Human factors/ergonomics to support the design and testing of rapidly manufactured ventilators in the UK during the COVID-19 pandemic

Abstract Background This paper describes a rapid response project from the Chartered Institute of Ergonomics & Human Factors (CIEHF) to support the design, development, usability testing and operation of new ventilators as part of the UK response during the COVID-19 pandemic. Method A five-step approach was taken to (1) assess the COVID-19 situation and decide to formulate a response; (2) mobilise and coordinate Human Factors/Ergonomics (HFE) specialists; (3) ideate, with HFE specialists collaborating to identify, analyse the issues and opportunities, and develop strategies, plans and processes; (4) generate outputs and solutions; and (5) respond to the COVID-19 situation via targeted support and guidance. Results The response for the rapidly manufactured ventilator systems (RMVS) has been used to influence both strategy and practice to address concerns about changing safety standards and the detailed design procedure with RMVS manufacturers. Conclusion The documents are part of a wider collection of HFE advice which is available on the CIEHF COVID-19 website (https://covid19.ergonomics.org.uk/).


Introduction
The COVID-19 pandemic has led to a massive demand for Intensive Care Unit (ICU) facilities, with healthcare providers working to increase the surge capacity of hospitals. To respond to the anticipated demand, the UK Government called for UK manufacturers to increase the number of available ventilators through a process of rapid manufacturing [1]. There were specific challenges, including manufacturers with little experience of healthcare or ventilators, a trade-off between regulatory control, international standards, rapid manufacturing and design for users with less experience of using ventilators.
In the UK, National Health Service design has been accepted as an important component in patient safety since the 2000s [2]. Internationally, a Usability and Human Factors/Ergonomics (HFE) standard for medical device development was established in 2007 [3] and was adopted in the UK in 2017 to address the 'errors in use leading to patient harm … Such errors may be due to poor device design, particularly where a complex user interface is involved. Medical devices, such as infusion pumps, ventilators, … are recognised as potentially having use-related design issues that can result in problems' [4]. To support the call for Rapidly Manufactured Ventilator Systems (RMVS; [5]), the Chartered Institute of Ergonomics & Human Factors (CIEHF) produced guidance to help and support manufacturers through the requirement for formative usability testing. It was 'accepted that full demonstration of compliance to ISO 80601-2-12:2020 is unrealistic in the time frame required for development' and that when 'the current emergency has passed these devices will NOT be usable for routine care unless they have been CE marked through the Medical Device Regulations' [5].
This paper describes the response process by the CIEHF to develop rapid advisory guidance documents, which was circulated by the UK Government to all RMVS manufacturers. Figure 1 provides a representation of how the CIEHF responded to COVID-10 for the design of ventilators and other projects (https://covid19.ergonomics.org.uk/).

Principles of HFE in ventilator design and operation
The first rapid project provided guidance on basic HFE principles ( Figure 2). The aim was to support RMVS manufacturers with a structured, yet simple, process for the design of the user interface and instructions for use, and with the development of training based on consideration of users and use environment, the tasks and the associated risks [6]. Each principle was explained clearly using plain language and key learning points. This was followed with a more detailed protocol for usability testing, including patient profiles and clinical test scenarios [7].

User interface
It was recommended that, where possible, the new ventilator designs should be aligned to existing designs to support existing operational mental models, allow rapid learning and reduce use errors. The user interface should be intuitive with buttons/controls spaced to minimise accidental operation. There should be informative feedback to users, which is informed by a risk analysis to identify any required warnings or alarms for critical steps and/or unsafe situations. Alarm design should consider the environment(s) of use and be audible in a noisy critical care environment, the potential for alarm fatigue due to multiple alarm systems, as well as lighting at different times of day/night [8]. Generally, if a situation does not require a user action, an alarm should not be used but should instead just display information indicator (feedback). Generic heuristics for interface design quality included consistency of the layout (e.g. colour-coding), transparency about device status and reducing the number of items a user needs to remember.  The physical design recommendations included ensuring that physical connectors were easily recognisable and worked across settings. To design for relocating the ventilator, the weight should be considered, with easy repositioning/adjustments to avoid musculoskeletal health risks to staff (including the adjustment of screens and displays). Retractable cables could reduce trip hazards in the bed space and for storage.
To reflect the different use during the COVID-19 pandemic, it was recommended that manufacturers design interfaces for users wearing personal protective equipment (PPE). This includes eye cover (safety glasses, safety goggles), face cover (surgical mask, face visor), body cover including surgical gowns (with and without sleeves), plastic aprons, one-piece disposable protection suit (and possibly a full gas-tight protection suit) and hand cover with two layers of gloves sticky taped onto the sleeves of gowns in between the layers.

