RC HCMG 3520 Notes PDF

Lecture 1 (082724) - An Introduction (Part 1) - Correlation ≠causation - **What is unique about healthcare** = **combination** and large number of distinctive features - 12 future trends and challenges in healthcare - Rising healthcare cost - Driven by technology th...

Lecture 1 (082724) - An Introduction (Part 1) - Correlation ≠causation - **What is unique about healthcare** = **combination** and large number of distinctive features - 12 future trends and challenges in healthcare - Rising healthcare cost - Driven by technology though it is beneficial - Aging population - Sus but other aging countries don't have this problem (Japan) - Increasing competition - Lots of M&A and PE - Volume-outcome challenges - Surgeons who do more surgeries have better outcomes - May be learning curve - May be that better surgeons command more demand (selective referrals and resources go towards these surgeons) - Downstream effect: healthcare inequity that are geological - Evolving regulatory landscape - Antitrust - Insurance regulations - Privacy law - Workforce shortages - Burnout and turnover rate - Population health management - Preventative care - Chronic disease management - Value based reimbursement models - Health equity and DEI - Discrimination - Health literarcy - Patient experience and satisfaction - Patient reported outcomes and shared decision making - Digital technological advancements - GenAI - HER - Telemedicine - Clinical Tech advancement - Personalized medicine - Collaborative care models - Interdisciplinary care - **Cost = Price \* Quantity** - - **Input price = price of resource (i.e. how much we pay physician)** - **X = not interested + little discussion** - **We are not willing to engage in difficult conversation (talk to physican and stakeholder) = squeezing from efficiency** - **Efficiency = how much resources and how much we getting out of them (service)** - Argument: although medical cost +200%, quality have increased - Counter: don't see this effect in other areas where tech / quality improved a lot (TV) - Life expectancy is going down but price is going up, also life expectancy is much lower than other countries with much higher spending per capita low ROI - Gap is getting more and more by the year - However, gap of life expectancy is huge within US -  - Iron triangle: cost, access, quality - Shifting from volume to value-based care (VBC) - What is VBC - Transition from fee for service (FFS) to anything else (value based bundles, shared saving, capitation) - Enable transition by shift risk from insurers to providers and patients - develop new payment models - alternative payment systems - Reduction in inefficient / unnecessary treatments - Focus on population health management via prevention + data analytics - A diagram of a cost reduction Description automatically generated - **FFS**: service has a capped fee when used - **Biggest problems:** a lot of necessary stuffs physican will do (unconsciously most of time like recommendation of surgery due to uncertainty in sitatuation) **incentive for volume** - **Capitation**: payment arrangement for health care service providers. It pays a set amount for each enrolled person assigned to them, per period of time, whether or not that person seeks care. - **Episode of care payment**: show up at hospitaldiagnostic related group that you fall into that comes with a payment if hospital get patient out quickly = money, if not = lose money - **Problem**: want to get patients out as quickly as possible = people are sicker when they come out quicker - Condition-adjusted capitation: FFS + episode of care capitation manage patient condition - Shift right of graph physician has less control over population - **Barriers to adoption and scaling of VBC** - Bargaining power - Payer think about value differently than provider - No first mover advantage: 第一个吃螃蟹会吃亏in healthcare Lecture 2 (082924) - An Introduction (Part 2) - Why is volume to value transition is possible now - Improvement in technology - Doctors are being included with cost saving push to ensure physicians to have financial responsibility via insurance - More medical literacy + consumerism (patients are more savvy and want second opinion) - Market consolidation: better population health management - Stakeholders in healthcare ecosystem - Regulators, manufacturers, providers, payers, patient - Want to break into ecosystem: technology companies, large employers, retailers - Many entrants failed, but they will be - Interested in healthcare because - **Lots of profit (20% GDP)** - Lots of data (retailers like Walmart) + combine data together to see pattern within customers - **Lots of frustration to healthcare, so opportunity to change** - **Quiz on healthcare spending** - Fraction of community hospitals that are part of system - 68% - Hospitals are organized systems, not individual hospitals - Fraction of bankruptcies due to medical bills - 50% - This is higher than mortgage and credit card bill - Fraction of healthcare spending for hospital care - 32% - Closer to 75% 10 years ago this means there are so many more different players - Fraction of healthcare spending by the top 1% spenders - 22% - 50% for top 5% spenders - Sick people spend most, rest are relatively healthy focus policy on the top 5% people - Fraction of households with **out-of-pocket** medical expenses that exceeded 20% of family income - 18% - 1/5 family experiences this, super high number - Out of pocket share reduced significantly for last century, but cost rose significantly - Shift from healthcare as business to healthier population - Medical errors = 3^rd^ highest cause of death in US - Hospitals react to patient safety issue = create chief safety officer -  - **Shift from volume to value (quantity to quality)** - **Things that make you go hmmm (Econ analysis on healthcare)** - **Connection between health and medical care is not that direct** - Health = medical care + habit + nutrient etc. - Economic relate output to input - Health and medical care are both input and output - **Patient = major influence on health, physician = major influence on cost of medical care** - Consumer of health have influence on quality of health - **Supplier (physician) can effect demand (consumer willingness to pay), demander (consumer) can effect supply (health service)** - **Hard to model financing in healthcare system** - Patients always who **pays** for healthcare, but don't know how to **finance** healthcare system - **Over romanticize relationship between payers (insurance) and provides (hospitals / physicians)** - Conflict of interests (insurance makes money when ppl don't show up, physician show up when ppl show up) - Insurance make money by (**manage financial risk**): - Processing claims - Spreading risk - Population heatlh management - Providers make money by (**clinical risk**): - Care delivery - Population health management - **HMO = financial risk managers tell how to manage clinical risks, good because low premium** - **ACO = financial risk managers let physician do whatever, but physicians need to manage clinical and financial risks** - A diagram of a risk scheme Description automatically generated with medium confidence - Health and healthcare markets have many distinctive features - Unique = combinations and large number of distinctive features, stand along = not special - **1. Uncertainty** - Demand (patients) - Quality: can't verify quality of care - Individual illness: hard to anticipate future health needs - Supply (physicians) - Effectiveness of medical treatment: hard to predict outcome by physicians - How to negate uncertainty: guidelines + licensing (everyone need degrees) + insurance - **2. Prominence of insurance** - Moral hazard (reduce price = higher consumption) - **3. Asymmetric information** - Patient vs insurance - Patient's private info: life style, family history, service propensity - Lead to **adverse selection problem** - Buyers or sellers of a product or service are able to use **their private knowledge of the risk factors** to **maximize their outcomes**, at the expense of the other parties to the transaction. - Provider private info: patient medical condition, quality of service - Lead to **induce demand / quality shrinking** - Physician takes an action **to shift the patient\'s demand curve** in the direction of the **physician\'s own interests,** often against patient interest - **4. Large role of nonprofit firms** - Many nonprofits hospitals - Objective for firm in econ = maximize profit - Is model of profit max