Insurance and Risk Protection

Why this matters in health economics

Insurance is the mechanism that turns the unbearable lottery of illness into a manageable, shared cost. Without it, a serious diagnosis is a financial catastrophe as well as a medical one, and families ration their own care by what they can find in their pockets on the day. With it, the cost of the sick is spread across the healthy and across a person's own healthier years, so that using care does not depend on being able to afford it at the worst possible moment. Whether the vehicle is a tax-funded entitlement, a social-insurance fund, or a private policy, every system in the world is in the business of insuring health risk — and the quality of that insurance's design decides who is protected and how well the money is spent.

For a director, the design of coverage is where abstract financing choices become concrete and consequential. The benefit package decides what is covered and what a patient pays for out of their own pocket. The cost-sharing structure decides who is deterred from care and whether the deterrence falls on waste or on need. The managed-care rules decide whether the insurer steers care towards value or simply towards denial. These are not actuarial details; they are the levers that determine whether coverage delivers on its promise or fails the people it was meant to protect, and they are chosen by someone — deliberately or by default.

This chapter takes the failures of insurance as given. Insurance markets suffer from adverse selection and moral hazard — the reasons voluntary pools unravel and the reasons insured people and their providers use more care than they otherwise would — and those failures are diagnosed in Chapter 1.3 — Market Failure; this chapter does not re-derive them. Nor does it re-teach the financing typologies of Beveridge and Bismarck, which belong to Chapter 3.1 — Health Systems, or argue the case for universal coverage as an equity goal, which belongs to Chapter 3.4 — Equity and Chapter 4.2 — Global Health and Trade. Its single question is narrower and more practical: given that insurance is both necessary and imperfect, how do you design it well?

Core concepts

Health insurance is a contract that pools the financial risk of illness across a group so that individuals pay a predictable, regular amount — a premium or contribution — in exchange for having some or all of their care costs met when they fall ill. The engine underneath it is the risk pool: a group large and mixed enough that the healthy majority in any period cover the sick minority, and the actuarially predictable total cost of the group can be financed by predictable contributions. A pool that is too small, or skewed towards the sick, cannot perform this trick — which is why pool design sits underneath every other decision in this chapter.

The benefit package is the definition of what the insurance actually covers: which services, drugs, and providers are included, under what conditions, and with what limits. It is the single most consequential design document an insurer or scheme produces, because it draws the line between what the pool pays for and what the individual bears alone. A generous package protects more but costs more and must be financed; a narrow package is cheaper but leaves people exposed to the very costs insurance exists to cover. Defining it well is an exercise in explicit priority-setting, drawing on the evaluation methods of Chapter 2.1 — Economic Evaluation and the rationing disciplines of Chapter 3.3 — Rationing.

Cost-sharing is the family of mechanisms by which an insured person pays part of the cost of the care they use, and it comes in distinct forms that behave differently. A copayment is a fixed amount per service — a set fee per prescription or per visit. A deductible is an amount the insured must pay in full each year before the insurer pays anything. Coinsurance is a percentage of each bill that the insured pays, so their exposure rises with the cost of care. Because these can accumulate to ruinous sums, a well-designed scheme caps them with an out-of-pocket maximum — a ceiling beyond which the insurer pays everything — so that the sickest patients, who use the most care, are not the ones most exposed. Without such a cap, cost-sharing quietly converts insurance back into the out-of-pocket payment it was meant to replace, precisely for the people who need protection most.

Managed care is the set of techniques by which an insurer actively influences the care its members receive, rather than passively paying whatever bills arrive. It emerged most visibly in the United States, in the form of the health maintenance organization (HMO), which combines insurance and provision and channels members to a defined network of providers, and the preferred provider organization (PPO), which offers wider choice at higher cost. Its tools include provider networks, gatekeeping through primary care, utilization management — reviewing whether requested care is appropriate — and prior authorization, requiring approval before certain expensive services are delivered. Managed care is a direct design response to moral hazard and supplier-induced demand, and its central tension is that the same tools that curb low-value care can, pushed too far, become obstacles to care that patients genuinely need.

How premiums are set is itself a design choice with large distributional consequences. Community rating charges everyone in a pool the same premium regardless of their individual health risk, so the healthy subsidize the sick and coverage stays affordable for those most likely to need it. Experience rating sets premiums according to the individual's or group's own claims history and risk profile, so the sick pay more — which is actuarially precise but undermines the protective purpose of insurance for exactly the people it should shield. Most systems that pursue universal coverage constrain or prohibit experience rating and rely on community rating, then solve the problem it creates — that insurers now have an incentive to avoid sick applicants — with the next tool.

