Broad DX charting improves mortality scores.
VANCOUVER, B.C. -- To improve hospital mortality scores, clinicians should include all possible diagnoses in patient charts, according to oncologist and palliative care specialist Dr. Thomas Smith.
U.S. News & World Report's "'Best Hospitals," the Centers for Medicaid and Medicare Services' Hospital Compare, and other rating programs compare outcomes to national averages based on how sick patients are, said Dr. Smith, medical director of palliative care at Virginia Commonwealth University Massey Cancer Center in Richmond.
"You get counted on your observed mortality rate. They compare that to how many people are expected to die'" according to the averages, he said at the meeting. Deaths in excess of the averages are considered unexpected and are assumed to be caused by poor quality care, which lowers hospitals in the mortality rankings.
By listing all possible diagnoses--especially at the end of life--"[you increase] your expected mortality, so that your observed-to-expected ratio improves. Even if you ignore improving mortality, if you change [how many people are expected to die] you'll improve your ratio" and hospital standing, he said.
The reason is that hospital coders pull the diagnoses off patient charts and include them in claims submissions, the data upon which the rankings are based.
The approach is not about "gaming the system," Dr. Smith said. "You are adequately documenting the severity of the patients that you take care of" to ensure accurate mortality scores.
The key is to list diagnoses, not "medical thoughts," he said.
For example, "'lytic lesion of vertebra on spine films" does not count. Write 'bone mets" [because] your coders can count that. Don't write 'neutrophils' with an arrow going down. It doesn't count. You have to write 'neutropenia.' If you write 'admit for chemo,' you can write 'admit for chemo, dehydration, chronic blood loss anemia" [as appropriate] and your severity index will go up a lot," said Dr. Smith, adding that each site of metastases should be listed because "you get points for each one."
Similarly, ""chest x-ray with pneumonitis status post radiation therapy" does not count. You have to write 'radiation pneumonitis.' [And] don't put 'total protein and albumin low.' Put 'cachexia or malnutrition, moderate or severe.' Don't write 'admit for pain control.' Write "intractable pain from malignancy," and you have to say 'lung cancer, breast cancer, bone cancer, liver cancer" [accordingly]," he said.
Such diagnoses have to be written only once during the admission to count, he noted.
The approach worked for Massey Cancer Center's palliative care unit.
"In about an 18- to 24-month period, we paid a lot attention to having most of the attendings and the fellows write out diagnostic statements, rather than medical thoughts," Dr. Smith said.
"When we started, our actual-vs.-expected mortality ratio was 150% higher than what it was supposed to be. We did not change the mortality rate one bit, but we changed the expected mortality. It came down to 55%, which is still high, but, heck, it's a palliative care unit. A lot of people are sick and they're going to die," he said. "The APR-DRG [All Patient Refined Diagnosis Related Groups] severity of illness doubled just by paying attention to having those diagnostic statements put in," he said.
Dr. Smith gave an example of how the strategy also increases payments: An elderly person was admitted upon diagnosis of urosepsis, dehydration, and chronic obstructive pulmonary disease, but the patient's condition allowed for more, including urinary tract infection; malnutrition; preexisting decubitus ulcer; and shock because the patient was a bit hypotensive. With the additional diagnoses, the patient's severity of illness (SOl) weighting increased from 0.5973 to 3.3739, resulting in a $5,613 payment increase.
Dr. Smith said he has no disclosures.
VIEW ON THE NEWS
Use the System, With Ethics Intact
What we write in the chart will have a direct impact on the DRG applied to that episode of care, which will in turn affect the expected length of stay In addition, the more accurately we reflect the patient's true condition--including all the patient diagnoses and comorbidities--the higher the patient's expected mortality will be. Accurate chart documentation is not gaming the system. It is an essential piece of the system itself.
As far as ethics goes, intent plays an important role. The medical literature already suggests that we enroll patients into hospice too late. If the intent is to identify hospice-appropriate patients early and to otherwise provide them with the benefits of hospice longer by early enrollment, I think that is fine. However, if the intent is to increase your exclusion pool by recruiting questionable hospice patients and manipulating them into enrolling into hospice--well then, that sort of speaks for itself, doesn't it?
FRANKLIN A. MICHOTA, M.D.,/s the director of academic affairs in the department of hospital medicine at the Cleveland Clinic. He reported no relevant conflicts of interest.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||PRACTICE TRENDS|
|Author:||Otto, M. Alexander|
|Publication:||Family Practice News|
|Date:||Apr 15, 2011|
|Previous Article:||Hospitals may rethink records storage policy: Massachusetts General Hospital paid $1 million to settle suit when records were lost on the subway.|
|Next Article:||Survey: majority of hospitals face serious drug shortages.|