applicable to nonprofit? What are other objectives? - **What are we getting out of non-profit** - **5. Restrictions on competition** -  - All economic rules in free market are violated in healthcare - **6. Natural features of health care services** - Economies of scale (average cost fall with output) = heart of natural monopoly - **7. Government subsidies and interventions** - Licensing, price controls, entry/expansion regulation, subsidies, taxes and insurances - One of the most regulated in world - **8. Externalities** - Action from producer / consumer that affects other producers / consumers, but not accounted for in market price - Positive externalities: vaccine = ppl avoid illness so benefit themselves and society (society part = externality) - Negative externalities: drunk driving - **9. Moral issues** - Is healthcare a need / right? Is healthcare a good / service and need price? Where do we draw the line? Main take aways - Healthcare is distinctive in many features, but each on is not unique - What is unique is the **combination** and **large number** of distinctive features - Lecture 3 (090324) - The Physicians: Supplier Induced Demand (SID) - Healthcare Value Chain - the entire production chain, from raw materials and labor to the final product or service consumed by the end user - Physicians induce demand (tells patients what they need) and supplies medical care - Intermediaries: - Suppliers - Group purchasing organizations, wholesalers - Regulation FDA - Physicians - Make critical decisions about - Spending of money - What tech to use - Public / consumers = money flow - Producer (tech creation) = goes to physician - **AI reads imaging better than physician, but one must sell that idea to physician (to use AI) in order to have it to work** - **Supplier induced demand** - Dual role of physicians - Suppliers of medical care (supply) - Agents for their patients (demand) - **Physician is unique because it influences both supply and demand** - **Principal Agent problem** - a situation in which a self-interested person or entity ("the agent") is able to make decisions on behalf of another person or entity ("the principal") - Principal (patient) lacks time or skills, and so delegates decision-making authority to an agent (doctor / physicians) - Principal can observe outcome, can't observe behavior + effort + sincerity - Latter impacts probability of outcome (may lead to more probable bad outcome) - **Physician agency**: a set of incentives that may cause physician behavior to deviate from that of a perfect gent - **Perfect agent** - **maximizes each patient's health status, utility and well being** - Utility includes and may impact health status - Health status vs utility vs well being (population vs individual) will conflict with one another - Example: cheaper drug with more side effect vs expensive drug with more side effect - Supplier induced demand - physicians will influence the demand of patients, deviating from perfect agent, may have other goals - Definition: **provider of an item or service stimulates or produces more demand for their offering than would otherwise exist in the market.** - **Induced demand = an increase in supply results in a decline in price and an increase in consumption** - **More service = less payment = more patient services** - - To promote doctor's own interest, the doctor recommends care that the perfect agent would not recommend - Doctor misuses patient trust - Can physicians shift demand curve up? - **Supplier induced demand = ability to shift curve up!!** - Want physicians to shift demand up -  - The role of physician - Don't have good understanding of meaning of perfect agency + deviation from perfect agency standard - Roemer's law: a bed built is a bed filled - Example: specialty hospitals like ambulatory surgical centers (ASC) → have grown tremendously, some work it both, some only ASCs - Specialty hospitals = single specialty hospital where physicians can invest in the hospitals themselves - ASC - Want to pay them less but want them to take less healthy patients - Because less payment, they will take more healthy patient because more profitable - Concept of specially hospitals provides profitable services for low risk people who are well insured - **Study: Physicians splitting time between ASC and hospital (HOPD/ASC splitters) vs those only in ASC (ASC)** - HOPD splitter takes higher risk than ASC splitter, but ASC only takes higher risk than ASC splitter - Explanation for results - ASC physicians are different than splitters - They may want to take more money and are just better - If you are trained in state, you are more likely to be splitter; if you are trained out of state, you can likely to be non-splitter - Splitters are less likely to take risk - HOPD splitters see risker patients in hospitals - ASC only treat risker patient - Greater distance between ASC and acute care facility - 1\) lower risk taken by ASC - 2\) Splitters and no splitters behave more similarly (don't want to a lot of risk) - **Supply induce demand: certain level of risk = hospital only, beyond that = not referring to hospital** - Theory behind supplier induced demand (not on final) - The target income model - assumptions - Physicians have some monopoly power - Consumer ignorance provides physicians with discretionary power to manipulate demand - **Income effect:** pay me less = income drop going to create inducement to bring income back - **Substitution effect:** perform same task = same utility, but less award over time - A diagram of income and income effect Description automatically generated - **Since a decrease in income makes demand inducement more desirable, while lower profitability from inducement makes it less desirable, it is the relative magnitude of the income and substitution effects that will determine the level of inducement.** - Not true that physicians work more + higher inducement when income drop (too absolute) - Punchline: supply often drive and induce demand, but we don't really see a lot of inducement - Many stat methods are false **\ ** **Supplier Induced Demand Main take-aways** - Physicians at the center of healthcare value chain (consumer money flow to physicians, producer innovation flow to physicians) - Supply = providers = R&D, demand = consumers = patients - Principal agent problem: Principal (patient) lacks time or skills, and so delegates decision-making authority to an agent (doctor / physicians) - Principal can observe outcome, can't observe behavior + effort + sincerity - Utility vs health status - Supplier induced demand: physicians will influence the demand of patients, deviating from perfect agent, may have other goals - ASC case: Supply induce demand: certain level of risk = hospital only, beyond that = not referring to hospital - Theories that generate supply induce demand - Target income model - **Income effect:** pay me less = income drop going to create inducement to bring income back - **Substitution effect:** perform same task = same utility, but less award over time - Inducement increases when income effect (less income = more induce) \> substitution effect (less award = less induce) - Not true that reimbursement lower, physician work more, higher inducement - **It's a net effect of income effect and substitution effect** - **Income effect = more inducement, substitution effect = less inducement** - Punchline: supply often drive and induce demand, but we don't really see a lot of inducement (statistics is flawed) Lecture 4 - Health Insurance - **Insurance**: contact between company and client, provides protection to illness/injury - Prevention against **health shock**: a sudden deterioration in health caused by illness/injury - Health insurance is unique due to uncertain nature of health - Uncertainty comes from random nature of health and illness - Risk aversion = when someone is willing to pay to reduce / avoid uncertainty even though they would be better off financially by living with the uncertainty. - Purpose of insurance = bear risk for us - Although it is socially efficient for the insurer to **bear all the risk**, **optimal solution = transferred some risk back to consumers offset potential perverse actions on their behalf.** - Common ways to share the risk with consumers include: - **Deductibles**: A specified amount patients **have to pay up** before insurance plan kicks in - **Co-pay**: A specific **flat fee** patients pay for each medical service - **Coinsurance**: Fraction of expenses an insured member has to pay - **Out-of-pocket limit**: upper bound limit of how much you can pay annually, after which bills are covered 100% by the insurance -  - **Missing the premiums from the graph!!