One of the questions I am often asked is “What is the typical distribution of CPT codes for hospitalists?” Prior to publication of the 2011 State of Hospital Medicine report, no one could answer that question with any authority. The Centers for Medicare & Medicaid Services (CMS) publishes some Healthcare Procedure Code (HCPC) distribution information by specialty, but because CMS does not recognize HM as a specialty, the closest proxies are the reported distributions for internal medicine (or pediatrics). And hospitalists argue that because their patient population and the work they do are different, typical distributions for those specialties might not be applicable to hospitalists.
“Coding for hospitalists has to be different from other internists,” says SHM Practice Analysis Committee (PAC) member Rachel Lovins, MD, SFHM. “Because we take responsibility for unfamiliar patients that we hand back to other providers, our level of admission and discharge documentation in particular needs to be higher, in order to ensure excellent communication between hospitalists and PCPs.”
We finally have information about hospitalist coding practices, because both the academic and non-academic Hospital Medicine Supplements captured information about the distribution of inpatient admissions (CPT codes 99221, 99222, and 99223), subsequent visits (99231, 99232, and 99233), and discharges (99238 and 99239). Figure 1 shows the average CPT code distribution for non-academic HM groups serving adults only.
The 2011 State of Hospital Medicine report also shows how CPT distribution varied based on some key practice characteristics. For example, HM practices that are not owned by hospitals/integrated delivery systems tend to code more of their services at higher service levels than do hospital-owned practices. And practices in the Western section of the country tend to code more services at higher levels than other parts of the country.
Other factors are certainly at play as well. “Whether a physician receives training in documentation and coding can have a tremendous impact on CPT distributions,” PAC member Beth Papetti says. “Historically, there has been a tendency for hospitalists to under-code, but through education and enhancements like electronic charge capture, hospitalists can more accurately substantiate the services they provided to the patient.”
Other committee members have speculated that a hospitalist’s compensation model might influence coding patterns, with those who receive less of their total compensation in the form of base salary (and more in the form of productivity and/or performance-based pay) tending to code more of their services at higher levels. But, in fact, the survey data don’t reveal any clear relationship between compensation structure and the average number of work RVUs (relative value units) per encounter.
Interestingly, coding patterns of academic HM practices were similar to those of non-academic practices for admissions and subsequent visits, but academic hospitalists tend to code a higher proportion of discharges at the <30-minute level (99238). PAC members speculate that residents and hospital support staff might perform a larger portion of the discharge coordination and paperwork in academic centers, and attendings can only bill based on their personal time, not time spent by others.
To contribute to a robust CPT distribution database, be sure to participate in the next State of Hospital Medicine survey, scheduled to launch in January 2012.
Leslie Flores, SHM senior advisor, practice management