregating financial information over time at the member level results in claims by member, on a monthly basis, referred to as claims per member per month (PMPM). Patient chart data were not available for this analysis. As before, risk adjustment techniques were employed to ensure comparability between the intervention and comparison cohorts. Comparisons were made between health plan members who participated in DSMT programs provided by a diabetes educator (procedure codes G0108 and G1009) and those who received no diabetes education (no evidence of any diabetes education CPT [Current Procedural Terminology] codes, listed in appendix 4 and available at the appendix link). Limiting the range of diabetes education services to those provided by diabetes educators (G0108, G0109) reduced the number of individuals in the diabetes education group by about one-fifth, as compared with the number of members with diabetes educa-

tion when the broader set of codes was used. Because the analysis was observational, we used a standard actuarial technique, risk adjustment, to ensure equivalence between the 2 populations. Risk adjustment is a method for reducing medical condition differences to a single number at the patient level, allowing us to construct average disease burden measures for different populations. Risk scores are calculated on the basis of demographic factors (age, sex) and diagnoses found on claims. The technique is described in more detail in appendix 5 (at the link cited above). The member counts for the longitudinal analysis cohort groups are as follows: To be included in the study, members need to be continuously enrolled for the study period. Because of this restriction in members eligible for inclusion in the study, the number of members included in the 4-year study is considerably fewer than that in our 3-year study, which required only 3 continuous years of eligibility (Table 2). Identifying Diabetes Education Claims

Members were identified for inclusion in the study according to the presence of DSMT services in their claims history. Every time a service is rendered to a health plan member (or Medicare patient), the service provider submits a claim for reimbursement. These claims represent a valuable source of nonclinician information about the patient’s health history and services received. More details are provided in appendix 4 (see the AADE website).

Volume 37, Number 5, September/October 2011 Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

Assessing the Value of the Diabetes Educator 647

Results Three-Year Longitudinal Analysis

The restated 3-year longitudinal analysis for DSMT delivered via accredited/recognized diabetes education programs indicates similar results to the earlier study10 of diabetes education in general. (In the earlier analysis, there were 8 CPT HCPCS procedure codes in the identification algorithm to designate an individual with diabetes with DSMT encounters. The restated analysis revised the DSMT assignment by including 2 procedure codes [G0108 and G0109], selected as the most relevant and direct DSMT codes by the AADE.) Health plan members who participate in diabetes education are more likely to be aligned with best practice treatment and have lower claims costs. Tables 3 and 4 show the use of different types of medical services by the diabetes education and non–diabetes education cohorts, indicating that these patients are seeking out and receiving more primary, preventive care and less acute care. Almost all the difference between the 2 commercial cohorts is accounted for by their difference in inpatient costs, with the non–diabetes education patients being heavier utilizers of inpatient services. Conversely, these patients use less prescription drug services. Differences are significant in each case. Differences are less marked in the Medicare cohort, but the same pattern emerges of lower inpatient and overall costs in the DSMT cohort. Table 5 compares the 2005-2007 annualized cost trends of patients who have taken diabetes education in any of its various forms (the prior study), patients who have taken DSMT (the current study), and patients with no diabetes education. The two columns identified as “No Diabetes Education” differ somewhat despite being applicable to the same period because, as discussed here, different comparison cohorts are involved in the 2 studies. Tables 5 and 6 show little difference between the results in terms of costs and cost trends whether the analysis is performed for “any diabetes education” or includes only individuals who have been provided with DSMT, within a commercial population and within a Medicare population, respectively. Four-Year Longitudinal Analysis

To further study the impact DSMT programs on diabetes patients over time, we constructed a 4-year longitudinal analysis based on data from 2005 through 2008. In

