This Article Abstract Free Full Text (Free PDF) Free Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via HighWire Citing Articles via Scopus Google Scholar Articles by Duncan, I. Articles by Boren, S. PubMed PubMed Citation Articles by Duncan, I. Articles by Boren, S. Social Bookmarking                         What's this?

Assessing the Value of Diabetes Education

From Solucia Consulting, Hartford, Connecticut (Mr Duncan, Mr Birkmeyer, Dr Coughlin, Ms Li); American Association of Diabetes Educators, Education and Content Development, Chicago, Illinois (Ms Sherr); and Department of Health Management and Informatics, University of Missouri, School of Medicine, Columbia, Missouri (Dr Boren).

Correspondence to Ian Duncan, FSA, FIA, FCIA, MAAA, Solucia Consulting, 220 Farmington Avenue, Suite 4, Hartford, CT 06106 (iduncan{at}soluciaconsulting.com).

	Abstract

Top Abstract Methods Results Discussion Limitations Conclusion References

 
Purpose

The purpose of this study was to evaluate the impact of diabetes self-management education/training (DSME/T) on financial outcomes (cost of patient care).

Methods

Commercial and Medicare claims payer-derived datasets were used to assess whether patients who participate in diabetes education are more likely to follow recommendations for care than similar patients who do not participate in diabetes education, and if claims of patients who participate in diabetes education are lower than those of similar patients who do not.

Results

Patients using diabetes education have lower average costs than patients who do not use diabetes education. Physicians exhibit high variation in their referral rates to diabetes education.

Conclusions

The collaboration between diabetes educators and physicians yields positive clinical quality and cost savings. The analysis indicates that quality can be improved, and cost reduced, by increasing referral rates to diabetes education among low-referring physicians, specifically among men and people in disadvantaged areas. More needs to be done to inform physicians about ways to increase access to diabetes education for underserved populations.

Diabetes educators are healthcare professionals who have specialized training in diabetes care, traditionally drawn from nursing and dietetics, and more recently involving registered pharmacists. The role of the diabetes educator may also be adopted by other members of a healthcare team including physicians, exercise physiologists, ophthalmologists, optometrists, and podiatrists. Studies suggest that an effective diabetes self-management education program includes a nurse, dietitian, and pharmacist as primary instructors and contributors to the curriculum.⁶ Diabetes educators provide DSME/T, but may extend beyond that to include case management, program management, educational activities, health and wellness promotion, and research. Most diabetes educators have undertaken advanced professional, educational, and credentialing requirements to become either certified diabetes educators (CDE) or board certified in advanced diabetes management (BC-ADM).

Evidence suggests that DSME/T is most effective when using a skills-based approach that is focused on making informed self-management choices,⁴ delivered by a multidisciplinary team with specialized knowledge in diabetes care management, and following a comprehensive plan of care using educational delivery skills^4,6-10 and behavioral and psychosocial strategies.^4,7,8

Despite its proven success, only around 50% of Americans with diabetes participate in formal diabetes education and the Healthy People 2010 policy goal is to increase the proportion of people receiving formal diabetes education from the 1998 baseline of 45% to 60% by 2010.^11,12 The utilization rates of certain preventive care practices by adults aged 18 and older in 42 states is generally high. In 2005, 89% had at least an annual doctor visit, around 70% of people had an annual eye exam, an annual foot exam, and at least 2 glycated hemoglobin (A1C) tests in the year, and 53.1% reported having attended a diabetes self-management class.^13,14 Attendance at a self-management class has increased from 51.4% in 2000 to 53.1% in 2005.^13,14 The value and worth of diabetes self-management education is recognized through reimbursement by the Centers for Medicare and Medicaid (CMS) and other third-party payers.

The terms diabetes self-management education (DSME) and diabetes self-management training (DSMT) are often used interchangeably to refer to "a formal process through which persons with or at risk for diabetes develop and use the knowledge and skill required to reach their self-defined diabetes goals."⁴ For simplicity, DSME/T is used throughout this article.

