PDF
Citation
Print
INTRODUCTION
As of 2020, coronavirus disease 2019 (COVID-19) has become an infectious disease with the highest number of reported cases worldwide. While COVID-19 vaccines were developed relatively quickly, returning to pre-COVID-19 daily life remains difficult due to the emergence of various variants of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), despite disease control efforts across the world. In South Korea, public health could be protected through successful disease control policies in accordance with citizens; however, the need for the development of timely and evidence-based clinical practice guidelines related to the treatment of pneumonia and infection of other organs among hospitalized patients remains high in actual clinical practice. Major international agencies, such as the World Health Organization (WHO), and the governments of the United States, Australia, etc., have developed and changed their methods for developing living guidelines.1 However, evidence-based national COVID-19 clinical practice guidelines have not been developed in South Korea.
Since the onset of the COVID-19 pandemic, clinical trials have been conducted globally, resulting in almost every publication of clinical findings. Therefore, from a methodological perspective, evidence-based guidelines with continued verification of additional evidence and updated recommendations need to be developed rapidly.2
The National Evidence-Based Healthcare Collaborating Agency (NECA), together with the Korean Society of Infectious Diseases (KSID), developed the first guidelines within a short period, from October 15 to December 31, 2020, focusing on medication treatments.3 In 2021, under a Memorandum of Understanding (MOU) with the Korean Academy of Medical Sciences (KAMS) and 7 affiliated academic societies, the scope of guideline development was expanded to include medication treatments, respiratory/critical care, pediatric care, and diagnostic tests and radiological exam. In 2022, NECA and KAMS developed guidelines together with 8 academic societies and expanded the scope of the guidelines to include emergency care, new issues in diagnostic tests, and new emerging medications.
These guidelines aim to improve the clinical outcomes of COVID-19 patients and help clinicians make decisions. We systematically reviewed evidence on the latest treatments for COVID-19 and outlined evidence-based clinical practice guidelines for healthcare professionals. In addition, in terms of the guideline development methodology, we would like to apply the principle of the guidelines development methodology of being prompt and up to date as strictly as possible in emergency situations such as COVID-19.
METHODS
Methodology for Korean COVID-19 living guideline development
Regarding the guidelines development methodology, we applied the living guidelines development methodology that is consistently being updated using the latest evidence.4 Currently, a significant number of practice guidelines are being published by major countries and organizations; therefore, adoption or adaptation has been considered in some areas. However, there was a difference between the time when the evidence search was completed; therefore, we decided to use the de novo development method in actual development and referred to existing guidelines or living systematic reviews for rapid development. The guideline-development process is described as follows.
Guideline development group (GDG)
The GDG consisted of several subcommittees and groups, including the steering committee, working group, advisory committee, and conflict-of-interest management committee. Each committee or group played a key role in the guideline development process. We referenced the opinions of an independent group of patients with NECA involvement (Fig. 1).
Fig. 1
Guideline development group.
KAMS = Korean Academy of Medical Sciences, NECA = National Evidence-based Healthcare Collaborating Agency, COI = Conflict of Interest.
Steering committee
The steering committee consisted of the principal investigator recommended by KAMS, a co-investigator from NECA, and an expert member, each recommended by professional societies under KAMS. The steering committee was responsible for general operations in guideline development, including guideline planning, review of directionality and continuity of development, prioritization of clinical questions, and review of recommendations. General meetings were held regularly through video conferencing or in-person meetings to discuss the direction of progress, problems, and solutions at various developmental stages.
Working group
The working group consisted of experts recommended by each academic society. These individuals were responsible for reviewing the evidence to develop clinical questions for each topic and drafting the proposed recommendations. Working group members were assigned to each clinical question together with members of the clinical evidence research team from NECA, who were responsible for the literature searches and article selection for each clinical question, risk of bias assessment, and writing recommendations. Moreover, a literature search specialist was invited to participate in the guideline search and construction and in the implementation of search strategies to obtain the latest literature. For efficient operation, the working group was divided into 6 subgroups based on major topics: medication treatments, respiratory/critical care patients, pediatric care, diagnostic tests, radiological examinations, and emergency medicine. Each working group member was assigned 1 or 2 clinical questions and worked with the methodology staff to review the evidence and derive recommendations.