Table 7, commercial members who did not participate in diabetes education began with costs slightly lower than those of the DSMT cohort (5.9%; $665.58 vs $702.51). However, over the 4 years of the study, the costs of members who do not participate in diabetes education increased at an average annual rate of 10.8% versus 6.5% for those members who participated in DSMT. Commercial members who participate in DSMT also exhibit higher rates of adherence to evidence-based process measures. The only area in which noneducation members exhibit higher adherence is in medication possession. Very similar results are observed in the Medicare population (Table 8). Initial cost of the population that did not participate in diabetes education is lower by about 1.6% than that of the DSMT population. The costs of the cohort not participating in diabetes education increases more rapidly than that of DSMT participants. With the exception of medication possession, all quality process measures are higher for the DSMT participating population. The results quoted above, while showing the association between DSMT and higher adherence and lower costs, do not indicate causation. It is possible, for example, that members who are already adherent are more likely to seek out and receive diabetes education. In the 4-year longitudinal analysis, a cohort of patients identified with diabetes in 2005 (the first year for which we have a complete data set for analysis) and continuously enrolled was followed for the next 4 years (ie, we omitted any members who terminated from our database before the end of the period). We also separated the participating patient population between those patients who participated in DSMT once (ie, codes G0108 and G0109) and those patients who had multiple encounters of this DSMT service. Patients who participated at the start of the period (2005) are recorded as DSMT = 1, and those who participated initially and had at least 1 follow-up episode of DSMT in a subsequent year are recorded as DSMT ≥ 2. The resulting sample sizes are as follows: no diabetes education, 31 075; DSMT = 1, 2044; and DSMT ≥ 2, 1050. (The patient populations are described in appendixes 1 and 3, which can be accessed via the AADE website.) Note that in this analysis, we have risk-adjusted the costs to ensure comparability between populations. In the commercial population (Figure 1), with the exception of 2005 for members who have education in 2005 only, the non–diabetes education population’s costs

Duncan et al Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

648

Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

285.38 202.54 301.34 141.54

 Inpatient

 Outpatient

 Professional

 Pharmacy

38.0 64.7 34.3 33.8

 2+

Lipid testing, %

Microalbuminuria, %

Eye exam, %

34.1

49.2

72.4

55.2

87.6

177.42

285.99

194.47

206.55

864.42

DSMT

a

.0005

P

.6411

< .0001

< .0001

< .0001

< .0001

< .0001

.0197

.1838

< .0001

DSMT, diabetes self-management training; PMPM, per member per month.

72.3

 1+

A1c tests, %

930.79

 Cost

PMPM, average ($)

No Diabetes Education

2005

37.2

39.8

69.0

41.3

76.5

157.79

342.98

208.34

303.56

1012.66

No Diabetes Education

Adjusted Costs and Service Measures of Patients With Diabetes: Commerciala

Table 3

37.6

51.1

72.9

53.6

87.9

198.06

297.95

188.00

203.44

887.44

DSMT

2006

.5999

< .0001

< .0001

< .0001

< .0001

< .0001

< .0001

.0077

< .0001

< .0001

P

37.8

40.0

66.4

39.6

73.4

169.66

384.75

216.42

314.61

1085.44

No Diabetes Education

38.2

51.2

71.3

49.4

85.8

209.72

315.20

200.08

188.22

913.21

DSMT

2007

.6411

< .0001

< .0001

< .0001

< .0001

< .0001

< .0001

.0674

< .0001

< .0001

P

649

Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

51.7 83.7 47.0 50.3

 2+

Lipid testing, %

Microalbuminuria, %

Eye exam, %

53.4

59.3

90.7

64.8

97.1

238.52

a

DSMT, diabetes self-management training; PMPM, per member per month.

91.7

 1+

A1c tests, %

222.72

 Pharmacy

254.52

174.55

153.65 235.10

 Outpatient

314.90

 Inpatient

 Professional

266.48

926.37

934.06

DSMT

 Cost

PMPM, average ($)

No Diabetes Education

2005

.0588

< .0001

< .0001

< .0001

< .0001

.0255

.1742

.0349

.0336

.8355

P

52.9

52.0

84.6

51.5

92.1

244.90

272.43

177.82

377.27

1072.42

No Diabetes Education

Adjusted Costs and Service Measures of Patients With Diabetes: Medicare

Table 4

55.8

61.6

91.0

60.0

95.3

254.50

278.17

173.33

321.95

1027.94

DSMT

2006

.0754

< .0001

< .0001

< .0001

.0002

.2436

.6627

.7262

.0336

.3027

P

52.8

50.6

82.6

51.4

90.1

265.10

342.51

205.48

480.19

1293.28

No Diabetes Education

57.1

58.9

87.3

55.9

92.0

266.75

341.60

197.62

418.64

1224.60

DSMT

2007

.0072

< .0001

.0001

.0053

.0443

.8557

.9606

.6289

.1287

.2877

P

The Diabetes EDUCATOR 650

Table 5

Costs per Member per Month: With and Without Diabetes Education—Commercial Prior Study10

