- Email: [email protected]
Trends in Care Practices and Outcomes Among Medicare Beneficiaries with Diabetes
Trends in Care Practices and Outcomes Among Medicare Beneficiaries with Diabetes Sylvia Kuo, PhD, Barbara B. Fleming, MD, PhD, Neil S. Gittings, MA, Lein F. Han, PhD, Linda S. Geiss, MA, Michael M. Engelgau, MD, MS, Sheila H. Roman, MD, MPH Background: While diabetes is a major issue for the aging U.S. population, few studies have described the recent trends in both preventive care practices and complications among the Medicare population with diabetes. Using the Medicare Quality Monitoring System (MQMS), this 2004 study describes these trends from 1992 to 2001 and how these rates vary across demographic subgroups. Methods:
Outcomes include age- and gender-adjusted rates of 15 indicators associated with diabetes care from 1992 to 2001, the absolute change in rates from 1992 to 2001, and 2001 rates by demographic subgroups. The data were cross-sectional samples of Medicare beneficiaries with diabetes from 1992 to 2001 from the Medicare 5% Standard Analytic Files.
Results:
Use of preventive care practices rose from 1992 to 2001: 45 percentage points for HbA1c tests, 51 for lipid tests, 8 for eye exams, and 38 for self-monitoring of glucose levels (all p ⬍0.05). Rates for short-term and some long-term complications of diabetes (e.g., lower-extremity amputations and cardiovascular conditions) fell from 1992 to 2001 (p ⬍0.05). However, rates of other long-term complications such as nephropathy, blindness, and retinopathy rose during the period (p ⬍0.05). Nonwhites and beneficiaries aged ⬍65 and ⬎85 exhibited consistently higher complication rates and lower use of preventive services.
Conclusions: The Medicare program has seen some significant improvement in preventive care practices and significant declines in lower-limb amputations and cardiovascular conditions. However, rates for other long-term complications have increased, with evidence of subgroup disparities. The MQMS results provide an early warning for policymakers to focus on the diabetes care provided to some vulnerable subgroups. (Am J Prev Med 2005;29(5):396 – 403) © 2005 American Journal of Preventive Medicine
Introduction
T
he economic cost of diabetes increased from $98 billion in 1997 to $132 billion in 2002,1,2 and has continued to increase in subsequent years. The disease has also become more prevalent. Both developments have made the tracking of complications and services received by individuals with diabetes imperative for informed policymaking. Understanding these trends is especially relevant to Medicare because the Medicare population bears a disproportionate burden of the disease in both preva-
From the Center for Gerontology and Health Care Research, Brown University (Kuo), Providence, Rhode Island; Veterans Health Administration (Fleming), Washington DC; Centers for Medicare & Medicaid Services (Gittings, Han, Roman), Baltimore, Maryland; and Centers for Disease Control and Prevention (Geiss, Engelgau), Atlanta, Georgia When the analysis was conducted, Kuo and Fleming were affiliated with Mathematica Policy Research, and the Centers for Medicare & Medicaid Services, respectively. Address correspondence and reprint requests to: Sylvia Kuo, PhD, Center for Gerontology and Health Care Research, Brown University, Box G-ST210, Providence RI 02912. E-mail: [email protected].
396
lence and cost. An estimated 20% of elderly individuals has diabetes compared with 5% of the general population.3,4 Elderly people living with diabetes incur healthcare expenditures ⱖ50% higher than those without the disease.5,6 Much work has shown recent improvements in the quality of diabetes care for Medicare beneficiaries (fee for service [FFS]7 and managed care8) as well as other groups,9 –11 although further improvement is necessary to achieve targets outlined in Healthy People 2010.12 These quality indicators emphasize types of preventive care, such as annual HbA1c testing or biennial dilated eye examinations, that may reduce the incidence or delay the onset of complications associated with diabetes. However, these studies provide only a partial picture of the burden of diabetes since they do not analyze trends in the complications that contribute to the high cost of the disease. Other work specific to the Medicare population has focused on late-stage outcomes by concentrating on death certificates, hospital discharges, and excess diabetes-associated mortality among the elderly.13–15 Al-
Am J Prev Med 2005;29(5) © 2005 American Journal of Preventive Medicine • Published by Elsevier Inc.
0749-3797/05/$–see front matter doi:10.1016/j.amepre.2005.08.010
though not focused on the Medicare population, the Diabetes Surveillance System16 provides self-reported preventive care and complication rates for the general population of individuals with diabetes, as well as by different subgroups (including age). Consequently, little has been done to assemble an overall picture of trends in diabetes care for Medicare beneficiaries based on indicators such as preventive services and the broad array of conditions complicating the disease. The purpose of this article is to develop such a picture by analyzing rates of preventive care practices and acute- and long-term complications among Medicare beneficiaries with diabetes over a 10-year period. The article also examines whether there is evidence that quality of care has improved, and whether racial or other subgroup differences exist. Results of the analysis will provide policymakers, researchers, and practitioners with information on trends in the quality of diabetes care and on potential areas of concern—such as specific complications or subgroup conditions and care disparities—where resources may be most beneficially directed.
Methods Study Sample The study sample comes from the Medicare Quality Monitoring System (MQMS), which the Centers for Medicare & Medicaid Services (CMS) maintains to monitor the quality of care delivered to Medicare FFS beneficiaries. MQMS uses Medicare administrative data to track national and state trends, patterns, and variations in the use of health care and outcomes of that care, preventable hospitalizations, and patient safety. Diabetes is among the five conditions (including acute myocardial infarction [AMI], heart failure, stroke, and pneumonia) common to the Medicare population specifically targeted by MQMS. The sample data represent random 5% annual cross-sectional samples of Medicare FFS beneficiaries enrolled for the full year from 1992 to 2001. Using an algorithm validated in earlier studies,17–19 the analysis, which was performed in 2004, defined beneficiaries with diabetes as those who had either one inpatient or emergency department (ED) visit or two ambulatory encounters coded for diabetes during the year (excluding individuals with gestational diabetes). The process yields annual sample sizes of 149,652 (in 1992) to 229,350 (in 2001). The study excluded beneficiaries with end-stage renal disease (ESRD) in order to avoid underestimating complication rates for those with ESRD resulting from the full-year participation criterion, as people with both diabetes and ESRD have an estimated first-year mortality rate of 25%.20 Beneficiaries with ESRD accounted for 1.6% to 2.3% of Medicare beneficiaries with diabetes across 1992 to 2001. However, because delivering quality care to beneficiaries with diabetes is a goal for the Medicare program, the analysis retained those eligible through disability, particularly because they comprise a substantial portion of Medicare beneficiaries with diabetes (14% in 2001). The resulting sample sizes
ranged from 147,217 beneficiaries for 1992 and 224,055 for 2001.
Indicators The 15 quality indicators used in the analysis represent a spectrum of disease stages and include preventive services and short- and long-term complications associated with diabetes (see Appendix). To ensure their acceptability to a wide audience, the measures were selected by Diabetes Quality Improvement Project members, which included representatives from the CMS, American Diabetes Association, American Academy of Family Physicians, American College of Physicians, Centers for Disease Control and Prevention (CDC), National Institute of Diabetes, Digestive, and Kidney Diseases, and Veterans Health Administration.21 Quality indicators were based on their relevance in terms of diabetes care to Medicare beneficiaries, and the belief that administrative data accurately represent the process or outcome of care. Each measure was generated by using specific ICD-9, Current Procedural Terminology (CPT)/Healthcare Common Procedure Coding System (HCPSC), and Revenue Center codes derived from standard sources, such as Health Plan Employer Data and Information Set (HEDIS) eye examination codes, or provided by appropriate technical experts, such as the American Podiatric Medical Association for lowerextremity complications. Measures of preventive services were obtained from any claim in the 5% files, including inpatient, skilled nursing facility, outpatient, home health agency, Part B carrier (physician and laboratory), and durable medical equipment files. The codes for short-term complications were obtained from the primary diagnosis associated with inpatient admissions or ED visits for metabolic crises. Long-term complication measures were drawn from all claims in the 5% files unless the measure was specified as a hospitalization rate (derived from inpatient claims only).
