Developing a Thai national critical-care allocation guideline during the COVID-19 pandemic: a rapid review and stakeholders’ consultation

Background: At the height of the COVID-19 pandemic, Thailand had almost depleted its critical care resources, particularly ICU beds and ventilators. This prompted the necessity to develop a national guideline for resource allocation. This paper describes the development process of a national guideline for critical resource allocation in Thailand during the COVID-19 pandemic. Methods: The guideline development process consisted of three steps (1) rapid review of existing rationing guidelines and literature; (2) interviews of Thai clinicians experienced in caring for COVID-19 cases; and (3) multi-stakeholder consultations. At steps 1 and 2, data was synthesized and categorized using a thematic and content analysis approach and this guided the formulation of the draft guideline. Within Step 3, the draft Thai critical care allocation guideline was debated and �nalized before entering the policy-decision stage. Results: Three-order prioritization criteria consisting of (1) clinical prognosis using four tools (Charlson Comorbidity Index, Sequential Organ Failure Assessment, Frailty Assessment, and Cognitive Impairment Assessment); (2) number of life years saved; and (3) social usefulness, were proposed by the research team based on literature reviews and interviews. At consultations, stakeholders rejected using life years as a criterion due to potential age and gender discrimination, as well as social utility due to a concern it would foster public distrust as this judgment can be arbitrary. It was agreed that attending physician is required to be the decision maker in the Thai medico-legal context while a patient review committee would play an advisory role. Allocation decisions are to be documented for transparency and no appealing mechanism is to be applied.This guideline will be triggered only when demand exceeds supply after the utmost efforts to mobilize surge capacity. Once implemented, it is applicable to all patients, COVID-19 and non-COVID-19, requiring critical care resources prior to ICU admission and during ICU stay. Conclusions: The guideline development process for the allocation of critical care resources in the context of the COVID-19 outbreak in Thailand was informed by scienti�c evidence, medico-legal context, existing norms and societal values to reduce risk of public distrust given the sensitive nature of the issue and ethical dilemmas of the guiding principle though it was conducted at recorded speed.


Background
The Coronavirus disease 2019 (COVID-19) outbreak was declared a global pandemic by the World Health Organization on 11th March 2020 (1).To date, there has been almost 15 million con rmed cases and over 600,000 deaths (2).Across the world, COVID-19 has overwhelmed healthcare systems and their capacity to respond.There have been extensive reports of inadequate supply of personal protective equipment for healthcare workers and shortages of intensive care beds and ventilators in many countries (3).The surplus in demand exceeding the availability of healthcare resources has led to the unavoidable rationing of medical equipment and interventions, notably critical care resources which are challenging to expand in a short time (3)(4)(5)(6)(7).As a result, several countries have been compelled to develop national resource allocation guides, speci c to COVID-19 and country context (8)(9)(10)(11)(12).
Thailand reported the rst case of the virus outside of China in early January 2020 (13,14).By the end of January 2020, the rst known person-to-person transmission in the country was documented and a total of 3,255 cases and 58 deaths have been con rmed as of 21 st July 2020 (15).The Ministry of Public Health (MOPH) maximized, expanded, and mobilized critical care resources from all public and private health facilities throughout the country, and earmarked 10% of all ICU beds for COVID-19 patients while keeping the remaining for non-COVID-19 critical patients(16).However, at the peak of the epidemic (22 nd March -3 rd April), there were between 91 to 188 new cases per day, causing the number of available ICU beds and airborne infection isolation rooms in the country to almost reach its maximum capacity, placing signi cant strain on health facilities (15,16).Recognizing that the current critical-care resource capacity may reach its threshold, the development of a national guideline for resource allocation was deemed necessary.
In the context of developing a guideline during a public health emergency, such as the COVID-19 pandemic, the standard timeframe of 6 months to 3 years for guideline development is not appropriate, therefore a rapid guideline development approach is needed (17,18).This paper aims to describe the rapid development of a critical care resource allocation guideline in Thailand during a public health emergency and share lessons learnt that may serve as a useful example for other countries.

Methods
In response to the COVID-19 pandemic, the Health Intervention and Technology Assessment Program (HITAP) and International Health Policy Program (IHPP), were tasked by the MOPH with developing a guideline for prioritizing critical care resources.The two units, partnered with the Thai National Health Commission O ce (NHCO), formed a technical team to lead the guideline development.In order to develop an evidence informed guide within a short time frame, the technical team designed three steps for guideline development: (1) rapid review of existing rationing guidelines; (2) interviews of clinicians experienced in caring for COVID-19 cases; and (3) multi-stakeholder consultations.