Current Study

Year

No Diabetes Education

Any Diabetes Education

No Diabetes Education

DSMT

2007

1072.36

923.25

1085.44

913.21

2005

919.58

865.18

930.79

864.42

8.0

3.3

8.0

2.8

%

DSMT, diabetes self-management training.

a

Table 6

Costs per Member per Month: With and Without Diabetes Education—Medicare Prior Study10

Current Study

Year

No Diabetes Education

2007

1321.81

1240.97

1293.28

1224.60

2005

946.81

947.24

926.37

934.06

18.2

14.5

18.2

14.5

%

Any Diabetes Education

No Diabetes Education

DSMT

DSMT, diabetes self-management training.

a

are always higher than those of the other populations. Members who have more than 1 episode of DSMT have uniformly lower costs than either of the other populations. The P values reflect the differences between DSMT ≥ 2 and no diabetes education. In the Medicare patient cohort (Figure 2), there is little difference in costs or cost trends between the no–diabetes education and 1-episode populations. The population that receives more than 1 episode of education is small but has lower costs than the other populations: no diabetes education, 23 342; DSMT = 1 762; and DSMT ≥ 2 136. Alignment With Clinical Process Measures

An important component of diabetes care is alignment with best practices and evidence-based approaches. Our data do not provide details of the results of tests, but we are able to study the frequency with which different populations receive different tests. Figures 3 and 4

c­ ompare the test rates between 2005 and 2008 of 3 populations: those without diabetes education, those who have 1 episode of DSMT, those with more than 1 episode. Recommended clinical process measure rates are higher in the DSMT population and almost always higher in the population with follow-up DSMT episodes. Among patients with DSMT, rates of alignment with process measures are higher in patients who participate more frequently in DSMT. Overall, the rates of alignment for Medicare patients with clinical process are higher than those of commercial patients. DSMT is associated with higher rates for nearly all measures, and as for the commercial patients, adherence rates generally improve over time (P < .05). Adherence to Medications

Adherence to prescribed medications is an important quality measure for people with diabetes. The medication possession ratio is defined as the number of days’ supply

Volume 37, Number 5, September/October 2011 Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

Assessing the Value of the Diabetes Educator 651

Table 7

Table 8

Costs and Service Measures of Commercial Patients With Diabetes: 2005

Costs and Service Measures of Medicare Patients With Diabetes: 2005

No Diabetes Education

No Diabetes Education

DSMT

Outcomes

DSMT

P

Outcomes

P

PMPM, average ($)

PMPM, average ($)  Cost

665.58

702.51

.0452

 Cost

864.32

895.80

.5908

 Inpatient

145.16

136.52

.5733

 Inpatient

282.92

252.10

.1909

 Outpatient

150.52

151.26

.6904

 Outpatient

139.29

174.08

.0490

 Professional

159.07

168.05

< .0001

 Professional

219.89

232.49

.2310

 Pharmacy

210.83

246.68

.0733

 Pharmacy

222.21

237.13

.0760

A1c tests, %

A1c tests, %  1+

89.3

95.0

< .0001

 1+

92.5

97.0

< .0001

 2+

47.8

61.2

< .0001

 2+

52.5

66.3

< .0001

Lipid testing, %

81.0

77.5

< .0001

Lipid testing, %

85.0

91.4

< .0001

Microalbuminuria, %

44.1

54.1

< .0001

Microalbuminuria, %

48.5

59.1

< .0001

Eye exam, %

27.9

31.3

< .0001

Eye exam, %

50.9

54.7

.0270

MPR

60.7

57.8

.0125

MPR

60.3

51.4

< .0001

MPR > 80%

45.5

39.6

.2118

MPR > 80%

47.6

34.8

< .0001

Average PMPM cost trend (2005-2008)

10.8

6.5

1.6

−0.8

Average PMPM cost trend (2005-2008)

DSMT, diabetes self-management training; PMPM, per member per month; MRP, medication possession ratio.

a

DSMT, diabetes self-management training; PMPM, per member per month; MRP, medication possession ratio.