** - Premiums - **Why pay premium for insurance:** Risk aversion **payment** to avoid uncertainty - Relationship between wealth and utility - A graph with a red line Description automatically generated - Different wealth have different utility - i.e. I will be happy to pick up \$50, but bill gates may not - Point A = expected utility = average of utilities (\>=2) - **Risk Aversion = always U(E) over E(U)** - **Risk Loving = prefer E(U) over U(E)** - **Difference between CE and E = premium** - Utility of the expectation will always exceed the average of the two utilities - U(E) \> E(U) as it\'s a concave function - Assume risk aversion, we need to find the same utility so that the patient is indifferent with purchasirng Insurance =\> point B: lowest wealth an individual can tolerate to get the same utility - Common insurance plan types - -  - Max spending for insurance = premium + out of pocket - HSA = saving account for health reasons / emergency - Seems like even in worst case scenario, HDHP is better - **However, we need to have money set aside to prepare for this worst-case scenario** - Distribution of Health Insurance Coverage in the US - Uninsured 8.6% - Any private plan 66.5% - Any public plan 34.8% - Medicare only pays up till 100 days, there are insurances covering after that (\~20 cents premium), but not one buy b/c the psyche of just thinking about that situation - A diagram of a medical care cost Description automatically generated - M = how much medical care I consume - Pm = real price of medical care - C \* Pm = consumption = how much I really pay - C = 0, it costs nothing to go see doctor - **Lower coinsurance rate, higher medical consumption** - Relationship between the coinsurance rate (C) and medical care consumption (M) - Price of an insurance is not the premium - Demand becomes more and more inelastic → perfectly inelastic when C = 0 - **Moral Hazard**: ↓ out-of pocket costs = ↑ medical consumption - Doesn't happen if decision is bundled with self (if C = 0, self has no responsibility) -  - Have more coverage (lower C) = higher med consumption (higher M) A diagram of a company Description automatically generated with medium confidence - Loading fee = logistic fees (processing claims, rent, sales force, ad, etc.) - **Higher loading fee = less attractive insurance fee** - Higher loading fee = demand less coverage - **If we want to know coverage, we need to know loading fee** - The money goes the salaries for employees, rent for office space, salesforce commissions, advertising -  - Employer based health care provides lower premium due to tax shield - (1 -t) = tax shield / income tax rate = offset loading fee - A math equations and numbers Description automatically generated with medium confidence - (1-t) (1+L) (1-C) (P\*M) - If tax (t) is greater than loading fee (l), the price will be negative, so premium will be lower - Higher tax = tax shield = more beneficial for people -  - ↑ taxes → ↓ effective loading fee → people choose lower coinsurance rates - This exacerbates moral hazard - A close-up of a sign Description automatically generated - Yes! Without program fire would cost \$1000, but with the program cost \$550 - 0.5%\*100,000 + 50 -  - \$1000, the owner is fully insured he has no expected loss, and no incentive to run the program. Therefore, when the probability of fire is 1%, and the insurance company makes 0 profit moral hazard - So the owner behaves differently with and without insurance. The hidden action is the **extent to which he is willing to invest in preventing fire**. - A screenshot of a document Description automatically generated - **Effort is not dichotomized, it is a continuum** - **Premium = always percentage of expected loss** - So this offset risk Rationale behind deductibles  - Two cases - Low demand = primary care - High demand = critical care - Low demand always in deductibles - Thought behind this: if serious shit happen, we got you; but if it's the small thing, we want you to think if its worthy - Welfare loss = moral hazard area - Questions: - U(E) vs U, the two graphs?? - What is saving option for HDHP? - What is loading fee, coverage - Tax shield of insurance, why premium equation true? - **Why premium = 1000 in fire is true?** - What is welfare loss, what does demand elasticity mean **Moral Hazard Main take-aways** - **Risk aversion:** when someone is willing to pay to reduce uncertainty even though they would be better off financially by living with the uncertainty - **Demand for insurance is linked with demand for medical care** - Share risk with consumers = change behavior - **Risk Aversion = always U(E) over E(U)** - **Risk Loving = prefer E(U) over U(E)** - **Difference between CE and E = premium** - HDHP always cheaper than PPO, **but need to have extra money set aside** - **Moral Hazard**: ↓ out-of pocket costs (price) = ↑ coinsurance rate ( more coverage) = ↑medical consumption - **Solution = shift risk back via coinsurance (partial coverage rather than full)** - Price of insurance - **Loading fee = actual price of insurance, premium = spread risk** - If need to know coverage, NEED to know loading fee - Low loading fee = high coverage = moral hazard - (1 -t) = tax shield / income tax rate = offset loading fee -  - **Premium = expected benefit** - A diagram of a cost of insurance Description automatically generated - ↑ taxes → ↓ effective loading fee → people choose lower coinsurance rates - This exacerbates moral hazard - Loading fee is high self insure high loading fee = low demand for coverage - **Effort is not dichotomized, it is a continuum (risk reduction is not black and white)** - See case study - Deductible: **Deadweight lose / welfare lose** (when serious shit happens, insurance kicks in, you pay fraction (almost none) of actual payment = moral hazard - People abuse high demand (critical care) because insurance covers everything Lecture 5 Speaker: Sushma Akunuru: Healthcare Trends and Role of Technology in Healthcare -- A Payor Perspective - CIO in independence blue cross / blue shield - Definitions - Consumer = patients - Payor = insurer - Provider = doctor - Regulator = policy maker - Producers, broker, suppliers, distributors = party altering buying decisions (pharma, tech companies) - Healthcare = 17-18% of GDP - Most money going to physician - Current state of healthcare - National healthcare spending is increasing every year - Personalizing healthcare - Wastage - Digitizing care delivery (telemedicine) - Interoperability - Overview of healthcare trends - Macroeconomic and regulatory context - Aging population will grow twice as fast - Shifts in market demand - Telemedicine and virtual health are first steps to digital health - Shift towards alternative sites of care for patient preference - Structural changes in healthcare market - New business model: Payers and providers are becoming more holistic (definition) - Healthcare landscape is fundamentally changing - Healthcare shifting from silos to highly interconnected and value oriented market (\*kinda bs) - Competitive responses - Shift in care delivery - Digitization = patient centric ecosystem + end to end virtual and in person solution - Shorten linkage between provider and consumer - Payer employ tech to transform stakeholder experience across healthcare spectrum - Member's health, payer-member relationship, member-provider relationship - History of healthcare tech - Electronic healthcare records (HER 1950s) telemedicine initiative (1980s) mobile health apps (mHealth 2000s) rise of AI and ML (2010s) GenAI-powered Co-pilots predictive (2020s) - Benefits of Ai in healthcare - Data mining - Identify sales and growth opportunities - Personalized patient experience - Minimize error - Accelerated decision making increase productivity and efficacy - \*no use if doctors refuse to use it due to conflict of interest Questions - What are some promising disruption you see happening that is addressing current state of healthcare (digitization)? What are your thoughts on the trends (good or bad)? - Biggest headwinds of incorporating AI in healthcare - Lecture 6 (091224) - **Adverse Selection** (Part 1) - The Supply of Health Insurance: The role of Adverse Selection - Moral hazard = hidden action - presence of insurance induces welfare decreasing behavior. Why? Because the presence of insurance alter individual behavior. - Adverse selection = hidden information - behavior is a function