Does Care at a Top-Rated Hospital Lower Your Chance of Dying?

Your choice of hospital has a large bearing on your risk of experiencing a complication of treatment, or even death, as a patient.

In our brief article on the rate of adverse events, we highlighted the enormity of this public health problem in the United States and elsewhere. Healthcare provider organizations can reduce medical errors and improve patient safety by addressing shortcomings in their processes of providing care.

But what if you were a healthcare consumer? Is there anything you can do to prevent accidental harm to yourself or a loved one?

Actually, you can do quite a number of things to improve the safety of the care you receive in a hospital. One of these things is to pick a good hospital.

In October 2009, HealthGrades, a healthcare ratings organization, released The Twelfth Annual HealthGrades Hospital Quality in America Study [PDF – 1 MB], which examined nearly 40 million Medicare hospitalization records for the years 2006, 2007 and 2008. This study looked at trends in inpatient death and complication rates at each of the 5,000 non-federal hospitals in the United States.

The study found a wide gap between the best hospitals (as rated by HealthGrades) and the other hospitals. For example, patients at low-rated hospitals were twice as likely to die than at top-rated hospitals for the same medical problems and procedures. HealthGrades gives each hospital a 1-, 3-, or 5-star rating for each procedure or diagnosis, reflecting the death or complication rates at that hospital.

Other findings of the study included the following:

  1. Overall inpatient deaths in US hospitals improved by 11% from 2006 through 2008.

  2. If all hospitals performed at the level of 5-star rated hospitals across the 17 procedures and diagnoses studied, more than 224,500 Medicare lives could have been saved from 2006 through 2008.

  3. More than half (57%) of the preventable deaths (127,488 of them), were associated with only four diagnoses: sepsis, pneumonia, heart failure and respiratory failure. (Clearly, the low-rated hospitals need to focus their improvement efforts on these conditions.)

  4. The chance of experiencing one or more in-hospital complications was more than 60% lower in a 5-star rated hospital compared to the US average hospital.

  5. Joint Commission stroke-certified hospitals have an 8% lower in-hospital death rate compared to hospitals that were not stroke-certified.

Features of the HealthGrades Study Methodology

Without going into too much detail, I think it might be useful to point out some key points about the methodology that HealthGrades used to rate the quality of procedures and diagnoses at each individual hospital.

  1. Data Acquisition
    HealthGrades analyzed patient outcome data for virtually every hospital in the US, with the exception of military and Veterans Administration hospitals. There were two main sources of data:

    • Medicare inpatient data from the MedPAR database, which is purchased from the Centers for Medicare and Medicaid Services. Accuracy of the data is regulated - hospitals are required by law to submit complete and accurate information, or they risk serious penalties.
    • Inpatient data for appendectomy provided by 19 states that provide all-payer data.
  2. Risk Adjustment
    Risk adjustment simply means that HealthGrades takes into account the differences in important characteristics between patient populations at different hospitals so that valid comparisons of patient outcomes can be made. For example, a hospital that treats patients who are more ill before admission, and therefore at greater risk of death or complication, may have a higher rating than another hospital with less ill patients, even if both hospitals have the same inpatient death and complication rates.

  3. Data Quality Checks
    HealthGrades also conducts so-called “data quality checks” for their “Multivariate Logistic Regression-Based Ratings.” I won’t describe the statistical methods used in this form of analysis here, but I will tell you that, as a result of these quality checks of the data, a small number of cases were excluded. This gives me even greater confidence in the ratings.

Overall, the analysis of the data is sufficiently robust and rigorous to ensure validity. The ratings are objective and are independently created; no hospital can opt-in or opt-out of being rated. No hospital pays to be rated, which I think is important to avoid any potential conflict of interest.


If you found the information above too heavy (even after my efforts to summarize it), that’s OK.

The key point is this: you can substantially reduce your risk of experiencing an adverse event by choosing a 5-​​star rated hospital.

Here’s to you having the safest possible care on your next hospital stay.

Andy Teh, MBBS, MSc, CPHQ
Global Healthcare Quality Consultant