During a recent phone call, I was asked if Extract does NLP (natural language processing). My answer was that Extract’s methods, rules and machine learning, and NLP are similar in that both are about finding information in data and then doing something with it.
So, you're thinking of migrating to Epic's solid organ transplant module Phoenix? Or did you recently switch? The Phoenix product has come a long way since its initial release around 10 years ago. It is a great way to effectively manage your transplant population within the fully-integrated Epic suite of products. I helped to support the first 40 or so Phoenix go-lives during my Epic tenure and take great pride in the application.
Can you benefit from OCR?
Optical Character Recognition (OCR) is powerful software that transforms images such as faxes and scanned documents into human readable text. Access to this text is very powerful and can be used for many purposes. The questions below will help you determine whether or not you could benefit from an OCR solution.
Once again I had the pleasure of attending the 24th Annual UNOS Transplant Management Forum for my 4th time earlier this year. As always, it was a flurry of learning, knowledge-sharing, networking, and well-deserved awards for leaders in the industry.
It was as apparent this time as it was every time before, that the transplant community is a close-knit group who all struggle with similar things regardless of their geographical location. These struggles span across many areas, including financial, staffing, regulatory requirements, lack of organs, information technology, reporting, managing the constant deluge of paper, and many more. While I can't claim that Extract can help with all of these, there are two specific struggles that we excel at fixing: extracting discrete results from faxed external lab results and intelligently splitting, classifying, and filing large documents (such as referral packets) into patients' charts.
If you've ever managed an EMR data conversion, you likely know how painful data conversions can be. They require someone with intimate knowledge of the old EMR to write complex queries to extract the data in the format that the new EMR requires it to be in. In addition, at some point in the process you have to transform the old values into the new system's values (assuming they can be mapped at all!). Even if you have experienced, intelligent people and excellent vendor support during this process it is expensive, time-consuming, risky, and can delay your go-live. So, what if you don't have experienced people and good vendor support for your healthcare data conversion? Believe me, it's gonna get ugly.
What I know for Sure:
Discrete, trending data is the bread and butter of a specialty clinic.
Hunting and pecking through the media tab to track down information on a patient is infuriating! And not only for the doctors. For nurses. For abstractors. For the patient! Trending a post-transplant patient's drug levels alongside their medication doses, rejections, infections, transplant history, UNOS data, procedures, and relevant transplant-related scores is of paramount importance to a clinician and is very time sensitive. Getting all patient data into the EMR is the holy grail when it comes to specialty medicine.
Specialty clinics, especially transplant clinics, are mini-ACOs.
When you are treating an acute, chronic disease it is critical that everything about the patient is known regardless of where they are being treated on a daily basis. Luckily, we now live in a world of Care Everywhere, CCD documents, and reference labs…BUT, despite what everyone wants to believe, these things are not a panacea.
Paper is very much alive and well in the healthcare world.
Sometimes clinicians are "closet paper users," other times they just lay it out there. But don't make any mistake about it…they are using. In the transplant world, you may be familiar with the "wall chart." Also known as "the flowsheet" or "the flowchart." You know the one. The monstrous grid that is the holy grail for the transplant clinic, but is the disdain of the HIM team and the project team trying to migrate clinicians to the EMR. But there are good reasons for this chart and the other paper being used. Many hospitals have not implemented effective document management strategies that classify documents in useful ways. And many hospitals don't have the resources to support entering (and QAing) important data discretely as it comes in from external sources (or even internal sources such as the pathology lab).
Specialty clinics are crazy busy.
There were times during my tenure at Epic that I felt stressed. That I felt my days were busy. That I felt it was hard to create work/life balance. And then I'd go onsite and spend a week in a transplant department. Wow. My workday was like a walk in the park! The chaos that is the life of a person in a specialty clinic is very hard to explain or quantify. It seems there is not a moment to breathe. And this isn't just for the doctors and nurses. Even the folks doing data entry are getting calls, being pulled into other things, being tapped on the shoulder constantly. It is nearly impossible to give something 100% of your attention.
Extract's products can help.