of the individual's type. Some people are low risk and some are high risk. - **In the moral hazard case** the insurance company **cannot observe** the **actions** people take once they have insurance (or it is too costly to do so). - **In the adverse selection case** the insurance company **lacks the information** required to tell the high risk from the low risk individuals apart (or find it too costly to do that). - **Death spiral:** a snowball effect referring to a condition of the insurance markets, where premiums are rapidly increasing because lower risk individuals are continuously priced out of the market (low risk ppl = not really any expected expenditure b/c health but super high premium **will self-insure but not pay for insurance uninsured**) -  - A diagram of a health insurance Description automatically generated - Ppl left of dotted line = risk lover = don't want to buy premium - **Raise premium = more people on the left of dotted line = bigger uninsured population = unstable insurance market** - **Solution:** offer multiple insurance plan, catering to low risk and high risk people - While we don't change prices for preexisting conditions, lot of debate surrounding why should we pay for other people's risks - If uninsured, more likely to get more diseases - Problem: too many variables for uninsured it is correlated with many variables like poverty - Correlation can't infer causation -  - - Y-axis = number of cases / surgeries / procedures - Negative control: Emergent vascular procedures and Cosmetic surgery procedures - Not insured - No major difference in trend pre/post 65 y/o - Outpatient lens procedures: symptomatic - Major jump in case rising - **Caused by influx of uninsured individuals** - Everyone has insurance before 65 - Maybe: people are retiring = have more time - Colonoscopy screening procedures: asymptomatic - Major jump in case rising post 65 y/o - **Solutions to adverse selection** - **Make the information more symmetric** - Allow insurers to collect health information and use that for pricing decisions - But...sicker people pay higher premiums, healthy people pay lower premiums, people with pre-existing conditions pay very high premiums or are denied coverage altogether - **Big problem:** purpose of insurance is to defend ourselves financially against sickness / risk if getting sick = higher premium, then defeat purpose of insurance - **Sell insurance to groups and not to individuals** - Employers purchase group plans, and risk is spread to a large pool of employees with a spectrum of health conditions - But...group health insurance ties labor participation and job mobility to health coverage - **Big problem:** diverse profile of employees, how do we unify background and make group purchase distort labor market - **Mandate insurance** - Low-risk individuals stay in the pool and help maintain affordable premiums - But...if forces people to buy insurance against their will - **Big problem:** healthy will subsidize the sick - **Offer multiple insurance contracts to separate high and low risk individuals** - The insurance company still cannot identify the high and the low risk individuals; it must offers two insurance plans that will make **the two groups self-select** into them, in a way that is profitable to the insurance company. - Low risk = low premium vs high risk = high premium - Problem: everyone will self select low risk and take low premium - The way to do it is by denying low risk individuals the insurance they would really want **(shift risk back only to low risk individual)** - **Penalize low risk individual to differentiate themselves from high risk individual** - Low risk people are more price sensitive and are more open to the possibility of partial coverage in return to low premium that high risk individuals - The insurance company will use this tradeoff between premiums and coverage levels to effectively separate the high risk individuals from low risk - Don't have insurance consequences - Burden to society: society have to find resources to take care - Behavior is different - The Rothschild and Stiglitz Model - 2nd degree price discrimination (cannot tell the WTP between groups) - Train example → prevent passengers who can pay the second class fare from traveling third class by making travel conditions worse off so people would pay for second class tickets - "It hits the poor, not because it wants to hurt them, but to frighten the rich" - Poor had to pay for the price of bad self selection of the middle class - 3rd degree price discrimination (students, senior citizen discounts has lower WTP) -  - Loss = cost for service = hit on finance - L = R, full insurance - L \> R, partial coverage - R \> L, insurance fraud - n = net payment = amount of money you get from insurance company - - W-pL = expected / average wealth wealth -- (probability of loss) (loss) - [*π*]{.math.inline} = maximum amount of money I am willing to put in to avoid risk = premium - Premium = want to shift the utility to a level that is expected of individual U(E) - CE = lowest wealth individual is willing to tolerate to avoid risk -  - If L = R and have insurance, you have w-[*π*]{.math.inline} at the end of day - A diagram of a graph Description automatically generated - Indifference curve = dots are the same on the curve - Inside indifference curve = better off - For instance, for B, getting in zone means more net payment (more money get from insurance company) but same premium - Direction of utility = towards better off zone -  - Insurance line ([*π*]{.math.inline} = -n + L, slope = -1) - On line = full coverage - Inside line = partial coverage - Outside line = insurance fraud - - Insurance company make 0 profit if market is competitive - [*π*]{.math.inline} = probability of getting paid (p) \* amount of payment - Zero profit line - On line = zero profit - Inside line = insurance company losing money - Outside line = company making money -  - Insurance company want to maximize utility E(U) - S.t. = state = insurance company at least need to make 0 profit (on the line) - Sub in [*π*]{.math.inline} into E(U) - A diagram of mathematical equations Description automatically generated with medium confidence - Solution = full insurance - Only 1 risk type, everyone should get full insurance -  - A diagram of a cost of insurance Description automatically generated with medium confidence - E = equilibrium = where zero profit line intersects with full insurance line - What's wrong with U1: firm is making a profit in a competitive market - What's wrong with U2: firm is out of business, no insurance at all - Solution is on U(E) Questions - Go over equation, especially Max one **Adverse Selection (1) Main take-aways** - **Moral hazard = hidden action** - Insurance company can't observe actions after insured - **Adverse selection = hidden information** - Insurance company lacks info to differentiate low risk vs high risk - **Death spiral:** snowballing of low risk individual don't buy insurance higher premium - If insurance company knows distribution of risk in society, it can offer one insurance contract based on average risk? - NO! This creates snowball effect rapid increase of insurance premiums with low-risk ppl self-insure (avg moves up as more people uninsured) - **Solutions to adverse selection** - **1. Make the information more symmetric** - Allow insurers to collect health information and use that for pricing decisions - **Big problem:** if getting sick = higher premium, then defeat purpose of insurance (defend financially against sickness) - **2. Sell insurance to groups and not to individuals** - Employers purchase group plans, and risk is spread to a large pool of employees with a spectrum of health conditions - **Big problem:** distort labor market (group insurance ties labor participation and job mobility to health coverage) - **3. Mandate insurance** - Low-risk individuals stay in pool and help maintain affordable premiums - **Big problem:** healthy will subsidize the sick + violate free will - **4. Offer multiple insurance contracts to separate high and low risk individuals** - Can't identify the high and the low risk individuals offers 2 insurance plans that make **the two groups self-select** into them - Low risk = low premium vs high risk = high premium - **Big Problem:** everyone will self-select low risk and take low premium - Solution to make (4) work - **Shift risk back only to low risk individual = Penalize low risk individual to differentiate themselves from high risk individual** - Rothchild & Stiglitz model - Make condition unbearable for high risk individual so they self-select into high risk, but low risk individual pays price - **Case 1: firm makes no profit, 1 risk type** - **Equilibrium = full insurance for everyone (R = L)** Lecture 7 - Adverse Selection (Part 2) - The Supply of Health Insurance: The role of Adverse Selection -  - A diagram of a firm profit line Description automatically generated - If \>1 risk factor, what will change? -  - Green line = high risk individual - Steeper because prob of loss is higher - Blue = low risk individual - Red line = maximum \$ willing to pay - High risk indiv willing to pay more than low risk indiv - **Nothing to do with risk aversion!!