Statistical Analysis The rates for the 15 quality measures express the proportion of Medicare beneficiaries with diabetes who had each procedure or diagnosis at least once in each year from 1992 to 2001. This paper analyzed the trends from 1992 to 2001 for each measure, and calculated 95% confidence intervals for each year. For scaling purposes, rates for preventive care practices are expressed per 100 Medicare beneficiaries with diabetes, and rates for complications are expressed per 1000 beneficiaries. Because not all trends were linear due to factors such as changes in diagnostic criteria and in reimbursement policies, the analysis compared 1992 rates to 2001 rates using two-tailed t -test statistics. The 2001 rates were also stratified by the following demographic and enrollment characteristics: age (0 to 64, 65 to 74, 75 to 84, and ⱖ85); race (white, black, and other [e.g., Hispanic, Asian, North American Native]); gender; and urban/rural residence (metropolitan statistical area [MSA] or non-MSA). All rates were age– gender standardized to the Medicare Part A FFS population in 1999 to control for variations in the age– gender distribution across the study period.22
Am J Prev Med 2005;29(5)
397
20
70 60 50 40 30
HbA1c test Lipid profiling Eye examination Self-monitoring
20 10 0
Per 1000 Medicare beneficiaries with diabetes
Per 100 Medicare beneficiaries with diabetes
80
15 10 5 0 1992 1993
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Results The number of Medicare beneficiaries with diagnosed diabetes (without ESRD) nationwide rose by 52% from 1992 through 2001, from an estimated 2,944,340 beneficiaries to 4,481,100. The estimated prevalence of diabetes among Medicare beneficiaries also increased over the period, from 9.2% in 1992 to 14.2% in 2001.
Trends in Diabetes Preventive Care Practices and Complications
Per 1000 Medicare beneficiaries with diabetes
The use of preventive care practices—including HbA1c tests, lipid tests, eye examinations, and selfmonitoring—increased significantly during the period, as shown in Figure 1. From 1992 to 2001, the rate of HbA1c testing rose by 44.7 percentage points, lipid profiling by 51.4, self-monitoring by 37.7, and eye examinations by 8.3 (all p ⬍0.05). Hospitalization rates associated with short-term complications from diabetes declined from 1992 to 2001, as shown in Figure 2. Rates of ED admissions for metabolic crisis and ketoacidosis dropped from 1992 to 2001, by 4.3 and 1.0 per 1000 beneficiaries, respectively (both p ⬍0.05). Most long-term microvascular complications, as illustrated in Figure 3, rose from 1992 to 2001, with the 15 12 9 6
Nephro pathy
1998 1999 2000 2001
Lo wer-limb amputatio n
B lindness
Figure 3. National trends in long-term microvascular complications, 1992–2001 (95% confidence intervals within ⫾0.62 per 1000 for all measures and years).
exception of lower-extremity amputations. Nephropathy rates increased overall by 10.3 per 1000 over the period, while blindness rates rose by 2.4 per 1000 (both p ⬍0.05). In addition, rates of diabetic retinopathy (not shown in Figure 3) increased consistently over the decade by a total of 41.0 per 1000 Medicare beneficiaries with diabetes (p ⬍0.05), reaching a prevalence of 141.1 per 1000 by 2001. However, lower-extremity amputation rates declined by 3.0 per 1000 (p ⬍0.05), or 26%, from 1992 to 2001. Figure 4 shows that the rate of hospitalizations for cardiovascular conditions—stroke, AMI, and heart failure— declined overall from 1992 to 2001. The net decrease over the period was 6.6 per 1000 for rate of stroke, 1.6 for AMI, and 4.7 for heart failure. Although not shown in Figure 4, the rate of coronary artery bypass graft (CABG) procedures decreased by 1.0 per 1000 beneficiaries (from 12.6 per 1000 in 1992 to 11.6 per 1000 in 2001), while that of percutaneous transluminal coronary angioplasties (PTCAs) increased by 10.5 per 1000 over the same period (from 11 per 1000 in 1992 to 21.5 per 1000 in 2001). (For all of the above statistics, p ⬍0.05.)
Per 1000 Medicare beneficiaries with diabetes
Figure 1. National trends in receipt of preventive services, 1992–2001 (95% confidence intervals within ⫾0.37 per 100 for all measures and years).
1994 1995 1996 1997
3
50 40 30 20 10 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Hospitalization f or stroke
0
Hospitalization f or heart f ailure
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Hospitalization f or AMI
Ho spitalizatio n fo r keto acido sis Emergency department admissio n fo r metabo lic crisis
Figure 2. National trends in short-term complications, 1992– 2001 (95% confidence intervals within ⫾0.55 per 1000 for all measures and years).
398
Figure 4. National trends in long-term macrovascular complications, 1992–2001 (95% confidence intervals within ⫾1.12 per 1000 for all measures and years). AMI, acute myocardial infarction.
American Journal of Preventive Medicine, Volume 29, Number 5
Rates for Subgroups Table 1 shows the 2001 rates for each of the measures overall and by demographic subgroups age, gender, and race/ethnicity. The analysis focused on the demographic characteristics because of their importance from a public health and clinical perspective. There were no systematic patterns in rates by urban/rural residency. Preventive care practices. Rates of preventive care practices—HbA1c tests, lipid profiling, eye examinations, and self-monitoring—were lowest among the oldest Medicare beneficiaries in 2001 (p ⬍0.05). White Medicare beneficiaries evidenced higher rates of preventive service receipt than nonwhites. The largest differences by race/ethnicity were in HbA1c tests and eye examinations, with blacks at least 7.6 percentage points and those of other race at least 6.4 less likely than whites to have received these services (p ⬍0.05). Short-term complications. In 2001, the rates of ketoacidosis and ED admissions were at least 8.0 and 2.7 times higher, respectively, for Medicare beneficiaries aged ⬍65 (and disabled) than for those aged ⱖ65 (p ⬍0.05). Black Medicare beneficiaries were also more likely to experience short-term complications than whites or those of other races/ethnicities, roughly at levels twice those of whites (p ⬍0.05). Long-term microvascular complications. The most marked differences in nephropathy rates occurred among Medicare beneficiaries aged ⬍65 compared with those aged ⱖ65. The group aged ⬍65 experienced nephropathy at a rate of 31.9 per 1000 in 2001 compared with the next highest rate of 18.9 per 1000 among those aged 65 to 74. Other than this case, the largest disparities in long-term microvascular complication rates occurred by race/ethnicity. Blacks accounted for the highest rates of lower-extremity amputation and blindness by at least 3.6 per 1000 compared with whites and those of other races/ethnicities. Although blacks also had significantly higher rates of nephropathy than whites, those of other races/ethnicities experienced the highest rates at 28.7 per 1000 beneficiaries in 2001. Male beneficiaries had lower-extremity amputation rates almost twice as high as those for females, while females were more likely to have experienced eye complications, such as blindness and retinopathy. (For all of the above statistics, p ⬍0.05.) Long-term macrovascular complications. The rates of hospitalizations for stroke, AMI, and heart failure increased with age, almost doubling from the youngest beneficiaries (aged ⬍65) to the oldest (aged ⬎85). Rates of AMI were greater by 4.9 per 1000 for male Medicare beneficiaries than for females (p ⬍0.05), while rates for stroke were 2 per 1000 greater (p ⬍0.05). The picture for cardiovascular complica-
tions by race/ethnicity was mixed. Although nonwhites and blacks in particular had higher rates of stroke and heart failure, the rate for AMI was highest among whites (p ⬍0.05). In addition, whites underwent CABG and PTCA more frequently (not shown in Table 1).
