Front Page

Next Story

NIH Record vertical blue bar column separator

Aged Pioneer To Retire — Gradually
New Clinical Research Information System Planned to Replace MIS

By Carla Garnett

On the Front Page...

In 1975, having a TV-like device that sat on the desktop and offered immediate access to hundreds of thousands of pages of patient records seemed like only a dream for the nurses, physicians, pharmacists and dozens of other professionals charged with patient care at the Clinical Center. But MIS — the computerized medical information system that made its NIH debut that year — turned the dream into a reality. Designed to collect, transmit and store information about patients, MIS was a pioneering system few other hospitals could boast having at the time. Now, however, with a PC atop every other surface one encounters, MIS — which has grown a lot, but changed only a little — seems to many users antiquated and limiting.

Continued...

"The perception is that we're way behind the curve," said Dr. Stephen Rosenfeld, chief of the Clinical Center's department of clinical research informatics and head of the group that is in search of a clinical research information system (CRIS), a modernized, more flexible cousin of MIS. "It's important for people to realize that only about 7 percent of hospitals even have physician order entry. We've had it for a long time."

Dr. Stephen Rosenfeld, chief of the Clinical Center's department of clinical research informatics and head of the group that is in search of a clinical research information system (CRIS)

Try to remember worklife before Windows, or the web, or even any common database programs. A patient, John Doe, is admitted to the CC. So begins his medical paper trail — records of X-rays, prescriptions, blood tests, and vital signs readings — that could be dozens of pages thick within a few days' stay here. What if it were computerized, however, and more easily transported from physician to nurse to pharmacist? Enter MIS, which with the blink of the lighted pen that accompanied the screen and keyboard, transformed the way medical information was managed.

A 1982 evaluation of MIS found the system "to be one of the most fully automated and comprehensive systems now operating in a hospital, and in the past, has served in the role of prototype for such systems considered or installed by other medical centers," according to an article in the NIH Record, Aug. 31, 1982.

But a lot has happened in the world of computers since the early 1980s. These days — the days of Access, FileMakerPro and other popular so-called relational database programs — patient care providers can generate all kinds of reports, depending on what questions they pose. "How many blood pressure screenings were taken within the last 7 days on patients with asthma?" a doctor could ask. "How many of those patients were over age 50?" "How many were men?" That is the kind of report the Information Age physician-researcher would like ideally, but MIS — never designed for that purpose — cannot easily produce.

Rosenfeld said that the increased complexity of clinical research has led to several reasons for phasing out MIS: The system cannot generate longitudinal data for studies conducted over time; its database is not relational, but proprietary — only its manufacturer can design how stored data can be reported and retrieved; and it has no ability to send out warning flags when errors — typos or misconstrued orders, for example — are detected. Complicating the issue is the fact that the company that originally designed MIS no longer supports upgrades of the 26-year-old product. Helping MIS adapt to the modern era of computers has been costly — in time, effort and other resources. In addition, the business aspects of patient care have never been incorporated into the operation of MIS, which was born in a different era of medical care and research.

In its heyday, MIS was "one of the most fully automated and comprehensive systems now operating in a hospital," and "served in the role of prototype for such systems considered or installed by other medical centers," according to an article in the NIH Record, Aug. 31, 1982.

"Twenty-six years ago, the cost associated with medical research was not a major factor," Rosenfeld pointed out. "That's not true anymore. We don't bill patients for their care here, but we still have to manage our resources carefully."

Rosenfeld's model of a dream CRIS looks a lot like a flower. In the center is a clinical data repository, where all patient data would be collected from various sources. The sources — pharmacy, nutrition or medical records, for example — are represented individually on the diagram by the "petals" surrounding the core. Between the repository and the sources is a ring of actions that may occur during a patient's care — doctors' orders, for example, or scheduling of procedures. All sources and actions feed into the repository. The repository, in turn, sends all of its data to a clinical data warehouse, where is it stored for the long term.

As Rosenfeld emphasized, the main benefit of such a CRIS would be its flexibility and ability to grow and change as does the clinical research it supports.

"CRIS is not meant to reduce staff power or replace people with computers," he assured. "I believe it will reduce redundancy and increase efficiency. It will allow us to do better research and make us run as a better hospital. That's the bottom line."

Under no illusion that the transition from MIS to CRIS will be problem-free, Rosenfeld explained that the months of effort and time invested in the first phases will be amply rewarded in the long run.

"The first 5 years will definitely be the most intense," he concluded. "That initial period is the critical time. You start with what is basically a shell, and you have to describe your hospital in the language of their new system. That is the really tough part. It's necessary, though. With MIS, we've developed work-arounds for 26 years. I'd like CRIS to evolve for the next 26 years."

CRIS developer Dr. Stephen Rosenfeld's model of a dream system looks a lot like a flower.

Up to Top