Rethinking measurement of health outcomes in Long COVID: complexities, challenges and considerations

Abstract

The reality of Long COVID emerged soon after the beginning of the COVID-19 pandemic. More than five years later, thousands of articles have been published with multiple case definitions, heterogenous populations, and numerous measurement instruments, yielding a massive amount of evidence. Health outcome measurement is vital for identifying health challenges, changes in health status and predicting future health states for people with Long COVID. Nevertheless, distinct issues of measurement require attention in the context of Long COVID. In this commentary, we discuss complexities, challenges and considerations associated with health outcome measurement in research and clinical practice with people with Long COVID. Specifically, we address: (i) identifying the population in the context of variable terminology, definitions and symptoms affecting people with Long COVID; (ii) identifying the complexity of health constructs, often multidimensional, to measure with numerous health-related consequences associated with Long COVID; (iii) identifying the purpose of measurement while taking into account the dynamic nature of Long COVID and (iv) identifying appropriate outcome measures used with people with Long COVID and their limitations. We highlight important considerations for measurement in research and clinical practice, including the impacts of the various symptoms and the dynamic nature of Long COVID. We provide examples of outcome measures used to date in the context of Long COVID to illustrate the complexities throughout, with a glimpse at wider consequences. We conclude with a brief discussion of considerations to help pave the way forward for the improvement in health outcomes measurement in Long COVID research and clinical practice. Advancing knowledge on Long COVID requires a return to the fundamentals of measurement science. It is critical to appropriately assess the measurement properties of existing instruments for their ability to accurately and reliably measure health-related constructs associated with this condition. Identifying limitations of currently used tools is also essential to prevent perpetuation of issues in the development of condition-specific measurement instruments for Long COVID. This, in turn, will help pave the way for more robust measurement and improved data interpretation in the context of Long COVID.

Debut of the maze

Within months of the onset of the COVID-19 pandemic, it became obvious to many that a SARS-CoV-2 infection was not to be a mere two-week hiatus from (their) life. “Long COVID” emerged as a name on social media in May 2020 [1]. In September 2020, Perego et al. described Long COVID as “a patient-made term that, in the absence of formally agreed definitions, [is used] to describe diverse symptoms persisting beyond four weeks after symptom onset suggestive of COVID-19” [2]. Public health organizations and the scientific community around the world progressively realized what persons living with post-infectious conditions dreaded from the early days: COVID-19 can trigger a complex, multisystemic, and potentially disabling condition, lasting for an indefinite period in many SARS-CoV-2 infected individuals [3]. This condition continues to affect millions of people worldwide [4].

Attempts to define Long COVID have led to the emergence of a plethora of names, labels and descriptions [5–9]. Most recently, the National Academies of Sciences, Engineering, and Medicine (NASEM) released its new Long COVID definition: “Long COVID is an infection-associated chronic condition (IACC) that occurs after SARS-CoV-2 infection and is present for at least 3 months as a continuous, relapsing and remitting, or progressive disease state that affects one or more organ systems (…)” [10].

While discussing details of various terms and definitions used to describe the aftermath of COVID-19 is beyond the scope of this commentary, in accordance with the community who coined it and NASEM, we use the term Long COVID. There is ongoing work examining terminology and definitions in more depth [11].

“If you cannot measure it, then it is not science.” – attributed to Lord Kelvin [12]

Any endeavor to understand and manage Long COVID must begin by appropriately defining and measuring it. However, accurately capturing Long COVID’s many facets with appropriate measurement instruments has proven to be highly challenging. Adding to the challenge, persons with lived experience, researchers and clinicians are still “building the plane while flying it”.

Hence, in line with the Measurement and Assessment Paradigm for Long COVID (MAP LC) project [13] within Long COVID Web [14], this commentary aims to spark a discussion, providing an overview of complexities, challenges and considerations in the measurement of health outcomes of people with Long COVID in research and clinical practice, and how they may have impacted current evidence on Long COVID. Specifically, we will discuss complexities, challenges and considerations of measurement in the context of Long COVID, which include: (i) identifying the population in the context of variable terminology, definitions and symptoms affecting people with Long COVID; (ii) identifying the complexity of health constructs, often multidimensional, to measure with numerous health-related consequences associated with Long COVID; (iii) identifying the purpose of measurement while taking into account the dynamic nature of Long COVID and (iv) identifying appropriate outcome measures used with people with Long COVID and their limitations. We conclude with a brief discussion of considerations to help pave the way forward for the improvement of measurement of health outcomes in the context of Long COVID research and clinical practice.

