The author shares highlights from the annual conference.
I am going to continue my series on what you should know from the sessions I attended at ACDIS. There was just so much excellent material.
Sam Antonios did a fantastic presentation on how CDI impacts quality. He explained that some programs pay for performance and some simply pay for reporting (even if your results are abysmal), and he focused on the hospital-based programs. The first thing I learned, and I am a tad embarrassed to confess that this was a knowledge deficit of mine prior to this session, was that with many Medicare programs, we don’t technically lose money. The way they work is that a certain amount of money is withheld initially, and you have the potential to recoup some or all of the money if you meet the requisites of the program.
He taught us that Medicare uses the term, “predicted,” instead of “observed.” He then used a very understandable cupcake model to explain risk-adjustment.
I attended a talk called “Using Data to Drive Program Success,” presented by folks from Allegheny Health Network. They were able to get great physician buy-in with 65 physician champions across seven facilities and nine institutes. Their process included dedicating specialists by service line, providing posters and photo cards of the CDISs to increase visibility, rounding with providers (which was the topic of my talking the second day), attending departmental staff meetings, and establishing a hotline so providers could get assistance at their convenience.
Their metric to determine need was one CDIS FTE for every 1,800 discharges per year. They talked about their report cards and how they reported to leadership quarterly with quarterly CDIS reviews as well. They essentially select five action items until they hit their benchmark consistently, and then those items are placed on a watchlist and revisited as needed.
In the year following implementation, they reaped an increase in case mix index of 3.45 percent and a decrease in reported complications by 50 percent. The severity of Illness (SOI) similarly increased and risk-adjusted mortality was favorably influenced as well. Their methods were systematic and interesting. My only concern with this presentation was I took issue with the cases they used as examples, and I discussed it with the speakers after the talk.
My friend Trey LaCharite did a terrific talk on CDI for medical subspecialties. His general tips included use the most specific ICD-10-CM-friendly language possible; ensure diagnosis-physical exam congruity, and linkage is imperative with significant SOI implications. He convinced me that keeping a diagnosis on the assessment and plan list with the qualifier, “resolved,” can prevent downgrade or being overlooked. It is also important for the attending service to adopt more specific terminology as laid out by the specialist.
Documentation of an in-stent stenosis is considered a complication; Trey’s recommendation was to reserve this terminology of a device failure for the occurrence up to a year post-placement. After a year, an occlusion should be linked to the progression of atherosclerotic disease.
He had recommendations for other subspecialists like oncologists being attuned to pancytopenia and malnutrition, and nephrologists giving the etiology of ATN when appropriate. Trey and I agree that hyponatremia should be reserved for sodium less than 130 and hypernatremia for greater than 150. We also both shy away from organ-specific documentation for critical care – you know, when they list the category and the action plan without a single codable diagnosis. At this time, once B20, always B20, that is, AIDS. It will not surprise me if, in the future, a patient with an AIDS-defining illness who gets treatment which leaves them with persistently undetectable viral load can be downgraded to Z21.
Trey’s, and my, final recommendations were about how to get this message to them and encourage participation. Get on their monthly staff meeting agenda, round with them, show them data and help the EHR and their minions help them.
