CalHospitalCompare.org

About the Ratings

As of 2010, more than 240 hospitals, representing over 86% of acute care hospital admissions in the state, have participated in this important voluntary effort. The conditions and procedures rated (heart attack, heart failure, heart bypass surgery, pneumonia, and maternity) are the five most common reasons for being admitted to a hospital. In addition, there are measures applicable to patients admitted to the Intensive Care Unit, and all surgical patients or all medical patients (patients admitted to the hospital for an illness that is not treated with surgery). Other conditions and measures will be included in future updates to this site.

Hospitals on this site are rated on quality and timeliness of care for several conditions, as well as patient experience with overall, medical, surgical, and maternity care, and adherence to recommended patient safety practices.

With the exception of some patient safety measures, hospitals are rated on a five-point scale, based on their performance:

• Superior: Hospital performed well above average compared to other hospitals on this measure.
• Above average: Hospital performed better than average compared to other hospitals on this measure.
• Average: Hospital performed within the average range compared to other hospitals on this measure.
• Below average: Hospital performed worse than average compared to other hospitals on this measure.
• Poor: Hospital performed well below average compared to other hospitals on this measure.

Time Periods for Data Collection

In all cases, we use the most recent data available. However, the time periods vary for different measures:

• Heart Attack, Heart Failure, and Pneumonia Care: October 2008 through September 2009.
• Heart Bypass Surgery: Mortality rates are from calendar years 2005 and 2006; Internal Mammary Artery Usage rates are from calendar year 2006.
• Maternity Care: Breastfeeding, C-section, and VBAC rates, 2007; Episiotomy rates, 2008; Regional NICU Status, 2009. 
• Critical Care: January through December 2009.
• Patient Safety: Surgical Care Measures, October 2008 through September 2009; Hospital-Acquired Pressure Ulcers, January through December 2009.
• All Patient Experience Surveys: July 2008 through June 2009.

How the Ratings Are Assigned

Hospital performance is evaluated compared to benchmarks of high, middle, and low performance. The benchmarks used were specific to each condition, but as a general rule, the top 10% of national performance on a measure was used as the high benchmark, and the national average on a measure was used as the middle benchmark. Performance 10% below the national average was used as the low benchmark.

Because CHART hospitals are releasing patient experience survey data and some patient safety measures before the rest of the nation, these hospitals can only be compared to each other, using their average performance, as well as top and bottom 10% performance levels, as the three benchmarks.

Applying a Margin of Error to Data

Choosing benchmarks is only the first step in assigning a rating. This is because it would be misleading only to compare a hospital's performance rate to the benchmarks, especially if that hospital has only treated a few patients with a specific condition or procedure. For example, a hospital that does only one heart surgery will have a death rate of either 0% (if the patient lives) or 100% (if the patient dies). Based on the outcome for one patient, it is difficult to determine how the hospital would perform for many surgeries.

Because it may not be possible to determine if the performance on a procedure or treatment has been measured accurately, the ratings on this site use a "margin of error," just as many news polls do. For hospitals that have only a few patients with a particular condition, the margin for error in the measurement of their performance in that condition is broad, while for hospitals with hundreds of patients the margin for error is narrower. (For this reason, for any single measure on this site, two hospitals with the same rating may have very different percentages shown.)

In rating the hospitals, the margin for error is taken into account by calculating an estimated range of hospital performance for each condition, with the range wider for hospitals with fewer patients and narrower for hospitals with many patients. The data for the specific condition are then checked to determine whether this range includes any of the performance benchmarks (low, middle, or high, where higher means better performance) and a performance score of superior, above average, average, below average, or poor is assigned.

The five scores were assigned using the low end (L) of the range of each hospital's estimated performance and the high end (H) of the range and comparing them to the low, middle and high benchmarks.

Below the Low Benchmark	Between the Low and Middle Benchmarks	Between Middle and High Benchmarks	Above High Benchmark	Ratings
			L & H	Superior
		L	H	Superior
		L & H		Above Average
	L		H	Above Average
	L	H		Average
	L & H			Average
L			H	Too Few Cases to Rate (Range too Wide)
L		H		Below Average
L	H			Below Average
L & H				Poor

Adjustments for Variations in Patient Mix

There are many factors that have an impact on how well a patient with a given condition or procedure will do, and these factors have an effect on how the performance ratings for those conditions are calculated as well. For instance, having diabetes can reduce your chances of surviving heart surgery, so a hospital with more diabetic patients than average should have its death rate adjusted. However, having diabetes does not prevent the nurse from checking your wristband before giving a medication, so there should be no adjustment for diabetes in that measure. Similarly, maternity patients nationwide report much better experiences with care (because they have plenty of time to identify a preferred hospital and they get to take home a baby) than routine medical or surgical patients, so a hospital that does not do any maternity care should have its overall rating adjusted up to allow comparisons of overall patient experience to hospitals that do have maternity care.

Therefore, for each condition, the most modern, state-of-the-art adjustment procedures available were used to account for variations in patient mix. The following organizations provided the rules for collecting the data and adjusting for patient differences when necessary:

• Heart Attack, Heart Failure, and Pneumonia Care: The Joint Commission and the federal Medicare program.
• Coronary Artery Bypass Mortality Rates: California Office of Statewide Health Planning and Development (OSHPD).
• Maternity Care (except Patient Experience Surveys): C-section and VBAC rates, California Office of Statewide Health Planning and Development. Breastfeeding rate, California Department of Public Health, Center for Family Health, Genetic Disease Screening Program, Newborn Screening Data. Regional NICU designations from the California Children's Services. Episiotomy rates, based on OSHPD Public Patient Discharge Data. 
• Critical Care: S. Lemeshow, D. Teres, J. Klar, J. S. Avrunin, S. H. Gehlbach, and J. Rapoport. "Mortality probability models (MPM II) based on an international cohort of intensive care unit patients." JAMA. 1993;270(20):2478-2486.
• Patient Safety: Surgical Infection Prevention measures from the Joint Commission and the federal Medicare program; and Hospital-Acquired Pressure Ulcer (HAPU) rates for the California Nursing Outcomes Coalition (CHART requires that a Stage 2 or above pressure ulcer be discovered within 24 hours of admission. Otherwise it is considered hospital-acquired and included in the CHART report. As of August 2008, this standard was changed from 48 hours to 24 hours.).
• Patient Experience Surveys: the National CAHPS Benchmarking Database (the national center for collecting patient surveys).

To reduce the impact of the different patient characteristics among the different hospitals, and thus to help assure the accuracy of the performance reports, each hospital's scores were adjusted to reflect the scores the hospital would have received had its patient response sample included an average mix of patient characteristics. In general, however, the effect of the adjustment process on participating hospitals' scores was small.

Maternity Data

The Cesarean-section measure is not assigned a performance rating because there is no generally agreed upon approach for rating this data. However, the data can be useful for an expecting mother or couple. If you want to do everything possible to have a vaginal delivery, you may prefer to use a hospital with a low C-section rate. You should discuss this concern with your obstetrician.

Episiotomy rates have fallen dramatically over the last decade as research has shown no advantage to the cut and a strong association with bigger tears into the anal muscles. On this Web site we do not assign hospitals a performance rating for episiotomy because doctors and hospitals have not yet determined what is an acceptable rate. However, in general, a lower rate is better.