Dr. Podolsky responds to Dallas Morning News coverage of hospital care quality

To the UT Southwestern Community:

The Dallas Morning News yesterday published a set of stories and graphics that made dramatic assertions about patient safety and the quality of care in Texas hospitals, most notably those in Dallas.  We believe that The News knowingly painted an inaccurate and incomplete picture, and I write to address several points in order to put these stories in their proper context.

After eighteen months of Dallas Morning News reporting based on unrepresentative patient care cases from Parkland Memorial Hospital, we were pleased to learn from reporters some weeks ago that they were at last willing to turn from a focus on anecdotes to a consideration of Dallas area hospitals based on publicly available patient safety data.  Unfortunately the articles in Sunday’s paper indicate that they have neither put aside their anecdotal approach nor have they been willing to use quality data appropriately to present accurate assessments of Parkland or, in our case, of University Hospital-St. Paul.  

As an academic medical center, UT Southwestern is committed to rigorous quality measurement, and we welcomed The Dallas Morning News’ invitation to review their data and discuss their conclusions in advance of the publication of Sunday’s front page story.  And, contrary to statements in their published article, we readily provided them with quality data about University Hospital-St. Paul, which is a public hospital but is not supported by state taxpayer funds, as The News also incorrectly stated.

However, as we considered their interpretation of patient safety data, we quickly realized that there were several critical flaws in their analysis which raised serious questions about the validity of their results.  We are very disappointed that despite spending many hours with UT Southwestern experts in quality measurement, The Dallas Morning News published this story, rather than taking the time to reconsider their approach or to acknowledge its limitations.

Major concerns that we brought to their attention included:

  • The fact that a full picture of quality – and especially any attempt to rank order institutions in regard to quality – must take into account at least four different types of quality measures:  clinical effectiveness, patient safety, patient satisfaction, and clinical efficiency.  For this article, The Dallas Morning News relied only on patient safety data and furthermore, only a subset of that data which supported their story.  Patient safety is obviously an important quality measure, but it is only one of the four quality domains essential to a thoughtful, credible analysis. 
  • Their work was based on inpatient discharge records from the Texas Department of State Health Services for 2007, 2008, and 2009 that were analyzed using a software program from the Agency for HealthCare Research and Quality (AHRQ).  The limitations of using administrative data, especially old billing data, which is the basis of the Texas State Health Services database, are well-known and are mentioned only at the end of the article.  What was not mentioned was that their own quoted expert, Dr. Patrick Romano, has discussed these limitations in detail in “Lessons Learned from PSI Validation and Demonstration Projects,” (University HealthSystem Consortium Webinar, May 6, 2010) – and that University Hospital-St. Paul now (in 2011) scores above the national average in ten of the fourteen patient safety indicators examined.
  • It is an axiom of scientific work that results must be reproducible to be valid, and we have not been able to replicate The Dallas Morning News’ findings, using their flawed methodology.  Their methodology started with observed administrative data; they then used an adjustment process not designed for that data – and then applied a data “smoothing” process that added incidents which did not actually occur to the totals.  As a final step, they aggregated the individual rates into a composite number and used that composite number to rank the hospitals.  This compounding of methods not only cannot be duplicated, it has no statistical validity.
  • Although The Dallas Morning News utilized a highly regarded software program, they applied it to a data set that lacked a key variable (called “present on admission”) that is needed to obtain accurate results from this program.  Using data that does not include this variable results in inaccurate calculations, especially for hospitals like University Hospital-St. Paul that are referral centers and typically admit sicker patients, with pre-existing medical conditions.  While the article superficially acknowledged that risk adjustment methodology has its limitations when applied to billing data, the reporters completely ignored this fact in drawing their conclusions.
  • If The Dallas Morning News had chosen to consider outcomes data, such as mortality rates, they would have been compelled to acknowledge that UT Southwestern looks very different from the picture they painted.  Hospital Compare, a publicly available website supported by the Centers for Medicare and Medicaid Services, shows that University Hospital-St. Paul, for example, has lower mortality rates for pneumonia and heart failure patients than the national average.  Mortality rates for many other conditions at University Hospital-St. Paul rank among the best in the country, another fact ignored by The News.
  • This analysis does not encompass the entirety of University Hospitals because it intentionally excludes University Hospital-Zale Lipshy.  The aging University Hospital-St. Paul serves a distinct service mix, with an especially high case mix index of severity.  As the newspaper is aware, an analysis that includes the full range of those patients cared for in our University Hospitals across both sites would yield substantially better outcomes.

In addition to not giving an accurate picture of the overall quality of University Hospital-St. Paul – or noting its current status in regard to patient safety indicators – The Dallas Morning News uses another anecdotal story of a patient cared for by our physicians at Parkland who had a highly unfortunate outcome to illustrate patient safety issues and to suggest reason for broader concern there.  However, it is important to note that:

  • Conclusions about overall quality of care cannot be drawn from an isolated, 8-year-old case, with selected excerpts from email and depositions dating from 2003, 2004 and 2006. There is not a hospital in the country that does not have unfortunate case outcomes, and it is wrong to suggest these reflect something unique about the hospital.
  • The attending surgeon supervising residents in the case was directly involved in the surgery and records show that he was “hands-on” working with the residents at the time it was alleged that the insertion of a trocar device caused injury to the patient. He remained present and actively involved throughout the surgery.
  • UT Southwestern’s decision to settle the lawsuit had absolutely nothing to do with The Dallas Morning News’ inquiry, despite their effort to claim credit.  UT Southwestern began weighing its options – to continue to trial or to pursue settlement – in October 2010 and had been in contact with the plaintiff’s attorneys before receiving any inquiries from The News.
  • The lengthy discussion of a research project that had been proposed to compare laparoscopic versus open hernia repair outcomes is a pointless and irrelevant diversion in this story, given that the patient was not a participant in any research study.  As explained to The News, the research project was terminated in 2006 due to a lack of participants and the inability to do necessary follow-ups with enrolled subjects. 
  • Absolutely no taxpayer dollars were involved in either the settlement or the legal fees in this case. Physicians are self-insured under the UT System Professional Medical Liability Benefits Plan, which is funded entirely from practice plan dollars.  The suggestion that any taxpayer money was involved in this case is another assertion that is simply wrong.
  • The implication that UT Southwestern as a public institution is motivated “to prolong litigation until it doesn’t make sense for victims’ lawyers” is erroneous and offensive.  As a state agency, UT Southwestern is obligated to assert all appropriate defenses in relation to lawsuits filed against it. 

While we initially supported The Dallas Morning News for taking on the complex and important topic of reporting on the quality of medical care in Dallas area hospitals – and as an institution are willing to acknowledge areas where we need improvement, as we did with the reporters working on this story – we believe that their results, at least for University Hospital-St. Paul, are based on an invalid methodology and therefore do not do justice to our hospital – or to this important topic of medical quality for all members of the greater Dallas community.

Knowing the information about quality measurement and the data about University Hospital-St. Paul that we provided the reporters in advance of this story, we can only conclude that they were determined to justify a predetermined bias and were willing to ignore or discount data that did not fit the story they were intent on writing.  

Daniel K. Podolsky, M.D.
President, UT Southwestern Medical Center