Exploring measurement-related dimensions will hint at their contribution to wider consequences, such as the erroneous recourse to mental health conditions as the sole explanation of symptoms and the failure to consider unfamiliar conditions, a situation that has endured for decades for many other complex chronic conditions. However, these consequences deserve to be covered as a specific topic that extends beyond the scope of this commentary focusing on measurement.

For the purpose of this commentary, we use health outcomes to broadly encompass “what” is measured. Similarly, we use outcome measures to encompass the various types of measurement instruments designed to discriminate or describe, predict and evaluate change: the “how” it is measured. We also use condition in a broad and inclusive meaning: a diagnosis, a group of signs and/or symptoms, a problem or a clinical state. This definition is broader than the term diagnosable condition used by NASEM, although it is not explicitly defined in the report [10].

Measuring Long COVID

Measuring a health condition appropriately requires careful identification of the population of interest (i.e., clinical characteristics, health conditions, demographics), the constructs to assess (i.e., the theories, concepts or outcomes) and the purpose of measurement (i.e., to discriminate or describe, predict, or evaluate change over time) [15, 16]. Moreover, for Long COVID, measurement requires a sound understanding of the complexities of the condition, and how its diverse symptoms affect the use and interpretation of outcome measures.

#1 - Identifying the population

Identifying the population in Long COVID research needs to be contextualized to appreciate the underlying complexity of outcome measurement and interpretation. Long COVID is a vast umbrella with numerous symptoms and highly variable presentations topped with varying terminology and definitions. Estimating Long COVID’s prevalence is challenging. For example, a meta-analysis on Long COVID in adults reported prevalence rates ranging from 6.2% to 57% [17], while another review in children reported rates ranging from 1.6% to 70% [18]. Understanding the factors leading to this wide range of prevalence estimates is crucial to appropriately interpret Long COVID literature [19].

Variability in terminology and definitions

Definitions and timelines of Long COVID show great variability, which has a direct impact on identifying the target population. A 2024 unpublished preliminary review by the first author of this commentary has identified more than 30 names for Long COVID in English alone, with or without accompanying definitions. In 2023, a descriptive study from Chaichana et al. examined 295 studies on “lasting symptoms of COVID-19”, of which 102 (34.6%) used either the Centers for Disease Control and Prevention (CDC – United States), the National Institute for Health Care Excellence (NICE – United Kingdom) or the World Health Organization (WHO) definitions, while 129 (66.8%) used their own definition and 64 (33.2%) did not define Long COVID at all [20]. In addition, many Long COVID studies use only partial definitions or confuse Long COVID with the period following COVID-19, which further limits interpretation and comparability within the literature.

Constellation of symptoms

Clinical presentations of Long COVID also vary dramatically, with descriptions of over 200 symptoms [21]. Evidence on symptom clusters, or phenotypes, potentially related to different pathophysiological mechanisms, is available [22]. However, description and operationalization of measurement of these distinct phenotypes still need to be achieved [19].

As a step in the right direction, the NASEM report provided detailed information on symptoms and diagnosable conditions under the umbrella term of Long COVID [23]. Until the mechanisms underlying Long COVID are fully identified, however, it is difficult to delineate direct COVID-19 sequelae (e.g. organ damage), sequelae of acute complications (e.g. post-intensive care syndrome), exacerbation of existing conditions (e.g. asthma), triggering of new conditions (e.g. autoimmune disease) and the range of other diagnosable conditions and symptoms frequently associated within IACCs (e.g. Myalgic Encephalomyelitis/Chronic Fatigue Syndrome [ME/CFS], dysautonomia) (Fig. 1).

Fig. 1.

