Contemporary Issues in Medical Informatics: Good Health IT, Bad Health IT, and Common Examples of Healthcare IT Difficulties
Chaos from MIS mismanagement of a clinical information system project

Chaos from MIS mismanagement of a clinical information system project

An informaticist took a full-time position as a Director of Informatics at the healthcare organization where he obtained his Medical Informatics education. He was to lead the clinical team working on a full-blown inpatient Computerized Patient Record. The project plan included nursing documentation, full online medical records, and physician order entry. The experience the project would provide and the prestige if the project were successful were very attractive to the informaticist.

For the first 6-9 months the project seemed to go well. The clinical data repository implementation was progressing and interface specs for lab results and data coming from ancillary systems were largely done with what seemed to be a few minor issues left to resolve. The hospital had just hired more clinical analysts to start the design and implementation of the physician order entry system, and that phase of the project was underway.

Then, due to muddled thinking and basic incompetence by the hospital’s MIS leaders, the bottom fell out.

The informaticist and his clinical team actually found themselves excluded from preliminary interface testing. Significant obstacles causing long delays in clinician participation in the project started to appear. Further, numerous essential aspects of the design had been delayed. As the clinical team finally began to see a test system, there were numerous clinically-important configuration settings that had not been addressed by MIS.

The delays in implementing the interfaces to other systems (e.g., lab) ran into months. Fortunately (or unfortunately, depending on one's point of view), there was a major network infrastructure problem that had to be fixed before new production systems could come on line. (Of course, that issue should been realized well beforehand by MIS, but we digress.)

The network issue, plus the Y2K issue, gave the organization a four-month ‘reprieve’ on the planned go-live date. Despite that, major interfacing problems continued to persist. What could only be described as outright incompetence by senior MIS personnel followed, both in understanding the technical issues and their clinical significance and in establishing consensus on how to resolve them. As a collateral injury, this caused even more delays in the informaticist and clinical team being informed of (and helping provide solutions to) these issues.

MIS wasted considerable resources and time with band-aid solutions, that, when the informaticist and clinical team found out about them, made them literally groan, as they usually were unacceptable and actually clinically nonsensical. MIS would then re-think the problem and possible solutions, often with still-unsatisfactory results!

Furthermore, the "go-live" date was by edict made a drop-dead date. Senior IS management and senior health-system management, lording it over the informaticist and clinical team, gave no heed to concerns that the clinical information system was not yet "ready for prime-time." Therefore, the project team was forced to either push ahead with ridiculous solutions to major interface problems, or simply postpone some interfaces crucial for full system effectiveness.

As it turned out, the clinical information system went "live" with only about 50% of the functionality and interfaces that had been planned.

The incompetence of the vendor's on-site personnel was another major factor contributing to delays. They quite literally could not understand the concerns of the clinical team or the clinical significance of the issues, and were instigators of several absolutely idiotic "solutions" to significant interface problems. Worse, many of the so-called "solutions" would actually not work with their data repository upon testing!

In other words, the vendor’s personnel came up with solutions not just clinically absurd, but also technically unfeasible! Also, they could tell the clinical team precious little about how the database and interface designs would influence the display and functionality of the GUI, which was another huge causative factor in delays and system rework. Between MIS and the vendor, it was like the blind leading the deaf.

As is typical in hospitals, MIS (not the project committee or clinicians) controlled the budget and the MIS resources that were responsible for executing the design and implementation of the data repository. Due to this, clinical concerns were often simply ignored.

The informaticist and clinical team compiled a list of about a dozen configuration changes that affected the GUI that were essential prior to bringing the system live ("go-live issues"). These dozen issues were simple preference settings that could be made with GUI tools provided by the vendor. They did not involve vendor enhancements or new code from the vendor.

At least half of these trivial configuration changes were not made for months, despite the fact that it would have only taken the informaticist or someone even minimally technically competent about an hour to do them. Of the remaining half, several were still not done at go-live. Senior MIS management simply never put them on the priority list, and they never got done, to the severe detriment of the success and ROI analysis of the system to date. It is amazing how MIS always seems to concentrate on "process" to the exclusion of needs and results.

The escapade that finally sparked this informaticist to resign his position is as follows: At a meeting of the senior MIS and clinical stakeholders of the project (the informaticist, the Medical Director of the project, the MIS project manager, and the MIS staffers), it was decided that since the current security functions did not meet the needs of the psychiatric hospital, the initial implementation would not include psychiatric data until the vendor redressed the security deficiencies. Several other options, all of which were "hacks," and would have compromised the functionality of the system for other users, were discussed and rejected at this meeting.

In fact, no users would be able to access psychiatric data even if it were included. The team did not plan to roll out the system to users with clearance to access psychiatric data for at least a year. Thus, the lack of psychiatric data would be of absolutely no clinical significance for at least a year after go-live.

However, when it came time to meet with the psychiatric stakeholders, a mid-level MIS manager sent an email listing "options" that were to be discussed at the meeting with psychiatry. To the informaticist’s horror, they were the very same "hacks" that had been ruled out at the previous meeting!

The informaticist and clinical team quickly challenged this MIS person, and the MIS senior managers, on this rather egregious breach of a consensus decision. They requested that the MIS manager and the manager’s superiors NOT present these options at the meeting.

Despite this, the MIS manager went ahead and presented the options anyway. When the Medical Director of the project stated at the meeting that those options were not acceptable to him, the whole Clinical Information System project quickly lost all credibility with the psychiatrists, and irreparable damage to morale had been done.

The options had been presented at the direction of the senior MIS person on the project. This person had apparently made promises to the psychiatrists that the informaticist and clinical team had not been consulted on, promises that were impossible to keep. It was that day that the informaticist realized that MIS was completely unresponsive to clinical end-users, and that the Clinical Information System project was in serious jeopardy as a result of its inappropriate leadership.

The informaticist sent resumes out that day. In less than two months he had a new position at another organization.

The MIS CIO, the Medical Director, and the academic Medical Informatics director pleaded with the informaticist to stay and even offered a salary increase and the promise that MIS would have more accountability to clinical members of the project. However, one senior VP in MIS was extremely reluctant to allow any changes, and did not want to remove the senior MIS manager responsible for the psychiatry debacle. The senior VP apparently felt a sense of loyalty to this MIS manager. It would have taken a serious restructuring of this person’s position to effect real change. The position would have to report directly to the Medical Director of the project. It was clear this was just not going to happen.

The promise of change was therefore window dressing, not the substantial change needed to make the project successful, in the informaticist’s opinion. He accepted a very attractive job offer with a healthcare IT vendor. Hospitals thus lost one more informaticist and abundant intellectual capital due to inappropriate leadership and organizational structures, structures that inappropriately put MIS in the pilot’s seat of an initiative to create a clinical tool.

The informaticist had hoped his leaving would send a strong signal and that real change would result. However, he later reported that he’d heard from a close friend (another clinician) on the project that things had actually got worse starting just a few weeks after his departure.