- Reporters can use hospital readmission data to explore key issues
- http://www.kaiserhealthnews.org/Stories/2012/August/13/medicare-hospitals-readmissions-penalties.aspx
- http://www.kaiserhealthnews.org/Stories/2012/December/21/medicare-hospitals-value-based-purchasing.aspx
- http://www.kaiserhealthnews.org/Stories/2012/May/09/Medicare-Hospitals-Costly-Patients.aspx
- http://www.kaiserhealthnews.org/Stories/2012/February/13/medicare-hospital-patient-safety-records.aspx
List date(s) this work was published or aired.
12/20/12, 8/13/12, 5/9/12, 2/13/12
Provide a brief synopsis of the story or stories, including any significant findings.
Beat: Jordan Rau covers federal health policy and health care providers for Kaiser Health News, focusing on hospitals. Our stories this year explored one of the most challenging endeavors in health care, the effort to measure the quality of hospitals and other providers and make them financially accountable for how their patients fare. We were the first news organization to analyze and report on Medicare’s new readmissions penalties, which affected 2,217 hospitals, finding that hospitals that treat low-income patients were more likely to be punished. We were the first organization to report and reveal how new bonuses and penalties affected nearly 3,000 hospitals, providing the first public list of how each hospital did. We were the first organization to analyze Medicare’s measure of hospital efficiency, finding that essentially similar patients cost the government far more at some hospitals than at others. Our examination of Medicare’s new measures of patient safety showed that they might be skewed against teaching hospitals. Our stories analyzed how hospitals in different states and regions in the country performed on some of these metrics and identified hospital outliers. We made our hospital data available through interactive graphics and download for other media to use to localize our story or do their own.
Explain types of documents, data or Internet resources used. Were FOI or public records act requests required? How did this affect the work?
Our stories used brand new Medicare data that provided insight into how specific hospitals performed on quality and cost metrics. We had to use multiple data sets to match up hospitals with their performance scores and their locations. In our readmissions stories we analyzed how the hospitals’ patient mix influenced their readmissions rate by using other Medicare data that identifies how many low-income patients hospitals treat. In our story on Medicare’s new quality bonuses and penalties, we incorporated a statistical analysis performed for us by a team of Harvard researchers, which showed how different types of hospitals fared. In our story on patient safety metrics, we analyzed how teaching hospitals fared against non-teaching hospitals, using data from the Association of American Medical Colleges when evaluating Medicare records. FOIs were not required for any of the data. Academic studies and technical reports were used to decipher the data and understand the measures. The data was the cornerstone of all these stories, without which we could not have examined major new government policies and determined how they impacted health care providers.
Explain types of human sources used.
Hospital administrators, doctors, nurses, academics, health care consultants, government officials, patients, consumers, nonprofit officials, communications workers.
Results:
As a nonprofit news organization, our goal is to help explain significant health policy issues for the public and assist existing news organizations that don’t have the time or resources to explore these topics on their own. We made our hospital data available through interactive graphics and download for other media to use to localize our story or do their own. Our readmission story was picked up, localized or redone by more than 90 news organizations. We localized our story on hospital patient safety for four major news organizations.
Follow-up (if any). Have you run a correction or clarification on the report or has anyone come forward to challenge its accuracy? If so, please explain.
We updated our Aug. 13 readmissions story and data after the Centers for Medicare & Medicaid disclosed that it had made a series of small errors in its own calculations. None of those errors altered our original work in any material way, but we did update the story with the adjusted figures for institutions we had mentioned. As we explained in an Oct. 2 story revealing Medicare’s mistake: “The Centers for Medicare & Medicaid Services has discovered errors in its initial calculations in August. As a result, 1,422 hospitals with comparatively high readmission rates will lose slightly more money than they were expecting, according to a Kaiser Health News analysis of the revised penalties. Fifty-five hospitals will lose less than were previously told. The changes were tiny, averaging two-hundreds of a percent of a hospital’s regular Medicare reimbursements.” We also published all the readmission data again, containing both Medicare’s initial penalties and its revised ones, and the difference between them. We consider the subsequent coverage of the readmissions penalties to be part of the story, not a clarification or mistake.
Advice to other journalists planning a similar story or project.
You need to fully understand how hospital measures and ratings are constructed in order to be aware of their limitations. Measures are like people: they all have value and they all have flaws, and the challenge in stories is to discern what they can tell you that is valuable while being able to discard conclusions that on the surface are sexy but upon close inspection are unfair. Also it’s very valuable to work with academics and other researchers who have analyzed the same data sets or similar ones, so that you’re not looking at it alone. Finally, the opportunity for mistakes when dealing with numbers is huge, so construct multiple ways of double-checking everything you plan to publish. In today’s media environment, where stories, snippets and conclusions can spread through a variety of social media, there’s no way to pull back an error or cover it with a simple correction, so you have to get it right the first time. You need to find someone, either inside or outside your organization, who can spot check the data for you and vet your approaches.