The umbrella of Long COVID. Long COVID currently encompasses the whole spectrum of long-term conditions and symptoms related to a SARS-CoV-2 infection. Using gears, this figure displays the complex interplay of the broad categories of the spectrum, with Infection-Associated Chronic Condition (IACC) in the centre. Each interacting gear can combine differently in each person, further complicating the clinical presentation. Among other (diagnosable) conditions, Myalgic Encephalomyelitis/Chronic Fatigue Syndrome (ME/CFS) and dysautonomia frequently increase the burden of symptoms and functional impact

Failure to consider this variability in terminology and definitions, phenotypes, and other factors when identifying and characterizing the studied population (and further downstream when conducting data analysis and interpretation) can lead to overgeneralization of findings. This can contribute to further confusion and misunderstanding around the aetiology and characteristics of Long COVID, and limit the ability to demonstrate the effects of potential treatments.

#2 - Identifying the constructs

With such a wide array of symptoms, it is difficult to determine what constructs should be measured in Long COVID research. Furthermore, the identification of constructs to measure should be informed by appropriate conceptual definitions for researchers to select appropriate outcome measures. In important foundational work, the Post-COVID Condition Core Outcome-Set (PC-COS) project identified important outcome domains and strived to identify appropriate outcome measures for adults and children through a Delphi study [24–26]. Unfortunately, consensus on outcome measures could not be reached for most domains. We appreciate the challenges associated with reaching a consensus on a core outcome set given the vast heterogeneity and complexity of Long COVID. These results further emphasize the need for more research that focuses on resolving these critical measurement challenges.

Essential or nice to have?

Including every potential health outcome in data collection to ensure comprehensiveness may at first glance be seen as a solution, but it is impossible to do for such a complex condition. Even without considering costs and research waste, length is a common barrier to the use of outcome measures [27, 28]. Furthermore, it could lead to harm. The energy required to complete questionnaires and performance tasks may trigger undesirable worsening of symptoms, such as fatigue and post-exertional malaise/post-exertional symptom exacerbation (PEM/PESE), a frequent condition in Long COVID [3, 21, 29–31]. Therefore, careful consideration must be given to determine what is essential to measure.

Unreported does not equal absent

As the number and types of symptoms assessed in Long COVID still vary greatly between studies, it is crucial to remember that an unreported symptom does not mean an absence of that symptom if it has not been assessed. PEM/PESE, a frequent condition in Long COVID that may cause severe functional impairment with implications for clinical management [3], is infrequently measured in the literature [32–34]. In a meta-analysis on symptom clusters which included 76 studies of people with Long COVID, Kuodi et al. found only seven that included PEM/PESE in the reported outcomes [35]. PEM/PESE is also infrequently assessed in rehabilitation studies [36]. Considering the impact of PEM/PESE on function and management, missing this key characteristic to classify participants may hinder data analysis and interpretation, and lead to inappropriate conclusions. Ignoring the absence of PEM/PESE without addressing the omission is still observed in 2025 [37].

Multidimensional constructs

Examining only the right foot when the wound is on the left is a problem. In complex chronic conditions such as Long COVID, many constructs may span across multiple health domains. A good example is shortness of breath (SOB). Most pre-existing patient-reported outcome measures (PROMs) addressing this construct, for example the Medical Research Council (MRC) breathlessness/dyspnea scale [38, 39], were designed and are interpreted from a cardiorespiratory perspective, familiar to researchers and clinicians. SOB experienced by people with Long COVID may be caused by cardiorespiratory limitations, dysautonomia, PEM/PESE, or a combination of mechanisms. However, many researchers and clinicians are less familiar with dysautonomia and PEM/PESE and may be unaware of how to differentiate between aetiologies. Some manifestations of SOB, such as those related to standing or after meals, air hunger sensation and exacerbation during PEM/PESE, are seldom assessed nor considered in the interpretation of pre-existing outcome measures leading to incomplete appreciation of the underlying aetiology of SOB. Furthermore, despite advancements in knowledge, some questionnaires are often used and interpreted from outdated or narrow perspectives. For example, the Nijmegen questionnaire [40] was intended to screen for hyperventilation complaints regardless of cause. Despite some clarification from the original authors [41], the questionnaire is often still interpreted solely from the psychological perspective of anxiety, without exploring other potential mechanisms that may fully or partially explain the symptoms [42].

