Estimating Risk Associated with Medical Device Malfunctions: FDA’s Warning letter to Medtronic Highlights an Under Appreciated Potential Source of Error in the Estimation

One of the more complex areas of device regulation is the management of risks associated with malfunctions reported from the field.  A critical task in this situation is for a firm to estimate the risk posed by the issue as accurately as possible.  This estimation of risk is important because it drives both decision making and regulatory compliance.

A December 2021 warning letter to Medtronic, Inc. (Medtronic Warning Letter) provides an important reminder from FDA as to the correct estimation of risk associated with malfunctions.  The warning letter is long and covers many issues associated with the management of risks associated with malfunctions and complying with the complex web of regulations potentially applicable when malfunctions occur.

Our focus, however, is on FDA’s criticism of Medtronic’s calculation of risk.  As background, Medtronic manufactures the MiniMed 600 series insulin infusion pumps.   During an inspection, FDA learned that Medtronic had opened a CAPA to address complaints of damaged retainer rings in these pumps (more than 74,000 between June 2016 and November 2019, with more than 57,000 reported to FDA as MDRs).  An internal investigation apparently had determined the device failures were caused by a malfunction involving the retainer ring.  According to FDA, a “damaged retainer ring may result in the over or under-delivery of insulin, which may lead to hypoglycemia or hyperglycemia.”

In Medtronic’s quality system, as FDA notes:  “Your CAPA System procedure (SOP114-01DOC, Effective Date 5/19/2015, Version L) requires your firm to determine whether distributed product is potentially affected by a nonconformance and to conduct additional risk assessment activities per the Product Risk Management Process (SOP104- 08, Version W).”  FDA acknowledged that Medtronic followed the procedures but disagreed with the assessment of risk.

Specifically, FDA criticized the risk calculation formula because it used the total shipment of affected product, which “underestimates the probability of occurrence because the number of products shipped includes devices not in use by patients (e.g., devices shipped to distributors that have not yet been distributed to customers).”

Is FDA’s criticism valid?  After all, it is fairly common for firms to use shipment (or sales) as a denominator in a risk estimate.  It would seem, though, that FDA is right that total shipments of affected product could potentially overstate the number in actual use by patients, thereby underestimating the risk.  Of course, it would not always be true that there is a significant gap between the number of medical devices shipped and the number in actual use.  Some products may move with high velocity to patient use.  Nonetheless, firms would be wise to at least consider the issue when calculating risk.

One last nuance:  As FDA noted toward the end of the warning letter, there was such a large gap between the number shipped versus the number in use (according to Medtronic’s own calculation) that the number held by distributors might not account for the entire gap.  Whether true or not in that case, the lesson here is that an estimate of the delay in reaching the patient should consider all possibly contributing factors and not simply time in possession of distributors.  For instance, some products could remain in boxes in hospital inventory for an extended period.

In a nutshell, when estimating risk of harm from a malfunction, be sure to consider whether the total shipments of affected product is an adequate proxy for the total number in use!  Firms should update their procedures to include this consideration.