Proof that high CPOE adoption reduces length of stay
Holy Spirit Hospital (Camp Hill, Penn., U.S.A.) wanted to know, “Does computerized provider order entry (CPOE) really make a difference?”
There are some obvious answers about the benefits of CPOE. When clinicians and pharmacists communicate electronically, they’ll reduce costs and errors associated with transcription and paper.
But Holy Spirit wanted to take it a step further: Does CPOE improve the overall measures of efficiency and quality of care? After conducting a rigorous study, the answer is a resounding yes, and is the subject of a recently published article in Applied Clinical Informatics.*
How we measured the correlation between CPOE and improved clinical outcomes
Length of Stay (LOS) is among the best proxy measures for overall efficiency and efficacy of care and clinical outcomes in hospitals. So we decided to study whether or not CPOE adoption correlates with LOS, which means we can assign causal attribution.
We gathered pre- and post-implementation data to quantify the impact of CPOE adoption rates on LOS. Prior to implementing Sunrise solutions from Allscripts in Holy Spirit Hospital, we gathered baseline data to represent the pre-CPOE state of operations. Then the organizations followed a pre-defined, shared protocol for quantification of LOS and CPOE adoption rates.
We analyzed the collective impact on the organization, but we also took into account the difference in LOS rates by specialty. For example, obstetrics specialties have predefined LOS for delivering mothers and their infants, which minimizes the impact of CPOE on LOS. However, for the majority of specialties, LOS has room for improvement.
We contrasted one year of pre-CPOE data with two years of post-CPOE data. That’s hundreds of thousands of orders and a significant sample size. We looked at the data at the organizational level and by 15 separate disciplines.
Results show LOS improvement across the board
The study found that CPOE and the corresponding changes in LOS were statistically significant for each organization and in every discipline—even obstetrics. In other words, as more clinicians used CPOE, length of stay decreased:
As you can see from the above chart, there’s a certain point of adoption that makes a huge impact on LOS. This inflection point hits at about 58.7% adoption—that’s when more benefits of CPOE materialize.
You might be wondering if the reduction in LOS is because the patient population was “getting easier.” To account for that, we measured the clinical severity of the patient population by looking at the CMI (case mix index).
We found the CMI actually went up during the course of the study, meaning Holy Spirit was addressing the needs of a more resource-intensive and clinically challenged patient population. So, not only did Holy Spirit implement CPOE, but patients were getting sicker and Holy Spirit was still able to reduce LOS.
This study is arguably the first to rigorously study and verify a predictive relationship between increased CPOE adoption and decreased LOS in a mature electronic patient record (EPR) environment. We can take away these key points:
- High-capability EPRs with CPOE can improve clinical results across specialties.
- Real benefits materialize when CPOE adoption reaches 60%.
Can you get these results with just any CPOE? The answer: no. We only see these kinds of results when CPOE lives within a mature or advanced clinical decision support environment. LOS and related costs decrease when the EPR assists clinicians in making the best, real-time clinical decisions with the most cost-effective impacts.