Rapid Literature Review
The main objective of the literature review was to identify existing national and international guidelines, recommendations and frameworks on critical care resource allocation, patient triage policies, and resource management during pandemic situations.A preliminary search on PubMed database using the keywords 'rationing' OR 'allocation' OR 'COVID' yielded few academic papers at the time of search, with mostly commentaries on ethical principles, challenges, and recommendations rather than allocation guides as intended, due to most usually being published as grey literature (19,20).Therefore, the technical team handsearched from the International Society for Priorities in Health website, which established a COVID-19 platform comprising of academic literature on priority setting, guidelines speci c or applicable to COVID-19 from various sources (17,18,21).Guidelines were screened for relevance and 11 guides from seven countries and one international agency was selected, based on an inclusion criterion of being published between 2010 and 2020 and exclusion criteria of not being available in a language that could be readily translated by the technical team.

Key Informant Interview
This process aimed to help the technical team understand the current practices on ICU admissions, patient triage methods and critical care resource allocation in Thailand as well as the need and feasibility of implementing a new guideline.Two ICU physicians working in Thai public hospitals, were able to participate in the interviews through telephone call.

Stakeholder Consultation
As public health emergencies pose several ethical, social and legal dilemmas, the stakeholder consultation aimed to incorporate the views and expertise of relevant groups to ensure the guidelines feasibility and acceptability to Thai society (17).Broad-base stakeholder groups were identi ed by the technical team, aiming for those with a technical expertise in a relevant eld and, also a sound understanding of health policy processes.To ensure full deliberation by stakeholders, two rounds of half-day consultations were convened, the rst for medicine and medical law stakeholders and the second policy makers and social science stakeholders (Table 1).Documents pertaining to the details of the proposed guideline developed by the technical team were shared with all stakeholders prior to the consultation.Participants attended either in-person or via teleconference call.The consultation consisted of a short presentation of the proposed guideline and courses of action to consider, followed by an open discussion.At each round of consultation an open forum approach was utilized and moderated by a senior member of the technical team to allow for the incorporation of all views and to reach consensus on key issue areas.The rst round of consultation informed a revised guideline presented during the second round.The nal text, after consultation, was circulated to all stakeholders for their further written comment or endorsement.

Data analysis and synthesis
All included documents from the rapid literature review were reviewed and key data extracted using a dataextraction form.Information from each interview was summarized by the interviewee and key issue areas were identi ed through discussion among the technical team.Immediately following each consultation, the technical team summarized and categorized all issues into similar thematic areas using a thematic and content analysis approach.

Existing guidelines
Data extracted from the rapid review of 11 included guides are summarized in Table 2. Contents were extracted into the following key themes: guiding principles, target patients, prioritization criteria, decision maker and process, and implementation.

Guiding Principles and Prioritization Criteria
The prioritization criteria and tools applied for each guide was positioned by the authors on a scale ranging from social function, which promotes and rewards instrumental value or bene ts to others, to clinical prognosis, which gives value to clinical success, the number of lives and life years that can be saved (See Figure 1).Social usefulness, such as patient's occupation, was applied only for the New York guide.Italy and Pittsburgh both utilized patient age, and it was indirectly assessed in the guides from Switzerland, Belgium and New York.Clinical frailty which involves assessing status of cognitive and physical function is applied in the United Kingdom, Belgium and Germany.Cognitive impairment assessment, measuring brain function and the patients' medical urgency were criteria only applied in the Belgium guide.Sequential Organ Failure Assessment (SOFA), a tool used to estimate and quantify the number and severity of potential organ failure, was used in New York, Pittsburg, Germany and Austria.Comorbidities, such as prior medical conditions were stated as criteria in the guides from Switzerland, Belgium, New York and Germany, and indirectly considered for United Kingdom, Italy, Pittsburgh and Austria guides.Additionally, rst-come, rst served method applied in New York and Belgium, as well as randomization, utilized in New York, were excluded from the scale.This wide-ranging scale obtained from the rapid literature review re ects differing societal and cultural values given on allocation criteria.

Target Patients and Guideline Application
All guides indicated the prioritization criteria was to be applied to all patients requiring critical care resources, which includes both COVID-19 and non COVID-19 patients, to ensure that everyone had the same chance of accessing to the scarce resources.The prioritization criteria for most guides was recommended to be applied on admission to intensive care units (ICU).Additionally, some guides recommended reassessment following ICU admission: daily for the Belgian and Italian guides, every 2-3 days for the Swiss guide, and every 48 to 120 hours for the New York ventilator allocation guide.