Discussion This analysis demonstrates significant improvement in the rate of preventive care practices. In addition, rates of both short-term and some of the most serious long-term complications have declined. Yet, this progress toward improving the overall quality of diabetes care is tempered by rising rates of other complications. The findings show progress toward meeting Healthy People 2010 targets for HbA1c testing, lower-extremity amputations, and eye examinations.12 The estimated rate of HbA1c testing for Medicare beneficiaries with diabetes was 76% in 2001, exceeding the Healthy People 2010 goal of 50%. Moreover, given that cardiovascular disease is the primary cause of death among people with diabetes, it is encouraging to see rising rates of hyperlipidemia screening, as well as falling rates of AMI, heart failure, and stroke for most subgroups. A large increase in PTCA rates and a decline in CABG rates occurred despite the literature’s suggestion that CABG should be the preferred procedure for a significant portion of the population with diabetes.23,24 The decline in lower-extremity amputation rates was another positive finding, particularly in light of the procedure’s associated health burdens and greater prevalence among Medicare beneficiaries with, rather than without, diabetes.25 Yet, despite improvement since 1992, estimated rates of lower-extremity amputation and eye examinations continue to fall short of target levels (8.7 vs 1.8 per 1000 and 57 vs 75%, respectively). The rate of eye exams also grew more slowly than those of other preventive services, and may be attributable to the comparatively high receipt of eye exams in 1992. The 1992 rate may be very close to the achievable ceiling rate within the control of the primary provider of diabetes care than other preventive services. At the same time, diabetes is the most common cause of kidney failure in the United States,26 and although the analysis excludes beneficiaries with ESRD, the rate of nephropathy, a precursor to ESRD, increased by 114% from 1992 to 2001. In addition, rates of blindness and diabetic retinopathy (a leading cause of acquired blindness) increased over the same period. The results also suggest possible disparities in care for vulnerable populations relative to other Medicare beneficiaries. In particular, nonwhite beneficiaries and beneficiaries aged ⬍65 and ⬎85 experienced higher rates of diabetic complications in 2001 compared with other beneficiaries. These vulnerable subgroups also tend to evidence lower relative rates of preventive Am J Prev Med 2005;29(5)
399
400 American Journal of Preventive Medicine, Volume 29, Number 5
Table 1. Age– gender adjusted 2001 rates among Medicare beneficiaries with diabetes by demographic subgroup Age (in years) Overall (nⴝ224,055)
<65 (nⴝ31,195)
65–74 (nⴝ96,462)
75–84 (nⴝ76,181)
Gender >85 (nⴝ20,217)
Preventive service receipt (per 1000 Medicare beneficiaries with diabetes) HbA1c test 76.0 71.3* 79.6* 76.7* 67.5* Lipid profiling 61.7 58.7* 69.3* 61.0* 39.7* Eye examination 57.3 43.1* 59.1* 63.4* 56.1* Self-monitoring 47.1 48.6 49.4 46.7* 38.2* Short-term complications (per 1000 Medicare beneficiaries with diabetes) e Hospitalization for 2.1 8.1* 1.0 0.9 ketoacidosis ED admission for 5.0 13.2* 2.9* 3.6* 4.8* metabolic crisis Long-term microvascular complications (per 1000 Medicare beneficiaries with diabetes) Nephropathy 19.3 31.9* 18.9* 16.2 12.8 Lower limb amputation 8.7 11.4a 7.4b 8.2c 10.3d Blindness 14.5 16.2 9.8* 15.9 25.7* Long-term macrovascular complications (per 1000 Medicare beneficiaries with diabetes) Hospitalization for stroke 27.2 17.0* 23.2* 33.6* 39.6* Hospitalization for AMI 18.6 13.5* 16.8* 20.6* 27.3* Hospitalization for heart 39.8 34.6* 31.4* 46.3* 63.0* failure
Race
Male (nⴝ906,953)
Female (nⴝ127,102)
White (nⴝ179,858)
Black (nⴝ30,722)
Other (nⴝ13,227)
75.5* 62.8* 55.3* 44.2*
76.6* 61.2* 59.3* 49.6*
77.6* 63.2* 58.6* 47.6
69.2 52.3* 51.0 47.1
68.3 61.8* 52.2 40.7*
2.0
2.0
1.9
3.4*
2.1
4.7
5.1
4.5
8.5*
4.8
20.8* 11.2* 13.7*
18.4* 6.6* 14.9*
17.9* 7.9 13.9
24.2* 13.9* 18.5*
28.7* 8.2 14.9
28.3* 21.4* 39.1
26.3* 16.5* 40.4
26.7 19.3* 39.0
30.0* 14.7 46.5*
25.6 15.9 36.6
Notes: Based on any such diagnosis from any of the 5% standard analytic files (inpatient, skilled nursing facility, outpatient, Part B carriers, home health agencies, and durable medical equipment) among full-year fee-for-service Medicare diabetic claimants. Exceptions are short-term complications and cardiovascular complications that are based on inpatient files only. Medicare claimants with diabetes are defined as those with a principal or secondary diagnosis of diabetes (not including gestational diabetes), with at least one inpatient encounter or at least two outpatient encounters. Beneficiaries who died during the year were excluded. All results reported are standardized to the age-gender distribution of the July 1999 Medicare fee-for-service population. Source: Centers for Medicare & Medicaid Services 5% Standard Analytic and Denominator files. a p⬍0.05 for difference between age groups (⬍65 and 65–74) and (⬍65 and 75– 84). b p⬍0.05 for difference between age groups (65–74 and ⬍65) and (65–74 and ⱖ85). c p⬍0.05 for difference between age groups (75– 84 and ⬍65) and (75– 84 and ⱖ85). d p⬍0.05 for difference between age groups (ⱖ85 and 65–74) and (ⱖ85 and 75– 84). e Results suppressed due to small sample size (n⬍25). *p⬍0.05 for difference between subgroup and all other subgroup(s) with subgroup category (bolded). AMI, acute myocardial infarction; ED, emergency department.
What This Study Adds . . . Diabetes is a major source of morbidity and healthcare costs in the United States, particularly for the elderly. As part of the Medicare Quality Monitoring System, this study describes trends from 1992 through 2001 in care practices and complications associated with diabetes for Medicare fee-forservice beneficiaries with diabetes. Although there have been improvements in many areas, increases in rates of some complications and evidence of subgroup disparities suggest potential areas for quality improvement.
services. For example, compared with whites, blacks had higher rates of stroke and heart failure and lower rates of lipid profiling in 2001. Although the analysis cannot attribute any causal links between preventive service use and the presence of complications, the analysis is consistent with other literature that has found racial/ethnic disparities in the quality and intensity of care, particularly in cardiovascular procedures and in poorer glucose control.27,28 The findings of racial disparities are particularly illuminating because the differences cannot be attributable to the presence or lack of insurance. The findings reported here should be interpreted critically, as the analysis had several limitations. First, although the indicators capture preventive service use and the range of complications associated with diabetes, causal links cannot be drawn across the two. The results stem from cross-sectional analyses rather than from following the same Medicare beneficiaries over time. Second, analyses based on administrative claims are sensitive to changes in billing and coding practices or practice patterns, which may undermine the accuracy of the estimates. Third, although diabetes is strongly associated with kidney failure in the United States, beneficiaries with ESRD were excluded to avoid misrepresenting their results (just as beneficiaries who died during the year were omitted). Therefore, it is likely that the complication rates reported here underestimate the rates for all Medicare beneficiaries with diabetes. Fourth, the analysis focused only on Medicare FFS beneficiaries, thus missing those covered by Medicare managed care. Because fluctuations over the past decade in Medicare managed-care plan characteristics and participation likely influenced the risk composition of those who chose managed care,29 it is difficult to speculate on the impact that including Medicare managed-care beneficiaries would have had on the rates. Fifth, rising complication rates may reflect better detection rather than increasing prevalence, particularly as preventive screenings became more
common over the study period. Sixth, preventive service rates are likely overstated, since Medicare beneficiaries with diabetes without a claim mentioning diabetes are not included in the analysis. However, the impact of their exclusion is likely to be small since persons with diabetes are frequent users of medical care. Finally, there have been changes in the denominator that may also affect our results. For example, the criteria for a diabetes diagnosis expanded in 1997,30 and better detection of diabetes may have contributed to more beneficiaries being diagnosed with diabetes who were less severely ill than in previous years. On the other hand, other factors such as improvements in comorbidities like cardiovascular health could work in the opposite direction by enabling some beneficiaries with diabetes who would have otherwise died to develop macrovascular complications to diabetes. The results point to clear successes in improving the quality of diabetes care for Medicare beneficiaries with diabetes. In particular, the study period saw across-theboard increases in preventive care practices, ranging from 8.3 percentage points for eye examinations to 51.4 percentage points for lipid profiling. The increase in lipid profiling for these beneficiaries with diabetes also appear to exceed that suggested by national estimates for the U.S. population aged ⬎65.31 The substantial rise in preventive care practices could presage improved outcomes in the future. The analysis found reductions in the rates of short-term complications as well as in some long-term complications such as lowerlimb amputations and cardiovascular conditions. Further, these reductions in cardiovascular conditions occurred during a period of increasing national rates.32 However, rates of other long-term complications, such as nephropathy and retinopathy, rose over time. Left untreated, these conditions may develop into more serious outcomes, such as ESRD or blindness. Further, there is evidence of persistent disparities both in complications and preventive care practices by race/ethnicity (nonwhite versus white) and age. These results highlight the importance of the MQMS surveillance system in providing early identification of potential areas for quality improvement. Despite important advances in the treatment of diabetes in a number of areas, increased rates of certain types of complications and continuing disparities for vulnerable subgroups are major concerns that will need to be addressed by policymakers, researchers, and practitioners. We are grateful to Hoangmai Pham, Tim Lake, Angela Merrill, and Myles Maxfield for their comments on earlier drafts, and to Nazmul Khan for programming assistance. The Centers for Medicare & Medicaid Services provided all funding for this study. The opinions herein are the authors’ and not necessarily those of Brown Medical School, the Centers for Medicare & Medicaid Services, the Centers for Disease Control and
Am J Prev Med 2005;29(5)
401
Prevention, Mathematica Policy Research, or the Veterans Health Administration. No financial conflict of interest was reported by the authors of this paper.
