The U.S. health care system has supported the use of quality measures to report and compare similar services among providers, health plans, and other health care organizations. Health plans have long reported quality measures to the National Committee for Quality Assurance (NCQA). Over 90% of managed care plans report annual quality measure results to NCQA, covering 81% of all insured persons (Burns 2017). Many quality measures focus on common chronic diseases and approved prevention or clinical management services whose benefits are supported by a solid evidence base. Some of these measures have been developed and promoted by several organizations and yet have a common basis in evidence.
To avoid variations among common measures, the National Quality Forum (NQF) has been designated as the only consensus-based health care organization in the nation by the Office of Management and Budget. As a result, the federal government and other users rely on NQF-defined measures as the best evidence-based approaches to improving care. NQF endorses measures submitted by various organizations designated as stewards of such measures. The NQF’s endorsed measures meet initial rigorous criteria and are regularly reviewed by experts. These measures are uniformly used within the industry to evaluate and report on performance and quality.
The set of measures constituting the Comprehensive Diabetes Care performance evaluation group have long been considered standard measures. In 1997, a coalition of public and private entities formed the Diabetes Quality Improvement Project (DQIP) to establish a set of diabetes-specific performance and outcome measures based on scientific evidence. Prestigious organizations such as the American Diabetes Association (ADA), NCQA, and American Academy of Family Physicians, among others, provided experts to develop a set of standard diabetes care measurements. The measures for comprehensive diabetes care were developed by and are owned by the NCQA as part of the HEDIS measure set, and NCQA is designated as the steward for these measures.
Here is the original set of seven accountability measures (McLaughlin 2000):
In 2002, the NQF convened the Diabetes Measures Review Committee both to review variations on the DQIP measure specifications promulgated by the American Medical Association (AMA), the Joint Commission, and NCQA, and to reach a consensus that was then endorsed by NQF (NQF 2002). The Consensus Report of 2002 included the recommendation for an annual microalbumin test for nephropathy, a biennial microalbumin test for low-risk patients, or evidence of medical attention for existing nephropathy. Microalbuminuria is considered an early indicator of adverse effects of diabetes on the kidney (Evans 2000). A 2004 report by the ADA specifies that a test for the presence of microalbumin should be performed upon initial diagnosis in type 2 diabetes and every five years for type 1 diabetes. Microalbumin testing can be performed with a random urine collection, 24-hour collection, or timed collection. The report also states that the need for annual microalbuminuria assessment is less clear after the diagnosis of microalbuminuria and initiation of angiotensin-converting enzyme (ACE) inhibitor or angiotensin-receptor blocker (ARB) therapy and blood pressure control (Molitch 2004). Most guidelines stress an annual test for microalbumin (NKF 2002), yet de Jong et al (2007) found that such screening is not carried out systematically in patients with diabetes.
ACE inhibitors and ARBs have been shown to diminish the risk for diabetic nephropathy in patients with type 2 diabetes (Gross 2005) and are used as a preventive measure (Roett 2012). However, their role in the prevention of diabetic nephropathy in patients with type 1 diabetes has not been defined (Gross 2005). The renal benefits of ACE inhibitors and ARBs were independent of blood pressure reduction, as they decrease the rate of microalbuminuria or more advanced diabetic nephropathy. Therefore, the use of either ACE inhibitors or ARBs was recommended as a first-line therapy for type 1 and type 2 diabetic patients with microalbuminuria even if they are not hypertensive.
Due to the findings of the positive effects of ACE inhibitors and ARBs on people with diabetes, the NCQA HEDIS 2007 Comprehensive Diabetes Care measure (for reporting year 2006) was revised to include an additional step for numerator compliance in the Medical Attention for Nephropathy measure, adding the use of ACE inhibitors/ARBs (NCQA 2007). From that date forward, a patient met the numerator if any of the following criteria were met: (1) a nephropathy screening test, (2) evidence of nephropathy through diagnosis codes, or procedure codes, (3) a nephrologist visit, or (4) evidence of ACE inhibitor/ARB therapy during the measurement year by an ambulatory prescription claim. Table 1 shows how the rates for health plans reporting to NCQA increased once this measure was put into place. Medicaid HMO plans, which were consistently >40% compliant from 2001 through 2005, jumped to 74.6% compliance in 2006, gradually increasing to 90% compliant in 2015.