National Standards Underlie Diabetes Self-Management Education
Evidence-based national standards for DSME/T define and address the quality and processes of diabetes self-management education.⁶ These standards are reviewed every 5 years to incorporate updated evidence-based knowledge by a task force including the AADE, the American Diabetes Association (ADA), industry organizations, and federal agencies. Based on evidence that diabetes education delivered from a behavioral change perspective achieves improved clinical outcomes and enhanced quality of life, 10 standards cover organizational structures and processes necessary to deliver a high quality service. The procedures of delivering quality education services including curriculum, educator credentials, and experience, and the outcomes achieved for individual participants including assessment, evaluation, and follow-up.⁶

Effective Diabetes Self-Management Addresses Seven Self-Care Behaviors
The national standards suggest that an individual's self-management goals be measured by progress toward 7 self-care standards known as the AADE7 self-care behaviors. Developed to complement the national standards,⁷ these self-care behaviors are considered to be "a useful framework for assessment and documentation" of an individual's progress.⁶ The AADE describes the AADE7 as "seven self-care behaviors that are essential for improved health status and greater quality of life."⁷ Five core outcome measures (reproduced below), which include the 7 self-care behaviors, form the framework for measuring the outcomes of DSME/T. The AADE7 cover skills and knowledge acquisition in key self-care areas of healthy eating, physical activity, monitoring, medication management, reducing risks of acute and chronic complications, problem solving of diabetes care related issues, and psychosocial adaptation to living with diabetes. In addition to providing a set of core measures of an individual's outcomes, this skill set was also intended to provide a key data set to establish the effectiveness of DSME/T at a population level in the management of diabetes.⁷

Outcomes of Diabetes Self-Management Education
There is a considerable amount of literature devoted to assessing the outcomes of DSME/T. The AADE 5 core standards for outcomes measurement of DSME/T are:

1. Behavior change is the unique outcome measurement for diabetes self-management education.
   
2. Seven diabetes self-care behavior measures determine the effectiveness of diabetes self-management education at individual, participant, and population levels.
   
3. Diabetes self-care behaviors should be evaluated at baseline and then at regular intervals after the education program.
   
4. The continuum of outcomes, including learning, behavioral, clinical, and health status, should be assessed to demonstrate the interrelationship between DSME/T and behavior change in the care of individuals with diabetes.
   
5. Individual patient outcomes are used to guide the intervention and improve care for that patient. Aggregate population outcomes are used to guide programmatic services and for continuous quality improvement activities for the DSME/T and the population it serves.
   

Financial Outcomes of Diabetes Self-Management Education
While DSME/T has been shown to improve quality of life and clinical outcomes, the impact of DSME/T on financial outcomes (cost of patient care) has not been similarly studied. We studied the value of diabetes education by testing the following 2 hypotheses:

1. Patients who participate in diabetes education are more likely to follow diabetes care standards than similar patients who do not participate in diabetes education.
   
2. Claims of patients who participate in diabetes education are lower than those of similar patients who do not participate in diabetes education.
   

We tested these hypotheses within administrative claims data from the Solucia database of multiple millions of lives of claims experience (nationally) over several years (a description of the data and the database may be found in Appendix 1, which is available on the AADE website⁵).

	Methods

Top Abstract Methods Results Discussion Limitations Conclusion References

 
Study Design
In a perfect world one would construct a randomized test of the hypotheses and compare results of equivalent groups of patients or would have access to patient chart information on which to build a complete health record for each patient. In a situation where it is neither possible to construct a randomized design nor to obtain patient chart data, the researcher is forced to use other available data, such as administrative claims. This study used administrative claims data to compare process measures and costs of those patients who participate in diabetes education and those who do not.

Study Population
The study population consists of members of commercial and Medicare Advantage health plans from a private national database of payer data. Medicare Advantage (risk-taking HMO) members are included; Medicare fee-for-service patients or Medicaid patients are excluded. Medicare members have access to diabetes education services because it is a covered Medicare benefit. It is likely that the commercial members in the database have access to reimbursement for diabetes education services (because it is generally a covered benefit under most employer plans).

Data
The data supporting the analyses is compiled from national healthcare payer data with 3 years of complete data. Data consists of claims of 8 749 569 health plan members who are employees and dependents of health plan purchasers (often employers), referred to as commercial, and 631 931 members who are eligible for Medicare benefits as enrollees in Medicare Advantage plans, referred to as Medicare. Table 1 presents the number of these individuals who are identified as having diabetes for each of the 3 years for which data were available.

In addition to the clinical (service and diagnosis) information included in claims records, claims also include financial information. Aggregating financial information over time at the member level results in claims by member. On a monthly basis, this is referred to as claims per member per month (PMPM).

Comparisons were made between health plan members who were subject to diabetes education and those who were not. Longitudinal analysis was also used. Because the analysis was observational, a standard actuarial technique, risk adjustment, was used to ensure equivalence between the 2 populations.

Identifying Diabetes Education Claims
Members were identified for inclusion in the study according to the presence of diabetes education services in their claims history. Every time a service is rendered to a health plan member (or Medicare patient) the provider of service submits a claim for reimbursement. These claims represent a valuable source of nonclinician information about the patient's health history and services received. Table 2 presents the procedure codes, developed in conjunction with AADE's research and professional practice committees, that were used to identify diabetes education in the dataset.