References 1. American Diabetes Association. Economic costs of diabetes in the U.S. in 2002. Diabetes Care 2003;26:917–32. 2. American Diabetes Association. Economic consequences of diabetes mellitus in the U.S. in 1997. Diabetes Care 1998;21:296 –309. 3. Centers for Disease Control and Prevention. 1999 Diabetes surveillance report. Atlanta GA: Centers for Disease Control and Prevention, 1999. 4. Mokdad AH, Ford ES, Bowman BA, et al. Diabetes trends in the U.S.: 1990 –1998. Diabetes Care 2000;23:1278 – 83. 5. Krop JS, Saudek CD, Weller WE, Powe NR, Shaffer T, Anderson GF. Predicting expenditures for Medicare beneficiaries with diabetes: a prospective cohort study from 1994 to 1996. Diabetes Care 1999;22:1660 – 6. 6. Weinberger M, Cowper PA, Kirkman MS, Vinicor F. Economic impact of diabetes mellitus in the elderly. Clin Geriatr Med 1990;6:959 –70. 7. Jencks SF, Huff ED, Cuerdon T. Change in the quality of care delivered to Medicare beneficiaries, 1998 –1999 to 2000 –2001. JAMA 2003;289:305–12. 8. Leid TR, Sheingold S. HEDIS performance measures in Medicare managed care. Health Care Finance Rev 2001;23:149 – 60. 9. Jha AK, Perlin JB, Kizer KW, Dudley RA. Effect of transformation of the Veterans Affairs health care system on the quality of care. N Engl J Med 2003;348:2218 –27. 10. Lojo M, Burrows NR, Geiss LS, Tierney EF, Wang J, Engelgau MM. Preventive-care practices among persons with diabetes—United States, 1995 and 2001. MMWR Morb Mortal Wkly Rep 2002;51:965–9. 11. Imperatore G, Cadwell BL, Geiss L, et al. Thirty-year trends in cardiovascular risk factor levels among US adults with diabetes. Am J Epidemiol 2004;160:531–9. 12. U.S. Department of Health and Human Services. Healthy people 2010. Understanding and improving health and objectives for improving health. 2nd ed. Washington DC: Government Printing Office, 2000. 13. Bertoni AG, Krop JS, Anderson GF, Brancati FL. Diabetes-related morbidity and mortality in a national sample of U.S. elders. Diabetes Care 2002; 25:471–5. 14. Wetterhall SF, Olson DR, DeStefano F, et al. Trends in diabetes and diabetic complications: 1980 –1987. Diabetes Care 1992;15:960 –7. 15. Bertoni AG, Hundley WG, Massing MW, Bonds DE, Burke GL, Goff DC Jr. Heart failure prevalence, incidence and mortality in the elderly with diabetes. Diabetes Care 2004;27:699 –703. 16. Centers for Disease Control and Prevention. Diabetes Surveillance System. Available at: www.cdc.gov/diabetes/statistics/index.htm. Accessed March 24, 2004.
402
17. Arday DR, Fleming BB, Keller DK et al. Variation in diabetes care among states: do patient characteristics matter? Diabetes Care 2002;25:2230 –7. 18. Krop JS, Power NR, Weller WE, Shaffer TJ, Saudek CD, Anderson GF. Patterns of expenditures and use of services among older adults with diabetes: implications for the transition to capitated managed care. Diabetes Care 1998;21:747–52. 19. Hebert PL, Geiss LS, Tierney EF, Engelgau MM, Yawn BP, McBean AM. Identifying persons with diabetes using Medicare claims data. Am J Med Qual 1999;14:270 –7. 20. National Institute of Diabetes and Digestive and Kidney Diseases. U.S. Renal Data System (USRDS). USRDS 2003 annual data report: atlas of end-stage renal disease in the United States. Bethesda MD: National Institutes of Health, 2003. 21. Fleming BB, Greenfield S, Engelgau MM, Pogach LM, Clauser SB, Parrott MA. The Diabetes Quality Improvement Project: moving science into health policy to gain an edge on the diabetes epidemic. Diabetes Care 2001;24:1815–20. 22. Anderson RN, Rosenberg HM. Age standardization of death rates: implementation of the year 2000 standard. Natl Vital Stat Rep 1998;47:1–16. 23. Ferguson JJ. NHLBI BARI clinical alert on diabetics treated with angioplasty. Circulation 1995;92:3371. 24. National Heart, Lung and Blood Institute. Seven-year outcome in the bypass angioplasty revascularization investigation (BARI) by treatment and diabetic status. J Am Coll Cardiol 2000;35:1122–9. 25. Wrobel JS, Mayfield JA, Reiber GE. Geographic variation in lower-extremity major amputation in individuals with and without diabetes in the Medicare population. Diabetes Care 2002;24:860 – 4. 26. National Institute of Diabetes, Digestive, and Kidney Diseases. Kidney disease of diabetes. Available at: http://kidney.niddk.nih.gov/kudiseases/ pubs/kdd. Accessed September 3, 2003. 27. Institute of Medicine. Unequal treatment: confronting racial and ethnic disparities in health care. Washington DC: National Academy Press, 2002. 28. McBean AM, Huang Z, Virnig BA, Lurie N, Musgrave D. Racial variation in the control of diabetes among elderly Medicare managed care beneficiaries. Diabetes Care 2003;26:3250 – 6. 29. Gold M, Achman L, Mittler J, Stevens B. Monitoring Medicare⫹Choice: what have we learned? Findings and operational lessons for Medicare Advantage. Washington DC: Mathematica Policy Research, August 2004. 30. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997;20:1183–97. 31. Centers for Disease Control and Prevention. No cholesterol check within 5 years. Available at: http://apps.nccd.cdc.gov/brfss/Trends/agechart.asp? qkey⫽10090&state⫽US. Accessed February 18, 2005. 32. Agency for Healthcare Research and Quality. HCUPnet, Healthcare Cost and Utilization Project. Rockville MD: Agency for Healthcare Research and Quality. Available at: www.ahrq.gov/HCUPnet/. Accessed February 20, 2005.
American Journal of Preventive Medicine, Volume 29, Number 5
Appendix Diabetes measures and code definitions Measure Preventive measures HbA1c test Lipid profiling Eye examination
Self-monitoring Short-term complications Hospitalization for diabetes with ketoacidosis Admission to an emergency department or observation bed for metabolic crisis Long-term complications Nephropathy Lower limb amputation Diabetic retinopathy Blindness Hospitalization for primary diagnosis of stroke Hospitalization for primary diagnosis of acute myocardial infarction Hospitalization for primary diagnosis of heart failure Hospitalization for coronary artery bypass graft Progression to percutaneous transluminal coronary angioplasty (including stents)
ICD-9, CPT/HCPCS codes 83036 80061 OR (82465, 83718, 84478 in the same day) 67101, 67105, 67107, 67108, 67110, 67112, 67141, 67145, 67208, 67210, 67218, 67227, 67228, 92002, 92004, 92012, 92014, 92018, 92019, 92225, 92226, 92230, 92235, 92240, 92250, 92260 OR provider specialty codes 18, 41 and (90000–90080, 99201, 99202, 99203, 99204, 99205, 99211, 99212, 99213, 99214, 99215, 90640–90643, 99241, 99242, 99243, 99244, 99245) 82962, A4253, E0607 250.1x (251.0x, 250.1x, 250.2x, 250.3x) AND (045x, 0762, 0981)
581.81, 583.81 84.10–84.17, 28800, 28805, 28810, 28820, 28825, 27880, 27881, 27882, 27886, 27888, 27591, 27592, 27594, 27596, 27598 362.0x 369.xx 362.34, 433.0x, 433.1x, 433.2x, 433.3x, 433.8x, 433.9x, 434.0x, 434.1x, 434.9x, 435.0, 435.1, 435.3, 435.8, 435.9, 436 Additional codes for 1992: 433.0, 433.1, 433.2, 433.3, 433.8, 433.9, 434.0, 434.1, 434.9, 435.2 410.00, 410.01, 410.10, 410.11, 410.20, 410.21, 410.30, 410.31, 410.40, 410.41, 410.50, 410.51, 410.60, 410.61, 410.70, 410.71, 410.80, 410.81, 410.90, 410.91 402.01, 402.11, 402.91, 404.01, 404.11, 404.91, 428.0, 428.1, 428.9 36.1x 36.01, 36.02, 36.05, 36.06, 92980, 92981, 92982, 92984
Am J Prev Med 2005;29(5)
403
Outcomes include age- and gender-adjusted rates of 15 indicators associated with diabetes care from 1992 to 2001, the absolute change in rates from 1992 to 2001, and 2001 rates by demographic subgroups. The data were cross-sectional samples of Medicare beneficiaries with diabetes from 1992 to 2001 from the Medicare 5% Standard Analytic Files.
