Investigation report: Piped supply of medical air and oxygen

An 85-year-old woman was admitted to hospital at 13:00 hours on a weekday afternoon via the Emergency Department (ED) after she fell at home. She suffered from chronic lung disease and was usually breathless at rest. On assessment in the ED her blood oxygen saturations were recorded as 93% breathing room air. A plan was made for her to receive supplementary oxygen as required if oxygen saturations fell below 92% breathing room air.

At 18:30 hours she was transferred to a ward where at 21:29 hours her oxygen saturation level was recorded as 98% breathing room air.

At 07:18 hours the following morning a nurse, who was coming to the end of her shift, was undertaking routine observations and recorded an oxygen saturation level of 84% breathing room air. In keeping with the previous plan, she administered what she thought was oxygen at a rate of 2 Litres per minute (L/min) via nasal cannulae connected to a wall-mounted supply. She then asked for a medical review.

At 08:35 hours another nurse, who had just come on duty, discovered that the patient’s oxygen saturations had fallen to 76% and that she was being administered air, not oxygen, from the adjacent wall-mounted supply. An air flowmeter flap was in place.

The nurse administered 2L/min oxygen and organised a medical review. The oxygen saturations quickly recovered to 93% and the patient suffered no apparent long-term adverse effects from this episode.

The Healthcare Safety Investigation Branch (HSIB) was made aware of a safety issue relating to a persistent risk in hospitals of connecting oxygen tubing to wall-mounted air flowmeters despite the release of a Patient Safety Alert by NHS Improvement (NHSI) in 2016 and a Rapid Response Report from the National Patient Safety Agency (NPSA) in 2009 intended to address this issue.

HSIB identified the reference event and with the support of the Trust commenced a scoping investigation. The findings were considered against the HSIB investigation criteria, and a decision was made to progress to a national investigation. The national investigation collected evidence relating to the reference event, published material associated with the 2016 NHSI Patient Safety Alert, and information related to implementation of the alert.

This investigation also considered the assurance processes there were to ensure that trusts undertook appropriate action in response to the alert, how they understood the requirements for piped medical air and oxygen in hospitals and how those requirements were met.

The national investigation found:

• There is a lack of clarity on the requirement for piped medical air in hospitals.
• Some trusts have misinterpreted the direction given in the NHS Improvement (NHSI) Patient Safety Alert to remove air flowmeters and cap air outlets in areas where there is no need for medical air and to deliver nebulised medication using nebuliser boxes.
• Financial considerations may be a barrier to trusts switching from piped air delivered through Medical Gas Pipeline Systems (MGPS) to nebuliser boxes, due to the cost of purchasing the required number of nebuliser boxes and the capping of the wall terminals, particularly in Private Finance Initiative (PFI) hospitals.
• Local assurance processes relating to the implementation of this alert were not effective in this case. The ability to nationally record trusts’ specific actions in response to this and other alerts exists but requires further enhancement to support trusts and national oversight.
• The movable flap fitted to air flowmeters and removal of the flowmeter when not in use has reduced the risk of human error but not prevented it.

1.1.1 The British Thoracic Society (BTS) guideline [1] identifies three main requirements for administration of oxygen in hospitals:

• administration to correct hypoxaemia (low blood oxygen levels).
• administration to prevent possible hypoxaemia.
• administration to alleviate breathlessness.

1.1.2 Since ‘oxygen is probably the commonest drug used in the care of patients who present with medical emergencies’, the guideline indicates that it should be prescribed by a doctor and a written record kept. [1]

1.1.3 Prescribing oxygen in advance enables nurses to take action if a patient’s oxygen saturation levels fall unexpectedly, without the need to wait for a doctor to attend and prescribe oxygen. [1]

1.1.4 Failure of oxygen saturation levels to respond to treatment may indicate that insufficient oxygen has been given, that the clinical situation has deteriorated or that a problem exists with the oxygen delivery system itself.