I'm a passionate person. I don't back something I don't believe in and I don't work for companies whose product doesn't excite me. When I first encountered the Extract product I was very skeptical. Optical Character Recognition (OCR) with clinical data? Fuggettabout it! However, I've been able to peel back the curtain. The magic isn't in the OCR, it's in the rules, logic, and processing that Extract has fine-tuned while working with numerous healthcare organizations. I've seen it in action. I've seen the product improve with features that allow more reliable mapping to patients and existing orders. I've seen it process large documents and auto-classify subsections of that document and route them accordingly (think referral packets, transplant folks!). I've seen it work. I believe in the product and think it can improve data quality, care quality, data entry efficiency, EMR user happiness, and much more.
Extract's products aren't restricted to specialty clinics.
Yes, it is very easy to see the benefit of using the product to discretely enter lab results or split/file referral packets in a specialty clinic. But once you've seen it in action, it's very hard not to let your imagination run wild. Have an HIM department that is backlogged and needs some help classifying and discretely filing data? Have a natural speech recognition engine that needs some intelligent processing and filing after the output is generated? Have Care Everywhere but wish that you could get some more discrete data from it, such as labs? Still have paper DNR, release forms, or patient surveys coming in and want them to be discrete?
Have any other ideas?
We want to hear them! You can email me directly to discuss your ideas further.
About the Author: Rob Fea
He has spent 12 years partnering with IT teams and clinicians at major hospitals and clinics worldwide during his tenure on the technical services team at Epic. For the vast majority of his time at Epic, Rob supported Epic's Phoenix product, playing a major role in project kickoffs, installation, data conversions, ongoing support, and optimization. During his tenure at Epic, he watched the Phoenix customer base expand from 0 to 55 live and installing transplant organizations. It was a terrific experience and he loved every minute of it. It gave him expansive insight into the healthcare world, especially the solid organ transplant industry. Rob has spent countless hours on the floor in transplant departments observing multidisciplinary visits, committee review meetings, data entry, data trending, reporting, medication dosing, and more.
A consultant who supports analytics for population health and quality of care recently told me that frequently, they can only access 80% or less of the total data needed for these initiatives.
If that data is truly random and characteristic of the whole body of data, than acquiring 80% of it is pretty good, perhaps even great. But what if that 80% comes largely from one population sub-group. What if it represents patients who are local - city-dwellers who live nearby and come directly to your facility for lab work and other tests - while the missing 20% is a completely different population. Perhaps this 20% is defined differently by lifestyle, geography or other variables because that population cannot easily come to your facility?
Tracking data in your transplant care software is a key component of QAPI programs, not only for CMS, but now also with UNOS debating the requirement of QAPI programs. One of the most challenging aspects of transplant program management is ensuring that your Quality Assessment and Process Improvement programs are measuring meaningful and actionable items that lead to program improvement.
Healthcare Informatics recently published an article most of us can relate to—that if you want to remember something important, it helps to write it down. Information is only useful if available. According to the article, Research: Access to Docs’ Notes Increases Medication Aherence, researchers at Geisinger Health System found that patients who had access to their doctor’s notes demonstrated an improved adherence to a medication regimen. [medical record software and information accessibility
Yesterday, the prestigious Institute of Medicine (IOM) announced a soon-to-be-released report highlighting diagnostic errors as a persistent “blind spot in the delivery of quality health care” and urges the healthcare industry to change in order to address the prevalence of diagnostic errors, which the IOM defines as “the failure to (a) establish an accurate and timely explanation of the patient’s health problem(s) or (b) communicate that explanation to the patient.”
Tracking Data is a Key Component of QAPI Programs, not only for CMS but also now with UNOS
One of the most challenging aspects of transplant program management is ensuring that your Quality Assessment and Process Improvement programs are measuring meaningful and actionable items that lead to program improvement. There are many factors that contribute to these projects but one thing in common is that they are all data-driven.
The ultimate juggling act: clinical labs in a hospital setting are required to maintain the highest operational standards. They complete their own inpatient testing while managing the logistics of send-outs and the returning results from reference and specialty labs. No matter where it’s coming from, comprehensive data needs to get back to the ordering physician - data required to make the best care decisions.