Table 1HEDIS screening rates for diabetic nephropathy, 1999–2015
|Commercial HMO||Commercial PPO||Medicaid HMO||Medicare HMO||Medicare PPO|
| Note: 2006 reported results are derived from HEDIS 2007 specifications |
The specifications for the Medical Attention for Nephropathy measure have remained similar to the 2007 revision aside from updates for code changes and other details (NQF 2015). While preparing some comparative quality reports, the authors of this article were struck with the apparent omission of a key criterion in this measure. The evidence noted above shows that ACE inhibitors or ARBs were beneficial to people with a diagnosis of microalbuminuria, but the measure specifications require only a prescription for the drugs, not a prescription in conjunction with a diagnosis of microalbuminuria. ACE inhibitors and ARBs are prescribed to treat or prevent cardiovascular conditions (Venkat 2004). As a result, patients on ACE inhibitors or ARBs for a common diagnosis such as hypertension, but without a comorbid diagnosis of microalbuminuria, would be compliant for this measure. Yet they may not have had a screening of any other kind. The authors thought this could lead to inflated reporting. This study evaluated the effect of adding ACE inhibitors/ARBs to the Medical Attention for Nephropathy measure and whether there is a risk for overestimating quality of care.
The Truven MarketScan Commercial Claims data set provides claims data for 28,348,363 persons enrolled in commercial health plans in 2015. The HEDIS 2015 specifications for Medical Attention for Nephropathy were applied to the data to derive the denominator and the count of numerator-compliant people with diabetes as identified by each step/option for compliance. This process identified the persons who did not satisfy compliance without documentation of a nephropathy screening test (Step 1) or a nephrologist visit or a positive urine macroalbumin test (Step 2). The remaining persons were then checked for evidence of ACE inhibitor/ARB therapy during the measurement year, which is the final step (Step 3) of evidence for nephropathy.
The clinical-history information available through administrative data of the individuals who attained compliance by way of this final step was evaluated to identify a prior diagnosis of microalbuminuria (ICD9 790.1). This analysis was completed for four treatment years and included one additional enrollment year, allowing us to analyze data from 2011 through 2015.
In general, the derived rates for the Medical Attention of Nephropathy measure in this population were consistent with the reported NCQA mean rates for the commercial population, 80% to 84%. Table 2 provides the numerator-compliant rates by each step for each year and demonstrates the contribution of each step to the annual rate. Across the four years, 47% to 48% of people with diabetes satisfied the compliance option of receiving a nephropathy screening test in the measurement year, as documented in claims data by a procedure code (Step 1). Another 18% to 20% of people with diabetes met compliance because of Step 2, evidence of nephropathy through a documented visit with a nephrologist or a procedure code for a urine microalbumin test or evidence of treatment for nephropathy using procedure codes, diagnosis codes, revenue codes, or diagnosis-related groups. The final step (Step 3) sought evidence of an ACE inhibitor or an ARB prescription (or HCPCS code) yielded between 14% to 16% of the numerator-compliant people with diabetes each year. Inclusion of Step 3 contributed between 18% to 20% to the final rate, bringing it to the 80%+ range each year.
Table 2Numerator results by measure specification steps
|Year||Step||Numerator count||Rate as %||Contribution to rate|
|2011 denominator: 1,548,638||Step 1||734,379||47.42||57%|
|2012 denominator: 1,726,727||Step 1||810,879||46.96||58%|
|2013 denominator: 1,409,950||Step 1||673,597||47.77||59%|
|2014 denominator: 1,685,753||Step 1||776,725||46.08||58%|
|2015 denominator: 1,047,855||Step 1||506,838||48.37||58%|
|Step 1=evidence of a nephropathy screening test in the measurement year, as documented in claims data by a procedure code. Step 2=evidence of nephropathy through a documented visit with a nephrologist or a procedure code for a urine microalbumin test or evidence of treatment for nephropathy using procedure codes, diagnosis codes, revenue codes, or diagnosed-related groups. Step 3=evidence of a prescription for an ACE inhibitor or an ARB.|
As previously noted, a prescription for an ACE inhibitor or an ARB is considered beneficial to people who have a diagnosis of microalbuminuria. If prescribed for that purpose, those prescriptions would satisfy the indication of nephropathy. The claims records of the persons who satisfied the numerator through Step 3 were analyzed only to find a diagnosis of microalbuminuria in the reporting year or the year prior.