Hence there is a need to consider perspectives beyond single medical specialties when planning measurement in Long COVID. This entails reconsidering the conceptualization of constructs, especially when multidimensional, as well as reconsidering the content validity of outcome measures and their interpretation.

#3 - Identifying the purpose of measurement

The purpose of measurement, i.e. to discriminate (or describe), predict or evaluate change over time [15, 16], needs to be clearly identified to ensure appropriate selection of an outcome measure and interpretation of the results.

Measurement purpose drift in practice

Habits can lead to overlooking the validated purpose of an outcome measure. A striking example is the Patient Health Questionnaire − 9 items – Depression module (PHQ-9), developed as a screening instrument for depressive symptomatology and validated in a general primary care population [43, 44]. As per the Diagnostic and Statistical Manual of Mental Disorders - fourth edition (DSM-IV) criteria for major depressive disorder, use of the PHQ-9 for diagnosis entails the presence of a depressed mood or anhedonia and should be combined with a clinical evaluation to assess functional impairment and exclusion of other causes for symptoms. The purpose of the PHQ-9 score was to correlate with the severity of the functional impairment of the condition and with the probability of a diagnosis of depression; it was not meant to be used as a diagnostic tool [45]. Also, even if a score of 10 or greater has been shown in a meta-analysis to maximize sensitivity and specificity for screening to detect major depression [46], the same team later found in another meta-analysis of 44 studies that, on average, the prevalence of a PHQ-9 score ≥ 10 was more than double the prevalence of depression found with a semi-structured diagnostic interview [47]. They also showed caveats in cutoff scores and accuracy estimates determination [48].

Nonetheless, a screener such as the PHQ-9 score alone is widely, and incorrectly, used to report depression prevalence among people with Long COVID [49]. In a recent meta-analysis on the prevalence of depression, anxiety and sleep disorder in Long COVID [49], the pooled prevalence of depression from 143 studies was estimated at 23% (95% CI: 20%—26%; I² = 99.9%). Not only it seems hazardous to draw conclusions with such a high level of heterogeneity, but the PHQ-9 or PHQ-8 scores were used in 34 studies with ten different cutoff values (average ≥10/24 or 27), including as low as five (cutoff values not reported for 11 studies).

Additionally, studies seldom consider that of the nine items, five (sleep disturbances, fatigue/lack of energy, appetite changes, cognitive dysfunction, slowness) are also common symptoms of Long COVID [50] and other conditions outside of depression. These items alone can contribute up to 15 points to the score and reach the cutoff values for “depression” in most studies, which may misclassify with depression someone who is not depressed. Even though experienced clinicians can properly distinguish the complex intertwining of Long COVID symptoms and mental health consequences, inaccurate conclusions about mental health diagnoses with screeners in studies are likely to distort views about the correlations with and real prevalence of depression in Long COVID. This is not without consequence, especially in conditions, such as Long COVID, where the tendency to dismiss symptoms as mental health issues only is still far too common [51–53]. One may also hypothesize that the PHQ-9 score could correlate with the severity of Long COVID, regardless of a potential co-morbid mental health condition.

Measuring fluctuations

It is critical to describe the presence, severity and impact of the multitude of symptoms associated with Long COVID, as well as assessing their evolution over time. However, Long COVID is often an unpredictable journey, with persons experiencing health challenges that are multidimensional in nature, characterized by fluctuations in both long-, mid- and short-term intervals. While some health challenges may be experienced as episodic, some may be more persistent, consistent or stable over time [54, 55]. This dynamic aspect of the condition adds to the complexity of timing and interpreting outcome measures.

For instance, in persons experiencing PEM/PESE and/or other associated conditions (e.g., Postural Orthostatic Tachycardia Syndrome – POTS, Mast Cell Activation Syndrome – MCAS, symptoms can vary substantially depending on the presence of PEM/PESE or the triggering of flare-ups. Knowing the status of individuals (e.g., whether they are in PEM/PESE) when measuring outcomes with PROMs or performance-based outcome measures (PBOMs) is crucial, as symptoms and performance may be very different on an “average” day from those during PEM/PESE or flare-up state. Furthermore, PEM/PESE onset can be delayed by hours or days after the trigger, further complexifying appropriate capture.

