Better Health's Diabetes Measures and Standards

1. MEASURES AND STANDARDS: NATIONALLY ENDORSED, LOCALLY VETTED
Nine clinical measures that are endorsed nationally for people with diabetes are included in this report. We distinguish measures, such as tests to be monitored, from standards, which are the desired frequency or target results of those tests. These nine measures include many of the most important and evidence-based indicators of care processes and outcomes in diabetes. Details and evidence for each standard were provided in the First Community Health Checkup, available here.

A. PROCESS OF CARE STANDARDS (4)
Our four process of care measures include actions that the physician should take to properly measure, monitor or manage diabetes or prevent its complications. As described in Table 2, standards for these four process measures are reported individually and as a composite standard, for each practice and for the region. The composite process “score” for the region, and for each practice individually, represents the percentage of patients who meet all four of the care standards.

B. OUTCOME STANDARDS (5)
Good results on the five outcome measures require active patient involvement in their care, as well as good treatment by the doctor. These measures (Table 2) include the results of blood tests, physical examination findings and behaviors associated with diabetic complications. They sometimes are called “intermediate outcomes” or “risk factors,” because they are not outcomes per se – such as diabetes-related blindness, kidney failure or leg amputation. But their results predict these outcomes, with better results lowering the risk, and poorer results raising the risk. For simplicity’s sake, and consistent with the medical literature, we refer to these measures as outcomes. As we have done with our care process measures, we report standards for these five outcome measures both individually and as a composite score within each practice and across the region. The composite outcome “score” for the region, and for each practice individually, represents the percentage of patients who meet at least four of the five of the outcome standards.
C. COMMENTS ON OUR STANDARDS AND COMPOSITE STANDARDS
POSITIVELY FRAMED STANDARDS.
We have constructed all of our individual and composite standards to be framed positively, so that higher percentages are better. This makes understanding our results easier for everyone – doctors, patients, and the public. It also keeps us focused on improvement.

ACCOUNTABILITY: SEPARATING PROCESSES AND OUTCOMES.
We have classified our standards into “process of care” and “outcomes,” because we believe that accountability for these standards are different. Our Clinical Advisory Committee and Leadership Team believe that Process Standards mostly depend on what doctors and their practices do in the care of their patients. Outcome Standards, on the other hand, are more complex. Good results on Outcome Standards not only require effective actions by physicians and their practices – such as counseling and treatment recommendations - but also behavior change on the part of their patients. Changing behavior is challenging, as we all know. For example, as physicians we can recommend exercise, maintaining an ideal body weight, and to avoid smoking – behaviors that would dramatically improve outcomes in diabetes. But to make these behavior changes happen is difficult. This is especially true for those patients who may live in neighborhoods that are not conducive to walking or close to sources of nutritious food. Likewise, as compared with care processes, outcomes also depend more heavily on patient adherence to sometimes complicated and expensive medication recommendations. This is especially difficult for patients with lower educational attainment and literacy levels, for the poor, and those who lack adequate health insurance coverage.

DIFFERENT ACHIEVEMENT THRESHOLDS FOR COMPOSITE OUTCOMES AND PROCESSES.
The Composite Process standard reports the percentage of patients (in a practice, or region-wide) who meet all four individual standards, while the more lenient Composite Outcome standard reports the percentage of patients who meet at least four of the five individual standards. The Leadership team chose the “all-or-none” criterion for processes because it believes that practice sites can be held accountable for continuously improving – towards 100% achievement – on this standard. By contrast, although the Leadership Team wants to motivate our practices to continuously improve on the (arguably more important) outcomes, it also recognizes that factors outside of the physician’s control may limit the upper bound of achievement, especially for practices with a larger proportion of disadvantaged patients. 


2. HOW WE OBTAIN OUR MEASURES: ADVANTAGES AND LIMITATIONS
All of the measures we report are obtained from medical records of our partner practices and systems, not from insurance claims. This approach has remarkable advantages, but it also has some limitations that we discuss below. Among participating practices with electronic medical records (EMRs), we obtain these measures on virtually all diabetic patients between the ages of 18 and 75 who visited the same doctors’ office at least twice during the 12 months between July 1, 2007 and June 30, 2008. (Patients with “diabetes of pregnancy” are excluded.)
For partner practices that do not have EMRs, the same measures were collected using a standard protocol from a random sample of 100 charts of diabetes patients. Two abstractors conducted the reviews at each non-EMR site between October and December.
Both kinds of practices, those with paper records and those with EMRs, provide information about their patients’ neighborhoods, including average household income and maximum educational attainment, using information from the year 2000 U.S. census. No identifiers for patients or their doctors are shared with anyone outside of their clinical practice or health care organization.

A. ADVANTAGES OF OUR RECORDS-BASED APPROACH
Advantages of our EMR-centered approach include our ability to:
• measure achievement on all patients, regardless of how or whether they are insured, or if their insurance status changes;
• accurately link doctors and their patients, appropriately assigning their care to specific doctors and group practices;
• obtain actual test results – not just whether the test was performed. This allows for flexibility in the standards (target values) we choose, to change standards as they evolve, to capture standards that may be useful to other users, and to benchmark our performance;
• obtain records of doctors’ prescriptions of medications, enabling us to report their intentions regarding treatments; and
• report information in a timely way (by avoiding delays of insurance claims adjudication). Timely data offers providers and patients the feedback they need to improve quality.

 
B. LIMITATIONS OF OUR RECORDS-BASED APPROACH
In general, limitations of the EMR-based approach, or to any practice-based approach in most organizations, relate to under-reporting achievement. Theoretically, these limitations can be important, and they can vary from standard to standard and from one health care organization to another.
Because many patients obtain care in different places, any approach that relies on practice-centered reporting relies on two factors: 1) whether the patient received a relevant service outside the home practice; or 2) if he/she did, whether the relevant “outside” service is documented in the medical record of the practice. Often, in our highly fragmented health system, details of a relevant test or service obtained elsewhere are not relayed to the home practice.
For our diabetes measures and standards, under-reporting is likely in documentation of eye exams, which frequently are performed at free-standing eye clinics. In addition, until our practices have information on medication prescriptions that are filled by patients, we are limited at this point to documented doctors’ prescriptions for medications.
Patients of small practices unconnected to large organizations are more likely to seek care outside its doors, creating challenges to comprehensively collect information that is important for performance measurement and reporting. Nonetheless, as documented by our region’s relatively resource-poor Federally Qualified Health Centers, committed smaller practices can collect and document relevant information in relatively inexpensive patient registries.
Exceptions to this “under-reporting” limitation occur in true systems of care, such as health maintenance organizations, where bills are generated for services obtained outside of the home practice and documentation of these services is simpler. Kaiser Permanente is an example of a system in which reporting should be more complete.