Advisory committee (consulting and review)
The role of the advisory committee was to consult on the development process for the contents included in the practice guidelines and to review the recommendations. The advisory board members consisted of experts recommended by the steering committee among the members recommended by the KAMS, KSID, and Korean Society for Laboratory Medicine. Committee members who reviewed the draft of recommendations for external review without directly participating in the development process were selected through the final recommendation by the steering committee from among members recommended by the associated medical societies, such as KAMS, Korean Society of Critical Care Medicine, and the Korean Society for Antimicrobial Therapy.
Conflict of Interest (COI) committee
The COI management committee consisted of 7 members, including the chairs of the GDG, members recommended by the chairs, and outside experts who confirmed the COI declaration form. In the early stage of the practice guidelines development, the COI declaration form was distributed to all committee members (including all members of the GDG) and a survey of the COI was conducted. Throughout the guideline development period, any changes could be voluntarily submitted to the COI management committee. For any reported COI case, at least 2 committee members independently reviewed each case and made determinations. The final decision was made during a general meeting with the COI committee. After the members were reviewed for possible COIs, the COI committee’s decision was held in person or via e-mail. If the COI committee decided that the reported COI impacted at moderate to high levels, members’ activities could be limited to reviewing evidence and not voting for decision-making. Over the 2-year period, 1 member’s reported academic COI was determined as ‘low’ and the other member who reported financial COI was determined as ‘moderate’ by the COI committee (Table 1).
Table 1
Conflict of interest level and decision in activity
Public involvement in NECA (PIN)
PIN is an independent group of representatives of 11 people from consumer and patient groups and 85 citizens. The PIN was established in 2018 to facilitate the involvement of patients in healthcare service research, health technology assessment, and guideline development to enhance the opportunity to reflect patient values and preferences in the evidence-based decision-making process.
Goals and targets of guidelines
Target population of guidelines
The guidelines are intended for patients with confirmed COVID-19, including adults and children. To classify COVID-19 severity among adult patients in 2020, the classification systems from each country were compared, resulting in the adoption of the COVID-19 severity classification system from the US National Institutes of Health (NIH) (Supplementary Table 1). In addition, in 2021, the National Institute of Allergy and Infectious Disease Ordinal Scale published clinical treatments performed on patients at each stage, classified using an 8-point scale ranging from asymptomatic (1 point) to death (8 points).5 In this guideline, we matched these criteria and used them for the classification and analysis (Supplementary Table 2). As pediatric patients show different disease progression, we established criteria for pediatric severity based on the NIH classification of pediatrics (Supplementary Table 3).
Guidelines users and setting
These guidelines are intended for healthcare professionals responsible for treating and caring for patients hospitalized with COVID-19 in South Korea. There were no limitations on the setting (size or type of medical institution); however, if there were any restrictions or resources in the setting, then clinical considerations were presented along with the recommendations.
Guidelines development process
Scope and development of clinical questions
The pilot COVID-19 treatment guidelines for 2020 included medication treatments that could help improve the prognosis of patients with COVID-19.3 In 2021, the scope of the guidelines was expanded to include respiratory/critical care, diagnostic tests, radiological examinations, and pediatric infectious diseases after the MOU with KAMS. The clinical questions were developed through brainstorming by the development group and a review of the therapies discussed in the existing practice guidelines, while the priority order of each therapy was determined based on the consensus reached by the development group. The domains selected by the informal consensus were medication treatments, diagnostic tests, radiological examinations, and emergency care. The domains in which multiple clinical questions were selected and prioritized by the working group by voting using a 5-point scale, by the working group were respiratory/critical care and pediatric care.
Among the Patient–Intervention–Comparison–Outcome (PICO) elements, “outcomes” were selected according to the Grading of Recommendations Assessment Development and Evaluation (GRADE) approach.6 Accordingly, outcomes were listed by clinical questioning through discussions by the working group and “critical” and “important” outcomes were classified by consensus. These are provided in the table summarizing the findings.
From late 2021 to early 2022, a preliminary survey was conducted with the steering committee and working groups to investigate the prioritization of clinical questions and the necessity of a living update to expand the scope of development and confirmation of clinical questions. Accordingly, the update cycle was set differently, depending on the priority. For recommendations subject to living updates through a priority review, searches were performed at 1-month intervals for updates every 3–4 months. The steering committee decided whether to update recommendations that did not require a living update based on a review by the working committee. If an update was necessary, it was revised at least once per year; otherwise, it was put on hold (Fig. 2).
