Why PSI-90 matters now

There is a lot of confusion among coders and CDI specialists about the status of PSI-90. I hope to clear up a little of the confusion in this blog post, which is based on a recent 3M Quality Webinar I co-presented with Margaret Schmidt about the current state of PSI-90. Consider this the condensed soup version.

PSI-90 was developed by AHRQ to provide a composite quality score for institutions. In addition to measuring performance on individual PSIs, the composite was meant to focus on those PSIs with the largest volume and/or impact. In my opinion, this was an attempt to level the playing field. For example, if an organization performed well in many measures and only showed poor performance in two or three PSIs, it could appear on the surface that the institution was performing well. What if, however, the measures where they performed poorly were associated with the highest morbidity and mortality? Having a composite of the most “important” PSIs ensures that organizations can be measured on what really counts for their patients.

Another unique aspect of PSI-90 is the determination of expected rates. Again, to level the playing field, the expected rates are not determined based on a single year of data collection, but on a multi-year set of data and rates. This benefits all organizations. Case in point: If over a one-year period, a hospital had an unfortunate patient population with many fatalities---all related to the patient’s presentation rather than poor care---that wild one-year performance could sway the data. That is why CMS measures performance over a two- or three-year performance period for most of the quality measures. AHRQ similarly needs several years of data to set expected rates.

So where are we today?  With the conversion from ICD-9-CM to ICD-10-CM/PCS, the data isn't "clean." The codes do not map one to one from ICD-9-CM to ICD-10CM/PCS or vice versa. This taints the accuracy of the metrics. Therefore, CMS has shortened measurement periods to only include ICD-9-CM data or only ICD-10-CM/PCS data. AHRQ has suspended PSI-90 until a sufficient multi-year set of data can be captured and analyzed. At this point, AHRQ will set expected rates for both the individual PSIs as well as the composite.  And due to AHRQ not setting rates, CMS used a transitional, recalibrated PSI-90 performance for the current quality reporting programs and will be suspending for FY19 forward until expected rates are available.

Do note there are some quality performance organizations utilizing a reverse GEM (general equivalency mapping) process to set expected rates on current data. 3M chooses to align itself with AHRQ and CMS, hence there are no expected rates for any PSI or PSI-90 in our software.

But if PSI-90 is suspended from both AHRQ reporting and CMS quality measures, why should it matter to organizations? That is the easy question.

Right now, organizations are submitting data that will be included in the data set AHRQ will utilize to set expected rates. The healthcare community is its own comparator and means of setting the base measurement. Accurate data submissions result in more accurate or “fair” metrics to be measured against. Also, this same set of data (FY16 forward) will most likely be what organizations are measured against for quality payment programs. So there is a unique chance to not only “clean up” the benchmark data, but our performance data as well.

And it would be remiss of me not to mention the most important reason PSI-90 matters now: The patient. AHRQ PSIs were originally intended to help organizations benchmark their performance and improve patient outcomes. Changes in care delivery should only be contemplated when decision-making is based on a clean data set. If anything, organizations seeking to improve their PSI performance shouldn't just focus on improving data, but establish goals of transforming care and improving patient outcomes.

Cheryl Manchenton is a senior inpatient consultant and project manager for 3M Health Information Systems.