Pricing the Invisible, Valuing the Cure

by Manuel Duenas | 2 weeks ago | in Cost of Illness: TOPx Submission
Approved

Quantifying the Human and Economic Costs of Invisible Illness

A Health Policy Decision Console by Healthful Data

Overview

Invisible illnesses are costly and poorly counted, so they are hard to fund or prioritize. We built a working policy console that answers one question for Long COVID: what would a cure be worth to the United States?

A policymaker or analyst sets a few levers: treatment effectiveness, share of the affected population reached, time horizon (1, 10, or 20 years), and discounted or undiscounted value. The console returns a national value with a range. In our base case, a hypothetical treatment that is 75% effective and reaches 65% of people living with Long COVID is worth $5.32 trillion over 20 years at a 3% discount rate. The range is $3.09T to $7.55T if any single parameter is taken to its evidence extreme. The widest driver is willingness to pay per QALY.

The model is built to be audited and disputed. Every number traces to a source, and an analyst can switch sources on or off, add their own, and see how the result changes.

Alongside the national burden and decision console, we built a community-led data-collection layer: the Community-Led Long COVID Burden Index (CLCBI) built off our team's expertise in developing and deploying similar models globally for the HIV pandemic and vetted by a half dozen individuals living with ME/CFS and Long COVID via the Science 4 Me (S4Me.info) forum and A Million Points of Light, a campaign and map that lets people with invisible illness be counted that includes several real lived experiences from individuals living with invisible illness from our team's sister nonprofit, Healthful Stories (HealthfulHeroes.com).

1. Results for the American People

The problem. Federal data systems capture invisible illness poorly. Claims and diagnosis codes miss most people with Long COVID, and the most severely affected are the least likely to appear in any survey. Without a defensible cost-of-illness figure, the burden is easy to overlook in funding and policy decisions.

What we delivered.

  • A national valuation of the human and economic burden of Long COVID. It covers value of health recovered (QALYs), medical spending avoided, national lost earnings, and wages recovered.

  • Three distinct definitions of the affected population, so policymakers can see how much the answer depends on who is counted:

    1. People who self-report Long COVID, whether or not they were symptomatic at the time of infection.

    2. People with symptomatic COVID-19 who self-report Long COVID.

    3. Diagnosed Long COVID, identified by U09 ICD-10 codes. We expect this scope to substantially undercount.

  • Quality-of-life loss per person, measured against age-matched population utility.

  • Burden modeled over time, including the annual rate of change in prevalence for each population definition.

Lived experience. The CLCBI is a community-led instrument. It invites caretakers to respond on behalf of people in their care, and we state openly that the most severely affected remain under-represented, so burden is likely under-quantified. A Million Points of Light adds a public, persistent map where people see they are not alone, with each of 17 invisible illnesses glowing a different color. Both are designed so lived experience feeds the model directly.

Of note, the Community-Led COVID Burden Index is designed for anyone who has ever been diagnosed with a COVID infection. For those who are unaware of Long COVID, it provides a knowledge assessment and offers recommendations on how they might increase their knowledge. It also offers a risk score based on modifiable and unmodifiable features of risk to ensure that individuals with higher risk of contracting Long COVID are aware of those risks. Uniquely, those who we establish as having features of Long COVID can be activated toward care.

Note that the CLCBI was also vetted with input from the Science 4 ME (S4Me.info forum) and after conversations with leading experts such as Professor Lenny Jason and Professor Matt Mazewski among others.

2. Data, Tech & Evidence-Driven Innovation

U.S. open data. We wired in MEPS and Household Pulse Survey data directly and reviewed each by hand to build the population scopes and prevalence trends. Our evidence pipeline screened roughly 6,500 data sources. We ingested the clinical trial landscape (ClinicalTrials.gov) to show where treatment development stands, and we connect the fragmented sources into one valuation.

AI-accelerated evidence runner. An AI-assisted literature-discovery and extraction pipeline (PubMed and ClinicalTrials.gov) pulls candidate estimates for each parameter. It combines multiple sources using inverse-variance pooling and records the disagreement range between them.

Source register and transparency. A searchable register shows every candidate source, its extracted value, the supporting quote, a link, and its review status. Analysts can include or exclude sources, add their own, and watch the headline update.

Sensitivity analysis. A tornado chart shows how much each parameter moves the result. This makes the evidence gaps visible. Several key multipliers still rest on a single source, and the console flags them as validation priorities.

Discounting. Discount rates are set by scenario and sourced through the same evidence runner. The headline range shows how far the answer moves under different assumptions.

