National guidance for the management of children with bronchiolitis and lower respiratory tract infections during COVID-19

These recommendations on the management of children with bronchiolitis and lower respiratory tract infections in hospital settings during COVID-19 are for clinicians to support winter planning in partnership with local infection control prevention teams.

While some recommendations describe organisational structures in England, services in the devolved nations are encouraged to adopt them to fit local models.
Last modified
12 April 2021


With concerns of a possible second surge of COVID-19 cases this winter, maintaining robust infection control processes is essential to keep patients, parents/carers and staff safe.

However, it is also necessary to ensure that the flow of patients through the hospital is maintained during the winter period when it is predicted that demand for paediatric services will increase significantly and the onset of the bronchiolitis / respiratory virus season will place services under considerable pressure.


  • The safety of patients and their families, and staff is paramount.
  • Recommendations are to be equitable irrespective of socioeconomic status, ethnicity, or geographic location.
  • An evidence-based approach is adopted, recognising recommendations will evolve with experience.
  • Recommendations should vary in line with current regional COVID-19 prevalence rates (see table 1 and table 2).
  • The potential COVID-19 status of an infant or child should not affect the initial assessment and management of the infant or child when they present to a healthcare setting. Key features of assessment are oxygenation, hydration and nutrition. If commencement of high flow nasal cannula oxygen (HFNCO) is being considered, a senior decision maker should be involved.
  • The personal protective equipment (PPE) recommendations within this guidance are based on the principles outlined in the current Public Health England (PHE) COVID-19 infection prevention and control guidance and care pathways (see Appendix 3). A testing based approach, including isolation of all high-risk/'red' patients on admission and repeat testing of low-risk/'green' patients remaining in hospital, enables most children presenting with lower respiratory tract symptoms to be designated to a low-risk/'green' pathway.

Recommendations have been reviewed and accepted by the NHS England/Improvement Infection Prevention & Control cell.

Aerosol generating procedures (AGP)

As per guidance from PHE, AGPs include:

  • tracheal intubation and extubation
  • manual ventilation
  • tracheotomy or tracheostomy procedures (insertion or removal)
  • bronchoscopy
  • dental procedures (using high speed devices, for example ultrasonic scalers/high speed drills
  • non-invasive ventilation (NIV), Bi-level Positive Airway Pressure Ventilation (BiPAP) and Continuous Positive Airway Pressure Ventilation (CPAP)
  • high flow nasal oxygen 
  • high frequency oscillatory ventilation (HFOV)
  • induction of sputum using nebulised saline
  • respiratory tract suctioning
  • upper ENT airway procedures that involve respiratory suctioning 
  • upper gastro-intestinal endoscopy where open suction of the upper respiratory tract occurs
  • high speed cutting in surgery/post-mortem procedures if respiratory tract / paranasal sinuses involved.

The delivery of oxygen via HFNCO is an AGP and needs to be carefully overseen to minimise the risk of nosocomial infection.

Summary flow chart

Updated 24 September 2020

Flowchart for bronchiolitis in infant or child during COVID-19 - infant with respiratory distress or bronchiolitis attending A&E or paeds assessment unit and requiring admission whether severely unwell or moderately unwell


  • ED – Emergency Department
  • POCT – Point of care testing
  • PICU - Paediatric Intensive Care Unit
  • HDU - High Dependency Unit


  • Staff must use appropriate PPE according to level of risk – high-risk/low-risk (see Appendix 3).
  • Transfers to wards / PICU / other hospitals for PICU should be expedited as rapidly as possible to facilitate patient flow.
  • Urgent review by a senior clinician is recommended before commencement of HFNCO / CPAP (see Appendix 1).
  • Where children require AGPs (HFNCO / CPAP etc.) rapid weaning protocols should be followed to minimise exposure to aerosols (see Appendix 2).

You can also download this flow chart as a poster below.

