Diagnostic challenges of long COVID in children: a survey of pediatric health care providers' preferences and practices

Abstract

Introduction: Given the challenges in diagnosing children with long COVID, we sought to explore diagnostic practices and preferences among clinicians.

Methods: A ten-question survey assessed pediatric providers' clinical decision making for identifying and evaluating long COVID in children. Of the 120 survey respondents, 84 (70%) were physicians, 31 (26%) nurse practitioners, and 5 (4%) physician assistants.

Results: The most common categories of symptoms identified as raising suspicion for long COVID in children included cardiopulmonary symptoms, selected by 119 (99%) of pediatric providers, and neurocognitive symptoms, selected by 118 (98%) of providers. However, there was more ambiguity on the primary feature of long COVID, with providers selecting a range of key symptoms. Of all physical exam findings, postural orthostatic tachycardia, was most suggestive of long COVID [identified by 49 (41%) of pediatric providers], whereas one-third of providers reported no specific identifiable exam finding.

Discussion: Pediatric providers report variable decision making in the clinical evaluation of long COVID, with patient demographics and clinical factors impacting whether a diagnosis of long COVID is considered. This variation in diagnosing pediatric long COVID reflects ambiguity in the definition of long COVID in children and the absence of clinical guidelines to support providers in the identification of disease and treatment. This study highlights an area of need for future clinical advances in pediatric long COVID.

Keywords: COVID-19; PASC; SARS-CoV-2; long COVID; pediatric; post-acute sequalae of COVID; post-viral illness.

Figure 1

Provider demographics and geographic distribution. Demographic information of providers including medical degree, specialty, years in practice, and workplace community type. Geographic distribution of providers is visualized on the map. *Other specialties include 1 sleep medicine, 1 psychiatry, 1 primary care, and 2 unspecified specialties. Created with mapchart.net.

Figure 2

Selected symptoms concerning for long COVID by each individual provider. Each row corresponds to an individual provider's selection of new onset symptoms concerning for long COVID. Rows are organized from top to bottom as least to most selected total number of symptoms. (Cardio/Pulm, cardiopulmonary; Endo, endocrine; GI, gastrointestinal; MSK, musculoskeletal; Neuro, neurocognitive).

Figure 3

Long COVID clinical presentation and provider diagnostic preferences. Percent of providers selecting (a) key primary symptoms and (b) exam findings concerning for long COVID. (c) Diagnostic tests ordered by pediatric providers including (d) specific blood tests. (e) Percent of providers referring patients to subspecialists if there is clinical concern for long COVID. (ECG, electrocardiogram; CPET, cardiopulmonary exercise testing; EMG, electromyogram; CBC, complete blood count; CRP/ESR, C-reactive protein/elevated sedimentation rate; ANA, anti-nuclear antibody; IgA-TTG, immunoglobulin A-tissue transglutaminase; SSA/SSB, anti-Ro/anti-La antibodies).

Figure 4

Provider understanding of long COVID symptom timeline. Percent of providers who would consider a diagnosis of long COVID based on (a) timing of symptom onset after acute infection and (b) symptom duration. (c) Reasons why providers chose timing of symptom onset does not matter.

Figure 5

Influence of patient demographics on provider diagnosis of long COVID. Forest plot representing percent of providers who are more (1.0) or less likely (−1.0) to consider a long COVID diagnosis based on age, sex, race/ethnicity, and COVID-19 vaccination status. Horizontal bars indicate 95% confidence intervals. P-values < 0.05 are considered significant.

Figure 6

Influence of other clinical considerations on provider diagnosis of long COVID. Forest plot representing percent of providers who are more (1.0) or less likely (−1.0) to consider a long COVID diagnosis based on other relevant clinical features.