Resources for literature search
For pilot practice guideline development in 2020, the databases (DBs) used for the literature search were limited to PubMed and KMbase for rapid development. However, a more comprehensive literature search has been conducted since 2021. For this round of comprehensive literature search, the search sources included international DBs such as PubMed, EMBASE, and Cochrane CDSR, and Korean DBs such as KMbase and KISS. In addition, the search results were supplemented by a manual search and reference checking. The use of preprint DBs such as MedRxiv and bioRxiv was limited to fields with very rapidly changing evidence (i.e., vaccines, delta variant).
Search strategy
The search strategy was devised by the informacists. The strategy reflected the term changes after the initial search including the confirmation of “COVID-19” as an official MeSH term and addition of drug names. Working group members proposed search terms for each clinical question, and a rough draft of the search strategy was prepared using PubMed. Working group members reviewed and sent feedback to the informacists. Thus, the final search strategy was confirmed. The search terms included those related to COVID-19, such as “coronavirus,” “novel coronavirus,” “novel coronavirus 2019,” “2019 nCoV,” “COVID-19,” “Wuhan coronavirus,” “Wuhan pneumonia,” and “SARS-CoV-2,” along with search terms for each intervention or therapy. The search strategy was established by selecting text words, considering controlled terms and synonyms, and a search was performed according to the characteristics of each DB.
Continuous evidence updates
The articles on COVID-19 therapies have been continuously updated. Accordingly, searches were updated at intervals of one month and key evidence was identified and incorporated into the revised recommendations. For medication treatments, searches were updated on the 10th of each month, starting in August 2021. For all other clinical questions, searches were updated every month from September 2021 on the 10th of each month. The search source was changed to Ovid-MEDLINE for search updates, which included preprint DB search results. To manage continuous evidence updates, we used Covidence, a commercial semi-automated tool for systematic reviews.7 In 2022, the update interval was confirmed by the GDG for each clinical question, considering the clinical situation and reflection of the article publication. For the new clinical questions that arose in 2022, the initial search for most questions was completed in April 2022, although the confirmation period for each topic varied.
Study selection criteria and process
For each clinical question, the inclusion and exclusion criteria were established based on the PICO elements of the clinical questions and study design. Considering that the search was performed monthly, the study selection flow diagram was revised. Title and abstract screening were performed by 1 reviewer because the agreement rate between 2 reviewers was over 80% for sample testing. A full-text review was independently conducted by 2 reviewers and a consensus was reached. Conflicts were resolved by a third reviewer or by discussion with the working group members.
Risk of bias assessment for selected studies
For assessment of the risk of bias which finally selected studies for each clinical question, tools appropriate for each study design were selected and each study was assessed independently by 2 reviewers. In the event of a disagreement, a third reviewer was consulted to reach a consensus. When possible, a NECA methodologist and a clinical working group member were paired as reviewers.
• Quality assessment tool for randomized controlled trials: Cochrane risk of bias (RoB)
The Cochrane RoB tool consists of 5 domains and 7 items, which are rated as “low,” “high,” and “uncertain” with “low” indicating low RoB, etc. These items were designed to assess random sequence generation, allocation concealment, blinding, handling of missing data, selective reporting, and other sources of bias.8
• Quality assessment tool for nonrandomized studies (NRSs): RoBANS 2.0
The tool for assessing RoB in NRSs is the RoB for NRSs (RoBANS).9 It was developed through the “Study on the Development of Quality Assessment Tool for Clinical Research Articles” in 2009 and was revised in 2013 to account for the latest research trends that were mentioned at the Cochrane handbook. This tool contains 8 items: the possibility of target group comparisons, target group selection, confounders, exposure measurement, blinding of assessors, outcome assessment, incomplete outcome data, and selective outcome reporting. Each item was rated as having a low, high, or an unclear risk of bias.