That tool is risk equalization, also called risk adjustment: a mechanism that transfers funds between insurers or funds so that those enrolling sicker, costlier members are compensated, and those enrolling healthy members contribute. Used as a design tool — the sense this chapter owns — risk equalization is what makes community rating and competition coexist. It removes the profit in selecting good risks, so insurers compete on efficiency and service quality rather than on avoiding the sick. The competing-sickness-fund systems of the Netherlands and Germany depend on it, and its sophistication — adjusting not just for age and sex but for diagnosed chronic conditions and prior costs — largely determines whether the competition it enables is healthy or corrosive. A related refinement, value-based insurance design, tunes cost-sharing to the clinical value of each service rather than its price, lowering or removing charges for high-value care such as maintenance drugs for chronic conditions while retaining them for low-value care.

Best practices

1. Design the benefit package explicitly, as priority-setting, not as an actuarial residual. The list of what is covered is a series of value judgements about which care the pool will pay for, and it deserves to be made openly using economic evaluation and a transparent process, not left to accumulate through historical accident and lobbying. Decide what the package includes on grounds of clinical and cost-effectiveness, state the criteria, and revisit them as evidence changes (see Chapter 2.1 — Economic Evaluation and Chapter 3.3 — Rationing). A package assembled without explicit criteria will cover the well-represented and omit the voiceless, and no one will be able to say why.

2. Cap out-of-pocket exposure before you set any other cost-sharing parameter. The first purpose of insurance is protection against catastrophic cost, so the out-of-pocket maximum is the design element that guarantees the product still insures. Set the annual ceiling at a level a household in your population can actually absorb, and only then decide how copayments, deductibles, and coinsurance fill the space beneath it. Cost-sharing without a firm cap re-exposes the sickest members — those who breach every threshold — to exactly the ruin the scheme was built to prevent.

3. Aim cost-sharing at low-value care and shield the poor and the chronically ill. The evidence from the RAND Health Insurance Experiment (see Chapter 1.3 — Market Failure) is that flat cost-sharing reduces effective and ineffective care alike, and that the poorest and sickest cut back most and are harmed. So a flat deductible is a blunt instrument that deters the wrong people; value-based insurance design is the sharper alternative, lowering or removing charges for care known to be high-value — maintenance medication, antenatal visits, chronic-disease management — while retaining them for genuinely discretionary use. Exempt low-income and chronic-disease groups explicitly rather than trusting a single threshold to sort them.

4. Choose community rating and make it viable with risk equalization, not wishful thinking. If you want coverage to stay affordable for the people most likely to need it, prohibit or tightly constrain experience rating so the sick are not priced out. But community rating alone gives any competing insurer a powerful incentive to avoid sick applicants, so pair it with risk equalization that compensates insurers for the risk they carry. Adopting one without the other predictably fails: community rating without risk adjustment invites covert selection, and risk adjustment without community rating solves a problem you have chosen not to create.

5. Invest in the sophistication of your risk-adjustment formula, because a crude one is worse than honest. Risk equalization that adjusts only for age and sex leaves large, predictable variation in cost unexplained, which means insurers can still profit by selecting the healthy within each age band. A formula that incorporates diagnosed chronic conditions, prior utilization, and pharmacy history closes most of that gap and shifts competition towards efficiency. The Netherlands and Germany refined their formulas over years precisely because early, crude versions left too much room for selection; treat the formula as a living instrument, audited and updated, not a fixed constant.

6. Use managed-care tools in proportion to the value they protect, and measure their friction. Provider networks, gatekeeping, utilization management, and prior authorization are legitimate responses to moral hazard and induced demand, but each imposes a cost — administrative burden, delay, and the risk of denying appropriate care. Reserve the heavier tools such as prior authorization for services where the evidence of overuse or high cost justifies the friction, and track the downstream consequences: denied-then-overturned rates, delays to necessary treatment, and the administrative cost borne by providers. A prior-authorization regime that blocks as much good care as bad is not managing care; it is rationing by hassle.