** - A diagram of a mathematical equation Description automatically generated with medium confidence - Compare zero profit line of high vs low risk indiv: slope of zero profit line (ZPL) - Red box = slope of ZPL -  - Slope of ZPL of low risk indiv \< high risk indiv - Hypothetically: have 2 plans, intersection of full insurance line and ZPL = premium - **But can't assign risk group (if high risk take low risk plan, lose \$)** - Net payment (n) = R - [*π*]{.math.inline}, so ∆[*π*]{.math.inline} is same on x and y axis (diff of H and L are [*π*]{.math.inline}) - A diagram of a graph Description automatically generated - A diagram of a graph Description automatically generated - **Where should EL contract be: intersection of indifference curve (red line) and ZPL (see below)** - A white paper with blue text Description automatically generated - El = partial coverage, but sig low premium - Don't want left graph, bc Eh want to shift to El (lower premium, but still full coverage - To prevent that from happening, we will let low risk indiv pay the price: low premium but partial coverage to low risk indiv high risk indiv would not want that cause only partial coverage = can't offset risk will pay for high premium but high coverage -  - Want: self select into contract (high into high, low into low) - Low risk indiv = pay price of lower utility (solid blue curve) instead of better utility (lower dotted blue curve), but still better than high risk indifference curve (much lower utility) - Solution: low risk indiv = low [*π*]{.math.inline} low coverage, high risk indiv = high [*π*]{.math.inline} full coverage - - Gamma = proportion of low risk indiv in society, p = prob of having an event (risk event) - If Gamma = 0, everyone is high risk pH / 1-pH \* n = π - Indifference curve of high risk indiv (green line) goes tangent to upper ZPL - If Gamma = 1, everyone is low risk pL / 1- pL \* n = π - Indifference curve of low risk indiv (blue line) goes tangent to lower ZPL -  - If we know risk of everyone in society, we can offer 1 plan to all - **Case studies** - A diagram of a line and a point Description automatically generated with medium confidence - If insurance company offer A, you will make money (above ZPL), but no body will want it because everyone is worse off in that ( -  - If offer B, high risk indiv will take because below ZPL higher utility and less premium - A diagram of a triangle with lines and points Description automatically generated - Both high and low risk indiv will love because higher utility for lower premium lose \$ -  - **Cream skim:** Low risk indiv would low because above ZPL for ONLY low risk indiv, but low risk indiv have higher utility with lower premium compared to high risk (although below ZPL for both high and low risk indiv, we don't care, cause only targeting low risk indiv) - **This would drive Ep out of business, meaning only option for high risk indiv = D D would lose money because high risk indiv would take it company bankrupt** - - This is why D shouldn't be offered (from a dynamic perspective, D is not sustainable... -  -...but we can shift D above ZPL - Can still be profitable even if pooling everyone into it - A diagram of a line Description automatically generated - D is not stable because it becomes Ep, so some other contract will try to undertake it (F will be the new D, and D will be the new Ep) - This will eventually make contract onto ZPL everyone going to Rothschild and Stigliz (EL in graph below) -  - What happen if we deviate from competitive model - A close-up of a paper Description automatically generated - Full insurance + 1 risk type = no selection problem (everyone would have same thing, so no moral hazard) - Monopolist would set up problem differently - If have competition - Maximize utility of individual, subject to zero profit condition (making sure firms show up) - If monopoly - Maximizing profit of company, subject to utility (making sure people come and buy) - Consumer surplus = individual are going to be indifferent about having or not having insurance (price so high so doesn't make difference) -  - Find the indifference curve = identify point that represent logical indifference origin (premium = 0, coverage = 0) = no insurance - Indifference curve = need to cross point where no insurance (origin) - Above full insurance = making money (where M indifference curve) - M = full coverage, no consumer surplus (monopoly) for high risk indiv - Between M and C = sharing surplus, no monopoly because companies get shares - A diagram of a graph Description automatically generated - Can't have two contracts on full insurance line (self selection problem happen) -  - Not gonna work, because can't offer contract where low risk indiv would rather not have contract at all (higher premium, lower utility) - A diagram of a graph Description automatically generated - This works - Rothschild and stigliz features - High risk indiv = full coverage, high premium - Low risk indiv = partial coverage, lower premium - This works, but need to give back some of consumer surplus to high risk indiv, but low risk indiv have no more room to be punished (no diff between have vs not having insurance = not buying insurance / squeezing out them by pricing them out of market) need to cater to high risk indiv - Should offer this contract: - If not a lot of low risk indiv (lots of high risk indiv) = doesn't make sense to cater to low risk - Should not offer this contract: - Pricing out a group doesn't mean this doesn't happen **(big risk)** -  - **Important: likely be on exam!!!** - **Results of Rothchild and stigliz model** - **Why monopoly can't take consumer surplus of ALL group** - **Either price out a group or cater to one group** - **Why is pooling equilibrium so unstable** - **Spiraling down to EL in Rothchild and stigliz** Questions: - What is pooling equilibrium - Why if we know risk for all, we can offer 1 plan to everyone - Why is high risk indiv indiff between 2 types of policies - Why is it hard to maintain a pooling equilibrium **\ ** **Adverse Selection (2) Main take-aways** - A diagram of a firm profit line Description automatically generated - Rothchild & Stiglitz model ([competitive market]) - **Case 1: 1 risk type, firm makes no profit** - **Solution: Equilibrium = full insurance for everyone (R = L)** - **Case 2: 2 risk types, don't know info on risk proportions** - High risk ppl = willing to increase to pay more premium for higher coverage (**high probability of loss**, steeper indifference curve) - **Willing to pay = probability of loss, no assumption in risk aversion!!** - But can't assign risk group (can't tell high risk vs low risk group, high risk will just take low risk plan (low premium, high coverage)) - **Solution: low risk indiv = low** [**π**]{.math.inline} **low [partial coverage,] high risk indiv = high** [**π**]{.math.inline} **full coverage** - Low risk have utility so low that high risk wouldn't want their plan, but to low risk, its still better utility than high risk - **Case 3: 2 risk type, know proportion of low vs high risk** - If we know risk of everyone in society, we can offer 1 plan to all that has full coverage - Problem: This is not stable equilibrium -  - A: earn profit, no one take - B: good for high risk, bad for low risk (proportion too small) - C: good for high and low risk, everyone take - **D: Cream skim:** only care about low risk (plan D) other firm with Ep (plan with higher premium and coverage that target high and low indiv) bankrupts high risk indiv will now take plan D (plan D becomes Ep) bankrupt - **D = profitable but not stable** - Stick with solution in case 2 - Monopoly (Non-competitive insurance market) - **Monopolist offer full insurance** (1 risk type) because no moral hazard, no selection problem - If have competition - **Maximize utility** of **individual**, subject to **zero profit** condition (making sure firms show up) - If monopoly - **Maximizing profit** of **company**, subject to **utility** (making sure people come and buy) - **Solution[: give consumer surplus back] (better utility for high risk, don't change low risk)** - In monopoly, no room to punish low risk (**if decrease utility for low risk, they will go uninsured**) - A diagram of a graph Description automatically generated  **Yes** **No** - However, if not a lot of low risk indiv (lots of high risk indiv) = doesn't make sense to cater to low risk just punish low risk all the way even if all go uninsured - On Exam!!! - A close-up of a document Description automatically generated - **Important: likely be on exam!!!