a

of a diabetes medication that the patient actually possesses, divided by the number of days in the period for which the patient should have possessed the drug, based on the diagnosis and appropriate medication. The specific drugs included in the study are insulin, sulfonylureas/ meglitinides, biguanides, thiazolidinediones, alphaglucosidase inhibitors, incretin mimetics, amylinomimetics, and dipeptidyl IV inhibitors. A patient was considered adherent if he or she had filled a prescription for 1 or more of the indicated drugs and had a medication possession ratio in excess of 80%. We measured the medication possession ratio of members in different years. The medication possession ratio

of the population with DSMT is in aggregate lower than that of the population that does not participate in diabetes education, for both the commercial population and the Medicare population. One contributing factor is the possession ratio of members who undergo initial education but receive no follow-up; this group’s possession ratio is lower than that of the no-education and repeat-education groups, and it accounts for the overall lower rate of the education cohort. Figure 5A and 5B shows the percentage of each population that we consider “adherent” with its drug therapy (ie, a medication possession ratio in excess of 80%) in commercial and Medicare plans. One important conclusion

Duncan et al Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

The Diabetes EDUCATOR 652

Commercial PMPM Cost Coverage (Adjusted) $925

$906.94

$875

$863.32

$825

$775

$810.47 $740.19 $740.94

$793.82 $762.89

$725

$675

$844.80

$723.31 $712.92

Conclusions

$665.58 $643.04

$625 2005

2008

2007

2006

P<..3668

P<.2984

P<.0778

P<.0497

No Diabetes Education

DSMT=1

DSMT=2

Figure 1.  Costs of commercial patients with diabetes: 2005-2008.

Medicare PMPM Cost Coverage (Adjusted) $1,342.88 $1,337.19

$1325 $1,192.99

$1225

$1,267.50

$1125 $1,129.34 $1025

$989.42 $915.13

$984.76 $941.61

$925 $825

showed the highest rates of adherence. Compliance behavior of members who had a single episode of DSMT in 2005 is uniformly lower than that of either the multipleepisode or no–diabetes education cohorts. Despite the fact that the single-episode cohort has a lower adherence rate, this cohort continues to improve faster than the no–diabetes education cohort. Patients who underwent multiple episodes of DSMT showed the highest rates of adherence change. Figure 6 shows the rate of increase in the medication possession ratio in different cohorts between 2005 and 2008.

$960.72 $864.32

$806.91

$725 $625 2005 P<.4828

2006 P<.9948

No Diabetes Education

2007 P<.0461 DSMT=1

2008 P<.7068 DSMT>=2

Figure 2.  Costs of Medicare patients with diabetes: 2005-2008.

from this figure is that overall adherence with medications has been rising over the period of the study in all populations. Patients who underwent multiple episodes of DSMT

The study extends previous research by examining the value of interventions performed by diabetes educators in accredited/recognized diabetes education programs, using a large database of payer-derived claims of services incurred over a 4-year period. In addition to examining financial impact, this study evaluated the associations between DSME/DSMT and alignment with HEDIS measures, clinical guidelines (A1C testing), and medication adherence, as well as the impact on outcomes of variations in the frequency of DSMT (dose-response analysis). To examine our first hypothesis—that diabetes education delivered in the Diabetes Education Accreditation Program or Education Recognition Program is at least as effective as diabetes education delivered outside an accredited/recognized program—we analyzed the source of differences between costs of patients who use diabetes education versus those who do not. Analysis of type of care confirms the quality of care associated with DSMT because differences in average costs are largely due to lower inpatient costs. Conversely, outpatient and pharmacy costs are higher for patients who use diabetes education, indicating that these patients are seeking out and receiving more primary, preventive care (consistent with a chronic care model) and less acute care. We also examined costs and rates of alignment with recommended care for a cohort of patients with diabetes continuously enrolled for 3 and 4 years. Diabetes education is associated with lower cost trends in both the Medicare population and the commercial population, leading us to conclude that diabetes education is an important contributor to “bending the cost curve” (downward). This study also evaluated whether alignment with accepted care guidelines and clinical process measures is

Volume 37, Number 5, September/October 2011 Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

Assessing the Value of the Diabetes Educator 653

Figure 3.  Commercial—diabetes clinical process measures by year: A, single A1c test compliance (sample size of the cohort is given in the first column); B, two or more A1c test compliance; C, lipid panel test compliance (≥ 30 years of age; sample size of the cohort is given in the first column, reduced from other clinical measures because of retention of members with diabetes ≥ 30 years of age); D, microalbumin test compliance; E, retinal eye examination compliance.

greater for patients who participate in diabetes education within an accredited/recognized program relative to similar patients who do not participate in diabetes education. Our findings indicate that DSMT is associated with higher and generally increasing rates of alignment with best practice treatment measures over time.

