The Future of Health Involves Human-Agent Collectives (Part 1 of 2)

Posted on February 2, 2015 I Written By

Andy Oram is an editor at O'Reilly Media, a highly respected book publisher and technology information provider. An employee of the company since 1992, Andy currently specializes in open source, software engineering, and health IT, but his editorial output has ranged from a legal guide covering intellectual property to a graphic novel about teenage hackers. His articles have appeared often on EMR & EHR and other blogs in the health IT space. Andy also writes often for O'Reilly's Radar site (http://oreilly.com/) and other publications on policy issues related to the Internet and on trends affecting technical innovation and its effects on society. Print publications where his work has appeared include The Economist, Communications of the ACM, Copyright World, the Journal of Information Technology & Politics, Vanguardia Dossier, and Internet Law and Business. Conferences where he has presented talks include O'Reilly's Open Source Convention, FISL (Brazil), FOSDEM, and DebConf.

Everyone understands that isolated interventions in the doctor’s office will not solve the chronic health conditions that plague developed nations and inflate health care costs. So as the health field shyly tries on new collaborative styles–including coordinated care, patient-centered medical homes, and accountable care organizations–participants are learning that the supporting technologies must also enable collaboration in ways vastly more sophisticated than current EHRs and devices.

EHRs are notoriously closed. Often the doctor must pay the EHR vendor a fee for every document exchanged. Even more outrageous is that the doctors often charge the patient for access to their own data, as revealed in a recent study. Medical devices often don’t communicate with EHRs, whereas consumer fitness devices communicate only with the manufacturer (who may offer an API to consumers sophisticated enough to access the data programmatically). No one is ready for the easy give-and-take of a collaborative health environment.

Although there’s a lot of talk about APIs and analytics, much more is required from the people and computers in our health care system. A cutting-edge concept developing in the computer field, sometimes known as human-agent collectives, may provide a theoretical basis for developing new tools, as well as the organizational evolution that inevitably tags along.

A Collective Responsibility

Human-agent collectives recognize that many different actors with different goals and agendas have to collaborate to coordinate their diverse and even conflicting needs. What could better describe the health landscape of providers, payers, and individuals concerned with their own health? The collectives respond to that need through sophisticated protocols for communications, negotiation, and catching anomalous behavior.

In the emerging connected health paradigm, a clinician will be interested in lowering the vital signs of a patient at risk of hospitalization, such as blood pressure and glucose levels. The patient will nominally agree to such a goal, but may want to fudge his eating and exercise habits. Another complicating factor includes fitness devices that make inexact measurements–which are fine for athletic endeavors, but render their data unreliable for the medical record. And the human factor includes medical specialists interested in demonstrating their indispensability, even though they might just add cost and encourage unnecessary interventions.

Each of these people and devices is an agent, meaning that it is independent and must be treated with more respect than a mere servant of another participant. This totally changes the emphasis in computing. Consider the assumptions behind the laptops, phones, and other computers we use daily:

  1. A person has a task he or she wants to accomplish.

  2. The computer is enlisted as a tool to help him or her reach the goal.

  3. The computer’s assignment is to carry out the task specified by the user.

In the more complicated human-agent collective, these assumptions are replaced by the following:

  1. A collection of people are engaged in a mission. The mission might not be a discrete task, but an ongoing goal such as to maintain the health of someone with a chronic condition such as diabetes or multiple sclerosis. The people may all work for an organization that assigns them a mission, or may encompass individuals with different goals and motives, including the patient himself or herself.

  2. The computer becomes not just a tool, but a mediator between individuals and often between multiple computer systems.

  3. The computer’s assignment is to convey the desires and needs of one individual to another, to harmonize goals that may conflict, to verify the user’s input, and to monitor and measure progress toward the ultimate mission.

Communication Is Only the Start

Basic communication among devices, EHRs, and PHRs is still in a primitive phase. But finally the manufacturers are making gestures toward overcoming the proprietary barriers that have kept them separate and mute up to now. It’s time to look further toward more complex types of interoperation.

Because the patient is central to any health improvement, the patient’s health record must be included in any health plan. New initiatives such as HealthKit make it easier to integrate patient and clinical data, posing device and EHR vendors with the imposing question of whether they can accept a common standard for patient data. At the same time, HealthKit uses a proprietary standard, which is less conducive to ubiquitous health care than an open standard.

Regarding devices, extensive consultation among device manufacturers, health care providers, and the FDA created the Integrated Clinical Environment (ICE) standard, and an open source implementation from MD PnP has demonstrated that it is more than an aspiration. OpenICE is based on many underlying standards from IEEE and other major standards organizations. No EHR vendors are using OpenICE yet, a significant barrier to hopes for interoperable health systems.

The next part of this series will range into more experimental ground, with devices that adapt to their collaborators and can handle bad input.