Decision making process
The recommendation for the primary decision maker for the guidelines differed.Decisions to be made by a team of healthcare professionals involved in the patient's care was suggested in Switzerland, United Kingdom, Belgium and Germany.While United States guides from Pittsburgh and New York recommended the formation of a triage committee or nominating a triage o cer to make decision, sparing those involved in direct patient care.Similarly, the Austrian and WHO guide suggested the nomination of an intensive care specialist as the decision maker.In addition to the primary decision maker, most guides recommended consultations to be made with an ethics committee (Switzerland, Germany, Austria and United States), technical or designated experts (United Kingdom, Belgium, Germany, New York, Pittsburgh, Italy) or with patients and/or relatives (Italy, Austria, United States).Three out of the 11 guides also provided information regarding the process of appeals against decisions made for the patient.The WHO guidance stated that mechanisms to resolve disputes are necessary.The Pittsburgh guide recommended the formation of a Triage Review Committee to review the appeal using majority vote, while the Italian guide suggested appeals be reviewed by a designated experts or regional health centers.Decision-making processes varied between settings re ecting diverse medico-legal and clinical practices across the world.

Implementation
Although all guidelines were developed for the pandemics or outbreaks, only the guides from Belgium, Germany, Pittsburgh, New York, Italy, Austria and WHO, clearly speci ed its application only upon demand exceeding supply.No guides stated legal mechanisms for enforcement; instead all were voluntary and nonbinding recommendations.This ensured the guides could be exible and adapted to suit the situation and context of each health facility and the changes in clinical data.However, the New York guideline discussed the concern of lack of statutory protection for healthcare workers and institutions.

Current Practices in Thailand
The key informant interview con rmed that no protocols or guidelines on allocating critical care resources currently exist or are being applied in Thailand though they agreed that such guideline would be useful to ensure consistent approach across patients and facilitate patient referral across hospitals during public health emergency.

Proposed Draft Guideline
Based on the rapid review and key informant interviews, a draft guideline was developed by the technical team to be proposed at the stakeholder consultations.At both rounds of consultations, the technical team proposed four areas for discussion 1) Criteria for patient assessment in allocating resources, 2) Decision maker and decision-making process, 3) Appeal and documentation mechanisms, and 4) Process of implementation and enforcement.

Prioritization Criteria
The proposed criteria comprised of three-orders to be applied sequentially to break ties in decisions between patients with the same level of priority.In the context of limited critical care resources, the rst order criteria aimed to assess patients based on short-term clinical prognosis and maximize the health outcomes.Four assessment tools, that would be chosen based on applicability for each health facility setting, were put forward including; 1) Charlson Comorbidity Index, 2) SOFA, 3) Clinical Frailty Scale (CFS) and 4) Cognitive Impairment Assessment.The second order criteria was the number of potential life years saved, favoring long term survival.Lastly, the third order criteria would prioritize those with higher social utility, such as health care and social service workers.

Decision maker and process
In Thailand, according to the Medical Profession Act 1982 (2525 BE), the decisions on a patient's prognosis and treatment can only be made by the patients' attending physician.However, in the situation of scare critical care resources, to alleviate stress and ethical dilemmas faced by physicians, the formation of a Triage Committee was proposed to assist physicians on allocation decisions.The committee of three healthcare professionals, such as a physician, nurse and/or technical expert, has the primary responsibility to apply the prioritization criteria to each patient upon consideration of ICU admission and reassessment every 48 hours and advise physicians on the decision to give, to continue critical care or step-down care, such as to palliative care facilities (Figure 2a).

Appeal and documentation mechanisms
The documentation of all patients' assessment results and decision allocation during the pandemic is necessary for transparency.In addition, a two-step mechanism for appeals to be made by patients, relatives or legal representatives, were proposed for the stakeholders to consider.Firstly, an immediate appeal of prior allocation decisions for cases where a decision is disputed by the health facility.Followed by the review by an established review committee to verify disputed decisions through a majority vote.

Implementation and Enforcement
To ensure the guideline is utilized in appropriate circumstances, the guideline was proposed to be triggered when only 10-20% of critical care resources remain available and to be applied to all patients requiring critical care resources, both patients affected by the pandemic or those with unrelated conditions.The guideline was proposed to be endorsed and issued as a legal document by Medical Council of Thailand, to ensure consistent application across all public and private facilities in Thailand and to provide legal protection to medical doctors who adhere to it.