1.2.1 Medical air is supplied through high-quality drying and filtration equipment. [2]

1.2.2 The main uses for medical air in hospitals are as follows:

‘Driving nebulisers. A nebuliser is a device that converts medication solutions into aerosols suitable for inhalation. Nebulisers are widely used in the treatment of respiratory diseases, e.g. asthma, chronic obstructive pulmonary disease (COPD) and cystic fibrosis. While piped medical air can be used to nebulise medication, machines that compress atmospheric air are widely available and some sites use ultrasonic nebulisers. Therefore, piped medical air has become obsolete in some areas that use these alternatives.
Driving ventilators(1) and resuscitaires(2). Patients sensitive to oxygen toxicity are given air to lower their exposure to oxygen, e.g. neonates, patients with chronic hypoxaemia (as in severe COPD) and those with acute pulmonary problems with severe hypoxia.
Power source for driving surgical tools. These tools require high flow and high pressure air (‘surgical air’). These power sources cannot be mistaken for oxygen outlets as they have different connections.
Carrier gas for volatile anaesthetic agents in anaesthesia.’ [3]

1.3.1 The delivery methods for air and oxygen are as follows:

• Pipelines built into the hospital infrastructure, called the Medical Gas Pipeline System (MGPS)(3).
  • Machines can connect directly to the terminal unit(4) to deliver the required gas.
  • Flowmeters (see page 12) connect to the terminal unit. An oxygen flowmeter has a connector specific for the oxygen terminal unit and cannot accidentally be connected to the air terminal unit and vice versa. In a ward a flowmeter is required to connect the tubing from the patient mask and to specify the flow rate of gas to be delivered.
• Portable air or oxygen cylinders with a flowmeter or a regulator.
• For air only, compressed or ultrasonic nebulisers.

1.3.2 Many hospital wards have an MGPS for delivering oxygen and air to wall outlets adjacent to each bed.

1.3.3 In 2006 the Department of Health and Social Care released a Health Technical Memorandum (HTM) in relation to MGPS. [2] [4] It stated: ‘An MGPS is designed to provide a safe and effective method of delivering medical gases, medical air and surgical air from the source of supply to the appropriate terminal unit [or ‘outlet’] by means of a pipeline distribution system.’ [2]

1.3.4 The HTM does not set out the need for MPGS to deliver air, rather the standards that must be applied for fitting and maintenance of systems if they are deemed to be required.

1.3.5 Ventilators, resuscitaires, surgical tools and anaesthetic machines can be connected directly to the air and oxygen supply and a flowmeter is not needed.

1.3.6 In the UK and Europe, oxygen flowmeters are coloured white and air flowmeters are coloured black (Figure 1).

Fig 1 Oxygen and air flowmeters

1.3.7 In hospitals the terminal unit for oxygen is situated next to the terminal unit for air. HTM (02-01) [2] states that in all installations the oxygen terminal unit should either be to the left of or above the air terminal unit.

1.4.1 The connectors between the wall outlets and the flowmeters for oxygen and air are different. They have been designed like this to prevent inadvertent connection of the oxygen flowmeter to the air outlet. However, the connectors between the flowmeters and the tubing which connects to the patient administration device are identical. Both have fir tree connectors (Figure 2) which means that oxygen tubing can inadvertently be connected to an air flowmeter and vice versa.

1.5.1 In 2009 the National Patient Safety Agency (NPSA) issued a Rapid Response Report in relation to oxygen safety in hospitals [5] recommending that ‘the risks of confusing oxygen and medical compressed air are assessed and action plans developed (e.g. removing the medical air flowmeter from the wall outlet when not in regular use).’ [5]

1.6.1 In 2016, NHS Improvement issued a Patient Safety Alert (PSA): NHS/PSA/D/2016/009 titled ‘Reducing the risk of oxygen tubing being connected to air flowmeters’ (see appendix). [6] The PSA stated: ‘Three barriers to human error have already been recommended by the NPSA and British Thoracic Society but continuing incidents suggest they have not been universally implemented.’ These three approaches to reducing the risk of connecting oxygen tubing to air flowmeters described in the alert were:

‘Medical air terminal units (wall outlets) are covered with designated caps in areas where there is no need for medical air. Medical air outlets were traditionally built into most clinical areas for the delivery of nebulised treatment but not all areas need them (e.g. they never have patients who need nebulisers, or they have access to electrically driven compressors or ultrasonic nebulisers).
Medical air flowmeters are removed from terminal units (wall outlets) and stored in an allocated place when not in active use. Removing unnecessary equipment is a more effective method of reducing human error than adding labels or warnings alone.
Air flowmeters are fitted with a labelled, movable flap. The lettering on the flap is larger and more visible than on the flowmeter itself and this flap has to be lifted to attach a tube. This acts as a further barrier to unintended connection if staff occasionally forget to remove medical air flowmeters after a period of active use.’ [6]

1.6.2 The alert required providers to complete the following actions by 4 July 2017:

• Identify a named individual who will take responsibility for co-ordinating the delivery of the actions required by this alert.
• Implement systems to ensure that the three barriers to human error described in this alert were all in place in all relevant clinical areas.
• Establish ongoing systems of audit or equipment checks to ensure the barriers are maintained.

