The Quality Collapse: The Aftermath

The Seven Pillars of Data Management, a novel in seven acts.

Chapter II, The Quality Collapse. Pillar II: Data Quality. Part 5 of 5.

The Price

The class action settled eighteen months after it was filed. The terms were not disclosed publicly, but court filings indicated a settlement fund of $189 million, the largest medical malpractice settlement in Ontario history. Of that amount, $62 million was allocated to direct compensation for patients who had suffered documented adverse events attributable to data quality failures. $47 million was allocated to the establishment of an independent health data quality oversight body for the province of Ontario. $38 million was allocated to the remediation program itself. $26 million was allocated to the establishment of data quality training programs for health information professionals across Canada. The remainder covered legal fees, administration, and a reserve fund for future claims.

Dr. Constance Armitage resigned as CEO of Lakeview Health Alliance on the day the settlement was announced. She did not give a press conference. She issued a statement of three paragraphs in which she accepted responsibility for the failures that had occurred under her leadership, expressed her commitment to the remediation program she had authorized, and wished the staff and patients of Lakeview Health Alliance well. She did not mention her twenty-seven years in medicine. She did not mention the cardiac surgeries she had performed, the lives she had saved, the residents she had trained. She did not mention any of this because she understood, as only a physician can, that the measure of a career in medicine is not the number of lives you save but the depth of your commitment to the principle that every patient deserves care that is safe, and that safety begins with truth.

The truth, at Lakeview, had been corrupted. And Constance Armitage believed that someone should answer for it.

The Physician, Revisited

Dr. Amir Hassani returned to clinical practice in April of the following year. The College of Physicians and Surgeons had cleared him of any wrongdoing in February, issuing a decision that stated, in the restrained language of regulatory adjudication, that Dr. Hassani had “acted in accordance with the standard of care by consulting the electronic health record prior to prescribing” and that “the data quality deficiency in the patient’s record was not reasonably discoverable by the prescribing physician through routine clinical means.”

He was exonerated. The word was correct and insufficient.

He returned to the Bank Street clinic, to the same examination room where he had seen Thomas Beaulieu, and he resumed seeing patients. But something had changed in him: not his competence, which had never been in question, but his relationship with the systems he was asked to trust. He now checked every allergy record twice. He asked every patient, verbally, about their allergies, even when the record was clear. He developed a habit of pausing before writing a prescription, a pause of perhaps three seconds, during which he reviewed not just the clinical indication but the provenance of the data he was relying on. His colleagues noticed the pause. Some thought it was anxiety. Some thought it was excessive caution. One colleague, a senior physician who had practiced for thirty years, understood it for what it was.

“That’s not hesitation,” she told a resident who had asked about it. “That’s a man who learned the hard way that the system can lie to you. He’s not being cautious. He’s being wise.”

The Boy

Thomas Beaulieu turned five in December. His birthday party was held at his grandparents’ house in Aylmer, across the river in Quebec, in a living room decorated with blue balloons and a banner that his grandfather had painted by hand: “Bon Anniversaire, Thomas!” The cake was chocolate, his favourite, and Capitaine the rabbit sat in the place of honour at the head of the table, propped against a juice box.

Thomas wore a medical alert bracelet now. It was small and blue, chosen specifically because blue was his favourite colour, and it bore the inscription: “Penicillin Allergy, Anaphylaxis Risk.” He did not fully understand what it meant. His mother had explained it to him in the way that parents explain the complicated cruelties of the world to young children: simply, honestly, and with the promise that the people who loved him would keep him safe.

He understood enough. He understood that there were medicines that could make him very sick. He understood that the bracelet told doctors about this. He understood that he must always wear it. He accepted these facts with the pragmatism of a child who has survived something terrifying and has decided, with the resilience that children possess and adults envy, to keep living.

What he did not understand, what he would not understand for many years, if ever, was that the bracelet existed because an institution had failed to do the one thing that should have been non-negotiable: ensure that the information in his medical record was true.

Not secure. Not private. Not compliant. True.