• Quality assessment tool for diagnostic testing: Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS 2.0)
The tool typically used to assess the quality of primary diagnostic accuracy studies is the QUADAS-2, which consists of 4 items that assess concerns regarding the bias and applicability of each study.10 Bias assessment consisted of 4 domains: patient selection, index test, reference test, and patient flow and timing. The applicability assessment consisted of 3 domains: patient selection, index testing, and reference testing. The RoB in each domain of bias and applicability is assessed as “low,” “high,” or “uncertain.”
Data extraction and evidence synthesis
The final studies selected for each clinical question were classified by study design, and data were extracted in a pre-defined format. A table showing the basic characteristics of the studies was created by a working group member and revised after a review by the research team. For the outcomes, data were extracted according to a pre-defined extraction method as needed for synthesis. Data extraction methods that could assess comparability were considered in studies comparing the 2 interventional methods. Data were extracted by the research team responsible for the NECA methodology and reviewed by working group members with specialty-related clinical questions. Based on the reviewed data extraction results, meta-analysis was performed if quantitative synthesis was possible, and qualitative description was used if meta-analysis was not possible.
If a meta-analysis was possible, the data heterogeneity was assessed. For data with high heterogeneity, a random-effects model was applied and subgroup or meta-regression analyses were performed to identify the cause of heterogeneity. Publication bias was assessed using Egger’s test and the trim-and-fill method if the number of studies included in the synthesis was 10. Review Manager (RevMan) Version 5.4 (The Nordic Cochrane Center, Copenhagen, Denmark), R 4.1.111 was used for single-arm analyses that were difficult to analyze using RevMan. STATA 17.0 (StataCorp.2021; StataCorp LLC, College Station, TX, USA) was used for diagnostic test meta-analysis.
Certainty of evidence and grade of recommendations
The certainty of evidence was assessed using the GRADE approach. In the GRADE approach, the importance of outcomes is assessed first, followed by determination of the certainty as “high,” “moderate,” “low,” or “very low.” The direction and strength of the recommendations were determined based on the 4 factors considered in the GRADE approach: certainty of evidence, magnitude of effects (weighing the benefits and harms), patient values and preferences, and resources used. The GRADE recommendations include “only in research” and “no recommendation,” but in these guidelines, the expression “inconclusive” was used, which also comprises the meaning of insufficient evidence. When recommending interventions based on clinical experience because of a lack of evidence, expert consensus was used rather than evidence-based recommendations. The certainty of the evidence and grading definitions are summarized in Supplementary Tables 4 and 5.
Consensus process
The draft of the recommendations was based on a review of the evidence by working group members in an attempt to reach an informal consensus during a general meeting with the majority of the development committee members in attendance. If a consensus could not be reached through discussion, the opinions were divided into 2 options and consensus was reached by voting using a 5-point scale with the majority in favor, with votes of 4 points (agree) or 5 points (strongly agree). There were no serious disagreements during the recommended development stage in 2021.
In 2022, the Delphi technique was applied because of the difficulty in reaching an informal consensus on some clinical questions (e.g., pediatric care) due to insufficient evidence. The level of agreement with the recommendations for the key questions in each field was rated on a 9-point Likert scale. The mean agreement scores for the recommendations were used as summary statistics for the survey results, and the coefficient of variation (CV) was calculated to check the level of agreement for each recommendation. Smaller CVs would indicate greater consistency in the level of agreement, and accordingly, CVs of < 0.5, 0.5–0.8, and > 0.8 were predefined as secondary survey not needed; stable, but secondary survey to be determined after open discussion among the research team; and additional survey needed, respectively.
Peer review and external review
The peer reviews were conducted in 2 stages. The recommendations drafted by the working-group members were peer-reviewed by other members in the same topic group. After the necessary revisions were made, the steering committee and secretaries of each working group topic reviewed and suggested revisions. Considering the expertise of the committee members, 2–5 recommendations were assigned to each expert for review. The level of agreement for each recommendation was rated on a 9-point scale (1 = strongly disagree, 9 = strongly agree), and the mean, median, and CV values were calculated. Subjective opinions were accepted and used as references for the final revision of the recommendations. The advisory committee conducted an external review of the final draft (recommendations and evidence profile). The final recommendations that were reviewed and revised were posted on the NECA website, and a public consultation was conducted for two weeks.
