Substantial health and economic burden of COVID-19 during the year after acute illness among US adults not at high risk of severe COVID-19

Nearly 20% of adults diagnosed with COVID-19 experience symptoms for ≥ 3 months after first contracting the virus [1, 2]. These highly variable signs and symptoms, often termed post-COVID conditions, can either begin at the time of initial infection and persist for several months or may be new symptoms or syndromes that develop only after the acute phase of COVID-19 [3‐5]. For many individuals, post-COVID conditions involve multiple organ systems and significantly impair daily functioning and productivity [6]. Although a universally accepted definition and timeframe of the condition has not yet been developed, a clinical diagnosis of post-acute sequelae of COVID-19 (PASC) was assigned an International Classification of Diseases 10 (ICD-10) code (U09.9) in October of 2021 [7, 8].

Although it is widely accepted that older age, belonging to racial and ethnic minority groups, and certain underlying medical conditions are associated with an increased risk of progression to severe COVID-19 upon initial infection [9], the characteristics associated with risk of developing post-COVID conditions are largely unknown. Some overlapping but distinct risk factors, such as female sex and older age, have been identified, as well as an association between acute COVID-19 severity and duration of symptoms [10‐12]. However, numerous reports from several countries have identified high rates of post-COVID conditions, even among patient cohorts with mixed disease severity or mild cases of COVID-19 [4, 13‐15]. In the current landscape where mild COVID-19 illness is becoming more common, owing both to vaccination [16] and the highly transmissible but potentially less virulent Omicron strain [16‐19], it is necessary to understand the burden on health and healthcare systems after an acute COVID-19 infection among patients who do not have underlying comorbidities and who did not require hospitalization for acute COVID-19. An understanding of long-term health effects after acute COVID-19 infection, the populations at risk, and the associated strain on healthcare systems is imperative to inform accurate estimations of the evolving clinical and economic burden of COVID-19.

We conducted a descriptive, retrospective analysis of morbidity, healthcare resource utilization, and costs associated with the post-acute phase of COVID-19 among adult patients aged < 65 years and without any underlying conditions placing them at high risk of progression to severe disease. In a companion report [20], we describe an identical analysis conducted among patients with ≥ 1 underlying condition or who were aged ≥ 65 years. The companion report and the current report were both purely descriptive with no formal statistical comparisons, with the goal to present a broad and unbiased dataset that can inform future hypothesis generation.

This descriptive, retrospective cohort study compared baseline healthcare utilization data from patients during the year before contracting COVID-19 with their healthcare utilization data during the year after recovery from the 30-day acute phase (from day 31 through day 390 after diagnosis). All patients served as their own control for evaluation of diagnoses, medications, healthcare utilization, and costs before compared with after acute COVID-19. Details regarding the study design, data source, and inclusion and exclusion criteria have been included in the companion manuscript regarding individuals at high risk of progression to severe COVID-19 [20].

Briefly, enrolled patients were diagnosed with COVID-19 (ICD-10 diagnosis code of U07.1) between April 1 and May 31, 2020 (the index period). Information regarding diagnoses, medications, healthcare utilization, and costs were collected from Optum’s de-identified Clinformatics® Data Mart Database (CDM), which contains patient-level information derived from administrative health claims of commercial and Medicare Advantage plan members. Data extracted from CDM were not sufficient to determine whether any post-COVID-19 diagnosis, medication prescription, or healthcare utilization was specifically caused by COVID-19; therefore, it was unknown whether any individual diagnosis or adverse health outcome was truly a “post-COVID condition” or was caused by unrelated factors [5]. Eligible patients were continuously enrolled in CDM (with gaps of ≤ 45 days permitted) over the 12 months before and 13 months after COVID-19 diagnosis and were aged 18 to 64 years on the index date. Patients were excluded if they had a diagnosis code (ICD-10-Clinical Modification [ICD-10-CM]), procedure code (ICD-10-Procedure Coding System [ICD-10-PCS], Current Procedural Terminology [CPT®], Healthcare Common Procedures Coding System [HCPCS]), or National Drug Code (NDC) for any condition placing them at high risk of progression to severe COVID-19, per CDC definitions as of October 14, 2021 [9], within the 12 months before the index date. Additional exclusion criteria were hospitalizations for ≥ 5 consecutive days during the baseline phase; any time spent at a long-term care facility, skilled nursing facility, inpatient rehabilitation, or hospice during baseline or at index date; an ICD-10 code for confirmed COVID-19 before the index period; or death during the acute phase of COVID-19.