Therefore, both the purpose of describing health status at one point in time (cross-sectionally) and evaluating change over time or in relation to an intervention (longitudinally) can be affected by fluctuation of symptoms. It is critical to ensure outcome measures accurately capture the fluctuating nature of Long COVID to avoid misleading results.

#4 - Identifying appropriate outcome measures

In the absence of validated biomarkers to assess disease activity, outcome measures, including PROMs and PBOMs, are essential for research and guiding assessment and management of persisting symptoms after COVID-19.

Like an old cozy sweater, using the same well-known outcome measures is comfortable and reassuring. But as an old sweater may have holes, familiar outcome measures may have unrecognized caveats for accurately measuring Long COVID. Thus, pre-existing measures may not be measuring what researchers intend to or may do so only partially.

Heterogeneity of construct measurement

Even when an appropriate construct is identified, selecting an appropriate outcome measure to assess it reliably and accurately may not be easy.

When reported, the prevalence of PEM/PESE in Long COVID varies according to the characteristics of the population, but also according to how it is measured [21, 22, 29–31, 35]. While researchers are increasingly aware of the importance of including PEM/PESE in outcome measurement, there is a paucity of short but comprehensive outcome measures to accurately screen for and assess the spectrum of exertional intolerance. The resulting heterogeneity may actually contribute to discrepancies between different rehabilitation studies [56–58], as well as observations in clinical practice.

Combination of outcome measures

Multiple outcome measures are often used together to comprehensively assess the many health domains of Long COVID. This lengthens assessment times, and may also result in unnecessary overlap, adding to the barriers of PROM use [27]. Furthermore, outcome measures often refer to different timeframes (e.g., today, over the last week or last month) as well as using varying vocabulary, question structures, and response scales. These are potential sources of confusion for responders, especially those with fatigue and cognitive dysfunction.

Myriad of outcome measures

A myriad of tools, mostly transposed from other conditions, is currently used to measure constructs for people with Long COVID. For instance, in a meta-analysis focusing on fatigue and cognitive dysfunction in Long COVID, 41 studies with PROMs/PBOMs used at least 45 different existing outcome measures in various combinations, representing 140 uses across all 41 studies. Furthermore, of 76 studies that used self-reported outcomes, 63 were unspecified/home-based questionnaires (with or without other formal PROMs/PBOMs) [59].

Measurement properties of pre-existing outcome measures

Measurement properties of the myriad of pre-existing outcome measures used in Long COVID have seldom been assessed. With such a complex and variable condition, constructs are often missing or assessed incompletely or inaccurately. The current use of pre-existing outcome measures not validated for Long COVID is based on assumptions from other conditions that may not be appropriate. These issues appear to be the norm rather than the exception.

Measuring change over time or in response to treatments is critical to determine the effectiveness of interventions. For example, without biomarkers, PROMs are used in clinical trials to measure the crucial difference between treatment and placebo arms. In Long COVID, the constellation and dynamic nature of symptoms further complicate the determination of criteria to demonstrate significant change. The frequent choice to measure constructs related to the functional impacts of the condition is warranted, but the heterogeneity of cohorts, as well as the episodic nature of Long COVID calls for careful considerations of assessments to ensure responsiveness (i.e., the ability to detect meaningful change if change truly occurs) across the spectrum of Long COVID. Otherwise, it may be impossible to determine if the lack of effect of an intervention is due to the inability of the measurement tool to detect true change, rather than absence of effect. For instance, a change on the Physical functioning sub-scale of the Medical Outcomes Study Questionnaire Short Form 36 Health Survey (SF-36) [60] is often used as a primary outcome in Long COVID studies. Although validated for use in other conditions [61], to our knowledge, the SF-36 has not been assessed in Long COVID. Meanwhile, in the absence of an assessment of its responsiveness, people with Long COVID among a group of health care professionals report notable improvement of their day-to-day activities while not eliciting change on the Physical functioning sub-scale of the SF-36, eloquently underlining the importance of formal assessment and the likely need to adapt measurement tools for this population.

