Teaching Machines to Talk: Edward P. Ambinder Interview, Part One

"We have international data and interoperability standards now that define every piece of medical information...whether it’s a lab report or a PET scan report or a discharge summary."

This is the first segment of a two-part interview originally published by OncLive.

Difficult as it may be to coordainate communication between physicians and patients, primaries and specialists, and departments within health system, the challenge of getting newer and more complex machines to communicate with one another efficiently and safely becomes increasingly important by the day. Edward P. Ambinder, MD, clinical professor of medicine at the Icahn School of Medicine at Mount Sinai, serves as chair of the ASCO Health Information Technology Work Group. He has a special interest in improving the interoperability of computer technology to facilitate the sharing and security of patient records for the purposes of furthering patient-centric, coordinated, and value-based care. He recently answered questions about the advances that are making this possible.

It’s often said that computer interoperability is not yet sufficient for the practice of oncology. Do you agree with this?

Ambinder: Gathering electronic medical data and planning treatment for a patient should really be no different from planning a trip on the Internet. What the Internet permits is for you to go to a travel site, identify yourself, define your preferences, and have your credit card accepted. You may want a flight somewhere, so you are connected to a database with flight information from every airline and you put in what you want. The computer knows that you are going to arrive at a given city, so it’ll ask whether you need to be met by a taxi or limo and help you to arrange that. You may also need a hotel room, so you’re connected to a site that has all the information about hotel rooms in that city.

So, basically, the problem is the way the healthcare system works with a patient’s data. Their data is on multiple different computer sites, and it’s impossible, really impossible, for patients to say “I want to have all my medical data.” And if they have laboratory tests, “I want to have access to data from the commercial laboratory, or the hospital laboratory, or my provider’s laboratory.” We’ve made it very difficult because, in the past, all of these organizations that controlled the data felt they had to control it because it kept the patient with them. What happened with governmental oversight of care and the different regulations on how doctors get reimbursed is that we were given a significant amount of money to use specialized electronic health records (EHRs) that could do everything except interoperate and simplify the sharing of data.

What’s the solution to the problem of siloed medical data?

We have international data and interoperability standards now that define every piece of medical information, every medical report, whether it’s a lab report or a PET scan report or a discharge summary. Health Level Seven International (HL7) has created architecture for defining almost every type of medical report we use. One is a consolidated Clinical Document Architecture (C-CDA), which summarizes all of your medical information in a machine-readable and human readable report. I sit on a committee at ASCO where we oversee health IT activities. We have created a CDA document to summarize the complete cancer journey of a patient who is being treated for colon cancer or breast cancer: the ASCO Clinical Oncology Treatment Plan and Summary. And we recently created a companion C-CDA document, called ASCO’s Survivorship Care Plan (SCP), to deal with survivorship care of any cancer type. In this document, we capture all of the basic clinical information necessary to understand what has happened to a particular patient during their cancer journey and how they should be monitored afterwards.

The Commission on Cancer has mandated that all cancer patients will need to have an SCP given to them that could be made available to other providers and caregivers or used for state registry reporting, rapid learning systems like ASCO’s CancerLinQ, or other research purposes. And once that’s arranged, then you have an efficient system where a patient’s primary care doctor creates a report which we call a Continuity of Care Document (CCD), a type of C-CDA, which can be sent electronically to any consulting specialist, like an oncologist, with all of the medical data that a general internist would want on the patient.

HL7 is also developing a more nimble, less expensive, more internet resource friendly, and granular architecture that can handle individual data elements, making data more available to research and discovery. It also works with existing C-CDA documents and is much easier for the technologically-shy to work with. It is known as Fast Healthcare Interoperability Resources (FHIR) that will, in the near future, become the leading data standard and interoperability enabler. The oncologist would have tools that combine FHIR and application programming interfaces (APIs). Basically, these pick up data from one computer or app and move them to another app or computer, or to API-enabled EHRs.

Those data can be brought to a health information technology app and can do wonderful things that an EHR can’t do. So, for example, if you have a patient who has genomic data on the computer of a genomic company, that particular information can be accessed by an app API and incorporated directly into an API-enabled EHR database, if everyone agrees to follow the rules of the system.

For this to work, the EHR has to be customized for the needs of cancer and to make sure that the data that’s collected is up-to-date in view of the dramatically increasing amount of data that we’re going to need to really monitor and measure a human’s “omic” data. This will overwhelm the comprehension of oncologists and their EHRs that can both barely keep up with the “classic” type of medical information that we used to scribble in our paper chart. So, we are going to be able to understand what technology can do and how standards can enable interoperability to become a seamless method of connecting anyone who has the right to receive medical information. That’s going to make a huge difference in saving time oncologists spend re-entering data. It’ll also save patients from having to fill out their medical history on paper every time they go to a new doctor.