Quality of Care Measurement

Process of care measures generally examine the appropriate order that services are delivered in patient care. Process measures may (or may not) have a proven link to actual patient outcomes of care. One methodology used by the Cave Consulting Group is to mine administrative claims data and measure practitioner quality of care by examining condition-specific, process of care measures. When available, we utilize patient outcomes variables for a medical condition.

Priority medical conditions are identified based on the prevalence rate of medical conditions, the overall charges to treat the medical conditions in a given time period, and on physicians' purported work effort in treating the conditions. Work effort is a function of the condition prevalence rate and the average episode charges to treat the specific medical condition. Some examples of priority medical conditions are hypertension, upper respiratory infection, sinusitis, rhinitis, back pain, bursitis, headaches, diabetes, asthma, gastroesophageal reflux disorders (GERD), gastroenteritis, and ischemic heart disease. Other conditions that may be considered include abdominal hernias, prostatic hypertrophy, glaucoma, hypothyroidism, seizure disorders, and congestive heart failure.

Criteria for Examining Quality Measures

The Cave Consulting Group has identified eight main criteria for examining appropriate practitioner outcomes and process of care quality measurement:

• Well documented in clinical literature and/or medical guidelines
  
• Clinically relevant to patient outcomes or medical treatment
  
• Easily identifiable using claims data
  
• Quantifiable using claims data
  
• Prevalent in the condition-specific episodes
  
• Variable use between physicians
  
• Cuts across service categories (lab, procedures, diagnostic tests, drug)
  
• Financially significant.

With respect to the first criterion, RAND has done an excellent review of the clinical literature and documented quality indicators at the medical condition level--RAND Health Quality of Care. This review and literature summarization was funded by the Agency for Healthcare Research and Quality (AHRQ) and was published in 2000. The Cave Consulting Group uses the RAND Health Quality of Care volumes as a source for obtaining meaningful outcomes and process of care measures.

Specialty-Specific Measures

Some of the priority health conditions treated by general internists and family and general practitioners are diabetes, hypertension, and back pain. We next present several suggested process of care quality measures with documented comments from the RAND Health Quality of Care study. This list is not meant to be exhaustive, but rather to show some potential quality measures.

• Diabetes:
  
  Patients with > 1 product for HTN (beta blockers, diuretic, ACE, calcium channel blockers). Many diabetic patients have hypertension (HTN). Evidence shows that pharmacotherapy reduces the progression of proteinuria and diabetic nephropathy.

  Patients with > 1 triglyceride lab test. This test searches for hyperlipidemia, which is shown to be more common in diabetics. This recommendation is based on expert opinion.

  Patients with > 1 urinalysis lab test. This test searches for renal disease, and is based on expert opinion.

  Patients with > 2 office visits. This measure is based on expert opinion, and is developed to reduce the probability of severe diabetic complications.

• Hypertension:

  Patients with at least three of the following blood chemistry tests - blood glucose, serum potassium, serum creatinine, serum cholesterol, and serum triglyceride. These lab tests support physician management of concurrent diagnoses, other cardiac risk factors, and end organ damage.

  Patients with > 1 urinalysis. This lab test searches for secondary causes, other cardiac risk factors, and end organ damage.

• Back pain:

  Patients with > 1 spine/bone x-ray in the first month of diagnosis. There is no proven value for these services in the first month of diagnosis for patients with routine lower back pain, and the patient is placed at some risk with x-rays. Consequently, the recommendation, based on expert opinion, is that this test should not be performed in the first month of diagnosis. In this sense, a higher rate (per episode of routine lower back pain) would be less appropriate than a lower rate.

  Patients with > 1 MRI/CAT scan/nuclear imaging in the first month of diagnosis. For patients whose symptoms persist longer than one month, consider additional diagnostic tests. There is no proven value for these services in the first month of diagnosis, and the patient is placed at some risk with MRI/CAT/nuclear imaging. The recommendation, based on expert opinion, is that this test should not be performed during the first month of diagnosis. A higher rate would be less appropriate than a lower rate.

Quality of Care Measure Information

Each quality of care measure should present the following information:

• Quantitative rate per episode or per patient (e.g., mean, mode, median)
  
• Quantitative range per episode or per patient (around the established rate)
  
• Rate and range by physician-specialty type--and not across all specialists
  
• Rate and range by severity-of-illness (SOI) class (per episode or patient)
  
• Formed based on statistical, literature, and expert panel review.

Using a large national claims database, the Cave Consulting Group has tracked about 20 condition-specific, quality of care measures over the past years. We next present some results.

• Diabetes and ACE inhibitors. Hypertension guidelines state a compelling case for using ACE inhibitors in treating diabetics who take insulin. ACE inhibitors have been shown to lower blood pressure, decrease the rate of progression of renal insufficiency, and reduce the need for renal transplantation or dialysis. We describe this performance measure as the number of diabetic patients with circulatory or renal involvement that had evidence of one or more ACE inhibitor prescriptions. The results are the following:

  1994: 34% of patients on one or more ACE inhibitors
  1996: 42% of patients on one or more ACE inhibitors
  1998: 44% of patients on one or more ACE inhibitors
  2002: 48% of patients on one or more ACE inhibitors (estimate).

• Ischemic heart disease and statins. Research has shown that in patients with ischemic heart disease (IHD) who followed a low-fat diet and took cholesterol-lowering drugs, the disease was reversed or arrested. Specifically, lowering high cholesterol through medical management reduced by one-third both the number of nonfatal heart attacks and the number of deaths from cardiovascular disease. We describer this prescription drug performance measure as the number of IHD patients with evidence of one or more cholesterol-lowering drug prescriptions. The results are the following for uncomplicated IHD patients only:

  1994: 25% of patients on one or more statin prescriptions
  1996: 32% of patients on one or more statin prescriptions
  1998: 40% of patients on one or more statin prescriptions
  2002: 44% of patients on one or more statin prescriptions (estimate).

• Back pain with imaging tests. The guidelines for uncomplicated low-back pain suggest to wait 30 days before ordering imaging procedures. We describe this performance measure as the number of low back pain patients that had evidence of one or more imaging tests or back surgery. The results are the following for uncomplicated low back pain:

  1994: 50% of patients had one or more imaging test
  1996: 47% of patients had one or more imaging test
  1998: 47% of patients had one or more imaging test
  2002: 45% of patients had one or more imaging test (estimate).

• Asthma with immunotherapy injections. Research shows that immunotherapy injections do not provide improved asthma control for pediatric asthma patients. We describe this procedure performance measure as the number of pediatric asthma patients that had evidence of one or more immunotherapy injections. The results are the following for pediatric asthma patients:

  1994: 21% of patients had one or more immunotherapy injections
  1996: 19% of patients had one or more immunotherapy injections
  1998: 17% of patients had one or more immunotherapy injections
  2002: 13% of patients had one or more immunotherapy injections (estimate).