We found that only 1% or less of the persons satisfying Step 3 had evidence of microalbuminuria in any of the four years prior to 2015 for which we had data. None had the diagnosis in the reporting year.
Records were reviewed again for the same persons to identify the number of persons who had a diagnosis for hypertension or another cardiovascular condition for which they might have received a prescription for an ACE inhibitor or an ARB. Table 3 reports the count and percentage of persons with a cardiovascular condition that may be associated with the prescribed ACE inhibitor or ARB medication, as well as the count and percentage of persons with a diagnosis of microalbuminuria. Between 73% and 81% of persons who satisfied Step 3 by being on a medication were diagnosed and treated for a cardiovascular condition.
Table 3Diagnosis check of persons satisfying numerator in Step 3
|Year||Persons meeting numerator by Step 3 (n)||Persons from Step 3 with cardiovascular diagnosis in reporting year (n)||Persons with cardiovascular diagnosis (%)||Persons from Step 3 with microalbuminuria diagnosis in reporting year (n)||Persons with microalbuminuria diagnosis in reporting year (%)||Persons from Step 3 with microalbuminuria diagnosis in prior year (n)||Persons with microalbuminuria diagnosis in prior year (%)|
The main limitation for this study and its conclusions is that the authors did not have access to the logic of the advisory committee that originally wrote the measure specifications.
The findings of this review indicate that the criteria to satisfy Step 3 of the Monitoring for Diabetic Nephropathy standard measure is inadequate and imprecise. Inclusion of the ACE inhibitor or the ARB prescription captures many persons who lack a documented diagnosis of microalbuminuria. The use of these medications does not, however, obviate the need for a nephropathy screening in people with diabetes. They are prescribed for hypertension and other cardiovascular conditions that are risk factors for nephropathy in people with diabetes. Therefore, the inclusion of prescriptions for ACE inhibitors and ARBs to satisfy the screening measure for nephropathy may be misleading and contribute to underscreening of a population at risk.
It could be argued that administrative claims data might not be the best data set to evaluate quality of care of the diabetic population, in favor of inclusion of clinical records. However, these are the data used to measure diabetic nephropathy screening as recommended by NQF and to assess and compare health plans and providers. Therefore, we believe it is of vital importance to create measures with undisputed evidence that are as accurate as possible.
Quality clinical practice for the management of diabetic patients should include screening for nephropathy. Based on the findings, the authors suggest that reliance on a prescription alone should not be considered evidence of effective screening for such patients when that drug is used for other clinical indications. Further, the authors propose that the medical directors of health plans support the suggestion that this standard measure be reevaluated by the expert panel. We recommend that Step 3 include a requirement that a diagnosis of microalbuminuria be present in the administrative records in the measurement year or the prior year. With these added criteria, the compliant rates are likely to be substantially reduced, from the 80% range to the 60% range, at least initially until further improvement can be achieved. The overstatement of screening results may lead to provider complacency and the erroneous impression for health plans and consumers that performance on this measure is sufficient.
These findings support the objectives of continued review and improvement of the standard quality and performance measures that are commonly used to evaluate health plans and providers. Accurate data and analytical processes are critical for fair reporting and accurate assessment of the state of compliance.
Trudy Millard Krause, DrPH
School of Public Health
1200 Pressler, RAS 905
Houston, TX 77030
Conflicts of Interest and Source of Funding: No conflicts of interest or sources of funding relevant to this study were declared.