Key Stakeholder Concerns and Considerations
During the consultations, ve major concerns were identi ed 1) Ethical principles, 2) Criteria to be used for prioritization, 3) Decision makers and decision-making process, 4) Transparency and process of appeal and 5) Implementation and enforcement.

Ethics
At both rounds of expert consultations, considerable time was spent discussing the ethical principles that the Thai guideline should comply with.Experts compared the COVID-19 pandemic with war time where eld resources were scarce, and the military goal was for the greatest utility of the society.Applying this utilitarian ideology can be at direct odds with the medical ethics, bene cence and non-male cence principles that health workers practiced in normal situations with su cient resource.Without clear and motivated communication between doctors and patients, the application of 'rationing' resources may trigger high levels of stress in some health staff as well as the patients and their relatives.Both stakeholder groups emphasized the need for citizen awareness initiatives to accompany the guideline.Pre-counselling and informing citizens of the necessity to potentially allocate critical care resources, prior to the implementation of the guide was recommended to gain high level of acceptance and limit distrust by the population.

Prioritization criteria
The inclusion of age and social values as prioritization criteria was highly contentious in the consultations.
Unlike many countries, including Belgium, Italy, and USA (11,(22)(23)(24), which use patient age as a prioritizing criterion, most Thai stakeholders did not support an age criterion, even in the form of number of life years saved.Number of life years saved, calculated by life expectancy at birth minus patient age, was seen as linked directly to patient age and also gender given females has higher average life expectancy at birth than males (25).Therefore, the second-order criterion was rejected on the basis that it would contradict the nondiscriminatory basis that health professionals are required to adhere to.
Similarly, the third order criterion on social usefulness greatly concerned both groups of stakeholders.Many argued that while the criterion itself carries good justi cation, to compare social values of individuals is not feasible, decision can be arbitrary which may lead to public distrust.Stakeholders from the Buddhist and Muslim communities felt that all individuals hold their own intrinsic value (either for their own family or to the society) and this cannot be compared to one another.It is worth noting that stakeholders preferred all criteria to be objective and quanti able to enable a transparent and veri able decision-making process.

Decision makers
A fundamental reason behind establishing a rationing protocol is to alleviate the psychological burden placed on frontline health workers who would be otherwise tasked with deciding who receives potentially lifesaving resources(26,27).While stakeholders acknowledged that most guidelines spare attending physicians from making allocation decisions and assign that role to a designated staff or a triage team; stakeholders agreed that in accordance with the Thai medico-legal context, decisions on diagnosis and treatment for patients are to be made by attending physicians.Stakeholders also agreed with the need for a committee to play an advisory role to the attending physicians.However, the 'triage team', originally proposed was amended to 'Patient Review Committee' as the word 'triage' may carry a negative connotation in the Thai society.Medical stakeholders suggested the committee consist of ve other healthcare professionals (such as physicians, nurses or relevant experts).The civil society and religious group highlighted that a trusted religious or community leader in the Patient Review Committee would also increase the likelihood of patients and relatives accepting the decisions made; though a counterargument was made that a religious leader in the committee may not be applicable to multi-religious communities.It was agreed that a highly respected member of the community should be selected as a committee member.

Appeal mechanisms
Whether or not to establish an appeal mechanism was diligently discussed.On the one hand, allowing patients or representatives to appeal immediately should they disagree with the decision was seen as recognizing patients' voice and helping ensure objectivity.On the other hand, stakeholders found that an established appeal process could result in delays in decisions, decrease trust in the process and may add an element of unfairness.Stakeholders raised a concern that patients' ability to appeal is also linked to socioeconomic class.Civil society and social science stakeholders were in favour of having an appeal committee; while the law stakeholders and medical experts were not due to the added pressure put on attending physicians.A consensus was nally reached not to establish an appeal committee.Instead, clear and motivated communication between physicians and patients or their relatives about the process and decisions must be actively practised.Together with regular review and a transparent patient registry of what decisions are made to allow for future evaluation of the guide to ensure the guide ts its purpose was decided.