Results:
Use of preventive care practices rose from 1992 to 2001: 45 percentage points for HbA1c tests, 51 for lipid tests, 8 for eye exams, and 38 for self-monitoring of glucose levels (all p ⬍0.05). Rates for short-term and some long-term complications of diabetes (e.g., lower-extremity amputations and cardiovascular conditions) fell from 1992 to 2001 (p ⬍0.05). However, rates of other long-term complications such as nephropathy, blindness, and retinopathy rose during the period (p ⬍0.05). Nonwhites and beneficiaries aged ⬍65 and ⬎85 exhibited consistently higher complication rates and lower use of preventive services.
Conclusions: The Medicare program has seen some significant improvement in preventive care practices and significant declines in lower-limb amputations and cardiovascular conditions. However, rates for other long-term complications have increased, with evidence of subgroup disparities. The MQMS results provide an early warning for policymakers to focus on the diabetes care provided to some vulnerable subgroups. (Am J Prev Med 2005;29(5):396 – 403) © 2005 American Journal of Preventive Medicine
Introduction
T
he economic cost of diabetes increased from $98 billion in 1997 to $132 billion in 2002,1,2 and has continued to increase in subsequent years. The disease has also become more prevalent. Both developments have made the tracking of complications and services received by individuals with diabetes imperative for informed policymaking. Understanding these trends is especially relevant to Medicare because the Medicare population bears a disproportionate burden of the disease in both preva-
From the Center for Gerontology and Health Care Research, Brown University (Kuo), Providence, Rhode Island; Veterans Health Administration (Fleming), Washington DC; Centers for Medicare & Medicaid Services (Gittings, Han, Roman), Baltimore, Maryland; and Centers for Disease Control and Prevention (Geiss, Engelgau), Atlanta, Georgia When the analysis was conducted, Kuo and Fleming were affiliated with Mathematica Policy Research, and the Centers for Medicare & Medicaid Services, respectively. Address correspondence and reprint requests to: Sylvia Kuo, PhD, Center for Gerontology and Health Care Research, Brown University, Box G-ST210, Providence RI 02912. E-mail: [email protected].
396
lence and cost. An estimated 20% of elderly individuals has diabetes compared with 5% of the general population.3,4 Elderly people living with diabetes incur healthcare expenditures ⱖ50% higher than those without the disease.5,6 Much work has shown recent improvements in the quality of diabetes care for Medicare beneficiaries (fee for service [FFS]7 and managed care8) as well as other groups,9 –11 although further improvement is necessary to achieve targets outlined in Healthy People 2010.12 These quality indicators emphasize types of preventive care, such as annual HbA1c testing or biennial dilated eye examinations, that may reduce the incidence or delay the onset of complications associated with diabetes. However, these studies provide only a partial picture of the burden of diabetes since they do not analyze trends in the complications that contribute to the high cost of the disease. Other work specific to the Medicare population has focused on late-stage outcomes by concentrating on death certificates, hospital discharges, and excess diabetes-associated mortality among the elderly.13–15 Al-
Am J Prev Med 2005;29(5) © 2005 American Journal of Preventive Medicine • Published by Elsevier Inc.
0749-3797/05/$–see front matter doi:10.1016/j.amepre.2005.08.010
though not focused on the Medicare population, the Diabetes Surveillance System16 provides self-reported preventive care and complication rates for the general population of individuals with diabetes, as well as by different subgroups (including age). Consequently, little has been done to assemble an overall picture of trends in diabetes care for Medicare beneficiaries based on indicators such as preventive services and the broad array of conditions complicating the disease. The purpose of this article is to develop such a picture by analyzing rates of preventive care practices and acute- and long-term complications among Medicare beneficiaries with diabetes over a 10-year period. The article also examines whether there is evidence that quality of care has improved, and whether racial or other subgroup differences exist. Results of the analysis will provide policymakers, researchers, and practitioners with information on trends in the quality of diabetes care and on potential areas of concern—such as specific complications or subgroup conditions and care disparities—where resources may be most beneficially directed.
Methods Study Sample The study sample comes from the Medicare Quality Monitoring System (MQMS), which the Centers for Medicare & Medicaid Services (CMS) maintains to monitor the quality of care delivered to Medicare FFS beneficiaries. MQMS uses Medicare administrative data to track national and state trends, patterns, and variations in the use of health care and outcomes of that care, preventable hospitalizations, and patient safety. Diabetes is among the five conditions (including acute myocardial infarction [AMI], heart failure, stroke, and pneumonia) common to the Medicare population specifically targeted by MQMS. The sample data represent random 5% annual cross-sectional samples of Medicare FFS beneficiaries enrolled for the full year from 1992 to 2001. Using an algorithm validated in earlier studies,17–19 the analysis, which was performed in 2004, defined beneficiaries with diabetes as those who had either one inpatient or emergency department (ED) visit or two ambulatory encounters coded for diabetes during the year (excluding individuals with gestational diabetes). The process yields annual sample sizes of 149,652 (in 1992) to 229,350 (in 2001). The study excluded beneficiaries with end-stage renal disease (ESRD) in order to avoid underestimating complication rates for those with ESRD resulting from the full-year participation criterion, as people with both diabetes and ESRD have an estimated first-year mortality rate of 25%.20 Beneficiaries with ESRD accounted for 1.6% to 2.3% of Medicare beneficiaries with diabetes across 1992 to 2001. However, because delivering quality care to beneficiaries with diabetes is a goal for the Medicare program, the analysis retained those eligible through disability, particularly because they comprise a substantial portion of Medicare beneficiaries with diabetes (14% in 2001). The resulting sample sizes
ranged from 147,217 beneficiaries for 1992 and 224,055 for 2001.
Indicators The 15 quality indicators used in the analysis represent a spectrum of disease stages and include preventive services and short- and long-term complications associated with diabetes (see Appendix). To ensure their acceptability to a wide audience, the measures were selected by Diabetes Quality Improvement Project members, which included representatives from the CMS, American Diabetes Association, American Academy of Family Physicians, American College of Physicians, Centers for Disease Control and Prevention (CDC), National Institute of Diabetes, Digestive, and Kidney Diseases, and Veterans Health Administration.21 Quality indicators were based on their relevance in terms of diabetes care to Medicare beneficiaries, and the belief that administrative data accurately represent the process or outcome of care. Each measure was generated by using specific ICD-9, Current Procedural Terminology (CPT)/Healthcare Common Procedure Coding System (HCPSC), and Revenue Center codes derived from standard sources, such as Health Plan Employer Data and Information Set (HEDIS) eye examination codes, or provided by appropriate technical experts, such as the American Podiatric Medical Association for lowerextremity complications. Measures of preventive services were obtained from any claim in the 5% files, including inpatient, skilled nursing facility, outpatient, home health agency, Part B carrier (physician and laboratory), and durable medical equipment files. The codes for short-term complications were obtained from the primary diagnosis associated with inpatient admissions or ED visits for metabolic crises. Long-term complication measures were drawn from all claims in the 5% files unless the measure was specified as a hospitalization rate (derived from inpatient claims only).