3. Execution, Efficiency & Delivery

Working today.

  • The live decision console runs end to end. It includes the calculation view, source register, breakdown, and a “How it works” explainer.

  • The CLCBI is live on the Healthful platform, and the Million Points of Light signup page and map are live.

  • Thrive.JoinHealthful.com has already served over 12,000 individuals in the US and with proper financing, we are excited to expand it to create a longitudinal cohort of 10,000 individuals living with Long COVID over the two years.

Near-term use. An analyst can use the console today to:

  • test how a cure or treatment scenario changes national value;

  • find which assumptions drive the answer;

  • justify where new data collection would reduce uncertainty most.

What we scoped deliberately. The MVP covers the core valuation and the time view. Additional modules (cure R&D return, cure timeline, downstream excess disease risk, program costs) are specified and sequenced on top of the same evidence-runner architecture.

4. Scalability, Reuse & Societal Impact

Reusable by design. Nothing in the architecture is specific to Long COVID. The condition selector, population scopes, utility estimates, cost parameters, and evidence runner are generic components. Adding ME/CFS or another chronic condition means supplying new population, utility, and cost inputs, not rebuilding the model. The selector already lists additional conditions as demonstration entries “in development.”

Policy application. The same logic supports:

  • research prioritization across conditions, comparing cure value per dollar of R&D;

  • coverage and program decisions, by showing medical spending avoided;

  • regulatory and benefit-cost analysis, using a transparent QALY valuation including those supported by US HHS.

Standards alignment. The CLCBI and the evidence register use structured, auditable records that can be mapped to federal data standards. Downloads are structured into FHIR format

Community scale. A Million Points of Light is an open invitation for people with invisible illness to be counted, and it works across all 17 listed conditions. It gives the model a growing, community-sourced base for future parameters. The Community Led COVID Burden Index runs on the same platform that has already proven valuable to over 12,000 users and delivers value regardless of Long COVID, ME/CFS, or other invisible illness status. That is, for the many millions who remain undiagnosed, a risk score helps to encourage them to take precautions or seek clinical evaluation for symptoms which persist 12 or more weeks after COVID infection. We can adjust these triggers over time as the science evolves.

Every figure on this page can be traced back to a document or a dataset, and every stage below is built so the trace cannot break quietly. That is the architecture in one sentence: it is designed to make an untraceable number impossible rather than unlikely.

Solution Overview (MVP Demonstration):

User-Facing Application

Live demo, prototype, or working product

Our approach is a complete, live user-facing data and AI system with interactive visualization and backend integration with 3 (Three) fully operational and deployed platforms that integrate and complement each other:

1) A Health Policy Decision Console built with AI, US Open Data, and after review of ~6,500 data sources - Pricing the Invisible, Valuing the Cure - Quantifying the Human and Economic Costs of Invisible Illness - Representing the team's combined 70+ years in HEOR and data science modeling and AI and ML engineering: https://pricing-the-invisible.thrive.joinhealthful.com ;; https://yellow-pebble-0a19a980f.6.azurestaticapps.net Password: sfmwk-umjrt-862qym

2) A Million Points of Light: A Campaign that Supports Sharing and Visualization of Lived Experience, with Several 20-minute or Longer Stories from Real People Living with Invisible Illness: https://thrive.joinhealthful.com/shine-light

3) The Community-Led Long COVID Burden Index: Built leveraging decades of team experience supporting similar work for the global HIV pandemic, expertise in patient-reported outcomes: https://thrive.joinhealthful.com/module/module-details/6aa8a7e4a0acc5a18d7373af

Access instructions (if needed)

https://yellow-pebble-0a19a980f.6.azurestaticapps.net Password: sfmwk-umjrt-862qym

Results for the American People

Target users. Federal health economists and budget analysts (ASPE, CMS, NIH, CBO), congressional and agency staff building the case for research funding, state health departments, and patient advocacy organizations who need a defensible number rather than an anecdote.

The workflow it fits. An analyst is asked what a Long COVID treatment would be worth, and has days, not months.

Today that means assembling scattered published estimates, inheriting each author’s population definition, and producing a figure nobody downstream can interrogate.

With the console, the analyst opens one page and:

  1. Picks the population definition — self-reported, symptomatic-and-self-reported, or U09-diagnosed — and sees how much the answer depends on who is counted.