Recommendations - prior to presentation at hospital

  • Integrated care systems spanning the entire urgent care pathway should be in place to ensure children with mild bronchiolitis and lower respiratory tract infections are managed in primary care settings where possible and to reduce the number of infants and children with respiratory symptoms presenting to hospital. Winter planning should include the implementation of locally appropriate models of care enabling secondary care clinicians to support primary care colleagues. The expectation should be that children with mild and moderate bronchiolitis or lower respiratory tract infection are initially reviewed in primary care settings.
  • Examples of clinical pathways supporting the management of children with shortness of breath by clinicians in primary care settings include the following:
  • Access to paediatric oxygen saturation monitor probes in primary care should be prioritised.
  • Optimise preventive treatment including influenza vaccines in children and palivizumab for children aged under 23 months that meet the criteria as specified in the Green Book. Children with risk factors for severe influenza outside of the ages of routine immunisation (2-12 years) should be actively identified and influenza vaccination promoted.

Recommendations - on presentation to ED or Paediatric Assessment Area

  • Although separate 'red'/'blue' areas are likely to remain in place this winter for both adults and children, these areas can be combined within one paediatric ED footprint if isolation facilities allows. If such a model is adopted, it is important to ensure that protective isolation can also be offered to RCPCH-defined clinically extremely vulnerable (CEV) children as well as other children routinely requiring protective isolation. 
  • Waiting areas should be organised to minimise the risk of nosocomial infection, by allowing adequate physical distancing, respiratory hygiene and hand hygiene. Adherence with face coverings as appropriate should be monitored and regular environmental cleaning performed according to national standards. A local risk assessment is required.

Testing of children with lower respiratory tract infections (including bronchiolitis)

  • Only children requiring admission need to be tested for SARS-CoV-2 in hospital. 
  • Capacity for performing rapid testing is currently extremely limited and this may remain the case over the winter. If a point of care testing platform is available (SARS-CoV-2 +/- RSV +/- influenza), local protocols should be followed and use should be prioritised for children who will most benefit from a rapid result (eg for PICU / HDU admission or surgery). Use can be considered in children requiring HFNCO / CPAP and where inpatient cubicle capacity is severely restricted.
  • Very few EDs have sufficient capacity to keep large numbers of children in their department awaiting virology results. Transfer of a child from ED to an inpatient setting should not be delayed whilst awaiting a test result.  However, testing should be performed in ED and processes should be in place to minimise the turnaround time of results. This is essential in order to step down patients from inpatient cubicles, enabling flow of children from ED.

AGPs (such as commencement of HFNCO or CPAP)

  • AGPs should only be performed or initiated when clinically indicated (see Appendix 1). A senior decision maker should be involved if commencement of HFNCO / CPAP is being considered. In addition, the infection control implications of transferring a child on HFNCO need to be considered. If starting HFNCO is unavoidable prior to transfer to an inpatient setting, a point of care test result may be useful.
  • The child must be managed in an appropriate isolation area by staff wearing the correct PPE as per level of risk (see Appendix 3). A local risk assessment of these isolation areas is recommended.
  • Procedures with no good evidence supporting their use should be minimised. Such procedures include administering nebulisers or hypertonic saline to infants with bronchiolitis.

