Preparing your intensive care unit for the COVID-19 pandemic: practical considerations and strategies

Designation of an isolation ICU ward, geographically separated from other clinical areas, allows for concentration and segregation of equipment and staff, contributing to more effective containment. Isolation ICUs should ideally consist of negative pressure airborne infectious isolation rooms (AIIRs). AIIRs are kept at negative pressure relative to surroundings and ventilated with at least 6–12 air changes per hour (Fig. 1a), with any recycled air filtered before recirculation [17]. For hospitals without AIIRs, containment may also be maximised with a dedicated ICU and strict infection control measures. If single rooms are unavailable, cohorted patients should ideally be nursed at least 2 m apart with engineering controls (separation with physical barriers) [18].

Rapid identification and isolation of suspected COVID-19 cases requires an efficient triaging protocol and access to rapid diagnostic testing. In our centre, all patients are screened at the emergency department, inpatient wards, and outpatient clinics. Based on screening questionnaires (travel and contact history), symptoms, and preliminary investigations (chest radiographs), patients are stratified into COVID-19 risk levels and admitted into open wards, cohort beds, single room beds, or negative pressure isolation rooms, respectively. With guidance from government authorities, our hospital’s ‘outbreak response command centre’ constantly updates necessary changes to our COVID-19 risk stratification protocols, which are immediately disseminated to HCWs. Diagnostic testing is performed on specimens (pharyngeal swabs, or lower respiratory tract samples if applicable) for SARS-CoV-2 with real-time reverse transcriptase–polymerase chain reaction (RT-PCR)—we emphasise that healthcare systems need to ensure that there is adequate infrastructure and facilities to enable rapid testing.

As with other respiratory viruses, SARS-CoV-2 is believed to be most contagious when patients are symptomatic. However, unlike MERS and SARS [19], transmission of SARS-CoV-2 may occur even with mild or asymptomatic illness [20]. This complicates identification of cases and contact tracing. In Singapore, comprehensive and coordinated contact tracing is done at the hospital and national level to maximise containment efforts [21]. Our suspect case definition is currently based on a combination of travel and contact history, and presence of respiratory illness [22]. Patients who do not fulfil these criteria but have respiratory symptoms or chest imaging suggestive of a respiratory infection are cohorted to designated ‘acute respiratory illness’ wards for surveillance and segregation. Each healthcare system will need to constantly evolve their protocols and suspected case criteria to balance safety, yet optimise resource utilisation [23, 24].

In general, ICUs should be equipped to expand immediately by at least 20% above baseline capacity [14]. During a pandemic, however, significantly higher surge ICU capacity is needed and critically ill patients may need to receive care outside of a traditional ICU [14]. The planned response will depend on available resources, and trigger targets for activation of each phase of response should be defined early.

Alternative sites to care for critically ill patients must be identified early. In our centre, plans are made to allow for the immediate conversion of all isolation AIIRs into ICU beds. Beyond this, patients will be cared for in existing ICU beds. We have also made comprehensive plans for the use of existing monitored beds (e.g. high dependency units, post-anaesthetic care units, operating suites, endoscopy suites) and cohorting of confirmed cases in designated areas. It is important that ICU bed planning should take into consideration the availability of oxygen ports, compressed air supply, clean water, and drainage systems [18]. Concurrent timely discharge to step-down care areas is critical and enables ICU resources to be used effectively. By this time, all elective surgeries and non-essential services need to be suspended so that resources can be quickly redirected.

At the extreme end of the surge continuum, ‘emergency mass critical care practices’ will have to be instituted which may come at a cost of suboptimal standards of care [14, 24]. This may mean that high-flow nasal cannula (HFNC) therapy or non-invasive ventilation (NIV) is used to mitigate acute ventilator shortages, even though they do not constitute preferred management due to infection control concerns (aerosolisation of respiratory droplets). Triage of critically ill patients may become necessary to prioritise ICU resources, and ethical principles must be carefully considered to ensure just and equitable delivery of care for all patients [25‐27]. Triaging protocols should not disadvantage patients without COVID-19 who need ICU care. Healthcare systems must find a balance between ‘saving the most lives’ and prioritising care based on the likelihood of clinical benefit [28]. These criteria should ideally be objective, transparent, and publicly disclosed. Authorities should engage the community in this process so that public trust exists when it is most needed [25‐27, 29].