1.7.1 The BTS guideline for oxygen use in adults in healthcare and emergency settings states: ‘Air flowmeters are never required in an emergency and should be removed from wall sockets or covered by a designated ‘hood’ when not in use.’ [1] The BTS guidelines also recommended that organisations implement measures to stop inadvertent connection to air instead of oxygen.

1.8.1 The supporting information that accompanied the 2016 PSA stated: ‘In theory, flaps would not need to be fitted if the other two barriers are always put into practice. If all ‘out of use’ medical air outlets are covered and all flowmeters are always removed from the outlet after active use, the risk of inadvertently attaching an oxygen tube to an air flowmeter would be low. However, we know that there will be occasions when air flowmeters might accidentally be left in situ, eg when staff are distracted or not fully aware of local arrangements.’ [3]

1.8.2 The PSA recommended that movable flaps were fitted to all air flowmeters that were still in use (see figure 3). The rationale was that ‘First, the flap is labelled ‘medical air’ in lettering larger than that on the flowmeter itself, making it more visible to staff. Second, the flap has to be lifted to attach a tube, which is an additional step to prompt the user to check the outlet supplies the correct type of gas.’ [3] This supporting information highlighted that this would have a ‘limited impact on reducing human error when implemented on their own.’

1.9.1 Nebuliser boxes and ultrasonic nebulisers are portable machines that convert liquid medicine into aerosols for inhalation. They are an alternative to using medical air through the MGPS for this purpose. The PSA stated that ‘unlike oxygen, flowmeters for medical air are never required in an emergency. Therefore, withdrawing ‘immediate’ access to medical air would not introduce risks to patients. Of course, any action plans to do this would need to ensure alternative arrangements are in place to meet patient needs (e.g. increased stocks of compression or ultrasonic nebulisers).’ [6] The main barrier to their widespread use has been cost; an individual nebuliser box typically costs £80-100. Individual trusts are responsible for identifying the number of nebuliser boxes required for each clinical area and ensuring that they are maintained and replaced as required.

1.10.1 Never Events are defined as incidents that are ‘wholly preventable, where guidance or safety recommendations that provide strong systemic protective barriers are available at a national level and should have been implemented by all healthcare providers.’ [7]

1.10.2 Unintentional connection of a patient requiring oxygen to an air flowmeter has been considered a Never Event under the NHS Improvement Never Events list, published in January 2018.

1.11.1 On 1 June 2018 The Medical Director and Nurse Director of NHS Improvement wrote to all acute trust medical directors and nurse directors stating: ‘Almost all trusts have stated they have complied with the NHS Improvement Alert that required removal of air flowmeters from routine use by July 2017. In almost all clinical settings, power driven nebulisers can be used instead, and all air flowmeters can be permanently removed and disposed of, with air outlets blocked off. In the rare clinical setting where it is impossible to use power-driven nebulisers due to lack of space, air flowmeters need to be separately stored and should be brought out of storage only for the ten minutes or so required to deliver the nebuliser. Unintended connection to an air flowmeter for a patient requiring oxygen therefore became a Never Event in February 2018. Since then, 15 trusts have reported Never Events related to unintended connection to air flowmeters, and what is particularly concerning is that most of these make no reference to how the trust had implemented the Alert and suggest air flowmeters had not been removed and remained constantly in wall outlet.’ [8]

1.12.1 ‘The Central Alerting System (CAS) is a web-based cascading system for issuing patient safety alerts, important public health messages and other safety critical information and guidance to the NHS and others.’ [9] All patient safety alerts are released through the CAS and each organisation will have a staff member who is responsible for receiving and disseminating the alerts. Where actions are required in response to an alert each organisation is required to sign off the actions as complete in the CAS system.

2.1.1 An 85-year-old woman was admitted to hospital at 13:00 hours on a weekday afternoon via the Emergency Department (ED) after she fell at home. She suffered from chronic lung disease and was usually breathless at rest. On assessment in the ED her blood oxygen saturations were recorded as 93% breathing room air. A plan was made for her to receive supplementary oxygen as required if oxygen saturations fell below 92% breathing room air.

2.1.2 At 18:30 hours she was transferred to a ward where at 21:29 hours her oxygen saturation level was recorded as 98% breathing room air.

2.1.3 At 07:18 hours the following morning a nurse, who was coming to the end of her shift, was taking routine observations and recorded an oxygen saturation level of 84% breathing room air. In keeping with the previous plan, she administered what she thought was oxygen at a rate of 2 litres per minute (L/min) via nasal cannulae connected to a wall-mounted supply. She then asked for a medical review.