Implementation and enforcement
Although existing guidelines state that guides should be implemented only when demand exceeds supply, stakeholders stressed that it is di cult for frontline health staff to know exactly when surge capacities are fully mobilized and all needed resources in the country have been occupied.Therefore, medical stakeholders suggested the guideline should only be triggered after all efforts have been made to mobilize resources and demand for the resources continue to exceed the supply capacity.Stakeholders added that only when a national public health emergency has been declared and critical resource have been exhausted, should the guideline implementation be prompted.This requires higher-level public health o cers to routinely monitor and update the resource situation during a public health emergency.
Another matter relating to guideline implementation is enforcement.Similar to the New York Ventilator Allocation Guideline which aged a concern that there was no legal protection for health staff who adhered to its guideline (23), stakeholders from the medical eld emphasized that a legally binding mechanism would be necessary to support and protect healthcare workers in their decisions.A consensus was reached among all stakeholder groups that the endorsement of the guide by the Medical Council should be sought to ensure harmonized implementation nationwide, and adherence with existing laws and practices.

Guideline Finalization
Following two rounds of consultations, a revised guideline was developed by the technical team and circulated to all stakeholders for con rmation and suggested revisions via email if needed.Table 3 summarizes the guideline evolution at different steps of development and Figure 2 compares the proposed and nal version of decision-making steps.

•
No cut-off score is applied.

•
Apply three-order criteria:

•
No cut-off score is applied.

•
Assess patients based on clinical prognosis using at least 2 of the following tools: Charlson comorbidity index

3)
Frailty assessment such as Clinical Frailty Scale (CFS)

•
Allocation decisions are based on relative scores.

•
No cut-off score is applied.

•
Each health facility must apply the same sequence of tools consistently across all cases.
Step 1: Rapid Review (First draft) Step 2: Key Informant Interview (Second draft) Step

•
The guideline is to be endorsed by the Medical Council of Thailand

•
The guideline is to be endorsed by the Medical Council of Thailand.Current status of endorsement is unclear due to the pandemics changing situation.
Upon nalization of the guideline's contents, the guide was presented at a high-level governmental meeting of decision makers and academics, including executive board members of the Medical Council of Thailand.The importance and necessity for the guide was acknowledged.However, the favourable outcomes of COVID-19 containment in the country leading to an average of less than 3 new daily cases of COVID-19 between June 6 th and June 25 th 2020, resulted in the announcement of the third and nal phase marking the end of lock down measures in Thailand in July 1 st , 2020.This has made the endorsement of the guideline no longer a matter of urgency, and therefore it was decided that commencing legal process with the Medical Council should be put on hold.However, recognizing looming threats of new waves of infection after the unlocking phase, the guide can be adopted should the necessary situation arises.