Recommendations - on admission to paediatric ward / HDU

  • All patients with lower respiratory tract infections need to be admitted into a cubicle until their virology results are available. However, if cubicle capacity is limited, a risk assessment is required to prioritise cubicle use. RCPCH-defined clinically extremely vulnerable (CEV) children must be prioritised to a cubicle. Children with respiratory tract symptoms in whom no AGPs are being performed (including those with viral induced wheeze) can potentially be cohorted awaiting SARS-CoV-2 results. However, 2 metre spacing between beds/cots should be maintained (with consideration of tape around the bedspace), use of curtains where possible (consider use of clear cubicle curtain), adherence with infection control procedures by parents/carers (use of face covering, maintaining 2 metre social distancing and complying with hand hygiene), ventilation of the bay reviewed and environmental cleaning optimised. AGPs must not be performed in this area. Staff looking after children in this cohort area can wear droplet PPE as per level of risk (see Appendix 3). 
  • If the patient has a non-SARS-CoV-2 pathogen identified which is consistent with their clinical phenotype, they can be moved into a bronchiolitis bay. Care needs to be taken with pathogens such as rhinovirus and bocavirus, which may be identified but may not be responsible for the clinical presentation. In addition, a risk assessment on the whole household needs to be conducted in terms of recent symptoms consistent with COVID-19 in parents/carers such as fever, prolonged cough and loss of smell/taste. If present, consider an urgent second SARS-CoV-2 test before moving the patient to a bronchiolitis bay.
  • RCPCH-defined clinically extremely vulnerable (CEV) children, as well as other children routinely requiring protective isolation, should not be managed in a bronchiolitis bay irrespective of their virology results. If cubicle capacity is limited, a risk assessment needs to be conducted.
  • If a patient is negative for all viruses (including COVID-19), a second SARS-CoV-2 test should be considered before moving the patient into a cohort bay. This is especially important if the child is on HFNCO / requiring AGPs, or members of the household have symptoms consistent with COVID-19. There is no need to delay performing this second test.
  • Patients who are confirmed to be SARS-CoV-2 negative can be managed in a cohort bay unless they require protective isolation. It is best practice to cohort children with the same pathogen. If this is not possible, for example due to bed pressures, then an organisational risk assessment should be undertaken and infection prevention and control precautions must be maintained to minimise the risk of nosocomial infection. These include adherence with hand hygiene, PPE and environmental cleaning.
  • Airborne precautions are not required for AGPs on patients in the low-risk COVID-19 pathway, providing the patient has no other infectious agent transmitted via the droplet or airborne route (see Appendix 3).
  • In order to enable a low-risk/'green' status to be maintained on a patient remaining in hospital, weekly SARS-CoV-2 testing should be performed or if further symptoms are identified.
  • If a patient is positive for SARS-CoV-2, they should be isolated in a single room applying the correct PPE (droplet/airborne) in accordance with care / procedures being performed as per the high-risk pathway. If AGPs are performed on a SARS-CoV-2 positive patient, they should be managed using transmission-based precautions ensuring safe systems of work are in place for donning and doffing PPE.
  • If HFNCO is initiated, a clear plan should be in place to promote rapid weaning (see Appendix 2).
  • Discharge of infants with bronchiolitis from an inpatient setting should be considered if oxygen saturations in room air is ≥90% and there are no other clinical or social indications for continued admission1.
  • Discharge from hospital should not be delayed if the SARS-CoV-2 result is not available. The child and family can continue to isolate at home until the result is available.

Recommendations - children being transferred to PICU

  • Virology samples should be sent from the referring hospital / ED, where possible. Consider a point of care test if routine laboratory results are not available. If local testing identifies a viral pathogen and the patient is negative for SARS-CoV-2, it may avoid unnecessary admission into a cubicle on PICU.
  • The principles outlined in the above recommendations on admission to paediatric ward / HDU also apply to children moving from cubicles on PICU to cohort areas on PICU.
  • Members of the retrieval team should wear the correct PPE as per level of risk (see Appendix 3).
  • Airborne precautions are not required for AGPs on patients in the low-risk COVID-19 pathway, providing the patient has no other infectious agent transmitted via the droplet or airborne route.
  • If the patient is positive for SARS-CoV-2, they should be managed in a designated COVID (high-risk pathway) area and staff must wear the appropriate PPE (see Appendix 3).
  • A child who requires repatriation from PICU to a local hospital should be given priority over an elective admission to facilitate flow of severely unwell children into and out of PICU. If a child has had a negative COVID-19 swab in the preceding 7 days, they do not require placement in a cubicle unless there is a separate indication for source or protective isolation.

Recommendations - parents and carers

  • Resident carers should not be in the hospital if they have symptoms of COVID-19. If both parents/carers are symptomatic, SARS-CoV-2 testing may be considered and a local risk assessment conducted.
  • All resident carers should wear a face covering whilst in hospital if away from their bed-space. Variations in local policy should be taken into account.
  • Resident carers should be minimised as far as possible, with ideally one accompanying each child. When children require an inpatient stay, local policy should be followed. Limiting changeover between named carers from different households should be considered. Ideally, resident carers should not have a co-morbidity that places them in a high-risk category.
  • Education and written information for resident carers should be made available regarding local policies, and include use of communal facilities, face coverings, hand hygiene, PPE and social distancing.

Guidance on escalating infection control processes if regional prevalence rates rise

Regional prevalence rate data are provided by the PHE modelling team to the Paediatric Critical Care Operational Delivery Networks on a weekly basis. It is the recommendation of PHE that regional prevalence data are used rather than local rolling period incidence data.

The dynamics of epidemics are such that high discordance between local and regional prevalence is unlikely; this is the rationale for basing decisions on regional prevalence data. It is hoped that low to moderate rates of regional prevalence will be maintained throughout the winter. The PHE modelling team will continue to look at local issues and in the event of an unusual localised threat, it will be reported as an exception to the network. 

