Skip to main content

One Hospital's $1M Savings in CDI Transcription Costs

Analysis  |  By Alexandra Wilson Pecci  
   May 15, 2018

Phoenix Children's CMIO reveals the IT-clinical collaboration behind its ambulatory disease-specific clinical documentation templates.

For many years, the ambulatory clinics at Phoenix Children's Hospital lagged in their EHR use.

"The ambulatory was a little step behind," says Vinay Vaidya, MD, vice president and chief medical information officer at Phoenix Children's. "Everything was on paper or dictation."

Now, the clinics not only use a clinical documentation platform in all 30 divisions of its ambulatory clinics, but the organization is in the process of building disease-specific templates that are part of a larger quality improvement initiative for chronic diseases.

Four years ago, the ambulatory clinics at Phoenix Children's adopted Quippe Clinical Documentation from Medicomp Systems, which delivers longitudinal patient information to clinicians using a clinical data engine containing 360,000 clinical concepts.

JIA templates

Now, the hospital is diving into disease-specific work that started with juvenile idiopathic arthritis (JIA), a complex disease to treat and document, thanks to the 71 joints that need to be monitored and many other measures about which rheumatologists must capture data.

For example, among those measures is one called "clinical inactive disease," which takes into account a number of factors, including the number of joints with active arthritis, a physician's overall assessment, and the duration of a patients' stiffness in the morning.

Unlike something relatively straightforward like high blood pressure, though, clinical inactive disease is not easy to document and track, and wasn't typically captured well.

The newly built JIA template, however, enables efficient documentation that is easy for providers to use. For example, the clinical inactive disease metric is now successfully recorded by the providers in 99% of notes, and nearly all documentation is done during office hours.

In addition, Phoenix Children's has virtually eliminated transcription in its ambulatory clinics overall, saving the organization over $1 million dollars annually on transcription costs.

Improving patient care

But cost-savings is far from the only way Phoenix Children's is measuring success.

"Well, we have achieved $1 million savings in dictation, but what have we done for patient care?"  Vaidya says.

Now, the data from those templates is, in fact, being used to improve patient care. The template not only records patient data, but also disease-specific, accurate, and detailed patient information, and features automated reminders and score calculations, patient- and population-level analytics, and tools that enable pre-visit planning.

What follows is HealthLeaders Media's conversation with Vaidya about the clinical and IT collaboration at Phoenix Children's.

HealthLeaders Media: So this template changes how you look at individual patients but also how you look at patients collectively, right?

Vaidya: That's the essence of population health.

[W]hen you treat individual patients, you lose the high-level, bird's-eye view, so you need to zoom out and look at your whole population.

For instance, three of our providers are going to see nine patients. And you can see every single patient, and with a quick, simple look, you can see who is doing well, who has had multiple visits, who has had just two visits, whose scores are going down.

[We can also see] how our entire patient population of 558 patients (which is real-time data) is doing.

We constantly zoom in at a patient level or a population level. All it takes is one click and it goes from one patient to nine patients.

HLM: Tell me about your IT and clinician collaboration.

VAIDYA: It started with the building of the templates [for each of the ambulatory divisions].

When we went live with our ambulatory EMR, we did it as a rolling go-live. We didn't do a big bang because we wanted to focus on about four or five divisions in a three-month period. It all started with meeting with each individual division.

Before we started meeting with physicians and saying, "What are you top five diagnoses?" we have their billing history from the past five years and we say, "Do these look like your top five diagnoses?" and they say yes.

We have reviewed their dictated notes. Our clinician and IT team have reviewed the notes, reviewed the literature, and then when we meet with the physicians, we say, 'What are your challenges, what are the things you want to capture and measure and improve?' and the to-and-fro exchange takes this from a draft template into a robust template that captures what they want without the chunkiness.

One of my rheumatologists said, "No EMR is going to be able to handle this, because in rheumatology we deal with 71 joints and it's going to be clutter." And we said, "Well, let's see," and this was the end result.

We are tightly joined at the hip with the physicians, as well as the disease. We don't consider ourselves as contractors where the physicians come with some specification and we build it. Because many times they do not know what is possible.

They know the clinical disease very well. So we understand the disease, we study, we attend their clinics and many times we build some things as a trial and show them [and ask], "How would this work? What about this table? Are you capturing this? Wouldn't it be nice if we auto-calculated your activity as low, moderate, severe?"

So the to-and-fro exchange makes this process a collaborative process, and by the time they're using the template, it is not something that is thrust upon them. It is a culmination of a well thought-out process.

HLM: What statistics can you share that show improvements, whether it's in documentation time, physician satisfaction, patient outcomes, etc.?

VAIDYA: [This is regarding] all 30 of our divisions. Since we went live in November 2014, we have generated about half a million documents using Quippe. And we wanted to measure the burden of documentation. [We asked clinicians]: 'Are you finalizing the document on Saturday and Sunday? Or are you doing it Monday through Friday? Are you completing your documents within 24 hours, or 48 hours, 72 hours, 96 hours, or greater than 96 hours? What portion of the documents are you completing by 5:00 pm on the day of service, the day the patient sees you?

[Overall] about 56% are completed by 5:00 pm on the day of service.

[In rheumatology] 97%–98% of the documents are completed within 24 hours, and about 86% are completed by 5:00 pm. So they are exceeding the standards and the median time to finalize their document from the time they open a note and they click final—including a coded diagnosis and billing. The median for the hospital is three hours. The median for our rheumatologists is one hour.

[Rheumatology documentation is] a complex thing, especially if you want to reach these kinds of metrics. Almost 100% of your notes have these kinds of documentation, where it's not just one checkbox for active or inactive disease. You are seeing that they are documenting the exact number of joints involved, what the provider's score was, what the parent's score was, what the overall score was, etc.

HLM: What about patient outcomes?

VAIDYA: One of the outcomes is the clinically inactive disease. It's the single most-important outcome that is measured by the national and international societies. So you want to validate this. This is a chronic, relapsing disease that doesn't get cured. I don't expect that we'll reach 90% ... We were already at an average over the last two years among 558 patients around 73%, and we want to keep on chipping at that. Can we go to 80? Can we go to 85?

We are seeing improvement—significant improvement. It is heartening to see this kind of thing. It's a difficult disease to manage…it's difficult to control. It's not just a checkbox in the note that our documentation looks squeaky clean. That is the process, but that's not the goal.

We are constantly looking at other metrics: Who has not shown up for the last six to nine months? A red line [in the tool] shows us how many days has a patient not come. A green bar—another example of squeezing the data—is telling me how many prior visits a patient has had. If a patient has come to you only once, then maybe they have gone somewhere [else], but I'm really worried about this [other] patient, who had seen us eight times and now this patient has not been there for 265 days.

As soon as we operationalize the clinic—manpower and the ability to handle this—we may be able to send automated texts to all these patients who have had three or more visits and who have not seen us for more than 180 days to as much as 260 days.

We've heard a lot of pushback about the EHR … yes, it is increased work. Yes, if you want to capture all the measures, if you want to do a good job, there are no shortcuts. But even with that we have shown that this is not increasing your time and effort. Within the clinic visit, you can really take your disease management to a new level given the right tools.

I think the message to the C-suite is that with dedication and collaboration and the right tools, just capturing a long narrative story without the key measures [isn't as valuable].

I think that with the right tools for documentation, with the right EHR in place, with the right governance and collaboration, we really can achieve what we set out to say what EMRs can do.

Alexandra Wilson Pecci is an editor for HealthLeaders.


Get the latest on healthcare leadership in your inbox.