7. Align provider payment with the coverage you are trying to deliver. Insurance design and provider payment are two halves of the same incentive system, and a mismatch defeats both. Paying network providers fee-for-service while asking them to restrain volume sets your payment against your coverage goals; capitation or an accountable care organization arrangement, which gives a provider group a budget and a stake in outcomes, aligns them but introduces its own under-provision risk. Decide the payment mechanism (developed in Chapter 3.1 — Health Systems) as part of the coverage design, not separately, and counterbalance whichever distortion you choose.

8. Keep the pool broad and stable, because every design choice above assumes it. Cost-sharing, community rating, and risk equalization all presuppose a pool large and mixed enough to spread risk; a pool that fragments or drifts sick makes every other lever weaker. Broad or automatic enrolment, a single national pool or well-equalized competing pools, and limits on cream-skimming products keep the pool healthy. Where risk is concentrated — a small scheme facing a few very high-cost members — reinsurance, by which the scheme insures itself against extreme individual claims, stabilizes finances without narrowing coverage.

9. Write the rules in language members can act on, and make appeals real. Coverage that a member cannot understand is coverage they cannot use; opaque benefit schedules, surprise exclusions, and impenetrable prior-authorization processes convert nominal protection into practical exposure. State clearly what is covered, what the patient will pay, and what to do when a claim is denied, and resource an appeals process that genuinely reverses wrong decisions. The underwriting and actuarial machinery behind a policy is the insurer's business; its consequences at the point of care are the member's, and design must serve the member.

10. Ground the whole design in actuarial reality and revisit it on a cycle. Every choice — package breadth, cost-sharing levels, the out-of-pocket cap, the risk-adjustment formula — has a price that actuarial science can estimate, and a design that ignores the numbers will either bankrupt the pool or under-protect its members. Model the cost of the package against the contribution base before committing, stress-test it against an ageing or sickening population, and set a standing review that adjusts parameters as costs and evidence move. Insurance design is never finished; it is a system to be steered.

Questions to discuss with your team

1. When one of our members faces a long, expensive illness, exactly how much could they pay out of their own pocket in a year — and is that a number a household here could survive? This question tests whether your coverage actually insures, because the promise of insurance is broken precisely for the people who use the most care. Trace a realistic high-cost patient — a cancer diagnosis, a premature birth, a chronic condition needing lifelong drugs — through every copayment, deductible, and coinsurance charge your scheme applies, and find the total before any out-of-pocket cap bites. The tension is that a lower cap costs the pool more and must be financed by higher contributions from everyone, so protecting the few has a price paid by the many. An honest answer produces a concrete worst-case figure, compares it against what households in your population can actually absorb, and states plainly whether the gap is acceptable and who pays to close it. If no one on the team can produce that figure, that is itself the finding.

2. Where in our cost-sharing are we deterring care we want people to use? Cost-sharing is meant to dampen low-value care, but flat charges dampen high-value care just as effectively, and the people who cut back most are the poor and the chronically ill. Look at where your copayments and deductibles fall on services with strong evidence of benefit — maintenance medication, antenatal care, follow-up after a serious event — and ask whether the charge is saving money now at the cost of avoidable illness later. The tension is real: removing charges on high-value care raises near-term spending and may increase use, and the saving from avoided complications is deferred and uncertain. An honest answer identifies two or three specific services where value-based insurance design would lower or remove charges, accepts that this costs money up front, and commits to measuring whether the downstream harm actually falls.

3. If a competing insurer in our system wanted to profit by attracting the healthy and avoiding the sick, how easily could they do it under our current rules? This question surfaces whether community rating and risk equalization are doing their job or leaving the door open to selection. Work through the concrete moves a profit-seeking insurer could make: designing a benefit package that appeals to the young and well, siting services away from poor neighbourhoods, marketing selectively, or exploiting a risk-adjustment formula too crude to capture the real cost of chronic illness. The tension is that the tools to close these gaps — a richer risk-adjustment formula, standardized packages, enrolment rules — add administrative cost and constrain the competition they are meant to make fair. An honest answer names the specific selection opportunities your rules currently leave open, estimates how much money is on the table for exploiting them, and decides which gaps are worth the cost of closing.

