Skip to main content

Prioritizing Covid-19 patients for intensive care, hospitalization and vaccines

Prioritizing Covid-19 patients for intensive care, hospitalization and vaccines

The Covid-19 pandemic put health systems worldwide under enormous pressure. Many Covid-19 patients needed to be hospitalized, with some so sick they required life-saving treatment with ventilators in intensive care units (ICU).

Also, Covid-19 vaccines, or the support services for administering them, were often initially in short supply. Low- and middle-income countries (LMIC), such as in Africa and the Pacific Islands, were potentially under even more pressure than richer countries with well-developed and well-funded health systems.

During this time of global crisis, 1000minds is proud to have been called on to help with creating tools for prioritizing: (1) patients for hospitalization, (2) patients for intensive care, (3) people to be vaccinated, and (4) countries to receive vaccine supplies.

2 1000minds tools for prioritizing Covid-19 patients

2 1000minds tools for prioritizing vaccines

200+ clinicians and infectious-diseases experts involved in creating the tools

As the pandemic raged around the world, decision-support tools to help with ethically difficult, often ‘life or death’, decisions were urgently needed – about who should be admitted to hospital and ICU respectively (and who not), who should be vaccinated and which countries should receive vaccines first.

Groups of clinicians and infectious-diseases experts in Italy, New Zealand, Tunisia and from around the world used 1000minds to create the four prioritization tools for the applications detailed below.

About the clinicians and infectious-diseases experts

Groups of clinicians and infectious-diseases experts used 1000minds to create prioritization tools for the following applications.

  1. Prioritizing Covid-19 patients who are non-critical but at risk of deterioration for admission to hospital, including in low-resource settings like low- and middle-income countries (LMICs). More than 100 Italian clinical experts with experience in treating Covid-19 patients were involved.
  2. Prioritizing critical Covid-19 patients for admission to ICUs in NZ hospitals. A dozen specialists from intensive care medicine and nursing, Māori health, infectious diseases and neonatology were involved, supported by decision analysts and overseen by an ethicist and the Ministry of Health.
  3. Prioritizing people in Tunisia to be vaccinated when vaccines were in short supply. Sixty experts and frontline physicians treating Covid-19 patients in Tunisia were involved.
  4. Prioritizing countries to receive vaccine supplies. From across 13 countries or with a global focus, 28 public health experts were involved.

The challenge

Prioritizing a waiting list of patients.

During the pandemic, in many countries there were not enough hospital beds or ventilators available, and so Covid-19 patients urgently needed to be prioritized.

For example, at the start of the pandemic in February 2020 New Zealand had only 4.6 ICU beds per 100,000 population – e.g. compared with 12.5 in Italy and 29.4 in the US. It was feared that even a low Covid-19 prevalence would quickly overwhelm NZ’s ICU capacity.

The Australian and New Zealand Intensive Care Society emphasized that the allocation of intensive care resources must be “consistent, transparent, objective and ethical” (Warrillow 2022, p. 98). The development of a prioritization tool based on valid and reliable criteria and weights if ICUs were to become overwhelmed was identified as a vital preparation.

Likewise, in many countries, especially LMICs, there were insufficient supplies of Covid-19 vaccines, or the support services for administering them, and so people needed to prioritized for vaccinations.

Finally, the global allocation of vaccines is largely determined by countries’ ability to pay and their vaccine manufacturing capacities, which favors rich countries over poorer countries. And yet, based on the risks facing their populations, countries have very different vaccine needs, and so a tool for prioritizing countries for vaccines was required.

All four groups of clinicians and infectious-diseases experts who used 1000minds (see above) had a very strong desire to create valid and reliable prioritization tools that were also fair and transparent.

Also, the objective for the two patient prioritization tools for hospital and ICU admissions was to reduce the burden and distress for clinicians making these extremely difficult ‘life or death’ prioritization decisions.

The solution

Which of these two imaginary Covid-19 patients should be admitted to ICU first (i.e. prioritized)?
Age (a risk factor, independent of other criteria)
19-39 yrs old
Pre-existing respiratory conditions
mild, e.g. FEV1 > 80%, mild asthma
Age (a risk factor, independent of other criteria)
≤ 18 yrs old
Pre-existing respiratory conditions
moderate, e.g. FEV1 40-80%, moderate asthma, heavy smoker (> 20/day)
This one
This one
They are equal

As an example, here is how the tool for prioritizing Covid-19 patients for hospital admissions (#1 earlier) was developed and how it works.

In early 2020, after reviewing the (at the time) emerging medical literature on predictors of outcomes for Covid-19 patients, five experts in infectious diseases specified 11 prioritization criteria suitable for use in LMICs.