Table 1

Escalation of infection control processes according to regional prevalence rates of COVID-19 during the pandemic

Low rates Moderate rates High rates
Prevalence <0.5% Prevalence >0.5%, but <2% Prevalence >2%
Follow guidance within this document

Watch the situation closely, including doubling time / growth rates, to judge whether the local situation is worsening

Ensure infection control measures in hospital (eg use of face coverings by parents/carers, hand washing) are being actively audited

Consider limiting visiting to one parent/carer for duration of admission (or swapping weekly)

Consider escalation of infection control processes including some or all of the following:

  • Increase the frequency of regular COVID-19 testing in patients undergoing AGPs
  • Mandate aerosol PPE for all staff managing low-risk patients undergoing AGPs
  • Tighter restrictions on visitors, such as limiting the frequency of changeover of resident parents/carers, or limiting carers to one resident parent for the entire stay
  • COVID-19 testing of resident parents/carers on admission and regularly during admission
  • Daily screening of symptoms in resident parents/carers
  • Regular COVID-19 testing of staff

Table 2

Temporal changes in prevalence of COVID-19 during the pandemic

Updated 12 April 2021 The table below shows how prevalence has changed between the initial peak, 23 March 2020, and now.

  Prevalence (Mean,95% CI)
Location 29 March 2020 02 April 2021 09 April 2021 16 April 2021 (estimated) 

3.7% (3.1-4.8)

North East 

2.5% (2.0-3.3)


Yorkshire and The Humber

2.5% (2.0-3.3) 0.22%
North West

3.8% (3.0-4.9)


East Midlands

3.5% (2.8-3.5)

0.33%                (0.14-0.79)

0.26%                    (0.08-0.84)

0.18%               (0.04-0.78)

West Midlands

3.5% (2.8-3.5)

East of England

3.0% (2.4-3.9)


9.0% (7.3-11.6)

South East

2.2% (1.8-2.9)


South West

1.1% (0.9-1.5) 0.04%

Mitigating risk if recommendations cannot be met

It is acknowledged that there is considerable variation between hospitals in terms of isolation capacity (cubicles), turnaround times for SARS-CoV-2 results and access to full respiratory virus panels. This may make it extremely challenging to comply with the recommendations made within this document whilst maintaining flow of patients.

In this situation, a pragmatic approach needs to be adopted based on infection control principles and relative risks. Weighing up of various factors including patient factors (extreme vulnerability, continuation of AGPs), staff factors (vulnerability of staff working within cohort areas), geographical factors (ventilation of cohort areas, distance between bed-spaces), regional prevalence rates and access to testing (turn-around time for SARS-CoV-2 testing, access to respiratory virus panels) is required.

It is recommended that local medical, nursing and infection control teams collaboratively develop and regularly review clinical pathways and contingencies based on local risk assessments. In addition, if virology samples are sent to regional virology units, it is recommended that discussions about prioritisation of samples and access to point of care or rapid testing take place.

Appendix 1 – Indications and contraindications for HFNCO in children and young people

Courtesy of North and South Thames Paediatric Networks and retrieval services

Indications (not exhaustive) Contraindications Cautions
  • High oxygen requirement
  • Signs of respiratory distress
  • Post extubation if clinically infected
  • Nasal obstruction or craniofacial abnormalities
  • Trauma/surgery to nasopharynx
  • Recurrent apnoea
  • Respiratory arrest or peri-arrest state
  • Undrained pneumothorax
  • Drained pneumothorax
  • Upper airway obstruction

Appendix 2 – Example guidance on commencing and rapid weaning from HFNCO

Courtesy of North and South Thames Paediatric Networks and retrieval services

Commencing treatment

  1. Select interface and equipment based on local availability and patient age and weight. Interface size should not exceed 50% of nares. If flow rate according to weight cannot be achieved on the correct interface, then use maximum flow for interface.
  2. On initiation a competent clinician should observe the patient for comfort and compliance. If necessary the flow can be increased to reach the maximum recommended range according to weight, over a five-minute period.
  3. Titrate FiO2 to maintain SpO2≥92% (or alternative patient range).
  4. Escalate or wean. To avoid rapid deterioration or unnecessary continuation on heated humidified high flow therapy (HHHFT), review response to HHHFT and follow the escalation or weaning criteria below.
<12 kg 2 l/min/kg
13-15 kg 20-30 l/min
16-30 kg 25-35 l/min
31-50 kg 30-40 l/min
>50 kg 40-50 l/min