The top 500 ICD-10 diagnosis codes were analyzed by ICD-10 code chapter (a system of organization based on the most affected organ systems or types of injury/disease), and medications were categorized according to Uniform System of Classification class. All diagnosis and medication categories applicable to < 2% of the overall population during the baseline phase were excluded from analysis. The “biologics” category was also excluded based on incomplete data capture. Standard medical cost means, standard deviations, medians, and quartiles were calculated using the number of patients with a related visit or service, and nonzero costs were calculated using the number of patients with a cost > 0 associated with that visit or service. No adjustments were made for patients who died during the 12-month post-acute phase; all deaths that occurred during the post-acute phase were accounted for and reported. Absolute and relative change from baseline to the post-acute phase were calculated for each outcome where possible using frequency counts. To better understand the relationship between post-acute outcomes and initial infection severity, all data were presented for the overall population and stratified by patient disposition during the acute phase of COVID-19 (within 30 days after diagnosis): not hospitalized, hospitalized without intensive care unit (ICU) admission, or admitted to the ICU. Analyses were performed using SAS version 9.4 (SAS, Cary, NC), and no statistical inference tests were conducted.

This retrospective analysis of adults aged < 65 years without any underlying high-risk conditions identified increases in diagnoses, medical prescriptions, healthcare utilization events, and associated costs during the year after the acute phase of COVID-19. As observed previously [10] and in our companion manuscript regarding high-risk patients [20], those who were hospitalized with or without ICU admission during the acute phase had the greatest increase in comorbidities and healthcare resource utilization burden. However, the burden was apparent across all 3 cohorts, including those who were not hospitalized for COVID-19.

Many of the changes observed in this cohort were similar to observations among high-risk patients [20]. Most notably, the greatest increase in both studies was in the percentages of patients with blood-related diseases, which has been identified previously as a feature of post-COVID conditions [21]. In the present study, new diagnoses of blood disorders increased by > 150% during the post-acute phase compared with baseline, even among the cohort of patients who were not hospitalized during the acute phase of illness. New diagnoses of neurological and psychiatric diseases also increased in the year following acute COVID-19 in both patients with and without any risk factors for severe COVID-19, regardless of the level of care setting during the acute phase of illness, which is consistent with previous reports of long-term COVID-19 sequelae [22, 23]. We observed a 27% decrease in respiratory disease during the post-acute phase; similar to results from the high-risk population, this decrease reflected a large increase in chronic lower respiratory diseases that was outweighed by a combination of smaller decreases in acute upper and lower respiratory infections, influenza, and pneumonia. Distinct from observations in high-risk patients, analysis of the cohort described here also revealed a 123% increase in endocrine, nutritional, and metabolic diseases and a 76% increase in diseases of the digestive system.

The analysis of medication use demonstrated the greatest increase in hormone prescriptions, including primarily injectable, oral, and topical glucocorticoids and corticoids (e.g., prednisone, methylprednisolone, and dexamethasone). These medications were typically not prescribed during the baseline phase and were likely being used for the treatment of persistent COVID-19 symptoms, such as joint stiffness and muscle pain [24]. Increases in hormone use were followed by vascular and musculoskeletal agents, and increases across the 3 classes of medications were observed across all levels of care during acute COVID-19. Those results were unique to this population and not observed among high-risk patients [20]. However, when the analysis was conducted including medication classes prescribed to < 2% of the baseline population, a 277% increase in hemostatic modifier prescriptions and a 106% increase in blood factor prescriptions were identified, consistent with results of the same analysis among high-risk patients. It is logical that several of the medication classes prescribed to ≥ 2% of the baseline population in the high-risk patient cohort were prescribed to < 2% of the baseline population in the cohort evaluated here because these patients were overall younger and healthier at baseline.

Healthcare resource utilization and costs were higher during the post-acute phase compared with the baseline phase for all evaluated measures, including outpatient visits, ED visits, inpatient visits and LOS, ICU admission and LOS, 30-day readmissions, and prescriptions. The likelihood of patients being discharged to another hospital department or long-term care facility also increased. The most striking cost changes were observed across measures related to inpatient resources, including an overall increase in inpatient visit costs of nearly 950% in the post-acute phase and resulting in a 178% increase in total medical costs in the overall population.

Cost increases varied in magnitude but were observed for every healthcare outcome across all categories of the acute COVID-19 level of care. Even among those who were not hospitalized during the acute phase of COVID-19, inpatient visit costs during the post-acute phase increased by 578%, outpatient visit costs increased by 139%, and total medical costs increased by 138%. This trend was similar to observations made in the cohort of high-risk patients [20] but was a more surprising result given that the patients in the cohort described here were aged < 65 years and lacked any comorbid conditions placing them at risk of severe COVID-19.

