Patients don't want to be treated at a hospital with a high mortality rate. Doctors and staff don't want to work at a hospital with a high rate. It's scary and embarrassing.
But there is more to achieving a low mortality rate than just hiring good staff and implementing evidence-based care. Hospital mortality rankings are heavily dependent on rating agency rules and documentation practices.
I was nine years old and obsessed with bowling. Maybe it was because Don Carter had become so popular. Perhaps it was the curiosity generated by the sound of bowling balls crashing into the pins at an alley only 2 blocks from my home. I had only bowled on a few occasions. But I knew that my dad had bowled in a league when he was a young man.
I begged him to sign me up for a league. On a lazy Saturday morning, after weeks of cajoling, he did just that and left me there to enjoy the games. I selected a ball and some bowling shoes.
I was assigned to a team and began bowling with the other kids in turn. I knock down a few pins and occasionally made a spare or a strike. But my score was hideous. When it was all over, I had three games with scores in the mid-double digits. Our team did poorly. I headed home disappointed and confused.
It turned out that there was something called a “foul line” and a rule against crossing it when throwing the ball. To this day, I don't understand why nobody explained that to me. It wasn't until much later that I figured it out.
But I learned a valuable lesson. Understand the rules or you will surely lose the game!
Hacking Mortality Rate Data
When you start digging, you'll find many rules when defining, measuring and reporting mortality rates. The rules vary depending on whether you're talking about rates published by CMS, Truven, HealthGrades or some other entity. And all of the reported scores use data sets originally designed for billing purposes.
Understanding and applying the appropriate rules makes a BIG difference in playing the game of published mortality rates.
Coding and Documentation
Hospital quality reporting is based on information from closed patient records. For Medicare patients, there are strict rules on the number and order of codes that can be used. In late 2007, “present on admission” (POA) modifiers were implemented. This helps to distinguish whether a co-morbid diagnosis is pre-existing or new (and possibly a complication of care).
The UB-04 Billing Form contains almost every determinant of the mortality rate calculation. Aspects that affect the patient's actual risk for mortality include:
- Patient origin (home, nursing home, hospice, etc.)
- Demographics (age, gender, race)
- Discharge status (alive to home, nursing home, etc. or expired)
- Admitting diagnosis
- Secondary diagnoses
- POA indicators
The first step is to assign the most appropriate admitting diagnosis. You don't want to mis-classify a patient with sepsis as having pyelonephritis or pneumonia. The latter conditions have a much lower expected mortality rate than sepsis patients.
Sometimes secondary diagnoses are missed. This is evidenced by a medication being listed in the patient's medical history without any corresponding diagnosis on the problem list.
When hospital mortality rates are compared and ranked, it is not based on raw mortality rates, but on risk-adjusted rates. Risk adjusted rates are calculated by taking into account each patient's demographic information, admission diagnosis, secondary diagnoses and so forth. Hospitals are ranked by comparing the ratios of actual to expected rates.
All things being equal, the organization that captures more co-morbid diagnoses in its documentation and coding will demonstrate a lower risk-adjusted mortality rate. And it is not just a matter of listing more secondary diagnoses. The POA indicators must be accurately assigned so that pre-existing conditions do not appear to be complications of care.
Inclusion and Exclusion Criteria
Inclusion and exclusion criteria also greatly influence quality metrics. If you don't know that a heart failure patient with an AICD (automatic implantable cardioverter-defibrillator) is excluded from mortality calculations, and you fail to document the presence of such a device in a deceased patient, your rate will look higher than it should. There are multiple such criteria for each diagnosis. There is some overlap, but each rating agency has its unique set of rules.
Discharge to Hospice
It is not uncommon for a terminally ill patient to be admitted to a hospital. When it becomes clear that supportive care is indicated, hospice care may be instituted. If the patient qualifies for in-patient hospice care, discharging the patient from an acute hospital status (general inpatient care or G.I.P.) to a contracted hospice bed will prevent the patient from being counted as an inpatient death. Without actually removing such patients from the hospital, you can put a process in place to change the status.
Your reported mortality rates can improve dramatically if you do the following:
- Create a physician-led, nursing-based documentation and coding initiative if you don't already have one. Do as much concurrent coding as possible so that you fully capture all diagnoses and the POA status of each. These initiatives more than cover their costs through improved reimbursements and play a major role in reducing published mortality rates.
- Create a team of coders, case managers, documentation specialists and quality professionals to review charts of patients prior to billing to clarify any documentation and coding issues. Assign one or two persons to become experts in the rules used to calculate quality metrics by the various quality rating agencies.
- Review the charts of all patients expiring in the hospital to ensure that all of the above concerns are addressed.
- Begin your efforts with heart failure, pneumonia, COPD, acute myocardial infarction and stroke. If necessary, develop quality committees for each diagnosis that address mortality, complications and readmissions. Expand them later to include overall hospital mortality.
By understanding the rules of the game, your hospital mortality rate will improve dramatically.
What other methods have you used to improve quality ratings at your facility?