Sunday, October 6, 2013

Correlation between hospital charges, patient satisfaction, and physician communication

The positive correlation between hospital costs and quality was debunked long ago, nonetheless it's still interesting to look at the data in new ways. Here I've combined two different data sets: first I took the hospital charge data that was published earlier this year by CMS and compared it against patient satisfaction or "perception" data from HCAHPS:


Let me explain the graph. Each circle represents an individual hospital. The x-axis represents patient satisfaction as measured by the HCAHPS question representing the percentage of patients giving their inpatient visit a score of a "9" or a "10." The y-axis represents the average charge (the amount the hospital billed Medicare, not necessarily the amount the hospital actually got paid) by the hospital for management of sepsis, which I am using as a surrogate measure for hospital charges overall. I chose sepsis because it is a common medical syndrome for which we have lots of data and recommendations, and it is a high-risk syndrome whose proper treatment requires all the different moving parts of a hospital to be well functioning. Therefore, I think it is a good indication of hospital functioning and efficiency. Or, in other words, hospitals that follow sepsis protocols really well will probably have shorter hospital stays and therefore smaller charges (this is my hypothesis). The color of the circle represents to what degree the patients perceived that doctors communicated with them (HCAHPS data, see my other recent posts); green means the patients perceived more and better communication with the doctor and red the opposite.

The boundary cases are really interesting: Oklahoma Heart Hospital and Hamlin Memorial Hospital may represent the best "values" when measured by my mechanism above. They each have very high satisfaction scores, along with some of the smallest charges, and very high rates of communication by doctors. On the other hand, Crozer Chester Medical Center in Pennsylvania appears to have much lower satisfaction scores and significantly higher charges, as well as poor physician communication.

The 20 hospitals with the highest dissatisfaction rates in America

Here are the twenty hospitals in America with the lowest recommendation rates by patients whom they serve, according to data from the official HCAHPS survey, accessed October 2013 from the Medicare data website:
What is "HCAHPS"? HCAHPS is best explained by their own fact sheet:
The HCAHPS (Hospital Consumer Assessment of Healthcare Providers and Systems) Survey is the first national, standardized, publicly reported survey of patients' perspectives of hospital care... While many hospitals have collected information on patient satisfaction for their own internal use, until HCAHPS there were no common metrics and no national standards for collecting and publicly reporting information about patient experience of care. Since 2008, HCAHPS has allowed valid comparisons to be made across hospitals locally, regionally and nationally. 
Why is HCAHPS important? The answer is "money". Again, from the fact sheet:
HCAHPS and Hospital Value-Based Purchasing. The FY 2014 Hospital Value-Based Purchasing (Hospital VBP) program links a portion of IPPS hospitals' payment from CMS to performance on a set of quality measures, which include the Clinical Process of Care Domain, which accounts for 45% of a hospital’s Total Performance Score (TPS); the Patient Experience of Care Domain, 30% of TPS; and the new Outcome Domain, 25% of TPS. The HCAHPS Survey is the basis of the Patient Experience of Care Domain. 

Friday, October 4, 2013

Alaska has highest rate of Central Line Associated Blood Stream Infections (CLABSI) in the country, and Maryland has highest rate of Catheter Associated Urinary Tract Infections.


According to the latest data I could find (last updated July 2013) Alaska has the highest rate of Central Line Associated Blood Stream Infections (CLABSI) in the country. CLABSI is only one of many hospital acquired infections. Here is the brief explanation of the data from CMS:
The Healthcare-Associated Infections (HAI) measures - state data. These measures are developed by Centers for Disease Control and Prevention (CDC) and collected through the National Healthcare Safety Network (NHSN). They provide information on infections that occur while the patient is in the hospital. These infections can be related to devices, such as central lines and urinary catheters, or spread from patient to patient after contact with an infected person or surface. Many healthcare associated infections can be prevented when the hospitals use CDC-recommended infection control steps.
There are four main Hospital Acquired Infections (HAI) being measured:
  1. HAI-1-SIR. Central Line Associated Blood Stream Infections (CLABSI)
  2. HAI-2-SIR. Catheter Associated Urinary Tract Infections (CAUTI)
  3. HAI-3-SIR. Surgical Site Infections from colon surgery (SSI: Colon)
  4. HAI-4-SIR. Surgical Site Infections from abdominal hysterectomy (SSI: Hysterectomy)
There is a very good explanation of the calculation methodology for Standardized Infection Ratios (SIR) on the Leapfrog website. Essentially, Alaska's score of 1.495 means that CLABSI is 49.5% more common in Alaska than expected number for the standard population.

Why are Hospital Acquired Infections important? There are many reasons, but the key reasons are that they cause harm to patients, they are preventable, and they cost a lot of money to treat.

I've created an interactive chart where you can look into the data more closely by either bar chart or map.

My piece on my experience with the healthcare system as an Emergency Medicine physician

My piece on the challenges and failures of the US healthcare system is being featured on the front page of Medium today. Would love to hear...