** - **Results of Rothchild and stigliz model** - **Why monopoly can't take consumer surplus of ALL group** - **Either price out a group or cater to one group** - **Why is pooling equilibrium so unstable** - **Spiraling down to EL in Rothchild and stigliz** Lecture 8 - Division of Labor and Specialization: The Physician Specialization: Hospitalists - Division of Labor: an economic concept which states that dividing the production process into different stages enables workers to focus on specific tasks, division of labor has done well! - Specialization - Taxonomy of medical specialties - Organ - Demographics (pediatrics, age groups, gender) - Diseases (oncology) - Site of Care (ED, Intensive Care, Hospitals) - Hospitalists: physicians whose primary professional focus is the general medical care of hospitalized patients - They specialize in everything in hospital - Provides information for care post-discharge and received info from PCP for essential information -  - Primary care physicians work with hospitalists - **Reason why this happen:** - primary care physicians can spend more time in office / clinic and give the portion of work to hospitals (division of labor) - **Major reduction time of stay in hospitalists because they have personal relationships with different doctors in hospitals that can speed up process** - A blue background with white text Description automatically generated - Limitations of division of labor - Market size - Adam smith: The division of labor is limited by the extent of the market" - Supply - \# of workers - Demand - \# of buyers - Increased productivity due to - Specialized knowledge - No wasted time - Isolated task sparked innovation - **Something else other than extent of market is limiting division of labor** -  - **More time learning = increase human capital but in decreasing fashion** - A graph and diagram of a function Description automatically generated - Get optimal h and l from equation (want best amount of time for learning + performing) -  - We don't have anything limiting process of specialization Adam Smith's theory doesn't tell us what limits division of labor (supposedly according to this model, more subdivision of labor = less learning time but higher performance = higher productivity) - **Coordination costs (major limitation)** - Increasing specialization in tasks raises coordination costs - A diagram of cost reduction Description automatically generated - More communication nodes = higher coordination needed = higher coordination cost (exponential increase) -  - Increase size of group = increase in benefit (concave), but the cost is also rising (convex) - **Optimal team size = max difference between benefit and cost (benefit needs to be \> cost)** - A diagram of a solution Description automatically generated - **Corner solution**: even adding \>1 person = higher cost than benefit - **Example**: psychologist - Skill complementarities - Investment in one skill lowers cost in investing in another skill - Given this, makes no sense to divide labor - Not likely to happen, but good to know - **Task adhesion (best theory on why we have hospitalist)** - A tendency or incentive to keep healthcare tasks bundled together or performed by the same provider or within the same healthcare entity. - Opposite of division of labor (want to multitask) -  - Investment in site specific skills - more different the skills needed for two tasks, less task adhesion / tendency to multitask - Reimbursement - If difference between reimbursement between 2 sites for same procedures, need to have internal mechanisms to do that - higher reimbursement rate = less likely for division of labor - same reimbursement rate = more likely to division of labor - Self-referral rate - the percentage of patients who refer themselves to a medical professional without being referred by another health provider - like OBGYN - Why primary care physician gave up hospital tasks - Fewer patients are hospitalized - Can't really help hospitalized patients because they are so severe - Only thing that primary care physician can do is to follow up meeting, but this can be done at clinic / office, not worth it to go to hospitals - Similar reimbursement rate - Benefit of hospitalists - Save money on testing and quick turnaround times - Cut length of stay by quicker testing - All of the hospitalists have lower diagnostic charges than do teaching teams. Questions: - Why reimbursement? See video **Hospitalists Main take-aways** - **Hospitalists:** physicians whose primary professional focus is the general medical care of hospitalized patients new form of specialist - Compared to primary care physicians**: more complex environment, similar reimbursement, fewer patients** - **Why hospitalist** - 1\. primary care physicians can spend more time in office and give up work in hospitals (hospitalized patients) form of division of labor / origin of specialization - **2.** **Major reduction length of stay in hospital because hospitalists have personal relationships with different doctors in hospitals speed up process** - **Limitation of division of labor** - 1\. Extent of market - Supply (how many workers) and demand (how many buyers / consumers) - 2\. Cost-benefit of specialization - More time learning = increase human capital but in decreasing fashion - Massive time needed for training, need talent, need to learn by doing rather than theory, boredom/burnout - 3\. Coordination cost - Increasing specialization in tasks raises profit and coordination costs - **Optimal team size** = benefit \>\> cost - **Corner solution:** some jobs only need 1 person (i.e. psychiatrist) - 4\. Skill complementarities - Investment in one skill lowers cost in investing in another skill doesn't make sense to specialize - **5. Task Adhesion** - A tendency to keep healthcare tasks bundled together or performed by the same provider(s) in same network - [Factors: investment in site-specific skills, reimbursement, self-referral rate (see patient in office and in hospital)] - **[Division of labor (want hospital + office) = high task adhesion = similar investment in site specific skills, different reimbursement, high self-referral rate]** - **Why PCP gave up hospital portions** - High investment in site specific skills (Hospitalized patients = more acute illness = need different skillsets) - Similar reimbursement - Low self-referral rate (because too general) - **Benefit of hospitalists** - Save money on testing - Quick turnaround times - Cut length of stay by quicker testing - Lower diagnostic charges - Lecture 9 (092424) - Changing Landscape of Primary Care - Definition: Integrated, accessible health care services by clinicians who are accountable for addressing a large majority of personal health care needs, developing a sustained partnership with patients, and practicing in the context of family and community - **Challenges** - **1. Undervalue and under pay of primary care physicians (PCP)** - Different compensation differences - No real correlation between actual vs expected compensation - Specialists earn 41.5% more, but working with the same amount of hours - Self employed physicians earn 20% more - Male physicians earn 31% more - Female are more concentrated in low paying specialties - Female are less likely to be self employed - Trend regardless of age and population - 2\. Challenging work life balance, high rates of burnout, workforce shortage - A close-up of a list of words Description automatically generated - For preferences, intellectual context = most important, relative income = not a lot of priority - Usually, more intellectual pursuit + drive = higher payment - **Not good explanation** - Subjective Discount Rates - Heterogeneity in time preferences - Can prob be PCP out of medschool - **Not good explanation** - Liquidity Constraints - Heterogeneity in access to