Weights for the criteria, representing their relative importance, were established by surveying 103 Italian clinical experts using 1000minds.

The criteria (and their levels) and their weights – the prioritization tool – are reported in the table below.

Table 1: Prioritization tool for admitting Covid-19 patients to hospital
Partial pressure of oxygen (PaO2)
>80 mmHg 0%
70-80 mmHg 8.2%
65-70 mmHg 16.3%
Oxygen saturation
>96% 0%
92-96% 8.3%
<92% 15.9%
Findings at chest X-ray
normal 0%
consolidation 7.2%
bilateral interstitial lung abnormalities 14.1%
Modified early warning score (MEWS)
0-2 0%
3-4 11.4%
Respiratory rate (breaths/minute)
<20 0%
>20 9.5%
no 0%
yes 6.5%
Living with vulnerable people
no 0%
yes 6.4%
Body Mass Index
<30 0%
31-40 2.8%
>40 5.6%
Duration of symptoms
<3 days 0%
>7 days 2.9%
4-7 days 5.4%
C-reactive protein
normal 0%
high 5.1%
18-50 0%
51-70 1.9%
>70 3.8%
Note: the bolded values for each criterion sum across the criteria to one (100%), and so they represent each criterion’s relative importance.

The tool is implemented by rating each patient on the criteria in the table, and then summing the corresponding weights to produce a score (in the range 0-100%). Patients are ranked (prioritized) by their scores.

For example, imagine two patients who have been rated on the 11 criteria like this:

  • Fatima: ‘70-80 mmHg’, ‘92-96%’, ‘consolidation’, ‘3-4’, ‘>20’, ‘no’, ‘no’, ‘31-40’, ‘>7 days’, ‘high’, ‘18-50’.
  • Fred: ‘65-70 mmHg’, ‘>96%’, ‘normal’, ‘0-2’, ‘<20’, ‘yes’, ‘no’, ‘<30’, ‘>7 days’, ‘normal’, ‘>70’.

Thus, applying the weights in the table, out of a possible maximum of 100%, Fatima’s and Fred’s scores are:

  • Fatima’s score: 55.4%, i.e. 8.2 + 8.3 + 7.2 + 11.4 + 9.5 + 0 + 0 + 2.8 + 2.9 + 5.1 + 0
  • Fred’s score: 29.5%, i.e. 16.3 + 0 + 0 + 0 + 0 + 6.5 + 0 + 0 + 2.9 + 0 + 3.8

Because her score is higher than Fred’s, Fatima has priority over Fred for hospital admission.

The results

A prioritized waiting list of patients.

All four prioritization tools were developed quickly as the need for them arose.

For example, the process for prioritizing patients for hospital admission outlined in detail above was performed in March-April 2020, led by doctors in northern Italy where the health system was close to collapse as the pandemic was at its worst.

Likewise, the tool for prioritizing patients for ICU in New Zealand was developed over a period of just 10 days as the country was preparing to go into national lockdown, with the tool ready for use just 27 days after NZ’s first reported Covid-19 case and one day after lockdown.

Fortunately for NZ, this tool was not used because the disease was successfully controlled by public health measures. Despite not having been tested ‘in the field’ (ICU), the tool’s rapid development clearly demonstrates its practical feasibility.

As well as being capable of being rapidly developed, all four Covid-19 prioritization tools are systematic, transparent and evidence-based. They differentiate between patients and countries respectively in a fair way, consistent with established ethical principles.


This is the first time multi-criteria decision analysis (MCDA) has been used during a pandemic to codify experts’ knowledge to rapidly create tools for prioritizing patients for hospital and ICU.

The speed with which the tools were created confirmed the efficiency of 1000minds. In the event of new emergencies, new prioritization tools could be created from scratch in a matter of just a few days.

Alternatively, the Covid-19 tools could be adapted to different settings or updated in just a few hours. Because all the data from the process are stored electronically by 1000minds, the criteria and weights can be easily revised – e.g. as new emergencies and evidence emerges – without repeating the whole process.

For example, if Covid-19 were to threaten to overwhelm ICUs and hospitals again, or a new virus emerged, the method could be rapidly reapplied to update or revise the prioritization tools to incorporate the latest evidence about the disease.

Peer-reviewed publications

About the use of 1000minds to create tools for prioritizing: (1) patients for hospitalization, (2) patients for intensive care, (3) people to be vaccinated, and (4) countries to receive vaccine supplies.


S Warrillow, D Austin, W Cheung, et al (2020), “ANZICS guiding principles for complex decision making during the Covid-19 pandemic”, Critical Care and Resuscitation 22, 98-102.

Back to top