  2. Sets effectiveness, reach, horizon, and discount rate to match the scenario being costed.

  3. Clicks any number to see the source behind it, the verbatim quote, and the analyst who verified it.

  4. Disagrees. Switches a source off, adds their own, and regenerates the total rather than rebuilding the model.

  5. Reads the tornado chart to see which evidence gap moves the result, which tells them where new data collection is worth funding.

  6. Exports a FHIR R4B bundle so the figure arrives in their own system with its derivation intact.

For patients and advocates, the Community-Led Long COVID Burden Index and A Million Points of Light turn lived experience into counted evidence, so the people the number describes help produce it.

Data, Tech & Evidence Driven Approach

Our model uses open federal data, calculated from primary sources rather than secondary reports. Seventeen connectors pull raw data from NHIS and MEPS microdata, the Census Household Pulse Survey, BLS earnings series, Census population projections, NIH RePORTER, openFDA, ClinicalTrials.gov, and other sources; each computes its own values. We calculate variance using each survey’s design, including published strata and primary sampling units or replicate weights. Each pull logs a SHA-256 hash. Automated tests continually re-pull and recalculate data, breaking the build if a source number changes.

AI is used only for peer-reviewed literature. It searches databases, screens abstracts against parameter definitions, and proposes values with verbatim quotes and page locators. In the latest run, it screened 6,019 papers across 6,492 decisions, logged 6,339 distinct exclusion reasons, and found four usable records. Rejections are visible in the tool. AI cannot decide: each proposal stays in a review queue until a named analyst verifies and promotes it, logging their name, timestamp, and reason. A separate check re-fetches documents to confirm quotes verbatim (25 of 25 in use; none unreachable). Nineteen more checks test whether valid numbers are believable, catching double-counted samples and dollar figures without a price year.

We also include one proprietary dataset: a COVID health-impact assessment hosted on Healthful. We query it read-only through an allow-listed crosswalk and aggregate outputs above strict minimum cell thresholds.

The pipeline exports as a 140-entry FHIR R4B Bundle, validated in tests and at runtime. Each Observation links to its source via derivedFrom, letting anyone trace final math to the raw survey file or exact sentence in a paper.

Execution, Efficiency & Delivery


The most useful thing the tool shows is where the money is not. Of a $635.2 billion annual burden for long COVID, medical spending is 7.2%. Lost earnings are 31.3% and lost quality of life is 61.5%, so any analysis counting only healthcare costs misses nine-tenths of the problem and goes looking for savings in the smallest layer. The second is proportion. Mean NIH funding for the condition was $251 million a year across FY2020 to FY2026, read from RePORTER. The annual burden is about 2,500 times that. For a decision maker the answer is not one number but a dependency. The horizon and the discount rate are controls, so the same evidence can be read over one, ten or twenty years, discounted at OMB's 3% or not at all, and the published range of $3.09 trillion to $7.55 trillion names willingness to pay per QALY as the input that moves it most. That tells a budget holder what to argue about. Because every figure clicks through to the sentence it came from, disagreement gets specific instead of general. An independently published model estimates 865,000 QALYs lost a year; ours implies 891,593, which we found afterwards rather than by tuning. The export is a FHIR R4B Bundle, so another system can ingest the result and walk the arithmetic behind it. Where an input is missing the tool shows the layers it can compute and names what is absent, which is what makes the numbers it does show worth using.

Scalability, Reuse & Societal Impact:

Chronic illness carries a hidden price tag. By counting only hospital bills and prescriptions, policymakers miss over 90% of the cost to American families: lost wages, daily function, and quality of life. We built this platform to make that burden visible and verifiable.

Rather than create a one-disease calculator, we built shared public infrastructure for chronic illness research. Our live MongoDB backend and automated validation engine run over 1,300 continuous checks. Because the engine is disease-agnostic, new conditions require configuration, not custom engineering. We expanded from Long COVID to Lyme disease and ME/CFS without changing core code. Open YAML and standard FHIR formats let patient organizations and research teams use our work.

Honest science means naming blind spots. Formal diagnoses make our numbers reflect healthcare access, not everyone who is sick. Black Americans, for example, face Lyme exposure risks comparable to White Americans but are diagnosed far less often. We also rely too heavily on cross-sectional snapshots and limited quality-of-life studies. Next, we’ll add validated quality-of-life instruments to patient surveys, state-level CDC data for local health departments, and evaluations of real-world clinical programs that help patients today—not just the abstract value of a total cure.

Videos
Idea image
Idea image
Idea image
edited on Nov 1, 2026 by Manuel Duenas
Public (8)
You will need to login to post a comment
Manuel Duenas 2 weeks ago

The Community-Led COVID Impact Assessment - the link above was a link to a different assessment.