4. Can we say, on the record, why each major service is in or out of our benefit package — and would that reasoning survive being published? This question tests whether your coverage is the product of explicit priority-setting or of historical accident and lobbying. Take three or four consequential inclusions and exclusions — a high-cost cancer drug, a fertility treatment, a class of mental-health support, a cosmetic borderline case — and try to state the criteria that put each where it is. The tension is that explicit criteria expose the scheme to challenge from every group that loses out, whereas a package that simply accreted attracts no such scrutiny precisely because no one can point to the decision. An honest answer admits which parts of the package no one can currently justify, commits to grounding inclusion and exclusion in economic evaluation and a transparent process (see Chapter 2.1 — Economic Evaluation and Chapter 3.3 — Rationing), and accepts that defensible decisions are contestable ones. If the reasoning would embarrass you in public, that is the finding.

5. For each managed-care control we operate, do we know how much good care it blocks alongside the bad? Prior authorization, utilization review, gatekeeping, and narrow networks are legitimate answers to moral hazard, but each imposes delay, administrative burden, and the risk of denying appropriate care. Pick your heaviest control and look for the numbers that reveal its true effect: the denied-then-overturned rate, the time patients wait for a decision, the provider hours consumed, and any evidence of necessary treatment forgone. The tension is that softening a control to reduce this friction predictably lets some low-value and wasteful care back through, so the saving and the harm move together. An honest answer distinguishes the controls that are aimed at evidenced overuse from those that simply throttle spending indiscriminately, and it retires or retargets any control that blocks as much good care as bad. A control whose friction you have never measured is rationing by hassle in all but name.

6. How exposed is our pool to fragmenting or drifting sick, and what happens to the scheme's finances if it does? Every other design choice in this chapter — cost-sharing, community rating, risk equalization — assumes a pool broad and mixed enough to spread risk, so the pool's stability is the foundation the rest stands on. Examine who can leave, who can be skimmed off by a competitor's leaner product, and what a run of a few catastrophic claims would do to a scheme your size. The tension is that the measures that keep a pool healthy — broad or automatic enrolment, limits on cream-skimming products, a single or well-equalized pool — constrain choice and competition, and reinsurance against extreme claims costs a premium that eats into funds for care. An honest answer names the concrete threats to your pool's breadth, models the finances under an ageing or sickening membership, and states whether reinsurance and enrolment rules are in place before the stress arrives rather than after. A pool that only looks stable because it has not yet been tested is not stable.

In practice: a health economics example

The fictional middle-income nation of Solraya is establishing a national health insurance scheme to replace a patchwork in which formal-sector workers have employer cover and everyone else pays cash at the clinic door. The scheme will be funded by a mix of payroll contributions and general taxation, with a single national pool, and it will contract both public and private providers. The design team must make three decisions that will define the scheme for a generation: what the benefit package covers, how cost-sharing is structured, and how providers are paid and steered. The finance ministry wants the scheme affordable; the health ministry wants it to actually protect people; the two goals meet in the design.

The team begins with the benefit package and treats it as explicit priority-setting rather than a wish list. Working with the country's nascent health-technology-assessment capacity, they define a package built around cost-effective essential services — primary care, maternal and child health, management of the major chronic diseases, and a defined set of hospital services — and they publish the criteria by which services are included or excluded, so the decisions can be defended and revised. They resist pressure to promise everything, knowing that an unfunded package is a false promise that will be rationed covertly through queues and stockouts. What the package leaves out is stated openly, with a route for adding services as the contribution base grows.

Cost-sharing is where the team spends the most care, because Solraya's population includes many households for whom a modest charge is a real barrier. They reject a flat deductible: modelling on the logic of the RAND evidence tells them it would deter the poor from effective care. Instead they set low fixed copayments for discretionary outpatient visits, exempt the poorest households and people with chronic conditions entirely, and remove charges altogether for high-value services such as antenatal care and maintenance medication for diabetes and hypertension — a value-based insurance design in miniature. Crucially, they cap total annual out-of-pocket payments at a fraction of typical household income, so that no insured family is bankrupted by a catastrophic illness. The finance ministry balks at the cost of the cap; the team shows that without it, the scheme fails its core purpose for the very people it most needs to reach, and that reinsurance against the small number of extreme claims keeps the pool's finances stable at a manageable price.

Because Solraya will contract private providers alongside public ones, the team confronts selection and payment together. Premiums are community-rated — everyone contributes according to means, not health risk — and to stop contracted insurers or provider groups from quietly avoiding the sick, they build a risk-equalization transfer that adjusts for age, sex, and a starter set of diagnosed chronic conditions, with an explicit plan to enrich the formula as data accumulate. Providers are paid a blend: capitation for a registered primary-care population to reward continuity and prevention, case-based payment for hospital admissions to reward efficiency, and a modest quality component. The team is candid with both ministries about the residual tensions. Capitation risks under-referral, so they will monitor it; the risk-adjustment formula is crude at launch and will leave some selection incentive until it matures; and the package will need disciplined, transparent expansion as revenue grows rather than ad hoc additions under political pressure. The design is not perfect, and the team says so — its virtue is that every imperfection is named, measured, and owned, rather than discovered later as a scandal.