COI management
Until 2021, the existing COI declaration form was prescreened and subsequently reviewed and revised by the steering committee. The COIs were divided into financial and nonfinancial COIs within the past three years. For non-financial COIs, intellectual property rights and employment (direct/indirect) were classified, and leadership-related items were added. For financial COIs, the nature of research funding and honorariums were clearly differentiated, and the upper limit of cost was determined based on consensus through discussions by the committee members. The final revised form was completed by all the participants and confirmed during the early stages of guideline development. The COIs that may affect recommendations and changes to the entire development process were voluntarily reported and reviewed by the steering committee.
Starting from 2022, an independent COI committee was established separately, and the levels of COIs were defined as “no,” “low,” “moderate,” and “high” COIs, which were used to restrict participation in various stages of guideline development (Table 1). These measures were incorporated into the operating terms and conditions. For any reported case of COIs, at least 2 committee members independently reviewed each case and made a decision. The final decision was made during regular meetings with the COI committee.
Financial and editorial independence
This practice guideline development project was funded by a research grant from NECA (Project number: NECA-P-21-004, NACA-A-22-008/009, NECA-A23-009/010). Research funding had no influence on the independence of the content of practice guidelines.
Living update of recommendations
These guidelines were designed to reflect recent COVID-19-related research trends. The search was updated monthly to incorporate the results. As an exception, considering that evidence selection, evidence synthesis, and certainty of evidence assessment would take at least 3 months, the difficulty of revising the recommendation every month was discussed. For living updates, evidence searching was performed at 1-month intervals and recommendations were updated every 3–4 months. Regarding recommendations that did not require a living update, the GDG determined whether to continue or withhold updates through discussion; for those that were determined to be updated continuously, updates were made at least once a year.
Dissemination and implementation
The final recommendations were published through the “COVID-19 living guideline” page on the NECA website.12 In 2022, the webpage was revised to enable quick recognition of updated recommendations by highlighting icons. Relevant academic medical societies have endorsed these recommendations. Relevant agencies, such as the Ministry of Health and Welfare and the Korea Disease Control and Prevention Agency, were notified, and the guidelines were disseminated through press releases. After publishing practice guidelines, the level of dissemination could be checked through press release monitoring. We have also published several systematic reviews and meta-analyses in international journals, such as the rapid antigen test (RAT), interferon (type I and type III), interleukin-6 receptor antagonists (tocilizumab and sarilumab), early intubation, and prone position for patients with severe COVID-19.1314151617 Moreover, for information sharing with the international community, the guidelines were published at the Guideline International Network (2021 and 2022) conferences. Although it may be practically difficult to obtain data by monitoring changes in the use of interventions, such as medications, after the publication of practice guidelines in clinical practice, changes in the use of interventions that are confirmed through discussions by the steering committee will be monitored using publicly available national open health data.
Patient values and preferences
For stakeholder participation in clinical practice guideline development, it is ideal that patient values and preferences be surveyed. In this practice guideline, a survey investigating clinical questions through an independent PIN was considered in the early stage of development. However, at the initial time, the steering committee and working groups determined that topics that had been prioritized for immediate development were necessary for clinicians, but the topics may require highly specialized knowledge, which would not be suitable for surveys conducted on the general population. An online survey on PIN for consumer preference reflection was conducted between October 22, 2022, and November 4, 2022, with 89 public participants to investigate the accessibility and usability of the practice guidelines. In addition, usability was investigated by approximately 80 members of the Clinical Practice Guideline Expert Committee of KAMS.
RESULTS
Recommendation summary
During the first year, 26 key clinical questions and recommendations were developed using a living evidence update process. In 2022, we continued revisions by updating living evidence and developed 11 new key clinical questions regarding new emerging treatments or unmet needs in clinical situations. A summary of these recommendations is provided in Table 2. And Supplementary Tables 6, 7, 8, 9, 10, 11 provides a Korean translation version of the recommendations.
Table 2
Recommendations for medications, respiratory critical care, pediatric care, emergency care, and diagnostic test/radiologic exams
A: strong recommendation; B: conditional recommendation; C: conditional against; D: strong against; I: inconclusive.