Reply 0

Manuel Duenas 6 days ago (edited)

Note that if we were to be fortunate to be recognized with a $1M award, our team would dedicate funding towards ensuring the map we created for A Million Points of Light ends up with at least 25,000 points of light and the Community-Led COVID Burden Index is completed longitudinally by a nationally representative sample of at least 10,000 individuals over two years.

We have removed the password requirement for the Pricing the Invisible, Valuing the Cure Decision Console:

https://pricing-the-invisible.joinhealthful.com/

As a reminder, if you have ever had a COVID infection, whether or not you had any symptoms or believe you have Long COVID, you are eligible to contribute to the Community-Led COVID Burden Index. 20 minutes can make the difference for how Long COVID is identified and burden is quantified in the US.

Light your light for invisible illness. Share what you wish others knew about living with your invisible illness through our map, A Million Points of Light. All that is required is a first name, a valid email address, and less than two minutes of your time. Together we can raise awareness, reduce stigma, and ensure that invisible illness is impossible to ignore.

If you would like a quick overview of what we have built, watch our 3-minute highlight reel here.

If an appropriate analyst would like to be added to our roster of vetted policy users, please email our team at contact@healthfuldata.com.

We are especially encouraged by the words of Long COVID advocate Gillian McCrea who is supportive of our roadmap to continue our work by extending our national model to the state level.

Reply 0

Uri Kartoun 1 week ago


Hi Manuel, the breakdown in your page, with medical spending at 7.2% of the Long COVID burden and most of the rest in lost earnings and quality of life, is a strong way to show why counting only healthcare costs misses most of the problem.

For this sprint I built a tool that turns published Lyme and Long COVID papers into runnable simulators. Simulator 25 covers a CDC claims study of extra medical spending after a COVID-19 diagnosis in privately insured people under 65 (Pike et al., Prev Chronic Dis 2023). Over six months, spending was $1,562 higher for adults and $1,011 higher for children than for matched people without the diagnosis, and costs were about 3.2 times higher for adults who had been hospitalized (chart attached). The authors describe it as a likely lower bound, partly because no Long COVID-specific diagnosis code existed yet, and it also excludes out-of-pocket spending and lost work. It’s one small piece of the medical layer your console models, and it reflects the same point about what claims can’t see.

I’d also be glad to hear the views of Bo Peng, Ian Liu and Hamad Al Ibrahim, whose entries deal with pricing care, Long COVID cost estimates and the link between burden and research funding.

My entry is here if you’d like to look: https://invisibleillness.crowdicity.com/post/896844

Thank you for sharing your work, and good luck with the community index.

Users tagged:

Avatar for Manuel Duenas Manuel Duenas
Avatar for Bo Peng Bo Peng
Avatar for Ian Liu Ian Liu
Avatar for Hamad Al Ibrahim Hamad Al Ibrahim

Reply 0

Sulamit Berkovits 1 week ago

Thank you for breaking this down, Uri. The 7.2 percent medical spending share really caught my attention, because it shows how much of the burden sits outside direct medical costs. I am curious how you think about that split when you compare population level traceability, where every number can be traced to a source, with patient level traceability, where the challenge is keeping conflicting records and unanswered questions visible. It feels like both matter, but they answer different questions. Thanks again for sharing your simulator entry so openly.

Reply 0

Uri Kartoun 1 week ago

Thank you, Sulamit. I think you've put it well: they answer different questions. Population-level traceability answers "what does the evidence say, and where does each number come from?" In my simulators every figure is tied to its paper and location, and the tool is deliberately built for published studies, not identified patient records. Patient-level traceability answers "what do we know about this person, and what is still unresolved?", which needs conflicting records and open questions kept visible, and that's the harder problem your approach addresses. I'd say the first gives a reference frame and the second keeps an individual's story from being flattened into it. The 7.2% share is Manuel's figure from his Pricing the Invisible entry, so I'd leave any detail on that to him.

Users tagged:

Avatar for Manuel Duenas Manuel Duenas

Reply 0

Sulamit Berkovits 6 days ago

Thank you, Uri, that distinction helps. Population traceability and patient traceability answer different questions, and trying to make one do the other job would flatten what matters.

Your simulators show where each figure comes from, tied to its paper and location, and they are built for published studies, not identified patient records. That gives a reference frame, a way to see what the evidence says at a population level.

Patient traceability is the harder problem, because it has to keep what is known about a person alongside what is conflicting and what is still unresolved, without smoothing it away.