Other medical nutrition therapy codes were considered, but were primarily follow-up codes. These codes identified only 15 additional members as potential patients for diabetes education and were therefore omitted from the identifying code set.

Controlling for Differences in Risk, Bias, and Confounding
In designing the study, efforts were taken to reduce potential bias that could occur if patients who were already better managed and/or educated are more likely to participate in diabetes education. Adjustments were made for known bias and confounding. Elements of bias that could arise from, for example, access to education programs, insurance coverage, availability of programs, or the participation of less-severe patients with less comorbidity in diabetes education simply because they enjoy a better (current) quality of life and are more able to participate in education. There is also a possibility that providers may discriminate in some way between patients, for example referring those that are more likely to be compliant to a diabetes educator. This tendency would also result in differential results when comparing patients with and without diabetes education.

To overcome the issue of potential bias due to self-selection by the patient, the results of patient panels of physicians who appeared to be relatively frequent prescribers of diabetes education were compared to those of relatively infrequent prescribers. In addition, the analysis considered the experience over time of a cohort of patients with diabetes. The advantage of looking at a cohort of patients with diabetes in a commercial payer database, such as the one used, is that it allows one to observe some dose-response reaction over time. This allowed for comparison of, for example, rates of compliance with best practice and HEDIS process measures (eg, A1C, lipids, microalbumin, foot checks, eye exams) over time.

This study controls for differences in severity of illness by applying risk adjustment, a technique found frequently in actuarial literature and used by, for example, CMS to assess the relative quality of physicians and in reimbursement for healthcare services. Risk Adjustment is a statistical technique frequently encountered in applications such as provider reimbursement in Medicare, Medicaid, and commercial populations. Risk Adjustment is a method for reducing medical condition differences to a single number at the patient level, allowing the investigator to construct average disease burden measures for different populations. Risk scores are calculated based on demographic factors (eg, age, sex) and diagnoses found on claims. A relatively high correlation exists between the risk score and the overall resource utilization (cost) of a population. Risk adjustment is a useful method when outcomes are related to consistent, measurable, administrative claims-based data.

	Results

Top Abstract Methods Results Discussion Limitations Conclusion References

 
Overall Outcomes
Commercially insured members who use diabetes education cost, on average, 5.7% less (P < .0001) than members who do not participate in diabetes education (Table 3). Participating Medicare members (Table 4) cost significantly less (14%, P < .0001). An important validator of these results is the source of the differences. Commercial members with diabetes education have lower claims for acute services (inpatient claims, P < .0001) and higher claims for primary and preventive services (outpatient, P = .0030; prescription drug claims, P < .0001). Claims for professional services are significantly lower (P = .0006) in nondiabetes education group.

Analysis by Provider Likelihood of Referring to DSME/T
To avoid selection on the part of patients, patient cost and adherence were analyzed by category of providers. Examination of the data indicated very different rates of diabetes education participation by physician category,⁵ even though physician panels appeared to be similar in other respects. The population was segmented according to percentage of diabetes education referrals in the provider practice, which are called least likely (0%-5% prevalence of diabetes education), middle (5%-10% diabetes education), and most likely (greater than 10% prevalence of diabetes education). Figure 1 shows results for commercial members. Figure 2 shows results for Medicare members.

View larger version (30K): [in this window] [in a new window]   Figure 1. Patient diabetes process measures by provider's likeliness to refer to DSME/T commercial. HbA1c 1 + test % refers to the proportion of patients with diabetes who had at least one HbA1c test per year. HbA1c 2 + test % refers to the proportion of patients with diabetes who had at least two HbA1c tests per year.

View larger version (33K): [in this window] [in a new window]   Figure 2. Patient diabetes process measures by provider's likeliness to refer to DSME/T Medicare. HbA1c 1 + test % refers to the proportion of patients with diabetes who had at least one HbA1c test per year. HbA1c 2 + test % refers to the proportion of patients with diabetes who had at least two HbA1c tests per year.

In the commercial population, the population that does not have diabetes education has initial costs that are slightly higher than those of the diabetes education population (2%). However, these values are not significantly different (P = .4226) to each other. Over time the costs of the 2 populations diverge significantly. What is particularly compelling about these results is that the gap between the cost of the diabetes education population and the noneducation population increases over time, so that by year 3 (2007) the nondiabetes population average cost is 12% higher (P < .0001). Similar results are seen in the Medicare population, although the differences are smaller. For the Medicare population, initial cost of the nondiabetes education population is 3% lower (P = .0914) than that of the diabetes education population. However, by 2007, this population's cost is 3% higher (P = .0587) than that of the diabetes education population. This analysis could, however, be affected by the relative risk of those patients who enroll in diabetes education and those who do not. The analysis was conducted again with risk adjustment; the adjusted results are presented in Figure 3.

