Progressively Widening Healthcare Costs in Long COVID Over Five Years

This retrospective cohort study of over 143,000 adults reveals that Long COVID is associated with progressively widening healthcare costs over five years, driven primarily by a threefold increase in visit frequency rather than care intensity, resulting in a cumulative excess burden of $7,124 per patient.

The Gist

The Big Picture: A "Silent Slow-Motion" Bill

Imagine you get a flat tire on your car. Usually, you fix it, pay the bill, and drive away. That's how we thought "Long COVID" worked: you get sick, you recover, and eventually, the medical bills go back to normal.

This study, however, found that for millions of people, Long COVID isn't like a flat tire. It's more like a car that has a slow, invisible leak in the engine. You fix the immediate problem, but over the next five years, the car keeps sputtering, needing more and more oil, more tune-ups, and more visits to the mechanic. The bill doesn't just stay high; it gets worse every single year.

The Study: A Five-Year Race

Researchers at Massachusetts General Hospital looked at the medical records of over 143,000 adults who had tested positive for COVID. They split them into two groups:

1. The "Recovered" Group: People who got sick but didn't develop Long COVID.
2. The "Long Haul" Group: People who developed Long COVID (symptoms lasting months or years).

They watched these two groups for five years (20 quarters) to see how much money they spent on healthcare.

The Shocking Result: The Gap Keeps Growing

Most previous studies only looked at the first 6 to 12 months. They saw that Long COVID patients spent more money, but they thought the gap would eventually close as people got better.

This study found the opposite. The gap didn't close; it exploded.

• At the start (Year 0): Long COVID patients spent about $79 more per quarter than the recovered group.
• By Year 5 (Quarter 19): That difference grew to $236 more per quarter.

The Analogy: Imagine two runners. One is running a normal pace. The other has a heavy backpack that gets heavier every mile. At the start, the backpack isn't that heavy. But by mile 20, the backpack is so heavy the runner is moving much slower and needs to stop for water, snacks, and medical help constantly. The "cost" of carrying that backpack triples over time.

Why Are the Bills Going Up? (The "Visit" vs. The "Price")

The researchers broke down why the bills were rising. They looked at two things:

1. How often they visited the doctor (Frequency).
2. How expensive each visit was (Intensity).

The Finding: The bills went up because Long COVID patients were visiting the doctor much more often, not because each visit became astronomically expensive.

• The Metaphor: Think of it like a house with a leaky roof.
  • Recovered Group: They fixed the roof once. No more leaks.
  • Long COVID Group: The roof keeps dripping. They don't need a new roof every time (high intensity), but they have to call the plumber more and more often (high frequency) to fix the small drips, manage the water damage, and check the gutters. Over five years, those frequent small calls add up to a massive bill.

Who Is Hit Hardest?

The study looked at different types of people to see if the "leaky roof" affected everyone the same way.

• Severity: It didn't matter if you were hospitalized for your initial COVID or if you stayed home. Both groups saw their costs rise over time.
• Gender & Age: Men, women, young, and old all saw the same pattern of rising costs.
• The "Super-Spenders": The study found that while everyone with Long COVID spent more, a small group of very sick patients spent a lot more. At the very top end (the 99th percentile), Long COVID patients were spending over $8,000 more per quarter than healthy people. This is like a few houses in a neighborhood having a massive flood while the others just have a few drips.

The Bottom Line

Long COVID is not a temporary phase; it is a chronic, progressive condition.

• The Old Myth: "You get sick, you recover, and the costs go down."
• The New Reality: "You get sick, and the healthcare needs slowly but steadily increase for years."

Over five years, a single Long COVID patient costs the healthcare system an extra $7,124 compared to someone who recovered fully. When you multiply that by the millions of people affected, it's a massive economic burden that is only going to get bigger if we don't find better ways to manage it.

In short: Long COVID is a slow-burning fire. It doesn't burn down the house immediately, but it keeps smoldering, requiring constant attention and money for years to come.

Technical Summary

1. Problem Statement

Long COVID (Post-Acute Sequelae of SARS-CoV-2 Infection, or PASC) affects an estimated 400 million people globally, creating a massive economic burden. While prior studies have documented elevated healthcare costs for Long COVID patients, they are limited by short follow-up periods (typically 6–15 months). These studies generally found elevated but stable excess costs, leading to the assumption that the economic burden would attenuate as patients recovered or adapted.

The Research Gap: There is a critical lack of evidence regarding the long-term trajectory of these costs. It remains unknown whether the economic burden stabilizes, attenuates, or widens over a multi-year period. Furthermore, existing studies often rely on ICD-10 codes (which have poor sensitivity) or lack granular decomposition of costs into utilization frequency versus care intensity.

2. Methodology

Study Design & Data Source

• Type: Retrospective cohort study.
• Data Source: Electronic Health Records (EHR) from the Mass General Brigham (MGB) health system, covering 10 hospitals and 20 community health centers in Massachusetts.
• Timeline: January 1, 2018, to December 31, 2024.
  • Baseline: Pre-pandemic period (2018–2019) to control for pre-existing utilization.
  • Analysis Period: Post-infection period (2020–2024), spanning 20 consecutive quarters (5 years).
• Cohort: 143,544 adults (≥18 years) with documented SARS-CoV-2 infection.
  • Long COVID Group (PASC): 27,986 patients.
  • Control Group: 115,558 patients without Long COVID.

Case Definition (Phenotyping)

• Long COVID was identified using a validated precision phenotyping algorithm (P2RC) rather than simple ICD codes.
• Criteria: Symptoms persisting >2 months post-infection, not explained by alternative diagnoses.
• Exclusion Logic: A symptom was classified as PASC only if it emerged after the index infection and no prior diagnosis (from 2017 onward) could explain it (e.g., new dyspnea without pre-existing asthma).