I am curious where you think the two should meet in practice. Should a patient view start from the population reference and then show where that person diverges from it, or should the two stay separate, so the individual's story is never read as just an example of the frame? And on the 7.2 percent figure, I will leave the detail to Manuel, since it is his figure from his Pricing entry.

Reply 0

Uri Kartoun 6 days ago

My lean is to keep them separate by default. If a patient view starts from the population reference and shows where the person diverges, the average quietly becomes the definition of "normal", and the individual's story can be read as a deviation. I'd rather let the person's record stand on its own and offer the population figure as optional, clearly labeled context beside it, never as the baseline. That's a design judgment, not something I've tested, so it's a question for people building patient views, like you.

Reply 0

Sulamit Berkovits 6 days ago

That separation makes sense to me. A population figure is useful as context, but it should never become the baseline a single person's story is measured against. Do you think Manuel should label that context right where the figure appears, so a reader never mistakes it for a target?

Reply 0

Uri Kartoun 6 days ago

I would, yes: a label right beside the figure, such as "population reference, not a target", is the surest way to keep it from being read as one. Manuel knows his console best, so that's his call.

Reply 0

Sulamit Berkovits 6 days ago

A label right beside the figure feels like the safest default, because a caption can get separated from the number so easily. Should that label travel with the figure even when it is shared outside the console, so the context is never lost in the sharing?

Reply 0

Uri Kartoun 6 days ago

Yes, that's the same idea as the in-image caption: the label should be part of the figure, so it travels wherever the figure goes.

Reply 0

Sulamit Berkovits 6 days ago

Making the label part of the figure itself feels like the fix, because a caption or a note alongside it can always get lost in sharing. Should that label be built into the exported figure too, so even a screenshot or a download keeps the context with the number?

Reply 0

Uri Kartoun 3 days ago

Yes. If the label is drawn into the figure itself, it survives a screenshot and a download too. I’ve given my view, and how to label the console is the entry owner’s call, so I’ll leave it with them. Thank you for the thoughtful questions.

Reply 0

Manuel Duenas 6 days ago (edited)

Hi @Uri and @Sulamit!

Thank you both for the thoughtful discussion.

@Uri -Thank you for sharing the Pike et al. CDC claims study analysis. It’s a great concrete illustration of how claims data captures only a fraction of the total economic impact of invisible illnesses.

@Sulamit & @Uri - Our decision console is designed specifically for policy and federal budgeting priorities where analysts need transparent, auditable reference bounds. The console is not intended to be used by individual patients. Philosophically, individual patients know what their illness is costing. The model helps makes this visible to decisionmakers.

At the same time, we built the Community-Led COVID Impact Assessment and A Million Points of Light specifically so that individual lived experiences feed into the ecosystem on their own terms.

Note that if we were to be fortunate to be recognized with a $1M award, our team would dedicate funding towards ensuring the map lights up with at least 25,000 points and the CLCBI is completed longitudinally by a cohort of a nationally representative sample of at least 10,000 over two years.

The purpose is captured well by Long COVID Advocate Gillian McCrea in the thoughts she kindly shared with us.

Reply 0

Sulamit Berkovits 5 days ago

Thank you, Manuel, for spelling out the split so clearly. Pricing as a policy and budgeting tool, with the individual lived experience stuff handled separately, that makes the roles way clearer in my head. What I wonder is what happens at the edges, like when a policymaker or an advocate quotes a big national figure in a local conversation. Is there anything in how it is presented that helps folks keep the national and the personal layers apart, or is it really just up to whoever is quoting it? Thanks for keeping those layers separate, I think that matters a lot for trust on both sides.

Reply 0

Uri Kartoun 5 days ago

Thank you, Manuel. Your reading is fair: the Pike study counts direct medical costs from claims, so lost work and other burdens fall outside it. A console for budget analysts, with a separate way for patients to add their own experience, seems like a sensible split.

Reply 0

Generic user icon
This comment has been removed
Generic user icon
This comment has been removed
Sulamit Berkovits 1 week ago

Manuel, what stands out to me is that your model goes beyond estimating the economic burden of Long COVID—it makes the assumptions, source evidence, and sensitivity of those estimates inspectable. The finding that medical spending represents only 7.2% of the modeled burden is particularly important, because analyses focused on healthcare costs alone can miss much of the impact on earnings and quality of life.

This connects to a related challenge in my project, LymeSignal AI by Ace. While your platform makes population-level economic assumptions transparent, LymeSignal focuses on patient-level evidence: organizing fragmented medical records into a source-linked visit brief while preserving conflicting information, patient clarification, and unresolved uncertainty rather than letting AI silently decide which version is correct.