View larger version (11K): [in this window] [in a new window]   Figure 3. Trend in cost of patients with diabetes (2005-2007).

Risk-adjusted, the nondiabetes education population begins with costs 6% higher (P = .0350) than those of the diabetes education population. At year 3 the divergence continues in the unadjusted claims, and the difference grows to 16.0% (P < .0001). Because the initial average costs are different, the rate of claims increase was analyzed. The analysis shows that for the nondiabetes education population, claims increased at 8% (P < .0001) per year on average. For the diabetes education population, the average rate of cost increase is 3.3% (P = .0131).

For Medicare, applying risk-adjustment results in initial cost of the 2 populations being the same (P = .4681). By year 3, the nondiabetes education population cost is 6% higher (P = .0264) than that of the diabetes education population. The average annual cost increase in the nondiabetes education population is 18.2% (P < .0001); for the diabetes education population it is 14.5% (P < .0001).

The analysis found rates of Healthcare Effectiveness Data and Information Set (HEDIS) process measures that are higher in the diabetes education population as compared to the population that did not receive diabetes education. The study indicates a positive correlation between the number of diabetes education claims in the population and adherence to process measures. Patients who have 1 or more claim for diabetes education in a year are more likely (P < .0001) to have an A1C test or a micro albumin test than those who do not have diabetes education. They are more likely (P < .0001) to have a lipid test (the lipid testing rate for all patients with diabetes is already high) and slightly more likely to have an eye exam (although the percentage of patients who comply with the eye exam is relatively low overall and P-values are not significant). With the exception of the measure 2 + A1C tests in the year, all measures improve over time (P < .0001) for the diabetes education group.

In looking at the overall HEDIS diabetes process measures by year, Medicare patient rates are higher (P < .0001) than those of commercial patients whose rates generally improve over time (P < .05).

	Discussion

Top Abstract Methods Results Discussion Limitations Conclusion References

 
The findings from this study indicate that diabetes education is associated with increased use of primary and preventive services and lower use of acute, inpatient hospital services. Overall, health plan members who participate in diabetes education are also more likely to follow best practice treatment recommendations (eg, HEDIS measures) and to have lower claims costs. The results quoted above show the association between diabetes education and the likelihood to follow treatment and experience lower costs. Diabetes education is associated with higher compliance rates for nearly all HEDIS measures, particularly for the Medicare population.

In all cases, claims for best practice treatment process measures are positively correlated with the extent of diabetes education prevalence at the provider practice level. It may be argued that higher rates of best practices are more likely in practices that prescribe diabetes education because these providers are higher quality. The higher diabetes education prescribers may in general be higher quality providers (analysis of this aspect was beyond the scope of our study). Nevertheless, if this is true, an important conclusion is that diabetes education is (like testing and eye exams) an important component, and possibly an indicator of best practice of diabetes care.

The advantage of looking at a cohort of patients with diabetes in a commercial payer database such as the one used is that it allows one to observe dose-response reaction over time. This allows for comparison of outcomes over time of a more homogeneous cohort with regard, for example, to rates of compliance with best practice, HEDIS process measures. It is noteworthy that the risk adjusted longitudinal analysis shows that for the commercial nondiabetes education population, claims increased at 8% per year on average while for the diabetes education population, the average rate of cost increase is only 3.3%. The average annual cost increase in the Medicare nondiabetes education population is 18.2%. For the diabetes education population it is 14.5%. The divergence observed in costs and diabetes care process measures over time in both the commercial and Medicare populations suggests that this divergence would continue with a longer series of data.

The indications are that diabetes education is helping to reduce the rate of increase in average cost of care. The strength of the correlations identified between diabetes education and both HEDIS process measures and cost suggest that it should be able to replicate this analysis in other datasets.

	Limitations

Top Abstract Methods Results Discussion Limitations Conclusion References

 
The findings from this study do not indicate causation but do provide strong findings based on a large number of covered lives of all ages included in the analysis. Some biases cannot be controlled (eg, perhaps patients who are already compliant are more likely to seek out and receive diabetes education). Information on provider prescribing behavior was not available because this study is based on payer data. Therefore, it was necessary to group providers into categories according to the extent to which their patients participated in diabetes education. Since diabetes education is ordered by physicians, this design corrects for selection on the part of the patient. It does not necessarily correct for selection of providers. As in any similar study, the results should be treated with some caution.

