What a Workaround Keeps Quiet

When American hospitals began using barcodes to safeguard the administration of medication, the idea made obvious sense. Before every dose, the nurse scans the patient’s wristband and the medication package. The system then checks the five “rights”: the right patient, the right drug, the right dose, the right route and the right time. What had depended on the memory and attention of a tired night shift was now to be secured by a machine.

Between 2003 and 2006, Ross Koppel and his colleagues studied how these systems were actually used day to day at five hospitals. They observed nurses, conducted interviews and analysed the logs of more than 300,000 medication administrations. They found 15 different types of workaround. One of them is so simple that it needs no explanation. Nurses attached extra copies of patients’ barcodes wherever they were easier to scan: at the nursing station, on computer carts, on doorjambs and clipboards. Some wore them on their belts or sleeves.

For the workarounds as a whole, the study lists causes, among them these: the computer carts were too large and too heavy, batteries ran flat, the connection dropped, some codes took several attempts to read, and the prescribed procedure was slower than other ways of getting the job done. The authors describe the barcode on the doorjamb as a shortcut that saves time and effort. It has one property, though, that cannot be overlooked: it disables precisely the check the system was bought for. Scanning the code on the doorjamb proves, at most, that you were at the right door. Nobody checks any more whether the right patient is in the bed.

Koppel and his colleagues state plainly that “none of these identified workarounds were malicious, and many were engendered by system design deficiencies.” Both are true at once. The workaround is sensible, and it is dangerous.

What a Workaround Achieves

That work depends on adaptation has been a theme of this magazine since its first piece. “Three Assumptions We Need to Leave Behind” argues that every functioning shift departs from the script every day, because the script never quite fits the situation. The barcode on the doorjamb shows, though, that “workarounds keep the system running” tells only half the story. A workaround always arises at a particular point: a scanner that can’t make it to the bedside, a store cupboard where the supplies have run out. That makes it a precise statement about where the system doesn’t fit the work. The real question is what happens to that statement.

What It Keeps Quiet

Anita Tucker and Amy Edmondson published an answer in 2003 that leaves little room for optimism. They shadowed 26 nurses at nine hospitals for a total of 239 hours and recorded every disruption that got in the way of their work. There were 194 of them: missing supplies, missing information, medications that had not arrived on the ward in time. A nurse on an oncology ward ran out of clean linen and took some from another ward’s supply. On average, the nurses lost 33 minutes per eight-hour shift this way.

For 93 per cent of the problems, the nurses did what it took to continue the patient-care task, “no more, no less”. Tucker and Edmondson call this first-order problem solving: you get hold of what is missing and finish the task. Second-order problem solving would mean also doing something about the cause, even if only drawing someone’s attention to the problem. Only 7 per cent of the nurses’ responses met even this lenient standard.

The authors describe the consequences matter-of-factly. “Workers rarely inform the person responsible for the problem.” As a result, “the failure, or one just like it, is likely to recur”. Sometimes the fix creates a new problem elsewhere. And first-order problem solving “can preclude improvement by obscuring the existence of problems and errors”.

This has little to do with the nurses’ character. Tucker and Edmondson name three factors, all of them located in the organisation and all of them normally regarded as strengths. The first is a professional ideal that rewards the individual who sorts everything out herself. The second is efficient staffing that leaves no time to think beyond the next task. The third is team empowerment that has moved managers so far from daily work that nobody is around to solve a problem that crosses ward boundaries. Of the twelve nurses the two researchers interviewed, 70 per cent believed their manager expected them to work through the daily disruptions on their own. “Speaking up about a problem or asking for help was likely to be seen as a sign of incompetence.” It is the same pattern as silence in teams.

For the organisation, this means it sees almost nothing of a workaround except its effect: the work got done. The message inside it does not get through. Each successful workaround even reinforces the impression that the system works, precisely because it succeeded.