Four sector lenses

Startup

A health-insurance start-up or "insurtech" venture usually enters with a narrow, well-defined product — a supplemental policy, a micro-insurance scheme for informal workers, or a digital front end that manages claims and prior authorization more cleanly than incumbents. Its opportunity is design agility: it can build value-based cost-sharing or a slick appeals process from scratch, unencumbered by legacy rules. Its danger is the thin pool: a small or self-selected membership is acutely vulnerable to adverse selection and to a handful of catastrophic claims, so reinsurance and a credible enrolment strategy are existential, not optional. Founders should be honest about which failure their model quietly relies on — a product that looks cheap because it has attracted only the healthy is not an innovation, it is a pool waiting to unravel.

Small business

A small but established insurance intermediary — a broker placing group cover for local employers, a mutual or community fund with a stable membership, or a clinic that runs a modest prepayment plan for its patients — lives with the design choices rather than inventing them, and its edge is knowledge of the members it already has. It cannot spread risk across millions, so its stability depends on reinsurance against extreme claims and on refusing products that would let its healthiest members drift to a leaner competitor. Its practical work is disciplined stewardship of a coverage it mostly did not design: keeping the benefit schedule clear, exempting the members who genuinely cannot bear a charge, and making appeals real for a population it can name. Unlike the start-up, it is not chasing growth or a novel model but protecting a going concern, so its risks are erosion and complacency rather than unravelling — the pool thinning quietly as the young and well find a cheaper home elsewhere.

Enterprise

A large insurer or integrated payer-provider lives inside these design choices as its core operating reality and its competitive battleground. It runs real risk pools, sets benefit packages and cost-sharing at scale, and operates the managed-care machinery — networks, utilization management, prior authorization — whose calibration determines both its costs and its reputation. At this scale the organization is large enough that its design choices shape provider and member behaviour across a whole market, and regulators watch whether it competes on efficiency and quality or quietly on risk selection. The strategic question is whether its managed-care tools are protecting value or merely denying care to defend margin, because members, providers, and regulators increasingly tell the difference and the cost of getting it wrong is trust.

Government

Government designs the coverage architecture that no single participant can set: whether the pool is national or fragmented, whether rating is community-based, whether risk equalization exists and how good it is, and what the mandated minimum benefit package must contain. Its comparative advantage is reach — it can compel broad enrolment, prohibit experience rating, standardize packages to limit selection, and cap out-of-pocket exposure across an entire population — and its discipline must match that reach, because every rule it sets has a characteristic failure that only stewardship catches. It also carries the long game: keeping the benefit package affordable against structural cost growth, maintaining the risk-adjustment formula as an honest instrument, and protecting the pool's breadth against products designed to fragment it. The wider goals of universal coverage and financial protection as matters of fairness belong to Chapter 3.4 — Equity and Chapter 4.2 — Global Health and Trade; the point here is that the machinery to deliver them is a design that government owns.

Common failure modes

• Cost-sharing without an out-of-pocket cap. Copayments, deductibles, and coinsurance that accumulate without a ceiling re-expose the sickest members to catastrophic cost. Fix: set the out-of-pocket maximum first, at a level households can absorb, and design everything else beneath it.

• Flat cost-sharing that deters need as much as waste. A single deductible or uniform copayment dampens high-value and low-value care alike and hits the poor and chronically ill hardest. Fix: use value-based insurance design, exempt low-income and chronic-disease groups, and aim charges at genuinely discretionary use.

• Community rating without risk equalization. Requiring equal premiums while letting insurers compete gives them a strong incentive to avoid the sick. Fix: build a risk-adjustment transfer, and keep enriching its formula so selection stays unprofitable.

• A benefit package assembled by accident. Coverage that accretes through history and lobbying rather than explicit criteria covers the well-represented and omits the voiceless. Fix: define and publish inclusion criteria grounded in economic evaluation, and revisit them on a cycle.

• Managed care that rations by hassle. Prior authorization and utilization review calibrated to block care generally, rather than low-value care specifically, deny good treatment and burden providers. Fix: reserve heavy tools for services with real evidence of overuse, and measure denied-then-overturned rates and delays.