Both approaches reflect an important principle: AI should make evidence easier to evaluate, not hide uncertainty behind a confident-looking answer.

If you're interested, here's my project: https://invisibleillness.crowdicity.com/post/896768

I'd welcome your thoughts on how these approaches to evidence traceability might complement one another.

Users tagged:

Avatar for Hamad Al Ibrahim Hamad Al Ibrahim
Avatar for Ian Liu Ian Liu

Reply 0

Manuel Duenas 6 days ago

Hi Sulamit,

Thank you for reading through our entry and for the kind words regarding our policy console's source transparency.

Our focus with Pricing the Invisible remains on national macroeconomic modeling and federal policy application, so we are keeping our scope tightly bound to those objectives for this competition.

Note that if we were to be fortunate to be recognized with a $1M award, our team would dedicate funding towards ensuring the map we created for A Million Points of Light ends up with at least 25,000 points of light and the Community-Led COVID Burden Index is completed longitudinally by a nationally representative sample of at least 10,000 individuals over two years.

Further, our sister non-profit organization (Healthful Stories / HealthfulHeroes.com) will commit to cataloguing and highlighting a hundred diverse lived experience with Long COVID, ME/CFS, and Lyme Disease, similar to how they have spoken with Jo Bruce living with ME/CFS.

Ultimately, the national economic burden, Community-Led COVID Burden, Million Points of Light, and the lived experience of a hundred diverse Americans living with invisible illness will, in our view, ensure that invisible illnesses are impossible to ignore, with national momentum, concerted strategy, and investment commensurate with the magnitude of burden we are seeing.

In Health, for Each Other,

Manny & Team Healthful

Reply 0

Sulamit Berkovits 6 days ago

I was thinking about someone actually opening Pricing for the first time, not as a policymaker, but as a person with Long COVID trying to make sense of their own situation. They would see $635.2 billion for the country and a 7.2% medical share, and none of that tells them what their own bills, lost shifts, or bad weeks add up to. Their own split could look completely different.

What I would want to see in the next few months is a small group of real people with Long COVID opening the console as it is, before any personal view is added, and watching where the national figure gets read as a personal estimate. If someone quotes it to a clinician or an employer and gets dismissed because it is not about her, that is the deployment result, not a labeling problem, and it would tell you what a personal view actually has to contain.

Reply 0

Manuel Duenas 6 days ago (edited)

Hi Sulamit,

We need to be completely clear about the scope, purpose, and audience of Pricing the Invisible:

Pricing the Invisible is a national health policy decision console. It was engineered for federal budget analysts, health economists, CMS, CBO, and agency leaders who need defensible, auditable macro-valuation to allocate billion-dollar research budgets and justify policy priorities. It was never intended, designed, or framed as a personalized health accounting tool or individual clinical assessment for a single patient in a doctor's office. Again, each individual patient knows their costs and does not need a tool to tell them what they have been.

A national GDP estimate doesn't tell a small business owner their monthly revenue. These answer completely different structural problems.

We already addressed the individual human layer by building separate, dedicated infrastructure:

The Community-Led Long COVID Burden Index (CLCBI): Captures individual and caregiver-reported burden directly from lived experience. This is a tool that was developed with our team who has spent years with such research instruments focused on the HIV pandemic in resource-constrained settings. The CLCBI also reverts real value to individuals by assessing their risk factors for long term sequelae of Long COVID.

A Million Points of Light: Ensures individual patient stories and qualitative realities remain persistent and visible without being flattened into an aggregate mean.

Note that if we were to be fortunate to be recognized with a $1M award, our team would dedicate funding towards ensuring the map lights up with at least 25,000 points and the CLCBI is completed longitudinally by a cohort of a nationally-representative sample of at least 10,000 over two years.

We purposefully separated these layers in our architecture so that macro-economic evidence remains transparent and rigorous for decision-makers, while individual burdens retain their full, un-averaged context.

The national console is deployed for the policymakers who hold the budgets for research and cures, ensuring that those who do not see the burden finally understand it in terms they can use to inform action.

Gillian McCrea, Long COVID advocate shared her perspective on the need for a national economic model extensible to ME/CFS and Lyme Disease with us. I encourage you to listen to her words.