Response to treatment

Sustained response to HHHFT

Nursing ratio 1:4 or 1:3 <2 years

Response to HHHFT

Nursing ratio 1:2 or 1:3 if cohort is ward level

Unresponsive to treatment
Wean FiO2 to 0.3-0.4 (depending on patient)

Moderate respiratory distress continues

and/or FiO2>0.4-0.6

In the first hour


Halve the flow rate


If no clinical deterioration is seen after 4 hours, HHHFT can be discontinued (or as soon as 1 hour if paediatric consultant confirms)


Restart at weaning flow rate if stopping HHHFT is not tolerated

Re-assess essential care considerations** and continue on current HHHFT settings until ready to wean


Continue to observe for any deterioration or red flags*

  • Re-assess essential care considerations**
  • Ensure paediatric consultant has reviewed the patient
  • Discuss with the retrieval service
  • Discuss/review with the anaesthetic registrar
  • Closely observe for any red flags*

After 2nd hour or with any red flags*:

  • Consider NIV or invasive mechanical ventilation (IMV)
  • Prepare patient, team and family for intubation
* Red flags for immediate escalation Immediate reaction
  • Any apnoeic/bradycardic episodes
  • Increasing respiratory distress after HHHFT commenced
  • Clinically tiring
  • The Paediatric Early Warning System (PEWS) indicates immediate escalation to resus team
  • FiO2 >0.6
  • Increase FiO2 to maximum
  • Call 2222
  • Prepare for intubation
  • Liaise with retrieval team or on-site Level 3 paediatric critical care
  • Communicate with the family

Monitoring and patient management

(with corresponding patient acuity)

  • Continuous oxygen saturations (green, amber, red)
  • Observation frequency and escalation according to PEWS (green)
  • Minimum hourly observations and escalation according to PEWS (amber, red)
  • Consider continuous electrocardiogram (ECG) if required (amber, red)
  • 2 hourly mouth and nose care including pressure area check (green, amber, red)
  • Hourly documentation of FiO2, flow rate, and temperature as well as equipment specific checks (green, amber, red)

** Essential care considerations

  • Optimised positioning (e.g. head elevation).
  • Consider referral for physiotherapy assessment.
  • Secretion clearance if indicated and safe to do so.
  • Consider feeding regime alteration according to risk and underlying disease:
    • High risk (red) should be nil by mouth (NBM) with intravenous fluids.
    • Medium risk (amber) should be assessed before feeding and fed with caution.
  • Psychosocial support, clear communication, play and distraction.
  • Minimal handling / cluster cares.
  • Blood gas analysis not essential and acidosis is a late sign of failure.

Patient transfer

If patient transfer is required then a suitable risk assessment tool should be used. Where portable HHHFT is not available, a senior clinician should assess the appropriate oxygen delivery based on direct patient assessment.

    Appendix 3 – PPE requirements based on risk stratification

    Adapted from guidance issued jointly by the Department of Health and Social Care (DHSC), Public Health Wales (PHW), Public Health Agency (PHA) Northern Ireland, Health Protection Scotland (HPS) / National Services Scotland, PHE and NHS England on 20 August 2020)

    Amber pathways are not included as patients presenting with a lower respiratory tract infection will either be designated to the high-risk/'red' pathway at presentation awaiting test results / following a positive SARS-CoV-2 test, or a low-risk/'green' pathway if SARS-CoV-2 has been excluded.

    High-risk Low-risk

    Confirmed SARS-CoV-2 (COVID-19) positive individuals 


    symptomatic or suspected COVID-19 individuals including those with a history of contact with a COVID-19 case, awaiting test results

    SARS-CoV-2 excluded following testing

    Droplet/contact PPE if contact with suspected/confirmed COVID-19 patient/individual:

    • single use disposable gloves
    • single use disposable apron (gown if risk of spraying/splashing)
    • FRSM Type IIR face mask for direct patient care
    • single use or re-usable eye/face protection (visor)

    Airborne PPE* when undertaking AGPs on confirmed or suspected COVID-19 patients:

    • single use disposable gloves
    • single use disposable gown
    • FFP3 or Hood for AGPs
    • single use or re-usable eye/face protection (visor)
    PPE** if contact with blood and/or body fluids is anticipated (all settings / all patients/individuals):
    • single use disposable gloves
    • single use disposable apron (gown if risk of spraying/splashing)
    • surgical mask Type II for extended use* / FRSM Type IIR for direct patient care*
    • risk assess and use eye/face protection (visor) if required for care procedure / task where there is anticipated blood / body fluids spraying/splashes

    * Extended use of facemasks in England/Scotland for healthcare workers when in any healthcare facility.