• A pool too small or too skewed to work. Fragmented or self-selected pools cannot spread risk and are destabilized by a few large claims. Fix: broaden or automate enrolment, pool nationally or equalize across pools, and use reinsurance against extreme claims.

• Insurance and payment designed in separate rooms. Paying providers to do more while asking coverage to restrain use sets the two halves of the incentive system against each other. Fix: design provider payment and coverage together, and counterbalance the distortion each introduces.

Maturity model

Checklist

• State the criteria by which your benefit package includes and excludes services, and confirm they are grounded in economic evaluation and published.
• Set the out-of-pocket maximum first, at a level households in your population can absorb, and design all other cost-sharing beneath it.
• Confirm cost-sharing targets low-value care, exempts the poor and the chronically ill, and removes charges on defined high-value services.
• Confirm premiums are community-rated (or state why experience rating is used) and that risk equalization makes community rating viable.
• Check that your risk-adjustment formula captures chronic conditions, not just age and sex, and that it is audited and updated.
• For each managed-care tool in use, confirm it is aimed at evidenced overuse and measure its friction and wrongful-denial rate.
• Confirm provider payment is designed together with coverage, and name the distortion each introduces and how it is counterbalanced.
• Confirm the pool is broad and stable, with reinsurance against extreme individual claims.
• Confirm members can understand what is covered, what they will pay, and how to appeal a denial, and that appeals genuinely reverse wrong decisions.
• Model the whole design against the contribution base, stress-test it against an ageing or sickening population, and set a standing review.

Key sources

• World Health Organization — health financing publications on pooling, prepayment, strategic purchasing, and benefit-package design for universal health coverage.
• OECD — Health at a Glance — comparative indicators on coverage, cost-sharing, and out-of-pocket burden across countries.
• RAND Health Insurance Experiment — the landmark randomised evidence on how cost-sharing affects use and health, foundational to cost-sharing design.
• ISPOR — good-practice guidance on value assessment relevant to benefit-package and value-based insurance design.
• Ellis, van de Ven & colleagues — the risk-adjustment and risk-equalisation literature underpinning competing-insurer systems in the Netherlands and Germany.
• Fendrick & Chernew — the value-based insurance design framework.

References

1. Health insurance — Wikipedia — https://en.wikipedia.org/wiki/Health_insurance
2. Risk pool — Wikipedia — https://en.wikipedia.org/wiki/Risk_pool
3. Copayment — Wikipedia — https://en.wikipedia.org/wiki/Copayment
4. Deductible — Wikipedia — https://en.wikipedia.org/wiki/Deductible
5. Coinsurance — Wikipedia — https://en.wikipedia.org/wiki/Coinsurance
6. Out-of-pocket expense — Wikipedia — https://en.wikipedia.org/wiki/Out-of-pocket_expense
7. Managed care — Wikipedia — https://en.wikipedia.org/wiki/Managed_care
8. Health maintenance organization — Wikipedia — https://en.wikipedia.org/wiki/Health_maintenance_organization
9. Preferred provider organization — Wikipedia — https://en.wikipedia.org/wiki/Preferred_provider_organization
10. Utilization management — Wikipedia — https://en.wikipedia.org/wiki/Utilization_management
11. Prior authorization — Wikipedia — https://en.wikipedia.org/wiki/Prior_authorization
12. Community rating — Wikipedia — https://en.wikipedia.org/wiki/Community_rating
13. Experience rating — Wikipedia — https://en.wikipedia.org/wiki/Experience_rating
14. Risk equalization — Wikipedia — https://en.wikipedia.org/wiki/Risk_equalization
15. Value-based insurance design — Wikipedia — https://en.wikipedia.org/wiki/Value-based_insurance_design
16. Accountable care organization — Wikipedia — https://en.wikipedia.org/wiki/Accountable_care_organization
17. Reinsurance — Wikipedia — https://en.wikipedia.org/wiki/Reinsurance
18. Underwriting — Wikipedia — https://en.wikipedia.org/wiki/Underwriting
19. Actuarial science — Wikipedia — https://en.wikipedia.org/wiki/Actuarial_science
20. World Health Organization — Health financing for universal coverage — https://www.who.int/health-topics/health-financing
21. OECD — Health at a Glance — https://www.oecd.org/health/health-at-a-glance/