Reply 0

Sulamit Berkovits 5 days ago

Thank you for clarifying that, Manuel, it really helps. I understand better now that Pricing is meant as a national policy console, not as a personal estimate, and that the individual layer lives separately so the macro evidence stays clean. What I am curious about now is that separate layer: when a community group or a small clinic first opens the community led index, what do they actually see, and how would they know whether an estimate is ready to use or still exploratory? That boundary feels important, because it is where a national number could still quietly turn into a personal expectation. Thank you for taking the time to explain the split so clearly.

Reply 0

Manuel Duenas 5 days ago

Hi Sulamit,

To clarify, the CLCBI is built specifically as a patient-reported cost and burden assessment integrated with a Long COVID sequelae risk model, capturing individual impact and downstream risk profiles directly.

Please review the instrument and what it delivers prior to further comments.

Kind regards,

Manny & Team Healthful

Reply 0

Sulamit Berkovits 5 days ago

Thank you, Manuel, I appreciate you pointing me to the actual instrument instead of letting me talk around it. A patient reported cost and burden assessment sounds like it fills a gap claims data just cannot, because claims show what was billed, not what it cost someone in time and daily life. Before I say more about it, I will take a proper look at what it delivers. What I am curious about is what it captures that surprised you most, like is there a kind of burden that shows up in patient reports and almost never shows up in the claims at all?

Reply 0

Latina Wright-Booker 4 days ago

Hi Manuel, I appreciate how Healthful distinguishes medical spending from the broader economic and quality-of-life impacts of Long COVID. That distinction resonates with our work on Visible Burden AI (VBAI), where we explore several dimensions of patient-reported burden beyond direct healthcare costs.

Your Community-Led Long COVID Burden Index particularly caught my attention. How do you envision using individual reports of employment disruption, caregiving demands, or healthcare navigation difficulties to strengthen the interpretation of population-level economic estimates?

Congratulations to your team on the work you've developed!

Reply 0

Manuel Duenas 4 days ago

Hi Latina,

First, I love your thoughts - refreshing that they are sincere and engaged!

At scale, we are targeting thousands (eventually 10k+ with funding, hopefully through the competition) longitudinal respondents, both with and without Long COVID, to allow for the sizes needed across phenotypes of interest to support our macroeconomic modeling. We are focused on three structured pathways:

EQ-5D-5L Crosswalks & QALYs: Mapping individual functional and symptom responses to EQ-5D-5L health states to isolate attributable health decrements and compute a second robust source for QALYs losses. Our review of the available literature suggests that there are very few US-focused utility estimates that could be applied to measure the loss in quality of life for Americans with Long COVID.

Empirical Cost Parameters: We will aggregate (and our platform anonymizes by design) individual reports of work disruption, lost hours, informal caregiving, and out-of-pocket spending into parameterized economic inputs scaled against BLS wage and MEPS data.

Severity & Covariate Stratification: We will group respondents into validated severity tiers and control for acute baseline covariates so that population-level estimates accurately reflect the real distribution of severity rather than an unweighted average. We will never capture the very severe population of individuals living with Long COVID until we have funding to send individuals out to homes to support those who might have difficulty in completing an online instrument. We are exploring ways to have AI read and voice response control for those who might be able to hear and speak but not look at a screen for 20 minutes. In the end, funding will be needed to make the CLCBI truly accessible in the broadest sense.

We are deeply grateful to the S4Me.info forum for their valuable feedback on our CLCBI. We received 20+ separate sets of feedback reviews and will continue to collect further feedback to refine our approach for accessibility.

Thanks again for stopping by our thread, and best of luck with Visible Burden AI!

Our team would welcome the opportunity to collaborate with good people like you post-sprint. Let's keep going for all individuals and their families who desperately need our energy and talents.


With deepest appreciation, Latina,


Manny & Team Healthful

Reply 0

Latina Wright-Booker 4 days ago

Thank you for such a thoughtful and comprehensive response! I especially appreciate your team's commitment to capturing the economic realities of Long COVID beyond traditional healthcare expenditures, including lost work hours, informal caregiving, and quality-of-life impacts.

Your acknowledgment of the challenges faced by individuals with severe Long COVID particularly resonated with me. Accessibility is such an important consideration, especially when those experiencing the greatest burden may also face the greatest barriers to participating in research.

I'm curious about one aspect of your proposed approach: As Healthful expands its longitudinal data collection, how do you envision accounting for participants whose illness severity or functional limitations make consistent reporting difficult? It seems particularly important that gaps in participation not unintentionally underrepresent those experiencing the greatest burden.

I also sincerely appreciate your generous invitation to collaborate after the Sprint. I would welcome the opportunity to stay connected and explore where our shared commitment to making invisible burdens visible might lead.