Getting through is not enough, though. In the barcode case, the message did get through. Koppel and his colleagues write that the causes and the related workarounds “are neither rare nor secret”. Hospital staff reported them frequently, vendors knew of many, and the committees overseeing the systems discussed them routinely. They were “hiding in plain sight”, as Koppel and his colleagues put it, obscured among other things by faith in technology. A report that is discussed and filed changes as little as one that was never written.

What the organisation learns from a workaround is that the work got done. It does not learn which barriers were bypassed in the process.

Four Questions to Ask of a Workaround

Whether a workaround carries the system or hollows it out can’t be read from its form. Four questions help. None of them sorts workarounds into good and bad. A workaround can pass one and fail the next.

Does it only bridge a gap, or does it also disable a barrier? The linen from another ward bridges a gap and may open a new one there, but it disables no check. The barcode on the doorjamb bridges a gap too, and in doing so disables precisely the check the system exists for. The second kind is often recognised only when someone asks what the bypassed step was originally meant to do.

Do the people who rely on the procedure know that it is no longer done that way? Anyone at an interface who assumes that the other side works by the book carries a risk they know nothing about. Scott Snook calls this gradual separation of practice from procedure practical drift: locally sensible, but dangerous as soon as a crisis makes everyone dependent on a procedure that is no longer followed.

Is it a stopgap, or has it been business as usual for months? A workaround for one night is different from one that has been everyday practice for years. The longer it lasts, the more others build on it, and the less anyone remembers the problem it once solved. The Costa Concordia shows how a routine can drift onto the rocks this way.

Does the information reach someone who can fix the cause, and does anything happen there? This is the most uncomfortable question, because it is aimed at the organisation rather than at the person using the workaround. In Tucker and Edmondson’s study, only 7 per cent of responses drew anyone’s attention to the problem at all. In Koppel’s, the reports got through and went nowhere.

What an Organisation Does With It

There are two obvious reactions to a discovered workaround. The first is to ban it. That removes the workaround but not the problem that caused it, and the next one appears where nobody sees it. The second is to make the workaround the official procedure. That sounds like learning, and the result can be right if the people doing the work have found a better way than those who designed the procedure. Without checking what problem the workaround solves and what it disables in the process, however, the adaptation simply becomes a new rule, and sooner or later the barcode on the doorjamb would end up in the manual.

The third option is to treat a workaround as what it is: a report that someone has filed with their hands instead of a form. Only this way does learning happen, in the sense of Tucker and Edmondson’s second-order problem solving: the report shows whether the workaround should disappear or move into the procedure. Koppel and his colleagues state a principle: “With rare exceptions, it should be easier to follow policy than work around it.” I would go further: where that is not the case, it is the procedure that needs examining.

Tucker and Edmondson draw a conclusion for the people in the organisation. For them, the ideal employee is not the one who quietly sorts out every problem herself. It is the “noisy complainer” who tells managers and others what isn’t working. For that to work, managers have to be close enough to the work to hear the complaint and do something with it. And they have to treat it as information rather than as a sign that someone isn’t up to the job.

In that sense, the barcode on the doorjamb is one of the most honest reports a hospital can receive. It shows to the centimetre where the system doesn’t fit the work. Whoever finds it can have it scraped off. Or they can ask why the scanner doesn’t make it to the bedside.

Sources

• Anita L. Tucker & Amy C. Edmondson – Why Hospitals Don’t Learn from Failures: Organizational and Psychological Dynamics That Inhibit System Change, California Management Review 45(2) 2003, 55–72 (main source)
• Ross Koppel, Tosha Wetterneck, Joel Leon Telles & Ben-Tzion Karsh – Workarounds to Barcode Medication Administration Systems: Their Occurrences, Causes, and Threats to Patient Safety, Journal of the American Medical Informatics Association 15(4) 2008, 408–423 (main source)
• Scott A. Snook – Friendly Fire: The Accidental Shootdown of U.S. Black Hawks over Northern Iraq, Princeton University Press 2000