CoE = certainty of evidence, GoR = grade of recommendation, COVID-19 = coronavirus disease 2019, IL = interleukin, IVIG = intravenous immunoglobulin, SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2, ECMO = extracorporeal membrane oxygenation, vv-ECMO = veno-venous extracorporeal membrane oxygenation, P/F ratio = ratio of arterial partial pressure of oxygen to fraction of inspired oxygen, PEEP = positive end expiratory pressure, HFNC = high-flow nasal cannula, ROX = respiratory rate-oxygenation, MIS-C = multisystem inflammatory syndrome in children, CPR = cardiopulmonary resuscitation, RAT = rapid antigen test, PCR = polymerase chain reaction, CT = computed tomography, NPIC = negative pressure isolation chamber, TNF = tumor necrosis factor.
aNirmatrelvir/ritonavir, remdesivir, or other monoclonal antibody effective against currently circulating variants.
bThe decision to place a patient in a prone position before ECMO should be based on a consideration of the patient's benefits and harms of the procedure, as well as the intensive care unit resources.
In 2021, medication treatment recommendations included 14 treatments, including antiviral agents (remdesivir, favipiravir, ribavirin, umifenovir, and baloxavir marboxil), steroids (systemic and inhaled), interlukin-6, interukin-1, kinase inhibitors, non-SARS-CoV-2 specific immunoglobulin, SARS-CoV-2 specific immunoglobulin, convalescent plasma, interferon, ivermectin, anti-SARS-CoV-2 monoclonal antibodies, and protease inhibitors (camostat, nafamostat). In 2022, nirmatrelvir/ritonavir (paxlovid) and molnupiravir were added to reflect new emerging treatments, and remdesivir, interlukin-6 (tocilizumab, sarilumab), kinase inhibitors, and anti-SARS-CoV-2 monoclonal antibodies were continuously updated quarterly. Other treatments such as steroids (systemic and inhalant), interukin-1, specific SARS-CoV-2 immunoglobulin, non-SARS-CoV-2 specific immunoglobulin, and convalescent plasma were also updated. This information was updated in 2022. Some topics have been determined to be pending update, such as protease inhibitors (camostat and nafamostat), interferon, ivermectin, non-SARS-CoV-2 Specific immunoglobulin, and other antivirals because of the lack of updated needs, as confirmed by the GDG.
For pediatrics reviewed in 2021, the target population was multiple inflammatory syndromes in children (MIS-C), and steroids, immunoglobulins, and antivirals were the interventions of interest. In 2022, we expanded the population to include general pediatrics, and medication treatments included remdesivir, steroids, tocilizumab, and nirmatrelvir/ritonavir (paxlovid).
In 2021, newly developed respiratory and critical care treatment topics included therapeutic doses of anticoagulants, early intubation, extracorporeal membrane oxygenation, positive end-expiratory pressure, and the prone position. We continued updates on all topics in 2022 and added a high-flow nasal cannula as a new topic. Mechanical cardiopulmonary resuscitation for cardiac arrest in patients with suspected or confirmed COVID-19 has been newly developed for emergency care needs. The diagnostic accuracy of the RAT by symptom (symptomatic or asymptomatic) and usefulness of contrast-enhanced chest computed tomography (CT) scan and chest X-ray follow-up were selected for the 2021 recommendations. In 2022, chest CT using a portable personal negative-pressure isolation chamber will be performed. Regarding updates, the contrast-enhanced chest CT scan continued; however, the chest X-ray follow-up was updated in 2022 owing to the low level of revision needs.
During these two years, we uploaded over 70 posts on the web board of the COVID-19 living guidelines. In addition, there were a number of official tweets, blogs, and official dissemination of recommendations through media (newsletters, TV, etc.) by both institutional communication channels.
At the end of 2022, we conducted a survey to investigate the target users and the general public’s cognition and attitudes toward usefulness. In the results from the PIN, 46/97 people responded and 69% selected good or very good to the question about the guidelines’ usefulness. To the same question, 31 clinicians of KAMS answered as 77.42% good or very good. Clinicians agreed on the scope of the guidelines (74.19%) and rigor of the methodological aspects (87.1%). However, PIN respondents described some difficulty in understanding the terms and research results of the open-ended questions.