Finally, this study examined whether participation in more educational contacts within an accredited/recognized diabetes education program, in which DSMT is provided by diabetes educators, and more regular DSMT is associated with better outcomes (ie, the dose-response effect) than individuals who had less DSMT or no diabetes

Duncan et al Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

The Diabetes EDUCATOR 654

Figure 4.  Medicare—diabetes clinical process measures by year: A, single A1c test compliance (sample size of the cohort is given in the first data column); B, Two or more A1c test compliance; C, lipid panel test compliance; D, microalbumin test compliance; E, retinal eye examination compliance.

education at all. Both commercial and Medicare database analyses showed that patients who had more than 1 episode of DSMT had lower costs than either of the other populations. Rates of alignment with recommended quality measures (eg, the frequency of receiving recommended tests) were almost always higher in the population with multiple years of diabetes education. Among patients with

DSMT, rates of alignment with process measures were higher for those who participated more frequently in DSMT. Examination of the percentage of each population considered “adherent” with its drug therapy indicated that overall adherence with medications has been rising over the period of the study in all populations. In addition, patients who participated in multiple episodes of DSMT Volume 37, Number 5, September/October 2011

Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

Assessing the Value of the Diabetes Educator 655

Figure 5.  Medication possession ratios: A, commercial; B, Medicare.

Figure 6.  Rate of increase in medication possession ratios (commercial and Medicare).

showed the highest rates of adherence. Medication adherence levels for individuals who had a single episode of DSMT was, as expected, lower than that of individuals with multiple episodes but, notably, was also lower than for the no–diabetes education cohort, for reasons that have not been verified as part of this study but may be due to a greater propensity to encourage education where individuals are not adherent to medication. Despite the fact that the single-episode cohort had a lower medication adherence rate than the no–diabetes education cohort, the data identify that medication adherence by the single education cohort improves at a faster rate than for the no–diabetes education cohort.

These longitudinal data, representing multiple outcomes and both commercial and Medicare populations, indicate that provision of recognized/accredited diabetes education provides numerous benefits and is costeffective. It also demonstrates that provision of more diabetes education contacts is even more beneficial. Note in Figure 1 that approximately 6% of patients with diabetes with commercial insurance attended an accredited/recognized DSMT program, with 3% attending more than 1 session. As shown in Figure 2, 3% of diabetes patients with Medicare attended 1 DSMT program, while less than 1% attended more than 1 session. The figures reinforce a statement in the recently revised

Duncan et al Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

The Diabetes EDUCATOR 656

document (June 2010) from the AADE, “Diabetes Education Services: Reimbursement Tips for Primary Care Practice,”16 which states, “Diabetes education is an underused service, perhaps because of a lack of clarity about how to obtain reimbursement.” The document is a tool that may clarify reimbursement issues for practitioners. Despite partial or full insurance coverage for DSME/ DSMT programs, it seems that many eligible patients either are not being referred to DSMT programs or are experiencing barriers to participation. Considering the lower patient cost for those who attend more than 1 session of DSMT, tracking systems that remind patients of follow-up education that is available to them may be a way to improve program participation. The reminders may come from the physician practice or from the diabetes education programs. In addition, efforts are needed to identify and remove barriers to participation in diabetes education programs. As noted in previous publications,9,11-15 the benefits of DSME/DSMT are positive and outweigh the costs of providing these services. Applicability to Other Populations

