Issuing a recall alert is always a fraught decision for any medical tech firm, but it is sometimes unavoidable.
In 2020, Medtronic recalled over 300,000 of its MiniMed 600 series insulin pumps after discovering a fault that could give the user the wrong dose of medication.
The FDA identified the event as a Class I recall, which suggests a fault could cause serious injuries or death. More than 2,000 people were reportedly injured, and one person died because of the faulty devices. The firm received over 26,000 complaints related to the pumps.
George Ball from Kelley School of Business at Indiana University knows how anxiety-inducing recalls can be – and the negative effect they can have on corporate costs, stock prices and a company’s reputation. Before his academic career, he spent over a decade as a manufacturing manager in global medical device firms AGA Medical and Guidant.
“In those roles, I experienced recalls very personally,” he reveals. “It was probably the worst part of the job.”
Since exiting the industry, Ball has authored several papers on the causes and effects of medical recalls, which have soared over the last decade. And they are going to get even more widespread as medical devices become ever-more complex, he says. “The competition is pushing firms to make devices smaller and more advanced.”
But the more complicated the device, the more serious a simple error can be. Ball believes the increasing reliance on contract manufacturing in the medical device sector may also result in more recalls.
“Sometimes you can have a supplier that will make what they think is a harmless process change to a piece of equipment or raw material,” Ball says. “If they don’t convey that properly to the manufacturer, they might not realise it can have an impact on patient safety.”
Most medical product recalls are not mandated by the FDA but voluntarily issued by the firms themselves, which means managers have a high level of discretion in the decision. “Every company faces failing devices at different levels,” says Ball.
“Sometimes they don’t require a recall and sometimes they do, but that decision is a fascinating thing to explore as an academic because there are all sorts of biases that get in the way.”
One study looked at healthcare professionals who purchase medical devices on behalf of patients. Ball and his colleagues found if medical tech firm managers know the doctors are likely to detect a defect in the device before using the product on the patient, the company is less likely to recall the product.
This might be because they trust the physician to screen out the detectable problems. Medical firms can now train decision-makers to be aware of this unwanted bias.
Another project looked at gender diversity in a medical firm’s senior leadership team. Kaitlin Wowak from the University of Notre Dame and Ball found that when a recall is needed to prevent a life-threatening situation, medical companies with female directors issue recalls more quickly than firms where men are in charge.
For less severe instances that require greater discretion, recalls occur more frequently for companies that have women on their boards. Ball says this suggests a more diverse team leads to a more responsive decision-making culture.
“Setting up these decision-making meetings where recall decisions are made in a way that ensures there’s a diversity of viewpoints and perspectives is really important,” he says. “You need to make sure every voice is heard.”
How to predict a recall
As well as learning what might motivate firms to pull the plug on a medical product, predicting recall events before they happen would be helpful for the industry.
“Recalls may be inevitable, but the impact can be mitigated by being aware ahead of time,” says Arshad Rahman, general manager, life sciences at Reed Tech. If you know which products are more at risk, focusing more of your attention on those devices could prevent some future recalls.
“By using predictive recalls methodology, manufacturers can have improved insight for proactive actions to protect patient outcomes.” He adds that as more medical devices enter the market, this will correspond with an increase in recalls.
One way to determine the safety and quality of a medical device is to look at historic data such as adverse events and recall reports. But it can be hard to extrapolate this information to determine what increases the probability of a future safety event.
Reed Tech’s new platform Navigator aims to offer a crystal ball for recall events. It leverages data from the FDA and several other public sources to train a machine-learning algorithm to spot medical products that may be recalled in future.
Some devices are more likely to fail than others, suggests the Reed Tech tool. Orthopaedic, radiology and cardiovascular products appear to have a much higher risk of recall than other devices on the market.
More than 22% of all products from these categories were predicted to be at risk of recall, compared to a value below 14% for devices in other categories.
“Manufacturers can ensure safety and quality before and after going to market by spotting potential risks as they occur,” says Rahman, who believes Navigator will help multiple stakeholders make decisions to improve patient safety, mitigate reputational risks and guide policy decisions.
Insurers and hospital systems, for instance, could spot trends that may impact product liability using the data to compare the safety profile of different devices.
Ball thinks predictive tools could be a great help to the medical technology industry for spotting failures ahead of time. “There’s no question that you’re better off using something than nothing,” he says. He reckons medical product recalls will continue to rise, impacted in part by the Covid-19 pandemic.
This life-changing event has left most people anxious and distracted, so it’s possible workers will make more mistakes than normal – and that includes medical manufacturing. “I wouldn’t be surprised if there’s a systematic downward shift in product quality which would lead to an increase in recalls based on Covid,” he predicts.