True in the way that a blood type must be true. True in the way that a dosage must be true. True in the way that the word “allergy” must mean allergy and the word “none” must mean none and the space between those two words must never, ever be filled by a system error, a migration shortcut, a budget decision, or the institutional assumption that data quality is someone else’s problem.

Data quality is not someone else’s problem. It is the problem. The first problem. The foundational problem upon which every clinical decision, every treatment plan, every prescription, every diagnosis, every act of care depends. Without quality, data is not an asset. It is a weapon.

Thomas blew out his candles. He wished for a dinosaur. The room erupted in applause. Isabelle held Marc’s hand under the table and closed her eyes and thanked a God she was not sure she believed in for the fact that her son was alive to make a wish at all.

Epilogue

The Lesson

Two years after the incident, Evelyn Chow was invited to present at the Canadian Health Informatics Association’s annual conference in Montreal. The audience was 600 people: CIOs, CMIOs, health informaticists, clinical leaders, and a delegation from the Ontario Ministry of Health that included the Assistant Deputy Minister responsible for digital health.

Evelyn stood at the podium in the Palais des congrès, a building of glass and colour that sits where the old city meets the new, and she told the story of Thomas Beaulieu. She told it completely, without redaction, because she believed that the only thing more dangerous than a data quality failure is the institutional instinct to minimize it.

She told the audience about Gerald Foss’s memo, the $12 million request that would have funded a comprehensive validation program, rejected because the budget was tight and the system was going live and there were quarterly targets to meet. She told them about the 347,000 records flagged for manual review and the twelve-person team that could not possibly review them and the merge-and-monitor policy that substituted hope for rigour. She told them about the allergy field that displayed “No Known Allergies” when the truth was “We don’t know,” and the distance between those two statements, which is the distance between safety and catastrophe.

She told them about Amir Hassani, a gifted young physician who did everything right and was punished by a system that had set him up to fail. She told them about Isabelle Beaulieu, a mother who sat in a yellow room for four days while her son fought for his life because an institution had confused “operational” with “trustworthy.”

And she told them about the seven steps, the framework that had taken Lakeview from a data quality score of 62 percent to 97 percent, from a system that nearly killed a child to a system that was now cited by the provincial auditor as a model for health data management.

She closed with seven sentences. One for each step. One for each dimension of a problem that she believed every healthcare organization in the country was either facing or pretending not to face:

“Define what quality means. Not in general terms. In specific, measurable, enforceable terms that your clinicians can rely on and your patients deserve.”

“Profile your data before you trust it. Run the scan. Face the results. The number you find will be worse than the number you assumed, and that delta is the measure of your institutional self-deception.”

“Establish rules and enforce them at the point of entry, not after the harm has been done. Every field in a medical record is a promise to a patient. Keep the promise.”

“Cleanse your data as though lives depend on it. In healthcare, they do.”

“Make quality visible. Put the numbers on the wall. If your organization cannot see its data quality, it will not fix its data quality. Transparency is not optional.”

“Trace every failure to its source and put a name on it. Systemic is not a synonym for blameless. Someone made a decision. Someone deferred a cost. Someone approved a shortcut. Find them, not to punish, but to learn.”

“And monitor. Continuously. Permanently. Because data quality degrades the way muscle degrades: slowly, silently, and then all at once. The day you stop measuring is the day the next Thomas Beaulieu enters your system.”

She stepped back from the podium. The room was silent. In the front row, a woman Evelyn did not recognize was crying: silently, without shame, the way people cry when they are not mourning something that has happened but confronting something they have allowed.

The applause, when it came, was not the polished applause of a conference audience responding to a polished presentation. It was the raw, uneven applause of people who had been shaken. People who ran systems like CareConnect. People who had signed off on migrations like Lakeview’s. People who had, at some point in their careers, accepted a data quality risk because the budget was tight or the timeline was short or the steering committee wanted the project closed.

People who understood, in that moment, that the distance between their institution and Lakeview was not competence or intention. It was luck.

And luck, in healthcare, is not a strategy.

End of Chapter Two