Infection prevention and control is essential to protect both patients and HCWs [30]. Experience from SARS demonstrated the vulnerability of HCWs, and an uncontrolled in-hospital outbreak can devastate an entire hospital’s services within days [31, 32]. Critical care HCWs are at high risk considering the higher exposure dose from aerosol-generating procedures and longer periods of patient contact. In addition, the reproductive number for SARS-CoV-2 (between 2 and 2.5) and therefore transmissibility is significantly higher than seasonal influenza [33]. As with other respiratory viruses, it is likely that SARS-CoV-2 spreads predominantly through contact (direct or indirect) and droplet transmission. Though faecal shedding of the virus has been demonstrated, faecal-oral transmission remains unproven [34]. Strict adherence to droplet and contact precautions including hand hygiene, eye protection, and safe donning and doffing of personal protective equipment (PPE) will be the main defences against transmission. A summary of recommended infection control measures is presented in Table 2.

Airborne transmission with respiratory viruses remains controversial [37], although it is likely that any airborne transmission with COVID-19 will be opportunistic in nature [38]. Studies on SARS-CoV and MERS-CoV have highlighted the possibility of airborne transmission with aerosol-generating procedures such as intubation [17], manual ventilation [39], NIV, and tracheotomy [40]. Until more is known about COVID-19 transmission, we recommend airborne precautions during aerosol-generating procedures in all patients (COVID-19 or otherwise) [17]. Disposable caps, gloves and fluid-resistant gowns, eye protection, and respiratory protection using a fit-tested National Institute of Occupational Safety and Health (NIOSH) certified disposable N95 particulate filtering facepiece respirator are used in our centre [17]. Powered air-purifying respirators (PAPRs) are alternatives, but careful doffing and disinfection is necessary to prevent self-contamination [41]. HCWs must doff and dispose of PPE safely within designated areas, with equipment and environmental decontamination enforced.

Other potential aerosol-generating procedures (e.g. airway suctioning, nebuliser treatment, bronchoscopy, chest physiotherapy, and HFNC) have not been conclusively established to increase the risk of viral transmission [40]. Nevertheless, we discourage nebulised therapy (metered dose inhalers with or without a spacer are used instead) and use closed in-line suction for tracheal aspiration in our centre. With no clear mortality benefit with HFNC therapy in respiratory failure, our preferred management is conventional oxygen therapy and early intubation [42]. NIV is considered only with evidence-based indications, in view of high failure rates and poor outcomes seen with MERS-CoV [43]. However, we do recognise that in the event of severe ventilator shortages, these modes of respiratory support may have to be utilised—ideally with strict infection control measures. In our institution, mechanical high-efficiency particulate air (HEPA) or pleated hydrophobic filters are fitted to mechanical ventilators at expiratory ports, as well as NIV mask and bag-valve-mask interfaces (Fig. 2). It should be noted that airway resistance and circuit dead space may be increased as a result [44‐47].

Education is important for effective infection control, upskilling HCWs, and identifying logistical and technical challenges with routine or ICU care. HCWs should be taught to inspect, disinfect, and dispose of PPE safely, and periodic refresher re-training is required to ensure staff readiness and proficiency. In our centre, an online critical care refresher course was organised for non-intensivist physicians, with voice-annotated lectures on ICU management and instructional videos on ventilator set-up and troubleshooting.

From our experience, simulation training has been invaluable in ensuring the preparedness of HCWs [48, 49]. We instituted ‘just-in-time’ inter-professional and in situ simulations. These included resuscitation and rapid response team training, extracorporeal membrane oxygenation (ECMO), transport of critically ill patients, and procedures like tracheostomy (Fig. 1), performed with full airborne precautions and PAPR. These helped to recognise and address unexpected problems, which led to downstream rectification. For example, cardiac arrest simulations revealed several violations of infection control during cardiopulmonary resuscitation, and challenges such as difficulties in auscultating for breath sounds while using PAPR. This triggered the recommendation for infection control re-training and posters for PPE doffing to be placed in all isolation wards. End tidal carbon dioxide monitoring was also used to confirm endotracheal tube placement for all intubations. Transport with full-feature ICU ventilators also revealed problems with limited internal battery lifespan, and physical bulk even with the largest available cargo lifts. Protocol changes to use transport ventilators fitted with HEPA filters were then made. Multidisciplinary peer-review processes were also held after each procedure with COVID-19 patients, where HCWs and relevant stakeholders were debriefed and process improvements discussed.

HCWs should be segregated to reduce the risk of cross-transmission. This allows for service continuity, should a team be quarantined. In our centre, HCWs caring for COVID-19 patients are also not permitted to care for other patients and are relieved of outpatient responsibilities. In our ICUs, we allocated ‘frontline’ teams of doctors, nurses, and allied health staff, with defined geographical coverage for each team. Teleconferences replaced face-to-face meetings. HCWs were instructed to record temperature surveillance readings twice a day and submit these to an online platform that is centrally tracked to facilitate early detection of any intra-hospital outbreak. HCWs with fever or respiratory symptoms are also required to report to designated staff clinics, with COVID-19 testing and/or stay home instructions implemented as required.

