Linking coding and mortality rates helps convince physicians to change

Tip: Point out Internet MD report cards

A major reason it’s so difficult to change physician behavior where coding and documentation are concerned is that many doctors see hospital coding as separate from their own practice and experience. So the key is to find a way to tie physicians’ coding to their own outcomes.

An HIM compliance manager with Catholic Healthcare West in Rancho Cordova, CA, has found a way to do exactly that by educating physicians about how popular web sites use hospital coding to rate both hospitals and the physicians who practice in them.

"These web sites use one to five stars for their ratings, and we wanted to explain to physicians where the web sites get their data and what the impact is specifically when the public has access to this information," says Mark Anderson, RHIT, CCS, corporate coding/HIM compliance manager for Catholic Healthcare West.

"The public can log on to these sites and look up physicians or five-star hospitals in their area and see which physicians are associated with them," Anderson explains. "Their documentation affects their rating, and the hospital’s."

When Anderson met with a group of physicians to discuss coding and documentation, he found that this link between coding and the Internet rating systems quickly got their attention.

"We told them we did a search in a certain town for physicians associated with a three-star or lower hospital, and we got 39 names of doctors," Anderson adds. "And then we asked for the physicians associated with a five-star hospital and came up with three names."

Anderson explained to physicians how these web sites come up with their ratings. Basically, the ratings are arrived at by measurements of predicted mortality against actual mortality.

From the Uniform Hospital Discharge Data Set, web site managers are able to find a hospital’s actual mortality rate. For predicted mortality rates, they use the Medicare Provider Analysis Review data, which is derived from ICD-9 codes, Anderson says.

"What the physician writes in documentation gets coded and goes into the government database," Anderson adds. "Web sites get hold of the data, which is public, and they go by the codes using a proprietary formula, and they don’t explain how they come up with it."

However, it is clear that the web sites make predictions of how sick a patient is based on the codes. If a doctor fails to document certain items, then that doctor’s patient might look healthier than another doctor’s patient, even though both patients are equally ill.

"If a doctor just writes down pneumonia’ and doesn’t mention the patient’s stroke, urinary tract infection, and aspiration pneumonia, then the risk of mortality for that patient will look lower than it really is," Anderson says. "If all of the information is documented, then the patient’s true severity of illness is reflected, the predicted mortality rate goes up, and both the hospital and physician look better."

By explaining this process and how the data could be used to rate the quality of care a particular physician provides, Anderson was able to get physicians to buy in to the idea of improved documentation and coding.

"Improving data quality may be a better incentive than asking physicians to consider financial implications for the hospital," Anderson notes.

"When I was a coder, we used to sit around after a consultant reviewed our coding for a week, and they’d say, If the physician stated this then you could have coded it this way,’ and it was always could have,’" Anderson recalls. "We’d look at each other and say, Why are they telling us this, why don’t they tell the physicians this, because it would sure be great if the doctors heard it.’"

This is why Anderson has jumped at the opportunity to meet with physicians face to face and talk with them about coding in a way that they understand and which motivates them.

Anderson says that after his presentation, several physicians came up to him and asked, "How do we know how well we’re doing on documentation?"

This opened the door for a very receptive audience to Anderson’s guidelines for improving physician documentation. He was even asked to meet with the physicians at the hospital’s nursing station and to review their notes on a particular day, telling them what they could do differently in their documentation.

This appeared to be such a good idea that Anderson asked the hospital’s media specialist to film these interviews with physicians. The unscripted sessions between Anderson and physicians use real charts that doctors had completed after rounds, and Anderson had no prior knowledge about what the diagnoses would be.

"I worked with five different hospitals and physicians, and we would meet at the nursing station at a certain time after they had finished rounding patients," Anderson says.

"We would open the chart and go to the progress notes, and as they wrote diagnoses one and two, I would let them write it the way they normally wrote it," Anderson adds. "And then I would inject commentary and let them know how we could make that diagnosis more specific to capture the patient’s true risk of mortality and severity of illness."

Teach coding guidelines to physicians

Anderson then gave the doctors a quick lesson on how what they traditionally have written in progress notes is not always reflected in ICD-9 classification because of coding industry guidelines.

For example, if a physician wrote the diagnosis of anemia, Anderson would ask whether that was due to blood loss or some other cause, and the physician might answer that it was due to gastrointestinal bleeding. Then Anderson would say that in the ICD-9 index, the blood loss anemia would default to chronic unless the physician also wrote down the word "acute" next to the diagnosis.

"I’d take them down the path and say, This is how you can be really specific and record a patient’s data more accurately,’" Anderson adds.

Other physician education methods used by Catholic Healthcare West include posters and flyers, articles in medical staff newsletters, inservice sessions, and continuing education credit presentations.

However, the health system’s most effective strategy has been a video of Anderson’s discussions with doctors as they were on the unit documenting ongoing cases.

"We had a media specialist edit the film, and using software, we put all the elements together with some commentary and PowerPoint slides in the background," Anderson says. "We showed a comparison of codes that would be assigned on the left side of the screen and codes that occurred after the documentation was improved on the right side of the screen."

The 38-minute film also showed the codes and risk of mortality.

Anderson personalizes the presentation for various Catholic Healthcare West hospitals by downloading a hospital’s Web site rating and public data and comparing these with competitors.

Another strategy might be to place copies of the film in medical staff lounges and the medical library and making them available for viewing at staff meetings, Anderson says.

Although one short film cannot teach physicians everything they need to know about improving coding, the idea is to give doctors some small bits of information that can be built upon, Anderson says.

"We give them one or two phrases they can change that will help them, and a week later we put more information in a newsletter," Anderson says.

"I just want to reiterate that it’s a multifaceted approach," Anderson says. "For years, hospitals have been trying to figure out a way to do this efficiently."

Other strategies might include using case managers or coders who will work with physicians on the floor and answer their questions about documentation before the files make it to the coding department, Anderson suggests.

"Any time you can get contact with physicians to get information across, you have to try multiple ways of doing it," Anderson adds.

(Editor’s note: For an example of an Internet hospital report card and physician listing, go to