Taken together, health and healthcare resource use results suggest that the risk factors associated with developing post-COVID conditions may be distinct from the risk factors that predict severity of acute disease. Similarly to how biomarkers have been characterized to predict the course of acute COVID-19 [25], there may be a unique set of biomarkers associated with the development of long-term adverse health outcomes. One recent study identified biomarkers associated with vascular transformation among long-COVID patients [26], which is consistent with findings in our present study and companion report regarding a high prevalence of blood-related diseases in the post-acute phase. Importantly, regardless of the mechanisms involved, results indicate that the health and economic impacts of COVID-19 may extend beyond the acute phase of illness even among the wide swath of the population that is relatively young and healthy and has mild symptoms upon infection.

Although limited data are available on strategies to reduce the risk of post-COVID conditions, vaccination against COVID-19 appears to be protective. In a recent prospective study from Antonelli and colleagues [27], fully vaccinated adults with breakthrough infections were less likely than unvaccinated controls to experience symptoms of COVID-19 lasting ≥ 28 days; the effect was observed among both older adults and adults aged < 60 years. In a retrospective cohort study from Taquet and colleagues [28], vaccination among adults who contracted COVID-19 was associated with steep reductions in risk of several long-term adverse health outcomes. Although authors of the Taquet study did not find COVID-19 vaccination to be associated with reduced risk of what they termed “long COVID features” (a collection of specific abdominal, respiratory, psychiatric, and pain-related symptoms), they did identify significantly lower risk of many of the diagnoses discussed in the present report, including blood disorders, muscle disease, certain neurological conditions, and chronic respiratory conditions. Because these outcomes were measured over a 6-month period after infection, they included diagnoses during both the acute and post-acute phases. Highlighting the importance of vaccination even among those not in a high-risk group, protective effects of vaccination against many post-acute sequelae were most robust among individuals aged < 60 years.

Recently, emerging data have also suggested that antiviral treatment for COVID-19, such as nirmatrelvir/ritonavir (Paxlovid®, Pfizer Inc, New York, NY, USA), may reduce the likelihood of developing PASC. In a large study of patients from the Veterans Affairs database [29], authors found that individuals prescribed nirmatrelvir-ritonavir during the acute phase of COVID-19 had reduced risks of prespecified sequelae (including cardiovascular, hematologic, and neurologic disorders), post-acute hospitalization, and post-acute death, regardless of vaccination status. A small case series has furthermore suggested that nirmatrelvir administered during the post-acute phase may alleviate long-term symptom burden [30], although more systematic study of these effects is warranted before conclusions can be drawn.

Our study had several strengths. Primarily, having all patients serve as their own control inherently adjusted for potential confounders, such as demographics and health-seeking behavior. To our knowledge, this was also the first study that evaluated the economic impact of post-acute COVID-19 in adults who were < 65 years of age and had no underlying comorbidities placing them at risk of severe acute COVID-19. The ability to pair these broad, descriptive data with the identical analysis in a cohort of adults with high-risk conditions [20] is of significant value in understanding differences between characteristics that predict acute versus post-acute COVID-19 outcomes. A limitation of our study was that the cohort was confined to individuals with commercial insurance; all enrolled participants were also diagnosed early in the pandemic and survived the acute phase of COVID-19. Additionally, there was no method of confirming that any adverse health outcomes reported here were related to COVID-19. There was also a possibility for incomplete data capture owing to nonbillable diagnoses, and our results may have been influenced by surveillance bias, whereby contracting COVID-19 led to higher medical scrutiny following diagnosis. Finally, the study was conducted in a period before vaccination and previous SARS-CoV-2 infection and before the emergence of SARs-CoV-2 variants of concern. Such factors could limit the generalizability of the findings to the present. Baseline assessments were also performed in the prepandemic period; whereas post-acute COVID-19 assessments were performed during a public health emergency in which healthcare practices could have changed. Thus, clinical burden and health costs during the post-acute COVID-19 period may have been impacted by altered heathcare practices.

Our data suggest that the health and economic burden of COVID-19 stretches well beyond the acute phase of illness, even among younger individuals without preexisting conditions whose acute infection did not merit hospitalization. Understanding the nature and extent of post-COVID conditions, as well as the unique populations at risk, is critical to evaluating the true societal cost–benefit of interventions such as COVID-19 vaccination and treatment.