The data supporting this analysis consist of national health care payer data. It includes members of Medicare Advantage (risk-taking health maintenance organization) but does not include Medicare fee-for-service patients or Medicaid patients. Medicare members have access to diabetes education services because this is a covered Medicare benefit. We believe that the commercial members in our database have access to reimbursement for DSMT services (because this is generally a covered benefit under most employer plans), but we are unable to demonstrate this conclusively. However, the strong correlations that we identified between DSMT and diabetes standards of care and cost suggest that we should be able to replicate this analysis in other data sets. We were also limited by the extent of data available (4 years). The divergence that we observed in costs and diabetes care process measures over time in both populations suggests, however, that we would observe this divergence with a longer series of data. Summary

Research indicates that DSMT that is provided by diabetes educators in accredited/recognized diabetes education programs is both clinically beneficial and costeffective. Repeated education encounters appear to

enhance these benefits. However, patients may not be receiving or may not have access to DSME/DSMT, despite the escalating epidemic of diabetes. Dissemination of the empirical support for diabetes education and the value of the diabetes educator to key stakeholders in diabetes care—including people with diabetes, their families and caregivers, health care providers, third party payers, institutions and government agencies—is critical to ensure the provision of cost-effective and beneficial services. References 1. International Diabetes Federation. IDF diabetes atlas. http://www .diabetesatlas.org. Accessed October 2010. 2. Mokdad AH, Ford ES, Bowman BA, et al. The continuing increase of diabetes in the US. Diabetes Care. 2001;24:412. 3. National Center for Chronic Disease Prevention and Health Promotion. Diabetes Successes and Opportunities for PopulationBased Prevention and Control: At a Glance, 2009. Atlanta, GA: Centers for Disease Control and Prevention; 2009. 4. Funnell MM, Brown TL, Childs BP, et al. National standards for diabetes self-management education. Diabetes Care. 2010; 33(suppl 1):S89-S96. 5. American Association of Diabetes Educators. AADE guidelines for the practice of diabetes self-management education and training (DSME/T). Diabetes Educ. 2009;35:85S-107S. 6. American Association of Diabetes Educators; Siminerio LM, Drab SR, Gabbay RA, et al. Diabetes educators: implementing the chronic care model. Diabetes Educ. 2008;34:451-456. 7. American Association of Diabetes Educators. The scope of practice, standards of practice, and standards of professional performance for diabetes educators. http://www.diabeteseducator.org/export/ sites/aade/_resoures/pdf/the_scope_of_practice_07_14_08_ update.pdf. Accessed October 2010. 8. Zahn D, Langelieer M, Moore J, Legendre Y, Edwards T. Certified diabetes educators in New York: findings from a statewide market analysis and recommendations for improving access to diabetes self-management education services. Executive summary. http://www/nysdiabetescampaign.org. 9. Clark M. Diabetes self-management education: a review of published studies. Prim Care Diabetes. 2008;2:113-120. 10. Duncan I, Birkmeyer C, Coughlin D, Li QE, Sherr D, Boren S. Assessing the value of diabetes education. Diabetes Educ. 2009;35:752-760. 11. Peyrot M, Rubin RR, Funnell MM, Siminerio LM. Access to diabetes self-management education: results of national surveys of patients, educators, and physicians. Diabetes Educ. 2009;35: 246-263. 12. Boren SA, Fitzner KA, Panhalkar PS, Specker J. Costs and benefits associated with diabetes education: a review of the literature. Diabetes Educ. 2009;35:72-96. 13. Urbanski P, Wolf A, Herman WH. Cost-effectiveness of diabetes education. J Am Diet Assoc. 2008;108:S6-S11. 14. Norris SL, Lau J, Smoth SJ, Schmid CH, Engelgau MM. Selfmanagement education for adults with type 2 diabetes: a meta

Volume 37, Number 5, September/October 2011 Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012

Assessing the Value of the Diabetes Educator 657

analysis of the effect on glycemic control. Diabetes Care. 2002;25:1159-1171. 15. Brownson CA, Hoerger TJ, Fisher EB, Kilpatrick KE. Costeffectiveness of diabetes self-management programs in community primary care settings. Diabetes Educ. 2009;35:761-769.

16. American Association of Diabetes Educators. Diabetes education services: reimbursement tips for primary care providers. https:// www.diabeteseducator.org/export/sites/aade/_resources/pdf/ research/Diabetes_Education_Services6-10-10.pdf. Published June 2010. Accessed October 2010.

For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.

Duncan et al Downloaded from tde.sagepub.com at American Association of Diabetes Educators on May 22, 2012