Statistical Analysis The rates for the 15 quality measures express the proportion of Medicare beneficiaries with diabetes who had each procedure or diagnosis at least once in each year from 1992 to 2001. This paper analyzed the trends from 1992 to 2001 for each measure, and calculated 95% confidence intervals for each year. For scaling purposes, rates for preventive care practices are expressed per 100 Medicare beneficiaries with diabetes, and rates for complications are expressed per 1000 beneficiaries. Because not all trends were linear due to factors such as changes in diagnostic criteria and in reimbursement policies, the analysis compared 1992 rates to 2001 rates using two-tailed t -test statistics. The 2001 rates were also stratified by the following demographic and enrollment characteristics: age (0 to 64, 65 to 74, 75 to 84, and ⱖ85); race (white, black, and other [e.g., Hispanic, Asian, North American Native]); gender; and urban/rural residence (metropolitan statistical area [MSA] or non-MSA). All rates were age– gender standardized to the Medicare Part A FFS population in 1999 to control for variations in the age– gender distribution across the study period.22
Am J Prev Med 2005;29(5)
397
20
70 60 50 40 30
HbA1c test Lipid profiling Eye examination Self-monitoring
20 10 0
Per 1000 Medicare beneficiaries with diabetes
Per 100 Medicare beneficiaries with diabetes
80
15 10 5 0 1992 1993
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Results The number of Medicare beneficiaries with diagnosed diabetes (without ESRD) nationwide rose by 52% from 1992 through 2001, from an estimated 2,944,340 beneficiaries to 4,481,100. The estimated prevalence of diabetes among Medicare beneficiaries also increased over the period, from 9.2% in 1992 to 14.2% in 2001.
Trends in Diabetes Preventive Care Practices and Complications
Per 1000 Medicare beneficiaries with diabetes
The use of preventive care practices—including HbA1c tests, lipid tests, eye examinations, and selfmonitoring—increased significantly during the period, as shown in Figure 1. From 1992 to 2001, the rate of HbA1c testing rose by 44.7 percentage points, lipid profiling by 51.4, self-monitoring by 37.7, and eye examinations by 8.3 (all p ⬍0.05). Hospitalization rates associated with short-term complications from diabetes declined from 1992 to 2001, as shown in Figure 2. Rates of ED admissions for metabolic crisis and ketoacidosis dropped from 1992 to 2001, by 4.3 and 1.0 per 1000 beneficiaries, respectively (both p ⬍0.05). Most long-term microvascular complications, as illustrated in Figure 3, rose from 1992 to 2001, with the 15 12 9 6
Nephro pathy
1998 1999 2000 2001
Lo wer-limb amputatio n
B lindness
Figure 3. National trends in long-term microvascular complications, 1992–2001 (95% confidence intervals within ⫾0.62 per 1000 for all measures and years).
exception of lower-extremity amputations. Nephropathy rates increased overall by 10.3 per 1000 over the period, while blindness rates rose by 2.4 per 1000 (both p ⬍0.05). In addition, rates of diabetic retinopathy (not shown in Figure 3) increased consistently over the decade by a total of 41.0 per 1000 Medicare beneficiaries with diabetes (p ⬍0.05), reaching a prevalence of 141.1 per 1000 by 2001. However, lower-extremity amputation rates declined by 3.0 per 1000 (p ⬍0.05), or 26%, from 1992 to 2001. Figure 4 shows that the rate of hospitalizations for cardiovascular conditions—stroke, AMI, and heart failure— declined overall from 1992 to 2001. The net decrease over the period was 6.6 per 1000 for rate of stroke, 1.6 for AMI, and 4.7 for heart failure. Although not shown in Figure 4, the rate of coronary artery bypass graft (CABG) procedures decreased by 1.0 per 1000 beneficiaries (from 12.6 per 1000 in 1992 to 11.6 per 1000 in 2001), while that of percutaneous transluminal coronary angioplasties (PTCAs) increased by 10.5 per 1000 over the same period (from 11 per 1000 in 1992 to 21.5 per 1000 in 2001). (For all of the above statistics, p ⬍0.05.)
Per 1000 Medicare beneficiaries with diabetes
Figure 1. National trends in receipt of preventive services, 1992–2001 (95% confidence intervals within ⫾0.37 per 100 for all measures and years).
1994 1995 1996 1997
3
50 40 30 20 10 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Hospitalization f or stroke
0
Hospitalization f or heart f ailure
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001
Hospitalization f or AMI
Ho spitalizatio n fo r keto acido sis Emergency department admissio n fo r metabo lic crisis
Figure 2. National trends in short-term complications, 1992– 2001 (95% confidence intervals within ⫾0.55 per 1000 for all measures and years).
398
Figure 4. National trends in long-term macrovascular complications, 1992–2001 (95% confidence intervals within ⫾1.12 per 1000 for all measures and years). AMI, acute myocardial infarction.
American Journal of Preventive Medicine, Volume 29, Number 5
Rates for Subgroups Table 1 shows the 2001 rates for each of the measures overall and by demographic subgroups age, gender, and race/ethnicity. The analysis focused on the demographic characteristics because of their importance from a public health and clinical perspective. There were no systematic patterns in rates by urban/rural residency. Preventive care practices. Rates of preventive care practices—HbA1c tests, lipid profiling, eye examinations, and self-monitoring—were lowest among the oldest Medicare beneficiaries in 2001 (p ⬍0.05). White Medicare beneficiaries evidenced higher rates of preventive service receipt than nonwhites. The largest differences by race/ethnicity were in HbA1c tests and eye examinations, with blacks at least 7.6 percentage points and those of other race at least 6.4 less likely than whites to have received these services (p ⬍0.05). Short-term complications. In 2001, the rates of ketoacidosis and ED admissions were at least 8.0 and 2.7 times higher, respectively, for Medicare beneficiaries aged ⬍65 (and disabled) than for those aged ⱖ65 (p ⬍0.05). Black Medicare beneficiaries were also more likely to experience short-term complications than whites or those of other races/ethnicities, roughly at levels twice those of whites (p ⬍0.05). Long-term microvascular complications. The most marked differences in nephropathy rates occurred among Medicare beneficiaries aged ⬍65 compared with those aged ⱖ65. The group aged ⬍65 experienced nephropathy at a rate of 31.9 per 1000 in 2001 compared with the next highest rate of 18.9 per 1000 among those aged 65 to 74. Other than this case, the largest disparities in long-term microvascular complication rates occurred by race/ethnicity. Blacks accounted for the highest rates of lower-extremity amputation and blindness by at least 3.6 per 1000 compared with whites and those of other races/ethnicities. Although blacks also had significantly higher rates of nephropathy than whites, those of other races/ethnicities experienced the highest rates at 28.7 per 1000 beneficiaries in 2001. Male beneficiaries had lower-extremity amputation rates almost twice as high as those for females, while females were more likely to have experienced eye complications, such as blindness and retinopathy. (For all of the above statistics, p ⬍0.05.) Long-term macrovascular complications. The rates of hospitalizations for stroke, AMI, and heart failure increased with age, almost doubling from the youngest beneficiaries (aged ⬍65) to the oldest (aged ⬎85). Rates of AMI were greater by 4.9 per 1000 for male Medicare beneficiaries than for females (p ⬍0.05), while rates for stroke were 2 per 1000 greater (p ⬍0.05). The picture for cardiovascular complica-
tions by race/ethnicity was mixed. Although nonwhites and blacks in particular had higher rates of stroke and heart failure, the rate for AMI was highest among whites (p ⬍0.05). In addition, whites underwent CABG and PTCA more frequently (not shown in Table 1).
