As a student, I worked for five summers on the assembly line at the Honda plant in Alliston, Ontario to keep my student loans to a minimum. It was a great life experience.
Thousands of Honda employees are involved in one way or another in the manufacture of each vehicle. I worked in Plant Two Final Assembly, building vehicles like the Honda Pilot, Odyssey, Ridgeline, and the Acura MDX. I installed parking brakes, seatbelts, roof liners, bed liners and a whole mess of screws, bolts and other hardware.
Sometimes I had little clue what I was installing, only that I had about 55 seconds to install it. And if I made a mistake, I’d be hearing about it pretty soon.
Drawing lessons from manufacturing
Since graduating, I’ve found a career in health care improvement, where I continue to draw lessons from my stint on the Honda line.
Much has been written about what health care can learn from manufacturing when it comes to quality improvement (QI): Lean (for eliminating waste) and Six Sigma (for process improvement and reducing defects) were developed by the manufacturing and engineering sectors, and have both been applied extensively in health care.
Yet one area where health care has been slow in applying manufacturing’s QI wisdom is post-market surveillance, particularly for medical devices.
While we routinely collect data on procedures, we do not track the brand or model of any implantable devices or prostheses in a systematic way.
By contrast, if a car company learns of a part defect, assembly error or safety issue, it can scan its database and know quickly and cheaply which vehicles may be affected. Whether it uses that information responsibly is a separate issue, but the point is that they have the information.
A recent example is the Takata airbag recall, the largest automotive recall in US history. Once the defect was identified by Takata, auto manufacturers that used the part were able to issue targeted recall notices to owners of affected vehicles and limit the potential impact of the defective part.
Real world evidence
In health care, we need to start thinking about the untapped potential of “real world evidence.” This, too, is something for which manufacturing is well known. Meticulous collection of data in the pre-market space (such as logistics, efficiency, safety and personnel) coupled with data from post-market surveillance in the real world is perhaps the most powerful QI tool in the manufacturer’s toolkit.
As a worker at Honda, it was never my job to figure out how to use the data, but data entry of some sort was built into most work processes. Sometimes data entry was manual, but to the extent possible it was automated to make it less burdensome. For example, each vehicle and part are assigned unique identifiers (numbers, barcodes, etc.). At the beginning of each process the vehicle and part-to-be-installed are scanned by the worker, linking them in Honda’s database.
The importance of these data cannot be understated: if an issue arises after a vehicle has left the lot, these data can be used to track exactly what part, from what lot, was installed by whom, in what vehicle, and at what time. Handy information in the event of a recall – to say nothing of the research potential.
It’s hard not to see the lesson here for health care.
Writing on the Health Affairs Blog about Essure, a permanent contraceptive device associated with harmful and sometimes serious side effects, researchers from the Geisinger Medical Center in Pennsylvania offer an example of how valuable real world data can be.
Since Essure has been the only device associated with a specific procedure code for the majority of its time on the market, researchers were able to examine the health records of their patients and, in a matter of hours, identify who had the procedure code in their health record — and hence the Essure device — and whether they potentially had unwanted side effects.
This is typically not possible for most other medical devices, which are coded only by procedure, not brand/model. While the Essure data are not perfect, they are a step in the right direction.
Data like these can be used in novel ways, by researchers reporting on the real world effectiveness and safety of medical devices, and by health officials needing to notify affected patients of potential risks/adverse events (discovered through research and surveillance) and what to do about them.
Opportunities, challenges
Fortunately, the US is embarking on a movement to build a National Medical Device Evaluation System, and to expand the use of unique device identifiers (UDIs) (which would allow for the routine surveillance possible in the Essure example). Draft guidance on how to use real world evidence to support regulatory decision-making was also recently released.
Piggybacking on these efforts in Canada would present both challenges and opportunities. Challenges might include building a system to alert patients to device safety concerns, likely requiring access to patient-level data and surmounting privacy barriers. Device manufacturers would also need to collaborate with providers and patients to create a complaints process that would identify device safety concerns.
An opportunity — indeed an advantage over the US, given their decentralized claims databases across multiple payers — might lie in adding UDI as a data element in Canadian Institute for Health Information national databases, such as the Discharge Abstract Database (DAD) and National Ambulatory Care Reporting System (NACRS).
These efforts could lead to meaningful, data-driven quality improvement in the medical devices field, a sector of growing importance to modern health care systems. It would be yet another prime example of how health care can shamelessly adopt successes from other industries.
Stephen Petersen would like to acknowledge Dr. Irfan Dhalla and Erik Hellsten for their feedback on earlier drafts of this essay.
The views expressed in this publication do not necessarily represent the views of Health Quality Ontario and should not be construed as representing an official position.
The comments section is closed.