	Conclusion

Top Abstract Methods Results Discussion Limitations Conclusion References

 
This analysis of a very large administrative claims dataset shows that patients participating in diabetes education are younger, more female, located in more affluent areas, and have lower clinical risk, higher adherence to diabetes standards of care, and lower average costs than patients who do not use diabetes education. The differences between average costs of patients who use diabetes education versus those that do not are entirely driven by reduced inpatient costs. Conversely, outpatient and pharmacy costs are higher for patients who use diabetes education, indicating that these patients are receiving more primary, preventive care and less acute, affordable care. Over time, diabetes education is associated with somewhat lower cost trends (Medicare) and significantly lower cost trends (commercial). Physicians exhibit high variation in their use of diabetes education. Patients with diabetes who are treated by high users of diabetes education are more likely to receive recommended care (eg, tests and exams) and have lower average cost.

	Acknowledgments

 
Work for this article was supported financially by the American Association of Diabetes Educators.

	References

Top Abstract Methods Results Discussion Limitations Conclusion References

 

1. 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/_resources/pdf/the_scope_of_practice_07_14_08_Update.pdf. Accessed December 2008.
2. Norris SL, Engelgau MM, Narayan KM. Effectiveness of self-management training in type 2 diabetes. Diabetes Care. 2001;24:561 -587.[Abstract/Free Full Text]
3. American Diabetes Association. Economic costs of diabetes in the US in 2007. Diabetes Care.2008; 31:596 -615.[Abstract/Free Full Text]
4. American Diabetes Association. Standards of medical care in diabetes—2008. Diabetes Care.2008; 31:S12 -S54.[Free Full Text]
5. American Association of Diabetes Educators. Http://www.diabeteseducators.org. Accessed August 2008.
6. Funnell MM, Brown TL, Childs BP, et al. National standards for diabetes self-management education. Diabetes Educ.2007; 33:599 -606.[Free Full Text]
7. American Association of Diabetes Educators. Standards for outcomes measurement of diabetes self-management education. Diabetes Educ. 2003;29:804 -816.[Free Full Text]
8. Booker S, Morris M, Johnson A. Empowered to change: evidence from a qualitative exploration of a user-informed psycho-educational program for people with type 1 diabetes. Chronic Illn.2008; 4:41 -53.[Abstract/Free Full Text]
9. Anderson RM, Funnell MM. The art and science of diabetes education: a culture out of balance. Diabetes Educ.2008; 34:109 -117.[Abstract/Free Full Text]
10. Renders CM, Valk GD, Griffin SJ, Wagner EH, Eijk Van JT, Assendelft WJ. Interventions to improve the management of diabetes in primary care, outpatient, and community settings. Diabetes Care.2001; 24:1821 -1833.[Abstract/Free Full Text]
11. Norris SL, Nichols PJ, Caspersen CJ, et al. Increasing diabetes self-management education in community settings a systematic review. Am J Prev Med.2002; 22:39 -66.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
12. US Department of Health and Human Services. Healthy People 2010: Understanding and Improving Health. 2nd ed. Washington, DC: US Government Printing Office; 2000. http://www.healthypeople.gov/data/midcourse/html/focusareas/FA05Objectives.htm. Accessed August 2008.
13. Centers for Disease Control and Prevention. Rates of preventive care practices per 100 adults with diabetes, United States. http://www.cdc.gov/diabetes/statistics/preventive/fAllPractices.htm. 2005. Accessed August 2008.
14. Centers for Disease Control and Prevention. Age-adjusted rates of ever attended diabetes self-management class per 100 adults with diabetes, United States. http://www.cdc.gov/diabetes/statistics/preventive/fY_class.htm 2000-2005. Accessed August 2008.

The Diabetes Educator, Vol. 35, No. 5, 752-760 (2009)
DOI: 10.1177/0145721709343609


CiteULike    Complore    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter    What's this?

This article has been cited by other articles:

				J. A. Fain The Value of Diabetes Education: Is It Cost-Effective? The Diabetes Educator, September 1, 2009; 35(5): 683 - 683. [Full Text] [PDF]

This Article Abstract Free Full Text (Free PDF) Free Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Saved Citations Download to citation manager Request Permissions Request Reprints Add to My Marked Citations Citing Articles Citing Articles via HighWire Citing Articles via Scopus Google Scholar Articles by Duncan, I. Articles by Boren, S. PubMed PubMed Citation Articles by Duncan, I. Articles by Boren, S. Social Bookmarking                         What's this?