Wishing you, Manny, and Team Healthful continued success!

Reply 0

Manuel Duenas 4 days ago

We have removed the password requirement for the Pricing the Invisible, Valuing the Cure Decision Console:

https://pricing-the-invisible.joinhealthful.com/

As a reminder, if you have ever had a COVID infection, whether or not you had any symptoms or believe you have Long COVID, you are eligible to contribute to the Community-Led COVID Burden Index. 20 minutes can make the difference for how Long COVID is identified and burden is quantified in the US.

Light your light for invisible illness. Share what you wish others knew about living with your invisible illness through our map, A Million Points of Light. All that is required is a first name, a valid email address, and less than two minutes of your time. Together we can raise awareness, reduce stigma, and ensure that invisible illness is impossible to ignore.

If you would like a quick overview of what we have built, watch our 3-minute highlight reel here.

If an appropriate analyst would like to be added to our roster of vetted policy users, please email our team at contact@healthfuldata.com.

Reply 0

Pranisha Karki 2 days ago

Hi Manuel, framing Long COVID as what a cure would be worth is a strong way to talk to policymakers. On the live console, the Population addressed slider says there are no sources and asks the user to set it. Do you plan to offer an evidence-based default for reach, for example from Household Pulse, so people can compare results against the $5.32 trillion base case more easily?

Reply 0

Manuel Duenas 1 day ago

Thank you, Pranisha! Incorporating historical adoption and reach curves is on our roadmap. One comparative reference we are currently exploring is looking at past therapeutic revolutions such as the rollout trajectory from first-generation Hepatitis C therapies, e.g., (peg-interferon) to the sofosbuvir generation of cures to model how a breakthrough Long COVID treatment reach might scale in practice.

We appreciate you taking the console for a spin and testing out the parameters. Thanks for the great feedback!

Best,
Manny & Team Healthful

Reply 0

Pranisha Karki 1 day ago

@Manuel Duenas Thanks, Manny. Using the move from peg-interferon to sofosbuvir as a guide is a smart choice, since Hepatitis C has years of real uptake data behind it. Will people be able to switch between a fast and a slow rollout in the console to see how much the $5.32 trillion base case changes? Since your console is live, a direct link-up could work well. ImpactAtlas shows what Long COVID costs one household, so a "What would a cure be worth?" button on our Long COVID card could open your console and connect the personal cost with the national picture. We'd be glad to set that up if you're interested.

Reply 0

Manuel Duenas 1 day ago (edited)

Hi Pranisha,

Thanks for the note. Our Community-Led Long COVID Burden Index (CLCBI) serves as our consumer-facing application. The CLCBI is purpose-built to translate individual patient-reported impacts into EQ-5D-5L utility decrements, QALY losses, and parameterized economic inputs for federal research valuation. We can talk post-sprint about a path to collaboration.

Thank you again, Pranisha!


Kind regards,

Manny & Team Healthful

Reply 0

Steven Forsythe 17 hours ago

I really like this idea and think that it could be applied to a much wider set of health issues. For example, what would be the economic benefits of finding a cure for ebola? What about a vaccine for HIV? It is also interesting how to value the altruistic benefits for disease prevention or cure. For example, men have significant value in terms of breast cancer, despite the fact that they almost never get breast cancer.

Reply 0

Manuel Duenas 16 hours ago (edited)

Hi Dr. Forsythe,

Our team is humbled that you have taken the time given your role as President of the International AIDS and Economics Network. We appreciate your thoughtful, expert perspective and for connecting our macroeconomic valuation framework to broader global health priorities.

You hit on a crucial extension of the model: because the underlying architecture of our Pricing the Invisible console is extensible wherein it relies on modular population scopes, utility estimates, and cost parameters rather than hardcoded logic for any one illness, expanding it to model prevention (e.g., an HIV or other long term vaccine) or other invisible illnesses is entirely feasible by swapping in the relevant epidemiological and cost inputs.

Your point about valuing altruistic benefits (such as broader population-level risk reduction) is also a vital frontier in HEOR. Capturing those externalities is essential for moving beyond pure disease-attributable utility to measure the full societal value of systemic eradication. Should we win a substantive prize, our aim will also be to capture the full downstream cost of COVID-19. Research has linked infection to higher risk of heart disease, diabetes, autoimmune disease, and other illness. More long-term effects will appear as research continues. The model will add new evidence when it becomes available.

Thank you so much for sharing your insights from decades of work in health economics and US and global HIV/AIDS modeling!

Kind regards,

Manny & Team Healthful

Reply 0