Discussion This analysis demonstrates significant improvement in the rate of preventive care practices. In addition, rates of both short-term and some of the most serious long-term complications have declined. Yet, this progress toward improving the overall quality of diabetes care is tempered by rising rates of other complications. The findings show progress toward meeting Healthy People 2010 targets for HbA1c testing, lower-extremity amputations, and eye examinations.12 The estimated rate of HbA1c testing for Medicare beneficiaries with diabetes was 76% in 2001, exceeding the Healthy People 2010 goal of 50%. Moreover, given that cardiovascular disease is the primary cause of death among people with diabetes, it is encouraging to see rising rates of hyperlipidemia screening, as well as falling rates of AMI, heart failure, and stroke for most subgroups. A large increase in PTCA rates and a decline in CABG rates occurred despite the literature’s suggestion that CABG should be the preferred procedure for a significant portion of the population with diabetes.23,24 The decline in lower-extremity amputation rates was another positive finding, particularly in light of the procedure’s associated health burdens and greater prevalence among Medicare beneficiaries with, rather than without, diabetes.25 Yet, despite improvement since 1992, estimated rates of lower-extremity amputation and eye examinations continue to fall short of target levels (8.7 vs 1.8 per 1000 and 57 vs 75%, respectively). The rate of eye exams also grew more slowly than those of other preventive services, and may be attributable to the comparatively high receipt of eye exams in 1992. The 1992 rate may be very close to the achievable ceiling rate within the control of the primary provider of diabetes care than other preventive services. At the same time, diabetes is the most common cause of kidney failure in the United States,26 and although the analysis excludes beneficiaries with ESRD, the rate of nephropathy, a precursor to ESRD, increased by 114% from 1992 to 2001. In addition, rates of blindness and diabetic retinopathy (a leading cause of acquired blindness) increased over the same period. The results also suggest possible disparities in care for vulnerable populations relative to other Medicare beneficiaries. In particular, nonwhite beneficiaries and beneficiaries aged ⬍65 and ⬎85 experienced higher rates of diabetic complications in 2001 compared with other beneficiaries. These vulnerable subgroups also tend to evidence lower relative rates of preventive Am J Prev Med 2005;29(5)
399
400 American Journal of Preventive Medicine, Volume 29, Number 5
Table 1. Age– gender adjusted 2001 rates among Medicare beneficiaries with diabetes by demographic subgroup Age (in years) Overall (nⴝ224,055)
<65 (nⴝ31,195)
65–74 (nⴝ96,462)
75–84 (nⴝ76,181)
Gender >85 (nⴝ20,217)
Preventive service receipt (per 1000 Medicare beneficiaries with diabetes) HbA1c test 76.0 71.3* 79.6* 76.7* 67.5* Lipid profiling 61.7 58.7* 69.3* 61.0* 39.7* Eye examination 57.3 43.1* 59.1* 63.4* 56.1* Self-monitoring 47.1 48.6 49.4 46.7* 38.2* Short-term complications (per 1000 Medicare beneficiaries with diabetes) e Hospitalization for 2.1 8.1* 1.0 0.9 ketoacidosis ED admission for 5.0 13.2* 2.9* 3.6* 4.8* metabolic crisis Long-term microvascular complications (per 1000 Medicare beneficiaries with diabetes) Nephropathy 19.3 31.9* 18.9* 16.2 12.8 Lower limb amputation 8.7 11.4a 7.4b 8.2c 10.3d Blindness 14.5 16.2 9.8* 15.9 25.7* Long-term macrovascular complications (per 1000 Medicare beneficiaries with diabetes) Hospitalization for stroke 27.2 17.0* 23.2* 33.6* 39.6* Hospitalization for AMI 18.6 13.5* 16.8* 20.6* 27.3* Hospitalization for heart 39.8 34.6* 31.4* 46.3* 63.0* failure
Race
Male (nⴝ906,953)
Female (nⴝ127,102)
White (nⴝ179,858)
Black (nⴝ30,722)
Other (nⴝ13,227)
75.5* 62.8* 55.3* 44.2*
76.6* 61.2* 59.3* 49.6*
77.6* 63.2* 58.6* 47.6
69.2 52.3* 51.0 47.1
68.3 61.8* 52.2 40.7*
2.0
2.0
1.9
3.4*
2.1
4.7
5.1
4.5
8.5*
4.8
20.8* 11.2* 13.7*
18.4* 6.6* 14.9*
17.9* 7.9 13.9
24.2* 13.9* 18.5*
28.7* 8.2 14.9
28.3* 21.4* 39.1
26.3* 16.5* 40.4
26.7 19.3* 39.0
30.0* 14.7 46.5*
25.6 15.9 36.6
Notes: Based on any such diagnosis from any of the 5% standard analytic files (inpatient, skilled nursing facility, outpatient, Part B carriers, home health agencies, and durable medical equipment) among full-year fee-for-service Medicare diabetic claimants. Exceptions are short-term complications and cardiovascular complications that are based on inpatient files only. Medicare claimants with diabetes are defined as those with a principal or secondary diagnosis of diabetes (not including gestational diabetes), with at least one inpatient encounter or at least two outpatient encounters. Beneficiaries who died during the year were excluded. All results reported are standardized to the age-gender distribution of the July 1999 Medicare fee-for-service population. Source: Centers for Medicare & Medicaid Services 5% Standard Analytic and Denominator files. a p⬍0.05 for difference between age groups (⬍65 and 65–74) and (⬍65 and 75– 84). b p⬍0.05 for difference between age groups (65–74 and ⬍65) and (65–74 and ⱖ85). c p⬍0.05 for difference between age groups (75– 84 and ⬍65) and (75– 84 and ⱖ85). d p⬍0.05 for difference between age groups (ⱖ85 and 65–74) and (ⱖ85 and 75– 84). e Results suppressed due to small sample size (n⬍25). *p⬍0.05 for difference between subgroup and all other subgroup(s) with subgroup category (bolded). AMI, acute myocardial infarction; ED, emergency department.
What This Study Adds . . . Diabetes is a major source of morbidity and healthcare costs in the United States, particularly for the elderly. As part of the Medicare Quality Monitoring System, this study describes trends from 1992 through 2001 in care practices and complications associated with diabetes for Medicare fee-forservice beneficiaries with diabetes. Although there have been improvements in many areas, increases in rates of some complications and evidence of subgroup disparities suggest potential areas for quality improvement.
services. For example, compared with whites, blacks had higher rates of stroke and heart failure and lower rates of lipid profiling in 2001. Although the analysis cannot attribute any causal links between preventive service use and the presence of complications, the analysis is consistent with other literature that has found racial/ethnic disparities in the quality and intensity of care, particularly in cardiovascular procedures and in poorer glucose control.27,28 The findings of racial disparities are particularly illuminating because the differences cannot be attributable to the presence or lack of insurance. The findings reported here should be interpreted critically, as the analysis had several limitations. First, although the indicators capture preventive service use and the range of complications associated with diabetes, causal links cannot be drawn across the two. The results stem from cross-sectional analyses rather than from following the same Medicare beneficiaries over time. Second, analyses based on administrative claims are sensitive to changes in billing and coding practices or practice patterns, which may undermine the accuracy of the estimates. Third, although diabetes is strongly associated with kidney failure in the United States, beneficiaries with ESRD were excluded to avoid misrepresenting their results (just as beneficiaries who died during the year were omitted). Therefore, it is likely that the complication rates reported here underestimate the rates for all Medicare beneficiaries with diabetes. Fourth, the analysis focused only on Medicare FFS beneficiaries, thus missing those covered by Medicare managed care. Because fluctuations over the past decade in Medicare managed-care plan characteristics and participation likely influenced the risk composition of those who chose managed care,29 it is difficult to speculate on the impact that including Medicare managed-care beneficiaries would have had on the rates. Fifth, rising complication rates may reflect better detection rather than increasing prevalence, particularly as preventive screenings became more
common over the study period. Sixth, preventive service rates are likely overstated, since Medicare beneficiaries with diabetes without a claim mentioning diabetes are not included in the analysis. However, the impact of their exclusion is likely to be small since persons with diabetes are frequent users of medical care. Finally, there have been changes in the denominator that may also affect our results. For example, the criteria for a diabetes diagnosis expanded in 1997,30 and better detection of diabetes may have contributed to more beneficiaries being diagnosed with diabetes who were less severely ill than in previous years. On the other hand, other factors such as improvements in comorbidities like cardiovascular health could work in the opposite direction by enabling some beneficiaries with diabetes who would have otherwise died to develop macrovascular complications to diabetes. The results point to clear successes in improving the quality of diabetes care for Medicare beneficiaries with diabetes. In particular, the study period saw across-theboard increases in preventive care practices, ranging from 8.3 percentage points for eye examinations to 51.4 percentage points for lipid profiling. The increase in lipid profiling for these beneficiaries with diabetes also appear to exceed that suggested by national estimates for the U.S. population aged ⬎65.31 The substantial rise in preventive care practices could presage improved outcomes in the future. The analysis found reductions in the rates of short-term complications as well as in some long-term complications such as lowerlimb amputations and cardiovascular conditions. Further, these reductions in cardiovascular conditions occurred during a period of increasing national rates.32 However, rates of other long-term complications, such as nephropathy and retinopathy, rose over time. Left untreated, these conditions may develop into more serious outcomes, such as ESRD or blindness. Further, there is evidence of persistent disparities both in complications and preventive care practices by race/ethnicity (nonwhite versus white) and age. These results highlight the importance of the MQMS surveillance system in providing early identification of potential areas for quality improvement. Despite important advances in the treatment of diabetes in a number of areas, increased rates of certain types of complications and continuing disparities for vulnerable subgroups are major concerns that will need to be addressed by policymakers, researchers, and practitioners. We are grateful to Hoangmai Pham, Tim Lake, Angela Merrill, and Myles Maxfield for their comments on earlier drafts, and to Nazmul Khan for programming assistance. The Centers for Medicare & Medicaid Services provided all funding for this study. The opinions herein are the authors’ and not necessarily those of Brown Medical School, the Centers for Medicare & Medicaid Services, the Centers for Disease Control and
Am J Prev Med 2005;29(5)
401
Prevention, Mathematica Policy Research, or the Veterans Health Administration. No financial conflict of interest was reported by the authors of this paper.