2.1.4 At 08:35 hours another nurse, who had just come on duty, discovered that the patient’s oxygen saturations had fallen to 76% and that she was being administered air, not oxygen, from the adjacent wall-mounted supply.

2.1.5 The nurse administered 2L/min oxygen and organised a medical review. The oxygen saturations quickly recovered to 93% and the patient suffered no apparent long-term adverse effects from this episode.

2.1.6 The air flowmeter had been fitted with a flap, and the flap was still in place.

3.1.1 The Healthcare Safety Investigation Branch (HSIB) was made aware of a safety issue relating to the risk of connecting oxygen tubing to air flowmeters.

3.2.1 Following preliminary information gathering, HSIB concluded that the safety issues represented by this event and supported by national data, met the criteria for investigation, which the Chief Investigator authorised.

3.2.2 A summary of the analysis against the high-level criteria is given below.

There is a risk of severe harm or death if air is administered rather than oxygen. This has been highlighted in previous alerts from the National Patient Safety Agency in 2009 and NHS Improvement in 2016.

Unintentional connection of a patient requiring oxygen to an air flowmeter became a Never Event on 1 February 2018. From February to June 2018, 32 cases were reported nationally.

This issue continues to occur despite the 2009 and 2016 national alerts and the designation of this as a Never Event in 2018. There is potential for HSIB to examine how national recommendations are developed and translated into local practice including the mechanisms available to redesign systems to reduce risks.

3.3.1 Methods used in this investigation included:

• Review of patient clinical records, hospital policies and procedures relating to management of Patient Safety Alerts (PSA) and relevant meeting records.
• Interviews with the staff involved in the incident and wider decision-making surrounding implementation of the PSA.
• Interview with the affected patient and her family.
• Review of the Trust’s serious incident reports for each of its four recent cases.
• Review of 32 relevant serious incidents reported to the Strategic Executive Information System (StEIS) on national incident reporting databases between 1 February and 30 June 2018.
• Research literature review.
• Communication with subject matter experts and those involved in the development of the PSA.

4.1.1 There is a lack of clarity about the circumstances in which piped medical air is required in hospitals. The Patient Safety Alert (PSA) states that ‘medical air outlets were traditionally built into most clinical areas for the delivery of nebulised treatment.’ [6]

4.1.2 NHS Improvement (NHSI) Patient Safety Team were confident that the only use for piped medical air via an air flowmeter was for the delivery of nebulised medications and that this need could be met using nebuliser boxes. Intensive Care Unit (ICU) consultants told the investigation that there could be other circumstances in which a piped air supply through a Medical Gas Pipeline System (MGPS) would be required on general wards to provide patient ventilation (for example, when ICU patients require care on a lower dependency ward, during a major incident, or pandemic flu).

4.1.3 There is a requirement for medical air in neonatology, operating theatres and high-dependency areas, such as ICUs. In these areas the medical air is delivered through an integral system(s) and does not require an air flowmeter.

4.2.1 NHS Improvement (NHSI) stated that the intention of the 2016 PSA was for trusts to remove all air flowmeters and replace them with nebuliser boxes except in areas where this was not feasible (for example emergency department (ED) cubicles without space for a nebuliser box).

4.2.2 Recognising that providing nebuliser boxes may not be feasible in all areas immediately, NHSI Patient Safety Team expected that trusts should remove air flowmeters from the outlet when not in use and have a moveable flap in place to act as a physical prompt.

4.2.3 The PSA contained an action that trusts should implement systems to ensure that ‘the three barriers to human error described in this alert are all in place in all relevant clinical areas.’ [6] Despite that, the investigation identified that some trusts interpreted the three barriers as options from which they could select those they felt were required locally.

4.2.4 A review of 32 Never Events reported to StEIS between 1 February and 30 June 2018 found that many trusts that experienced Never Events opted only for the third barrier; installing a flap over the air flowmeter connector. While others did report removing air flowmeters and/or replacing these with nebulisers after the Never Event, there was no consistent response. Other responses to the Never Events included raising awareness of the issue with staff, capping of piped medical air supplies, replacing piped medical air with nebuliser boxes and removing air flowmeters when not in use. Evidence from these events showed that the removal of air flowmeters after each use was not sustained and so did not provide a strong barrier to preventing further incidents.

4.2.5 The supporting information that accompanied the PSA [3] provided three case studies which showed that the wall outlets were capped. It also included illustrations on the installation of movable flaps. These illustrations may have led some trusts to focus more on the installation of flaps than the other barriers.