financial markets - Smoothing consumption becomes more challenging when residents have substantial debt payments (how to pay for edu) - **Not good explanation** - Static Income Expectations - Are students capable of incorporating future trends? - How financially literate are medschool students - Med students are very good at this - **Not good explanation** - Barriers of entry - Residency review committees (RRC) = cartel that controls flow of physician, who to let into this specialty or not - Control difficulty of tests, how many surgeries to do, how many years of residency, etc. gate keep professional capacity shortages in profession higher price of service - **Doesn't explain all, but some** - Specialists, Spending and Quality - More earning for specialists = more spending - Higher rate of specialization = less PCP = lower healthcare quality in community -  - 1^st^ column - mix of demand and supply - Lots of PCP works for another jobs (consulting, pharma, CMO, insurance companies) - They think they have more impact that way something really wrong in primary care - 2^nd^ -- 3^rd^ column - Next consequences - 4^th^ column - Eventual consequences - Low value bc low utilization of time - Data analytics and disease management - - If patient = very sick (high severity), no meaning to intervene = low value - If patient = not sick, not going to treat / intervene = low value - Middle = patients were there will be intervention, and will be treated successfully = high value - **Create predictive model based on these** - Likelihood of hospitalization (LOH) program - Use algorithm to predict who is gonna be in hospital, rank them disease management = distribute and contact patients - Algorithm = extremely predicative + good at sorting high risk ppl into yes vs no treatment - Testing value of disease management - New: Look at all ppl who got treatment, but some sorted by human some by algorithm, compare outcome - Changing landscape of primary care -  - Retail clinic = more convenient - Patient-Centered Medical Home (PCMH) - improve quality through tech and process restructuring - What is good - Comprehensive care - Prevention to chronic - Accessible services - Shorter waiting times, Alternative communication methods - Coordinated care - Accountable + continuous care - Improve patient experience and population health - ↑ in NCQA qualification in PCMH, lots of PCMH have individual programs - - Questions - Subjective Discount Rates vs liquidity constraints? - LOH study main takeaway? **\ ** **Primary Care (1) Main take-aways** - **Challenges in primary care** - **1. Low payment / undervalue PCP skills and roles** - Specialists earn much more than PCP because - Specialists self employ more - More male in high earning specialties - 2\. Challenging work-life balance / burnouts / shortages - **Why do people go to specialization vs PCP?** - 1\. Preferences (No) - intellectual context \> relative income - 2\. Ability to do diff specialties (No) - 3\. Subjective discount rates (No) - More specialized = spend more time and \$ (poor for long time) - PCP = can practice straight out - 4\. Liquidity rates (No) - How to pay for education - 5\. Static income expectations (No) - Do medical students know which specialty earns money (yes) - **6. Barriers to entry (maybe)** - Residency review committees (RRC) gatekeep to control physician flow - PCP = easier to enter - **More specialization = less PCP = more spending = lower prevention / quality of care** - **Primary care crisis** - Begin: High demand for PCP (more patients), but decreasing supply (PCP exit + not join) - Primary care has longer wait time (PCP over worked) - shorter patient visits \_ heavy reliance on referrals - more expensive, inadequate, low value care - **Data analytics and disease management** - If patient = very sick, no meaning to treat = low value - If patient = not sick, not going to treat = low value - Middle = patients were there will be intervention, and will be treated successfully = **high value but hard to predict establish model on these, NOT the sickest** - **[Predictive algorithm for disease management] performs better than normal** - **Changing landscape for primary care** - A diagram of a patient and a patient Description automatically generated - 1\. less "fighting" between these departments b/c free up time for other work - **2. PCMG vs RBM** - Patient-Centered Medical Home (PCMH) - Put patient in middle, providers/payers/healthcare providers form around - **What is good** - Comprehensive care (prevention to chronic, everything under same roof) - Accessible services (shorter waiting times, Alternative communication methods - Coordinated care (different physicians around patient) - Accountable + continuous care - **What is missing** - Resources + incentive for providers Lecture 10 (092624) - Changing Landscape of Primary Care (II)  - Patient centered medical home (PCMH) - Improve quality through tech and process restructuring - National Committee for Quality Assurance (NCQA) - Key opinion leader (KOL) drive changes - PCMH is a recognition, not a certificate - 2008 is the model year for PCP-PCMH recognition process - - Created hierarchy to give score that translates into levels of recognition, 0-100 points - Standard = number, element = letter - **Increase access: turn into walk in clinic at the end of the day** - How to get NCQA recognition -  - Almost everyone primary care practice is recognized - Insurance company try to influence processes in primary care try to incentivize in diff way (bonuses) - They liked NCQA PCMH immediately provided better reimbursement and incentives - How insurance company pay primary care practices - Combination of **FFS** and **capitation** - A white paper with black text Description automatically generated - Base capitation: certain amount of \$ / patient - Bill above: additional payment for certain services - Capitation boost: primary care clinics receive more \$ / patient based on higher level -  - Cascade of elements in Venn diagram doesn't tell you differences and similarities between practices need to use cluster analysis (maximize correlations between clusters of variables) - A diagram of a clustering algorithm Description automatically generated - Classification on dendrogram = closer distance - Give hierarchy of closeness = hierarchy of clusters -  - Having 3 clusters is good bc PCMH has 3 levels - A diagram of a cluster Description automatically generated - Observations of real data - 3 clusters - Cluster 1, 3 = high performing - Cluster 2 sucks - Cluster 3 = physician facing, Cluster 1 = patient facing - **Which one produces better results?** -  - X= no change - With higher level, should have more practice visits and lower professional expenditures - Found no meaning results regarding diff levels producing different results - Once looking at clusters, more significant differences - A white paper with black text Description automatically generated - Retainer based medicine (RBM) - Reduce patient panels by introducing membership fee -  - - Retainer fee: **mount of money paid upfront** to secure the services of PCP - Concierge = type of primary care that offers personalized medical services in exchange for a membership fee - Increase quality of activity - Direct primary care = patients purchase a membership that allows them unlimited access to certain primary care services - Reduce cost of activity - Let patient pay out of pocket to reduce cost - Insurance works well if they insure against things that have less probability of happening - If insure against things that happen often, market fails -  - RBM basically paying to exclude other people, not can't legally do so, how to do? - Come up with activities not covered by insurances Concierge = great vehicle for carebut most people that have access to this type of care are wealthy Question - RBM vs PCMH - Concierge vs direct primary care **Primary Care (2) Main take-aways** - **Patient Centered Medical Home (PCMH)** - National Committee for Quality Assurance (**NCQA**) = accrediting organization - Recognition data belong to physicians, not NCQA - **PCMH = recognition, not certification** - Recognition process = key opinion leader (KOL) driven - **Insurance company** - Incentivize primary care clinic to be recognized as PCMH through **capitation boost** and **covering 10-20k for recognition** - Reimburse primary care clinics with **capitation** and **fee for service (FFS)** - **Base capitation:** certain amount of \$ / patient - **Bill above:** additional payment for certain services - **Capitation