Cost Estimation

• Metric: Quarterly healthcare costs per patient.
• Standardization: Costs were derived from institutional billing records but standardized using CMS (Centers for Medicare & Medicaid Services) national average payment rates (OPPS and Inpatient PPS) mapped via CPT/HCPCS codes. This removes payer-specific negotiation biases.
• Scope: Includes hospital, ambulatory surgery, and Medicare payments. Note: Does not include pharmacy, lab tests, or professional E&M fees.
• Adjustments: Costs were adjusted for inflation (2024 USD) and winsorized at the 99.5th percentile.

Statistical Analysis

• Primary Model: Two-part Generalized Estimating Equation (GEE) models to handle the skewed distribution of healthcare costs (many zeros, heavy right tail).
  1. Part 1 (Logistic): Probability of incurring any cost (utilization).
  2. Part 2 (Gamma): Magnitude of costs conditional on utilization.
• Decomposition: Separate models analyzed Visit Frequency (Poisson GEE) and Cost-per-Visit (Gamma GEE) to distinguish between increased utilization vs. increased intensity.
• Covariates: Age, sex, race/ethnicity, Charlson Comorbidity Index, baseline costs, and hospitalization status.
• Sensitivity: Survival analysis (Kaplan-Meier) was used to adjust for differential mortality, calculating population-level expected costs to account for terminal care costs.

3. Key Contributions

1. Longest Follow-up: Provides the longest longitudinal analysis of Long COVID healthcare costs to date (5 years/20 quarters).
2. Trajectory Discovery: Challenges the "stable cost" hypothesis by demonstrating a progressively widening cost gap over time.
3. Mechanistic Insight: Decomposes costs to reveal that the widening gap is driven by increasing visit frequency (utilization) rather than increasing cost-per-visit (intensity).
4. Precision Phenotyping: Utilizes a rigorous, exclusion-based algorithm to minimize misclassification bias compared to code-based studies.
5. Distributional Analysis: Highlights that excess costs are heavily concentrated in the upper tail of the distribution (high-cost patients).

4. Key Results

Demographics

• The Long COVID group was older (mean 56.2 vs. 53.0 years), had higher comorbidity burdens (Charlson Index 3.14 vs. 2.25), and had higher baseline utilization than controls.

Primary Outcome: Widening Cost Gap

• Baseline (Quarter 0): Adjusted excess cost was $79 per quarter (95% CI: $48–$118).
• Quarter 19 (End of Study): Adjusted excess cost widened to $236 per quarter (95% CI: $176–$287).
• Trend: The cost differential tripled over the 5-year period. There was no evidence of plateauing; the divergence continued to accelerate.

Drivers of Cost (Decomposition)

• Utilization (Visit Frequency): Long COVID patients had a 20% higher odds of any utilization. By Quarter 19, visit frequency was 44% higher than controls. This gap widened over time.
• Intensity (Cost-per-Visit): While Long COVID patients had ~30% higher costs per visit, this premium remained stable over time (no significant widening interaction).
• Conclusion: The economic burden is driven by the need for more frequent care, not more complex/expensive individual encounters.

Cost Distribution

• Excess costs were highly concentrated in the upper tail.
• At the 99th percentile, the absolute difference between groups was $8,482 per quarter.
• The 99th percentile Long COVID patients incurred ~$21,321/quarter vs. $12,839 for controls.

Subgroup Consistency

• The widening trajectory was consistent across:
  • Severity: Both hospitalized and non-hospitalized acute infection groups showed widening gaps.
  • Sex: Similar patterns for males and females.
  • Comorbidity: Widening occurred in both low and high comorbidity groups, though absolute costs were higher in the high-comorbidity group.

Mortality-Adjusted Cumulative Costs

• After accounting for differential mortality (survival bias), the cumulative 5-year excess cost was $7,124 per Long COVID patient.
• Population-level costs peaked around Year 3 (Quarter 12) and declined slightly in Years 4–5 due to survival effects, but the excess cost remained substantial ($340–$480/quarter).

5. Significance and Implications

Clinical & Health System Planning

• Chronic Disease Model: The findings suggest Long COVID behaves as a progressive chronic condition rather than a transient post-acute state. The assumption that costs will naturally attenuate is incorrect.
• Resource Allocation: Health systems must plan for increasing utilization needs over time, not just initial spikes. Static cost projections based on short-term data will severely underestimate future burdens.
• Care Strategy: Since the driver is frequency rather than intensity, interventions should focus on care coordination, longitudinal management, and multidisciplinary clinics to manage frequent monitoring needs, rather than just acute exacerbation management.

Economic Impact

• With tens of millions of cases in the US, the aggregate annual cost is projected to be in the tens of billions of dollars.
• The $7,124 per-patient excess cost over 5 years implies a massive strain on insurance risk pools, employer health plans, and public programs.

Policy & Research

• Policy: Urgent need for sustained investment in Long COVID infrastructure and disability support.
• Research: Future studies must prioritize extended follow-up (beyond 12 months) to capture the true trajectory of chronic conditions.
• Limitations: The study notes that costs are likely underestimated as they exclude pharmacy and lab fees. Additionally, the single-system design may limit generalizability, though the large sample size and rigorous phenotyping strengthen internal validity.

Conclusion
This study fundamentally shifts the understanding of Long COVID's economic impact, demonstrating that the condition imposes a substantial, growing, and persistent financial burden on the healthcare system, driven primarily by the escalating frequency of healthcare encounters.