Discussion
Clinical practice guidelines must be evidence informed, to ensure consistency with the best practices of the scienti c community, and the development process must involve a wide range of relevant expertise and stakeholders who may be affected by the recommendations to ensure applicability, feasibility, and relevance to the situations and contexts (17,28,29).As a result, the process of developing health guidelines often takes time, ranging from 6 months to 3 years to produce (17,18).This poses signi cant challenges for the development of guidelines during public health emergencies, where fast evidence informed, and context speci c recommendations are required.
In the context of the COVID-19 pandemic, especially at the peak of the outbreak where resources were quickly depleting, there was uncertainty in depleting resources in which the allocation recommendations may be needed.Additionally, in Thailand, as with many countries, the situation in which some lives have to be sacri ced due to insu cient resources is alien for many health professionals and therefore guidance on how to make these stark decisions is essential.The idea of 'rationing' healthcare is immensely sensitive and may cause alarm and foster unconstructive discourses in the general public.Therefore, the guideline development process must be fast while maintaining methodological and scienti c rigor, transparent, suitable to medico-legal context and be carefully planned to aid public acceptance.The development of Thai guideline aimed for an evidence-based and context-relevant recommendations to be executed within a short time frame(18).
First, we applied a rapid review to ensure recommendations were based on the best available evidence in a month's time frame (17,18).Although guidelines ideally apply a full systematic review to inform recommendations, rapid reviews can provide adequate advice for both clinical and policy decisions and inform on emergent issues when a full systematic review is not feasible (17,18).De ned as an accelerated or modi ed systematic review method, rapid reviews apply a more targeted scope and search criteria, while allowing for modi cations in search strategy as the evidence base is explored, though remaining systematic, transparent and explicit in both the identi cation and use of evidence (17,18).Similar to the 2014 Ebola outbreak in West Africa where no relevant data could be obtained through systematic review of published literature (17), we found limited relevant information using standard systematic search methods.Therefore, adapting our rapid review strategy and searching in the International Society for Priorities in Health website for both academic and grey literature as well as existing national and international guidelines, and explicitly specifying inclusion and exclusion criteria was most applicable.
Second, we applied a qualitative research methodology, using key informant interviews and stakeholder consultations, to ensure adaptation of the recommendations for acceptability, feasibility and stakeholder buy-in (17).Due to the absence of guidelines from low-and middle-income countries (LMICs), most LMICs resort to international guidelines or those developed by high income countries which may be less likely to be accepted or successfully implemented because of the differences in health systems, cultural and social norms (30,31) The key informant interviews allowed us to compare best practices from the rapid review with current norms and practices of physicians in caring for COVID-19 and ICU patients in Thailand.This is important because less variation from the current practice means less investment on capacity building interventions for healthcare workers especially in the urgent COVID-19 situation.Also, we con rmed the lack of existing protocols and guidelines, to prevent duplication; the key informants con rmed the need for a COVID-19 speci c guide.
The engagement of stakeholders has become an essential component of guideline development and implementation to improve acceptability and potential concerns such as inequities in recommendations on varying populations (32).Stakeholder consultations, have been utilized in the development of similar guidelines for scare resource allocation in other countries, including the United States and Canada (33,34).There are four levels of stakeholder engagement: communication, consultation, collaboration and coproduction (32).In our process, we aimed to engage with stakeholders on two levels: consultation to gather stakeholder views, feedback and experiences; and collaboration, through which stakeholders can in uence the decisions on the guideline contents.We used a consensus approach to decision making on key issues of each recommendation in the guide, a process strongly encouraged for WHO's guideline development groups (17).Multi-stakeholder engagement allowed recommendations to be adapted to be both medically sound, applicable to existing laws and practices as well as acceptable to citizens including majority and minority populations.
Given the time and resource constraints during the COVID-19 state of emergency and travel restrictions in Thailand at the time of data collection, the technical team was only able to interview two ICU physicians from public hospitals in Bangkok city.As a result, a limitation we faced was not being able to physically visit facilities, obtain information on practices more representative of the whole countries situation and of practices in private facilities.In engaging with stakeholders, we were unable to organize consultations with large number of participants due to social distancing measures.As some stakeholders did not nd teleconferencing to be convenient, we selected a mixed consultation approach, allowing options for both inperson or teleconference participation.This allowed us to engage with a more diverse group of stakeholders from throughout the country, who would otherwise not have been able to participate.To ensure all voices are equally heard, a capable moderator is necessary, to call on various stakeholder groups, regardless of whether they were attending in-person or remotely.

Conclusion
The attempt to develop critical care resource allocation guideline provides useful lessons for policymakers, public health practitioners, professional bodies, and academic scholars in Thailand and other LMICs that may nd it necessary to establish a similar allocation guideline.Reviews of international experience are important to provide solid scienti c ground for the contents of the guideline to be adapted in line with the country context.Multi-sectoral stakeholder consultation provides constructive comments and contextualized views to adapt international experience to local contexts.To the best of our knowledge, the process of developing a rapid national guideline for rationing of resources during the COVID-19 pandemic, particularly in LMIC settings, has not yet been shared or publicly available.Therefore, this paper is among the rst of its kind, and can provide an insight for the development of similar guidelines, especially in other LMICs.

Figures
Figures

Table 1
Stakeholders involved in multi-stakeholder consultation

Table 2
Data extraction of existing guidelines on critical care resource allocation The decisions on allocating critical care resources are primarily made by ICU doctors, usually in consensus and based on several qualitative and quantitative factors including medical urgency, SOFA score and comorbidities such as Charlson Comorbidity Index etc.In addition to clinical prognosis, patients' cognitive function may be evaluated through relatives; examples provided included the Modi ed Informant Questionnaire on Cognitive Decline in the Elderly (modi ed IQ CODE) and Function AssessmentStaging Test (FAST).It was maintained that cognitive impairment assessments are not usually undertaken by patients themselves given their critical conditions; though discussions with relatives and caregivers may be utilized by physicians to assess patients cognitive function.While re-assessment following initial admission may be necessary, many doctors are reluctant to step down treatment or withdraw care except upon patient's prior consent not to treat.In addition to lack of allocation protocols, advanced care plans are not routinely practiced, and as attending ICU physicians are on rotation, adequate communication between ICU doctors and patient relatives is lacking.Overall, key informants expressed the need for the development of a national guideline especially in the public health emergency and also stressed the importance of the guideline being endorsed by the Thai Medical Council and various Royal Colleges of physicians to ensure successful implementation.

Table 3
Summary of guideline evolution at each development step Step 1: Rapid Review