NHS Improvement issued a patient safety alert on medical devices in January 2018. It highlighted 400 incidents - including six deaths - over three years that involved the incorrect operation of oxygen cylinder controls. This investigation reinforces that alert and makes further safety recommendations.

Investigation summary

The reference event in this investigation is an 83-year-old man who failed to receive oxygen from a portable system whilst having cardiopulmonary resuscitation (CPR) for approximately 10 minutes. The lack of oxygen is thought unlikely to have affected the outcome for this man, but it was recognised that oxygen is widely used throughout the NHS and we found clear evidence of an ongoing risk to patient safety. 

This investigation focuses on:

  • Reviewing how the design of portable oxygen systems is regulated by the Medicines and Healthcare products Regulatory Agency (MHRA).
  • Reviewing the design of portable oxygen systems used in other industries to determine if there are appropriate lessons for healthcare.

The investigation also looks at the design and regulation of medical devices in general, in particular looking at how people interact with systems and devices (referred to in the report as human factors) in both the pre and post-market regulatory stages. 

Download and read the full report.

Safety recommendations

We made four safety recommendations to the Medicines and Healthcare products Regulatory Agency (MHRA) as a result of this investigation. We received their response to our recommendations within 90 days of publication of the investigation report.

It is recommended that the Medicines and Healthcare products Regulatory Agency evaluate how its Human Factors guidance document is used in practice by manufacturers and by Notified Bodies. Based on the review, the MHRA should make any changes necessary to the document or use other mechanisms to improve the implementation of Human Factors in the pre-market approval process.

It is recommended that the Medicines and Healthcare products Regulatory Agency require oxygen manufacturers to submit evidence of Human Factors summative testing of the complete product as part of the market authorisation process for medicinal licence.

It is recommended that the Medicines and Healthcare products Regulatory Agency review its documentation to determine whether more specific guidance is required on how to incorporate Human Factors into post-market adverse event investigations.

It is recommended that, when reviewing manufacturers’ Field Safety Notifications, the Medicines and Healthcare products Regulatory Agency discourages the use of weak barriers as defined in ISO 14971 (Risk Management for Medical Devices) particularly as long-term solutions.

Response from the Medicines and Healthcare products Regulatory Agency

Combined response to these recommendations and safety recommendations as a result of our Insertion of an incorrect intraocular lens investigation, received 5 February 2019:

These recommendations overlap with each other and so I [Dr Ian Hudson, Chief Executive, Medicines and Healthcare products Regulatory Agency] intend to respond to all of them together. Please also note your report, I2017/010, Implantation of wrong prostheses during joint replacement surgery, didn’t specifically direct recommendations for us, however it did prompt actions by us relevant to the response to these recommendations.

Overall, we agree these are good recommendations and are areas we have also identified where further systemic safety improvement effort should be directed. We intend to implement them as fully as we can, within the constraints of our regulatory abilities. We supplied a considerable amount of information to your investigators during these investigations and I am confident this contributed to the recommendations being as well targeted as they are.

We think the recommendations you have made are logical conclusions from the specific investigations you have undertaken. However, we think these are examples of wider problems, and so we intend to implement improvements to the whole category of problem we identify, rather than concentrate on the specific examples you give. For example, we suspect usability of all medical devices can be an issue and following our human factors guidance would help improve all medical devices, so we will work to improve this whole issue rather than focus on improving it just for ophthalmology electronic patient record systems that include an ocular biometry module that is a medical device. In support of this, the usability and human factors design of labelling of orthopaedic implants was identified as an issue by your report on that area.

I will set out some of the things we have done already and have planned, indicating where they contribute to enacting your recommendations:

- In September 2017 we published Guidance on applying human factors to medical devices. This is aimed at all manufacturers of medical devices, Notified Bodies and users of devices. It was promoted at the time with a news story.

- We have planned to evaluate how this guidance has been received (recommendation 2018/021)

  • We had identified usability of labelling as an area of interest and report I2017/010 helped us focus this on labelling of orthopaedic implants. We planned a study of these labels to see whether they follow our guidance, have started gathering data for that study and plan to analyse and publish the results by the end of 2019.
  • We intend to expand our evaluation to include other methods such as a survey of Notified Bodies and manufacturers to ascertain awareness and use of the guidance. This is planned to occur during 2019.
  • This evaluation will include whether sufficient weight is given to post market surveillance of devices within the guidance (recommendation 2018/023) and whether the strength of our recommendation that this guidance be used is understood (recommendation 2018/014).

- Further actions, such as specific action on recommendation 2018/14, depend on the results of this evaluation. Therefore, we will plan and take action as necessary after we have completed the evaluation and have a clear picture of what action would be most helpful.

- We consider we have made good progress toward recommendation 2018/024 because it is already widespread practice to consider whether field safety corrective actions planned by manufacturers are likely to be effective, and weak barriers are likely to be ineffective. We have limited power here because the field safety notice and corrective actions are determined by the manufacturer of the affected device and our role is supervisory. We also plan to make further progress on the related point of actions that we recommend in our own publication of Medical Device Alerts. The procedure for MDA writing is due for revision in September 2019, and during this review we will consider whether an extra step to explicitly consider strength of actions recommended is needed.

- We consider we have met recommendation 2018/022. For the last two years, such requirement has been implemented in new marketing authorisation applications for drug-device combination products that are regulated as medicinal products. Results of summative human factors and usability testing studies, or other evidence of usability for such products, are actively requested for review during the assessment process. For example, for oxygen cylinders with integral valves the usability study of the CE marked integral valves is reviewed to ensure that the study is appropriate with respect to the medicinal product as a whole. If it is not sufficient further results of usability testing are requested to support the use of the integral valves with the oxygen cylinders under assessment. We have not received any applications for oxygen cylinder and integrated valve combinations in this time.