With the need for strict infection control measures, routine procedures in the ICU will require additional manpower and time [49]. In our isolation ICU, the nurse-to-patient ratio was increased to approximately 1.5:1, with critical care nurses focused on providing care and supported by non-critical care nurses who assisted in preparing drugs and equipment outside the AIIRs. Infection control specialty nurses assisted ICU staff in appropriate donning and doffing of PPE. In addition, doctors and nurses with previous critical care experience were also reassigned from other departments to extend frontline coverage to 24 hours a day.

In the event of widespread community transmission, critical care manpower capabilities will be easily overwhelmed. It will then be necessary to draw manpower from various departments and divisions in the hospital. In our centre, critical care crash courses are being held continuously for non-critical care nurses with online learning materials, instructional videos, and hands-on practicums supervised by critical care nurses in non-isolation ICUs. Manpower is also being drawn from nurses with critical care experience, recently retired ICU nurses, and private hospital nurses. Nurses from various subspecialty centres within the extended hospital campus can augment non-critical care nursing roles, such as infection control [50]. Critical care intensivist- or nurse-to-patient ratios should be re-evaluated frequently, guided by clinical demands, provider experience, and available support by ancillary staff [50]. With critical shortages in manpower, intensivists and ICU nurses may have to take up a supervisory role, with non-critical care HCWs providing direct care to patients.

High workload and anxiety over disease transmission can result in significant physical and mental fatigue [51]. To mitigate this, we established open communication channels and rapid dissemination of information to keep staff informed of new developments. Psychological support resources and a helpline are also made available. Sections of a community hospital building were repurposed as optional temporary staff quarters. Despite best intentions, ‘presenteeism’, the act of coming to work whilst ill, potentially increases the risk of intra-hospital transmission [52]. Policies must ensure that HCWs seek early medical attention and stay home when ill.

We still have a poor understanding of clinical features of critical illness in COVID-19. Based on the current evidence, the spectrum of disease severity appears to vary significantly from asymptomatic or mild symptoms to critical illness with a high mortality rate [4]. Critical illness appears to develop most commonly between 1 and 2 weeks after symptom onset [3, 5]. Ground glass opacities are typical findings on chest imaging, with basal consolidation seen in some patients [59, 60]. However, chest radiograph or computed tomography (CT) imaging can be normal, even in patients who subsequently develop critical illness [61]. In addition, patients can develop rapid deterioration and progression to ARDS [62]. Risk factors for critical illness include advanced age (≥ 60 years) and pre-existing comorbidities (cardiovascular disease, cerebrovascular disease, chronic pulmonary disease) [3, 61, 63]. Biochemical ‘markers’ including significant lymphopenia, elevated lactate dehydrogenase or d-dimer, and neutrophilia are associated with more severe disease [3, 5, 63]. Nosocomial or secondary infections appear to be common (14–31%) in critically ill patients [4, 5]. It remains unclear if COVID-19 can result in fulminant myocarditis [63].

Until more evidence emerges, management of COVID-19-induced ARDS should not differ from conventional ARDS management [64, 65]. These include lung protective ventilation, a fluid conservative strategy when appropriate, and early initiation of prone ventilation for severe ARDS [66‐68]. When appropriate, early neuromuscular blockade may be considered for moderate-severe ARDS [69]. Referral to an ECMO centre, for refractory hypoxemia or if lung protective ventilation cannot be applied, remains an option. Corticosteroids should not be routinely administered [70], with concerns of delayed viral clearance or worse patient outcomes extrapolated from studies with SARS and MERS-CoV [71, 72]. Whilst there is no proven anti-viral therapy for COVID-19 to date, ongoing trials include the use of remdesivir, which has shown some promise in in vitro studies [73]. More research is urgently needed to fill knowledge gaps in terms of disease clinical course, prognostic markers, treatment options, and transmission dynamics, which will be critical for ICU management and resource planning.

Effective communication is essential in a pandemic response. Information should be disseminated to all HCWs in a timely fashion, and two-way communication should be established. In our centre, information and protocol updates are disseminated to all HCWs daily by email and hospital intranet. In addition, ICUs will need to work closely with nursing services, infection control, supplies/engineering/cleaning services and other departments to harmonise efforts for a coordinated response. In our institution, an ‘outbreak response command centre’ involving key appointment holders discusses anticipated challenges and reviews protocols regularly. In Singapore, several ICU-led inter-hospital video conferences have also been organised to share knowledge. Finally, good coordination between local or regional hospitals, guided by government or national organisations, is a key component of the pandemic response with effective resource allocation [77].

Family engagement is important as visitors are largely excluded from isolation wards [78]. Effective communication is necessary to address concerns and manage expectations, or to help determine goals of care in the critically ill. Public relations departments and social workers should be involved to facilitate communication with family members or the public media, to minimise misunderstanding and build trust and confidence in the healthcare system [79].