4.2.6 The Trust involved in the reference event told the investigation that at the time it received the 2016 PSA, it considered the alert and decided to install moveable flaps on all its air flowmeters but to take no other actions. The Trust believed that was all the alert required it to do. It recorded the alert as ‘action completed’ in the Central Alerting System (CAS). The Trust then had a Never Event involving the unintended connection of a patient to an air flowmeter. As a response to this the Trust reminded staff that air flowmeters should be removed when not in use and to ensure that movable flaps were in place. Following two further Never Events, the Trust instigated audits to ensure that air flowmeters were removed when not in use, and hourly checks by ward leaders. A fourth Never Event occurred; the Trust declared ‘air permissible areas’ where air supply via MGPS could still be used, and purchased nebulisers for all other areas. The Trust successfully secured funding from its charity to enable this purchase.

4.3.1 The investigation found that cost may be a barrier to trusts implementing all recommendations in the PSA.

4.3.2 The investigation observed that nebuliser boxes varied in cost when purchased through NHS Supply Chain or directly from manufacturers. Trusts explained that each box can cost between £80-100, and one Trust reported that it relied on charitable funding for the purchase of these to cover all clinical areas.

4.3.3 The investigation was made aware that in trusts with Private Finance Initiative (PFI) buildings, capital works, such as capping air wall outlets, could require renegotiation of contracts which would incur additional expense.

4.4.1 The internal assurance process for taking the appropriate action in response to a PSA has not been effective in this instance. The Trust had declared compliance.

4.4.2 The CAS provides a free-text feedback box which healthcare providers can use to detail what actions they have taken in response to alerts issued through the system. However, whilst the system mandates that a response is selected (for example ‘action completed’) it does not mandate that the provider supply any detail. There is no current option for documents to be uploaded (for example action plans) by providers, and uploading documentation rather than completing a free-text box may be a more practical way of gathering this type of feedback. NHSI Patient Safety Team, supported by literature from the Care Quality Commission, [10] stated that systems within healthcare providers for acting effectively on alerts are currently not always robust.

4.5.1 Air and oxygen flowmeters connect to tubing using the same connector design; this represents a further source of risk.

4.5.2 An air flowmeter cannot be connected to an oxygen outlet and an oxygen flowmeter cannot be connected to the medical air outlet as the connectors are incompatible. However, the fir tree connectors that connect both the oxygen flowmeter and air flowmeter to the tubing are the same.

4.5.3 The International Organization for Standardization (ISO) draft standard ISO-80369-2 relates to breathing systems and for driving gases applications.

4.5.4 At the time of writing, the standards for ISO-80369-2 remain in draft form. Initially the proposal was for the fir tree connectors, tubing and mask to be different between medical air and oxygen, but product manufacturers considered this would not be achievable. The current proposal is for a specific air mask and oxygen mask with a unique connection to the tubing, although the end user would still be able to attach the same tubing to the fir tree on the flowmeter. The end user would also have to select the correct type of mask for the required use.

4.5.5 The investigation concluded that the new ISO standards for piped medical air connectors are unlikely to fully prevent the risk of unintentionally connecting the patient’s oxygen tubing to medical air instead of oxygen.

4.6.1 The moveable flap was designed and developed by a hospital team in Australia [11] and marketed under the brand name AirGuard. The team recognised that the flap was an error-reducing solution rather than error-proofing solution.

4.6.2 A small-scale study at the institution where it was developed found a reduction in the number of reports of misconnection incidents when the AirGuard was used, [12] although some incidents still occurred. It is unclear to what extent the AirGuard has been further tested and evaluated.

4.6.3 Staff from the Trust involved in the reference event told the investigation that the moveable flap was not an effective barrier to reduce the risk of connecting oxygen tubing to the air flowmeter.

The national investigation found:

• There is a lack of clarity on the requirement for piped medical air in hospitals.
• Some trusts have misinterpreted the direction given in the NHS Improvement (NHSI) Patient Safety Alert to remove air flowmeters and cap air outlets in areas where there is no need for medical air and to deliver nebulised medication using nebuliser boxes.
• Financial considerations may be a barrier to trusts switching from piped air delivered through Medical Gas Pipeline Systems (MGPS) to nebuliser boxes, due to the cost of purchasing the required number of nebuliser boxes and the capping of the wall terminals, particularly in Private Finance Initiative (PFI) hospitals.
• Local assurance processes relating to the implementation of this alert were not effective in this case. The ability to nationally record trusts’ specific actions in response to this and other alerts exists but requires further enhancement to support trusts and national oversight.
• The movable flap fitted to air flowmeters and removal of the flowmeter when not in use has reduced the risk of human error but not prevented it.