    **Airborne precautions are not required for AGPs on patients in the low-risk/'green' COVID-19 pathway. 

    Methodology for developing recommendations

    Key stakeholders representing national groups (Royal College of Paediatrics and Child Health, British Paediatric Respiratory Society, Association of Paediatric Emergency Medicine, Paediatric Intensive Care Society, British Paediatric Allergy Immunity & Infection Group, NHS England/Improvement Infection Prevention & Control Cell), and professional groups (paediatric infectious diseases, infection control, virology, general paediatrics, PICU) were identified to support the development of these recommendations.

    The group met virtually on 21 August 2020 and again on 3 September 2020. Each step in the patient pathway was discussed systematically by the group, in terms of place of admission / patient flow, virus testing, PPE requirements and use of HFNCO, prior to developing the consensus recommendations.

    Final consultation included executive committees from all national groups mentioned above. Publication was approved by the RCPCH Winter Pressures Clinical Advisory Group and Senior Officers.

    Steering group


    • Dr Sanjay Patel, Paediatric Infectious Diseases Consultant, Southampton Children’s Hospital

    Clinical Advisors:

    • Dr Conor Doherty, Paediatric Infectious Diseases Consultant, NHS Greater Glasgow & Clyde
    • Dr Danielle Eddy, Paediatric Specialty Trainee, Gloucestershire Hospitals NHS Foundation Trust
    • Helen Dunn, Lead Nurse for Infection Prevention Control, Great Ormond Street Hospital
    • Dr Hermione Lyall, Paediatric Infectious Diseases Consultant, Imperial College Healthcare NHS Trust
    • Dr Ian Maconochie, Paediatric ED consultant, Imperial College Healthcare NHS Trust 
    • Dr Ian Sinha, Paediatric Respiratory Consultant, Alder Hey Children’s Hospital
    • Dr John Criddle, Paediatric ED Consultant, Evelina London Children's Hospital
    • Dr Julian Legg, Lead for Paediatric Respiratory Medicine, Southampton Children’s Hospital 
    • Dr Liz Whittaker, Paediatric Infectious Diseases Consultant, Imperial College Healthcare NHS Trust
    • Dr Matthew Clarke, NHSE National Specialty Advisor for Children and Young People 
    • Dr Mike Linney, General Paediatric Consultant and Registrar for RCPCH 
    • Dr Padmanabhan Ramnarayan, PICU Consultant, Imperial College Healthcare NHS Trust
    • Dr Paul Randell, Consultant Virologist, Imperial College Healthcare NHS Trust
    • Dr Poonamallee Govindaraj, Paediatric Consultant, Cwm Taf Morgannwg University Health Board
    • Dr Raymond Nethercott, General Paediatric Consultant and RCPCH Officer for Ireland
    • Dr Ruchi Sinha, PICU Consultant, Imperial College Healthcare NHS Trust
    • Samantha Matthews, NHSE/I Infection Prevention & Control National Clinical Lead
    • Dr Sean O’Riordan, Paediatric Infectious Diseases Consultant, Leeds Children’s Hospital
    • Professor Simon Kenny, NHSE National Clinical Director for Children and Young People


    24 September 2020 - update to recommendations on admission to paediatric ward / HDU and revised flow chart with clarification on risk assessment and 2nd SARS-CoV-2 PCR testing.

    19 October 2020 - update to table 1 column 3 (high rates, prevalence >2%) and addition of guidance for mitigating risk if recommendations cannot be met.

    • 1. Cunningham S, Rodriguez A, Adams T, et al. Oxygen saturation targets in infants with bronchiolitis (BIDS): a double-blind, randomised, equivalence trial. Lancet. 2015;386(9998):1041-1048. doi:10.1016/S0140-6736(15)00163-4