Identified limitations of pre-existing outcome measures

When limitations of measurement instruments in the context of Long COVID are eventually identified, they take time to be widely acknowledged, and previous literature is not amended. For instance, there has been debate about the disparity between patient-reported cognitive dysfunction in clinical history or in research versus performance-based evaluations [62]. However, many cognitive PBOMs, such as the Montreal Cognitive Assessment (MoCA) [63, 64], may not be suitable for use in Long COVID, especially if limitations of the tool are not recognized. Clinicians have expressed concerns that the MoCA, intended to screen for mild cognitive impairment in the elderly, may not capture all phenotypes of Long COVID cognitive dysfunction. In 2022, Lynch et al. [64] showed that the MoCA may not be an appropriate screening tool for cognitive dysfunction in Long COVID, yet it is still commonly used. Failing to address limitations and potential underlying reasons [65] for discrepancies contributes to the lack of understanding and even disbelief from clinicians towards patients’ experiences.

Measurement properties of condition-specific outcome measures

Promising condition-specific instruments such as the COVID-19 Yorkshire Rehabilitation Scale (C19-YRS) and its shorter modified version (mC19-YRS) [66, 67], and the Symptom Burden Questionnaire for Long COVID (SBQ-LC) [68] have emerged. Condition-specific instruments will help address some limitations of non-specific instruments by being directly aligned with the reality of Long COVID. While some evidence on their psychometric properties is available, the crucial assessment of content validity for various phenotypes and episodic nature of Long COVID is still lacking [69].

The above examples are only a glimpse of aspects to consider when identifying appropriate outcome measures for Long COVID.

Considerations to pave the way forward

There remains much work to be done to reach coordinated and appropriate measurement for all Long COVID phenotypes. Several themes should be explored to achieve a relevant improvement of measurement framework. While assessing measurement properties of existing outcome measures and developing new measures specific for Long COVID will take time, many adjustments can be implemented rapidly, even for projects already in progress.

When using outcome measures, going back to basics [70] is always relevant: What is the target population? What needs to be measured? For what purpose? What are the measurement properties of the intended outcome measures? Have they been assessed with adults, children or young people with Long COVID? Is the interpretation appropriate for all dimensions of Long COVID? What are the perspectives of people with Long COVID and persons seeking equity on the outcome measure? What are the potential caveats, limitations or blind spots of the measure to consider for interpretation?

While some issues may not preclude the use of an outcome measure, careful appreciation of the limitations and caution when interpreting and drawing conclusions are of the utmost importance. A visual summary of a compilation of issues in measuring health outcomes for people with Long COVID is provided in Fig. 2. It emphasizes consideration of the heterogeneity of Long COVID within the psychometric properties of outcome measures (circles) in combination with an overview of the many challenges (lozenges) of this endeavor. This funnels down to the collective abilities of persons with lived experience, researchers and clinicians to tackle Long COVID measurement. Building on examples of development of instruments involving lived experiences is crucial [71] to consider for the future. Table 1 provides general perspectives and a more detailed list of short- and long-term ideas to explore while addressing the complexities and challenges for measurement of health outcomes in Long COVID.

Fig. 2.

Schematization of complexities, challenges and considerations for measurement of health outcomes in Long COVID

Table 1.

Considerations for Health Outcomes Measurement in Long COVID: Paving the way forward