Tasks
A range of operational tasks were considered, including frequently occurring and safety critical tasks, exceptional or emergency responses, tasks where novice users may make mistakes or where errors are known to be common, and maintenance/inspection and moving tasks. The task [9] requires a thorough understanding of the work, so when developing the Usability Testing Protocol, existing procedures and documentation from three different models of ventilators were used to generate hierarchical task analyses which were used by the clinicians to develop the task scenario (Table 1).
Errors were identified from previous research [10] to use both as prompts during the task scenario walk/talk through and to develop the evaluation proforma (Table 2). Key error types identified included: • Failure to set up correctly: including ability to use, despite failure to pass self-test; ability of novice to set up ventilator circuit according to on-screen instructions; interchangeability of circuit with other types of ventilator circuitry that look similar. • Failure to find a setting site or display site: difficulty with indirect adjustment of a requested setting; difficulty manipulating multiple controls of different types; difficulty making basic adjustments; confusion and error for the new or occasional user when adjusting for advanced parameters. • Setting site identified correctly but inappropriate setting: illogical default settings, not necessarily immediately obvious to user; errors in adjusting the inspiratory trigger; unclear indication on the controls of the trigger sensitivity where changing one parameter leads to change in other parameters which is not immediately recognised. • Failure to confirm settings: poor tactile and visual interface design/feedback. • Errors of interpretation: difficulty in reading/interpreting display linked to information design and mode presentation (thresholds, configuration, default values, etc.). • Errors of cleaning: risks associated with poor cleaning or failure to replace contaminated parts, missing parts during reassembly.  • Errors of maintenance: lack of knowledge (i.e. training/qualifications) of technical support staff; lack of awareness of common failures and failure modes.

Formative usability testing
As this was a rapid manufacturing project, the Government specification only allowed for one day of formative usability testing [5]. To support the manufacturers, the CIEHF produced a task scenario (Table 1) and patient profiles to provide end users with the opportunity either to undertake simulated tasks with the physical prototype (walk-through) or to talk-through for an online evaluation. The development of the usability protocol included telephone assistance by CIEHF expert group members with the RMVS manufacturing teams. The task scenario and patient profiles used previously published templates [11] and were developed by the clinicians on the CIEHF writing team (AW, A-MB and PM). The task scenario was designed as a pathway to reflect an individual patient requirement for ventilator use. It depicts a combined set of patient pathways to test the ventilator across a range of circumstances that would be unlikely to occur in an individual patient experience. The scenario starts from admission and initial testing of the ventilator, initiation of mechanical ventilation, mandatory modes (likely to be used in the initial phase), switching to spontaneous/triggered modes, monitoring and then finishes with the weaning process.
Patient profiles were developed to reflect common issues and patient presentations. Each profile included details about the patient, the task and the equipment to be used, for example:

User evaluation questionnaire
A user evaluation questionnaire was developed based on previous research [10,[12][13][14] to provide a standardised template for gathering the required formative feedback from end users. The questions were checked against the task scenario and professional practice by the clinical authors and adapted to align with the MHRA Specification [5]. Finally, an issue reporting template was designed to support the systematic collection and recording of issues, including:

Conclusion
This was a global crisis; everyone was trying to help and to adapt. As new players entered the field (i.e. manufacturers with engineering knowledge but unfamiliar with healthcare), it was important that efforts to respond to the crisis were based on established practice and structured approach [15]. Clinical staff working in ICUs and at the new National Health Service field hospitals could have been asked to use different types of ventilators with known risks of accidently pressing the wrong buttons or misreading information on screens.
The CIEHF community responded by providing structured guidance to help manufacturers with the novel requirements and challenges. The CIEHF guidance was issued to RMVS manufacturers to support the design and testing of new machines and to encourage standard designs and protocols to prevent avoidable harm to patients. The usability testing protocol supported realistic testing (work-asdone), including operability whilst wearing a range of PPE. The guidance and usability evaluation protocol are simple tools with the potential to make a significant contribution and could be adapted to other medical devices or equipment. This opens debate for national policymakers and others about the role and contribution of HFE in healthcare, which should be sustained beyond the immediate COVID-19 pandemic.