References 1. American Diabetes Association. Economic costs of diabetes in the U.S. in 2002. Diabetes Care 2003;26:917–32. 2. American Diabetes Association. Economic consequences of diabetes mellitus in the U.S. in 1997. Diabetes Care 1998;21:296 –309. 3. Centers for Disease Control and Prevention. 1999 Diabetes surveillance report. Atlanta GA: Centers for Disease Control and Prevention, 1999. 4. Mokdad AH, Ford ES, Bowman BA, et al. Diabetes trends in the U.S.: 1990 –1998. Diabetes Care 2000;23:1278 – 83. 5. Krop JS, Saudek CD, Weller WE, Powe NR, Shaffer T, Anderson GF. Predicting expenditures for Medicare beneficiaries with diabetes: a prospective cohort study from 1994 to 1996. Diabetes Care 1999;22:1660 – 6. 6. Weinberger M, Cowper PA, Kirkman MS, Vinicor F. Economic impact of diabetes mellitus in the elderly. Clin Geriatr Med 1990;6:959 –70. 7. Jencks SF, Huff ED, Cuerdon T. Change in the quality of care delivered to Medicare beneficiaries, 1998 –1999 to 2000 –2001. JAMA 2003;289:305–12. 8. Leid TR, Sheingold S. HEDIS performance measures in Medicare managed care. Health Care Finance Rev 2001;23:149 – 60. 9. Jha AK, Perlin JB, Kizer KW, Dudley RA. Effect of transformation of the Veterans Affairs health care system on the quality of care. N Engl J Med 2003;348:2218 –27. 10. Lojo M, Burrows NR, Geiss LS, Tierney EF, Wang J, Engelgau MM. Preventive-care practices among persons with diabetes—United States, 1995 and 2001. MMWR Morb Mortal Wkly Rep 2002;51:965–9. 11. Imperatore G, Cadwell BL, Geiss L, et al. Thirty-year trends in cardiovascular risk factor levels among US adults with diabetes. Am J Epidemiol 2004;160:531–9. 12. U.S. Department of Health and Human Services. Healthy people 2010. Understanding and improving health and objectives for improving health. 2nd ed. Washington DC: Government Printing Office, 2000. 13. Bertoni AG, Krop JS, Anderson GF, Brancati FL. Diabetes-related morbidity and mortality in a national sample of U.S. elders. Diabetes Care 2002; 25:471–5. 14. Wetterhall SF, Olson DR, DeStefano F, et al. Trends in diabetes and diabetic complications: 1980 –1987. Diabetes Care 1992;15:960 –7. 15. Bertoni AG, Hundley WG, Massing MW, Bonds DE, Burke GL, Goff DC Jr. Heart failure prevalence, incidence and mortality in the elderly with diabetes. Diabetes Care 2004;27:699 –703. 16. Centers for Disease Control and Prevention. Diabetes Surveillance System. Available at: www.cdc.gov/diabetes/statistics/index.htm. Accessed March 24, 2004.
402
17. Arday DR, Fleming BB, Keller DK et al. Variation in diabetes care among states: do patient characteristics matter? Diabetes Care 2002;25:2230 –7. 18. Krop JS, Power NR, Weller WE, Shaffer TJ, Saudek CD, Anderson GF. Patterns of expenditures and use of services among older adults with diabetes: implications for the transition to capitated managed care. Diabetes Care 1998;21:747–52. 19. Hebert PL, Geiss LS, Tierney EF, Engelgau MM, Yawn BP, McBean AM. Identifying persons with diabetes using Medicare claims data. Am J Med Qual 1999;14:270 –7. 20. National Institute of Diabetes and Digestive and Kidney Diseases. U.S. Renal Data System (USRDS). USRDS 2003 annual data report: atlas of end-stage renal disease in the United States. Bethesda MD: National Institutes of Health, 2003. 21. Fleming BB, Greenfield S, Engelgau MM, Pogach LM, Clauser SB, Parrott MA. The Diabetes Quality Improvement Project: moving science into health policy to gain an edge on the diabetes epidemic. Diabetes Care 2001;24:1815–20. 22. Anderson RN, Rosenberg HM. Age standardization of death rates: implementation of the year 2000 standard. Natl Vital Stat Rep 1998;47:1–16. 23. Ferguson JJ. NHLBI BARI clinical alert on diabetics treated with angioplasty. Circulation 1995;92:3371. 24. National Heart, Lung and Blood Institute. Seven-year outcome in the bypass angioplasty revascularization investigation (BARI) by treatment and diabetic status. J Am Coll Cardiol 2000;35:1122–9. 25. Wrobel JS, Mayfield JA, Reiber GE. Geographic variation in lower-extremity major amputation in individuals with and without diabetes in the Medicare population. Diabetes Care 2002;24:860 – 4. 26. National Institute of Diabetes, Digestive, and Kidney Diseases. Kidney disease of diabetes. Available at: http://kidney.niddk.nih.gov/kudiseases/ pubs/kdd. Accessed September 3, 2003. 27. Institute of Medicine. Unequal treatment: confronting racial and ethnic disparities in health care. Washington DC: National Academy Press, 2002. 28. McBean AM, Huang Z, Virnig BA, Lurie N, Musgrave D. Racial variation in the control of diabetes among elderly Medicare managed care beneficiaries. Diabetes Care 2003;26:3250 – 6. 29. Gold M, Achman L, Mittler J, Stevens B. Monitoring Medicare⫹Choice: what have we learned? Findings and operational lessons for Medicare Advantage. Washington DC: Mathematica Policy Research, August 2004. 30. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997;20:1183–97. 31. Centers for Disease Control and Prevention. No cholesterol check within 5 years. Available at: http://apps.nccd.cdc.gov/brfss/Trends/agechart.asp? qkey⫽10090&state⫽US. Accessed February 18, 2005. 32. Agency for Healthcare Research and Quality. HCUPnet, Healthcare Cost and Utilization Project. Rockville MD: Agency for Healthcare Research and Quality. Available at: www.ahrq.gov/HCUPnet/. Accessed February 20, 2005.
American Journal of Preventive Medicine, Volume 29, Number 5
Appendix Diabetes measures and code definitions Measure Preventive measures HbA1c test Lipid profiling Eye examination
Self-monitoring Short-term complications Hospitalization for diabetes with ketoacidosis Admission to an emergency department or observation bed for metabolic crisis Long-term complications Nephropathy Lower limb amputation Diabetic retinopathy Blindness Hospitalization for primary diagnosis of stroke Hospitalization for primary diagnosis of acute myocardial infarction Hospitalization for primary diagnosis of heart failure Hospitalization for coronary artery bypass graft Progression to percutaneous transluminal coronary angioplasty (including stents)
ICD-9, CPT/HCPCS codes 83036 80061 OR (82465, 83718, 84478 in the same day) 67101, 67105, 67107, 67108, 67110, 67112, 67141, 67145, 67208, 67210, 67218, 67227, 67228, 92002, 92004, 92012, 92014, 92018, 92019, 92225, 92226, 92230, 92235, 92240, 92250, 92260 OR provider specialty codes 18, 41 and (90000–90080, 99201, 99202, 99203, 99204, 99205, 99211, 99212, 99213, 99214, 99215, 90640–90643, 99241, 99242, 99243, 99244, 99245) 82962, A4253, E0607 250.1x (251.0x, 250.1x, 250.2x, 250.3x) AND (045x, 0762, 0981)
581.81, 583.81 84.10–84.17, 28800, 28805, 28810, 28820, 28825, 27880, 27881, 27882, 27886, 27888, 27591, 27592, 27594, 27596, 27598 362.0x 369.xx 362.34, 433.0x, 433.1x, 433.2x, 433.3x, 433.8x, 433.9x, 434.0x, 434.1x, 434.9x, 435.0, 435.1, 435.3, 435.8, 435.9, 436 Additional codes for 1992: 433.0, 433.1, 433.2, 433.3, 433.8, 433.9, 434.0, 434.1, 434.9, 435.2 410.00, 410.01, 410.10, 410.11, 410.20, 410.21, 410.30, 410.31, 410.40, 410.41, 410.50, 410.51, 410.60, 410.61, 410.70, 410.71, 410.80, 410.81, 410.90, 410.91 402.01, 402.11, 402.91, 404.01, 404.11, 404.91, 428.0, 428.1, 428.9 36.1x 36.01, 36.02, 36.05, 36.06, 92980, 92981, 92982, 92984
Am J Prev Med 2005;29(5)
403