DISCUSSION
Previously, there was no attempts to implement living guidelines in Korea, and in a survey of the Korean guidelines, the guideline update period was usually 3 to 5 years.18 Before the COVID-19 pandemic and after the introduction of the living guideline methodology,19 the Australian Stroke Living Guideline Group presented the feasibility of living guideline.20 They showed that the feasibility of the national-level rigorous rapid/living guideline model and living evidence synthesis methods could be applied to public crises such as COVID-19. During the COVID-19 pandemic, living systematic reviews and recommendations have been a major reflection of the current evidence status. The Australian COVID-19 Living Guidelines of the National COVID-19 Task Force published weekly recommendations during the early COVID-19 pandemic.21 The United Kingdom has also published COVID-19 rapid/living guidelines that have been updated monthly during the pandemic.22 The intervals of revision varied but were usually shorter than the standard update period of general practice guidelines.1 These living guidelines were provided by online and shared platforms, such as MAGICapp23 and the COVID-19 RecMap platform.24 The WHO published living recommendations for drug therapy during the early pandemic.2526 Japan has also published two English versions of rapid/living recommendations, including medication treatments.2728 The WHO and Japanese recommendations included infographic figures (Supplementary Figure 1) that are useful for understanding recommendations and accelerating feasibility. These activities increased the accessibility of guidelines during the pandemic and reduced the efforts for evidence synthesis.29
However, deciding on an appropriate update period for recommendations remains a challenge. A shorter update period requires more incorporated resources and labor; otherwise, the rigor of the methodology may not be assured. There are some barriers to the short interval of updating, such as lack of human resources, insufficient public funding, and lack of methodologists. Prioritization and the decision to update the interval were difficult issues in continuing the living process.30 Regarding medication treatments, clinical trials have been published rapidly, which warrant monthly monitoring, whereas in other fields, such as respiratory/critical care, new articles, such as medication treatments or vaccine studies, were not frequently published. Furthermore, some cases no longer required continuous evidence review, such as hydroxychloroquine, where the controversy had ceased, whereas other cases required the addition of clinical questions with newly emerging oral therapeutics for treating COVID-19. Therefore, the GDG continuously monitored changes in evidence and priority discussions.
For dissemination, we updated our webpages to help users rapidly find recommendations and recognize that the update status and content have changed. The time required to reach the desired number of views was reduced by a factor of 4. However, only half the respondents answered that they were aware of these guidelines. Reasons for this include the lack of channels for the public, limited use of social media, and the difficulty of contents related to health literacy. Plain language for the public may be the key solution to this issue; however, we did not have sufficient resources for dissemination, planning, and implementation.
The GDG clinicians’ knowledge and understanding of the systematic review methodology, especially the GRADE approach, was low in the initial stage of guideline development. We planned a series of educational programs for the GDG members. Most sessions were webinars, and links were recorded during the social distancing period. Even though there were several duplicate education sessions, the clinicians still expressed that it was difficult to understand the GRADE method.
In this guideline, an implementation plan for each recommendation was devised and considerations were applied to the strength and direction of the recommendations. Overall, there is a need to identify the actual barriers and facilitators. The biggest barriers to this guideline were the explosive increase in the number of patients with COVID-19 and the shortage of resources in clinical practice. As resource-related issues were mostly impacted by policymakers, it was difficult to propose actual alternatives. For example, considerations were given to aspects related to human resources in the recommendations for intensive care patients, and efforts were made to reflect domestic and international statuses. The differences in resource considerations for the setting (hospital size, equipment, etc.) still limit implementation.
The lessons learned from the COVID-19 pandemic recommend government support for future crises. In terms of national-level clinical practice guidelines, all successful rapid/living guidelines were supported by the government and public, including sufficient budgets, networking and training of clinical experts and methodologists, participation of various stakeholders, and international collaborations. These rapid/living guidelines can not only provide decision support for clinicians but also increase the quality of patient outcomes and manage care resources.
In conclusion, this project was a national-level COVID-19 living guideline involving more than 40 COVID-19 clinical experts and methodologists over 2 years. We intend to improve the clinical outcomes of patients and help overcome the national crisis caused by COVID-19 by continuously updating the evidence. We provided timely recommendations on living schemes and disseminated them to the public, policymakers, and various stakeholders using webpages and social media. To respond to the next pandemic or a similar crisis in public health, clinicians and methodologists should collaborate to provide evidence to improve trustworthiness and reduce uncertainty in clinical practice. We recommend that public policymakers procure funding and cooperate closely to implement these recommendations.
XML Download