Overview of avenues to collectively explore to address the complexities, challenges and considerations for measurement of health outcomes in Long COVID
Rapid	Research Teams	Persons with lived experience	Involvement early in development of study protocols is increasing [72].
			Benefits shown across research stages [73].
			Meaningful collaborations could help prevent many frequent caveats.
		Involvement of researchers with experience with IACCs.
	OMs and data collection	Identify psychometric (measurement) properties and characteristics of outcome measures	Context of validation.
			Scoring and interpretation.
			Assessment in Long COVID, including content validity and considerations for original interpretation that may not apply directly.
			Wording and translations.
			…
		Capture status pertaining to answers (usual symptoms, worse than usual – PEM/PESE/other flare-ups, before illness) to ensure appropriate assessments and comparisons.
		Timelines/questions capturing potential episodic nature.
	Adaptation	Adapt to energy-limiting conditions	Careful consideration for length to mitigate burden.
			Allow to answer in multiple sittings with reminders during progression.
			Careful screening for PEM/PESE and evaluation of threshold for PBOMs.
	Data analysis	Careful description of population.
		Scope of health domains assessed (and missed).
		Avoidance of overgeneralization of findings.
	Knowledge Mobilization	Classify and organize knowledge on Long COVID.
		Contextualization of populations and findings.
		Considerations of study variability for the interpretation of the evidence and limitations.
		Avoidance of overgeneralization of findings.
Future	Directory	Outcome measures used in Long COVID and considerations for use, including interpretation.
	Psychometric (measurement) properties	Assessment of pre-existing and condition-specific outcome measures for Long COVID.
		Consider various phenotypes and episodic nature in the assessment of content validity.
	Specific OMs for Long COVID	Consider using questions to screen for each domain, triggering more questions when positive.
		Including outcome measures for PEM/PESE in collaboration with researchers with experience in the field of ME/CFS.
	Collaboration	Promote collaboration and avoid silos.
	Knowledge Mobilization	Address knowledge gaps in Long COVID assessment and measurement (researchers and clinicians).

IACC: Infection-Associated Chronic Condition, ME/CFS: Myalgic Encephalomyelitis/Chronic Fatigue Syndrome, OM: outcome measure, PBOM: performance-based outcome measure, PEM/PESE: post-exertional malaise/post-exertional symptom exacerbation

Conclusion

This commentary highlights complexities, challenges and considerations of measurement in the context of Long COVID. Current issues impair appropriate assessments, comparison of populations and studies, development of clinical practice guidelines, clinical trials and comprehensive understanding of Long COVID with the constellation of conditions and symptoms currently encompassed under its umbrella (Fig. 1). It is also a source of confusion among researchers and clinicians.

We recognize this commentary only scratches the surface and focuses primarily on measurement issues, with only glimpses at wider consequences on the appropriate assessment and management of complex chronic conditions such as Long COVID.

There are no simple solutions to complex situations. Some of these issues are the result of ingrained habits. Once challenges have been recognized, we, persons with lived experience, researchers, and clinicians, can move forward together to lay stronger foundations for understanding the respective limitations of existing outcome measures that will need to be used while working collectively to develop a coordinated approach to guide appropriate and meaningful measurement in Long COVID. This may serve as a benchmark for tackling the measurement of other complex chronic conditions in the future.

Abbreviations

COVID-19

Coronavirus Disease 2019

ME/CFS

Myalgic Encephalomyelitis/Chronic Fatigue Syndrome

NASEM

National Academies of Sciences, Engineering, and Medicine

PBOM

Performance-based outcome measure

PEM/PESE

Post-exertional malaise/post-exertional symptom exacerbation

PHQ-9

Patient Health Questionnaire − 9 items – Depression module)

PROM

Patient-reported outcome measure

SOB

Shortness of breath

Author contributions

AB led the conceptualization of this commentary with collaboration from KM, LL, MOH, ZK and KKO. AB drafted the manuscript with KM, ZK and KKO. All authors reviewed and contributed to refinements of the commentary. All authors reviewed the final version of this commentary. AB, LL and MOH are persons living with Long COVID. FN is a person living with chronic pain and ADG is a person living with multiple chronic conditions and rare diseases. All embedded their lived experiences within this work.

Funding

The Measurement and Assessment Paradigm in Long COVID (MAP LC) study is funded by Long COVID Web, funded by the Canadian Institutes of Health Research (CIHR) – FRN 185352. KKO is supported by a Canada Research Chair in Episodic Disability and Rehabilitation (CRC-2022-00510) and KLL by a Canada Research Chair in Behavioural Medicine (CRC-950-232522), both from the Canada Research Chairs Program.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests with this commentary. DM acknowledges a leadership role in the development of a Core Outcome Measurement Set for post COVID-19 condition for adults and children as a part of the PC-COS project.