boost:** primary care clinics receive more \$ / patient based on higher level of PCMH - **Major conclusions** - Theory: ↑primary care visit = ↓specialist / ED visit = ↓risk and cost - PCMH model is carried out, but [little evidence that it improves care] - **PCMH levels** has **no impact** difference on primary care visits ([healthcare utilization]) and costs ([expenditures]), but specific path to implementation (clusters) matters!! - **B2C = patient facing (population health management (ok)** - ↑primary care visits + ↑hospital admission - PCP found problems = patient go more to hospital (short term). Should decrease in long term - **B2B = physician facing cluster (great)** - ↑primary care visits + ↓specialist, ED, and cost - Compared to levels of recognition, detailed info on **practice capabilities**, **investments**, **improvements**, and **processes** = valuable for regulators and **[insurers]** - **PCMH = not in best interest for providers (physicians) + 3rd party credentialing organizations (NCQA)** - B/c level / checkbox system doesn't work - **Retainer Based Medicine Model (RBM)** - Concierge vs Direct primary care model - **Similarities** - Both have retainer fee - **Differences** - **Concierge** =↑personal attention/meeting time **= ↑quality of care** = ↓patient volume - Benefit - strong patient-physician relationship - Weaknesses - **↑PCP shortage** ↑demand **patient without doctors** - **no scalability** - **Paying to exclude other people** (time of physician not attending others) - **Physician are not ultra rich, just trading volume for peace of mind** - **Direct primary care** = ↓cost by stop working with insurance and let patient **[pay out of pocket]** (young + healthy people with high deductible plan) - **No insurance b/c** ↓ operation cost + ↑leverage with facilities + ↑ saved \$ - [Shouldn't use insurance for primary care (probability too high), but only for catastrophe] - **Concierge model health-wealth relationship** - **Hella sick** and **wealthy people** like concierge model - As a society, we want concierge to **attend sick people** rather than **healthy + wealthy people** - **Reality:** - concierges see **wealthy people \> sick people** - No change in mortality, but works well for **sick people** - ↑wealthy people because they are sick more often (i.e. first class flyers = fly more often) - **What causes primary care shortage** - **Large patient volume** - Burnout - Mounting bureaucracy - **Dissatisfaction with impersonal care (shift to concierge)** - Rushed appointments - **\*Big problem:** Concierge = ↓patient volume ↑patients go to traditional PCP practices add more stress/burnout to PCP ↓quality of care ↓PCP patient = no PCP Lecture 11 - Nonprofit Hospitals: Ownership Dynamics, Objectives, and Benefits for Society (Part III) - On the Objectives and Behavior of Nonprofit Hospitals in the US - **Main questions: what are the objectives of nonprofit, what are they maximizing, why do they co-exist with for profit hospitals, do they justify tax exempt status** - **History of Hospitals:** - sick and poor increase middle class + urbanization + industrialization monetization + physicians control hospitals and new tech poor and sick get bad quality of care + become lab rats rise of insurance (less out of pocket) rising cost and regulations + certification of needs shift from volume to value + healthcare system for M&A - **Hill Burton Act:** only to governmental / non-profit hospitals  provided construction **grants** and **loans** to build hospitals - Demographics of hospitals - Mostly short-term general hospitals, about 30% non profit. 16% for profit, 17% government - **Less beds and hospitals (all capacity goes down), but expenditures 10x** - All beds in systems instead of stand alone hospitals - Decreasing beds - A white paper with black text and red text Description automatically generated - **Medicare = gov says don't want to provide care, want to finance care** - Increase for profit beds, decrease gov beds - Before 1940s: - What is non-profit -  - A table with text on it Description automatically generated - **Nonprofit Hospitals (1) Main take-aways** - **Non-profit hospitals** - Tax exemption, but profit can't be distributed (profit = reinvested back) - Legally, no residual claimant (no owner of non-profit can claim profit) - **Beds / capita in US are decreasing at steep rate** -  - **!! Bed days = patient days** - **Medicare (1965) = gov don't want to provide care, but want to finance care** - Pre 1940: nonprofit bed \>\> profit beds - post 1965: ↓gov beds, ↑for profit beds - Trend: For profit hospital size converges to non-profit hospitals - **What account for decline** - A diagram of numbers and symbols Description automatically generated with medium confidence - **1. ↓number of admissions to hospitals** - Short inpatient procedures outpatient - ↑ Outpatient in physician office / freestanding facilities - 2\. ↓average length of stay - **Shift to outpatient [doesn't] ↓length of stay!!** (short procedure) - **Only shifting patient population does** - ↓Short length of stay, but gap between non-profit vs for profit hospitals gap is smaller ([become more similar]) - **Reason: Patient population ([Elder + poor]) becomes similar between non-profit and for-profit hospitals** - Length of stay for poor patients = higher b/c [longer discharge time] (Keep in hospital b/c its safter) - 3\. Both - Why is there **↓** non-profit hospital size (**↓**nonprofit to for-profit ratio) - 1\. ↓average length of stay - Medicare/Medicaid = \$\$ for poor / old patient = for profit grabs them from non profit similar patient population - 2\. ↑ for profit hospital [chains], ↓standalone small for-profit hospitals - 3\. M&A of hospitals = ↓ hospitals - **Punchline: Ownership status is very sensitive to changing economic conditions** - Poor times (low margin) = ↑ non-profit (no incentive for for-profit) - Rich times (high margin) = ↑ for-profit - [Market responds to good and bad times with ownership switches effect non-profit vs for-profit ratio] - **Non-profit = not white horse, ppl are actively thinking about switch between non-profit vs for-profit** - **What do for-profit vs non-profit hospitals maximize / objectives?** -  - Stakeholders = trustees, physicians, administration, patients - A diagram of a medical model Description automatically generated with medium confidence - Lecture 12 - Nonprofit Hospitals: Ownership Dynamics, Objectives, and Benefits for Society (continued) - **Average cost pricing = profit is 0 not maximizing profit** -  - First 2 questions = previous lecture - **Split of for and non profit (Q3)** - A chart with numbers and text Description automatically generated - High expectation for hospitals - Why nursing home / out patient = profit based - In hospice = physician certify that patient has only 6 months left to live - Used to be dominated by non-profit, but shift to for profit because condition going into hospice have changed - Cancer = easy to predict, AD = hard - **Tax exemptions (Q4)** - Although demand for emergency care is increasing, closure is also higher - Charity care + bad deft = uncompensated care (care hospital never receive compensation) - **Nonprofit Hospitals (2) Main take-aways** - Non-profit = large revenue from patients (undistributed profit), charity = only construction - **Q1. What do for-profit vs non-profit hospitals maximize / objectives?** - **Quantity-quality model** - Non-profit hospitals **maximize [quantity and quality]** - Reinvestment of profit = ↑quantity and quality = ↑demand b/c ↑quality + ↑ avg cost - **(naïve) Non-profit = ↑quantity (larger)+ ↑quality + ↓ price** - **Physician cooperative model** - Non-profit hospitals **maximize [earnings of physicians]** [ ] - Partner / owner = run firm = no separation of ownership / control - Partners **will not be** added if value of [marginal product] (single contribution) exceeds [marginal cost] (cost of adding) need to be compared to [average] - Depending on how profit is split, **partner will only be added if contribution \> [average] contribution** - 2 forces of partnerships - \(1) ↑physicians = ↑profit total (size of pie) - \(2) ↑ physicians = ↑ ppl split pie = ↓profit individual - **↓Marginal product of labor (indiv contribute), (2) \> (1)** - **↑ reimbursement (↑ demand = ppl pay more) = ↓partnership size** - **Gain of pie \< loss via sharing pie** - Normally, ↑ reimbursement = ↑ size (scale) - Errors - **1.** Predicts **non-profit size \< for-profit size** - **2.** Predicts **bad time = ↑non-profit size**, good time = ↓non-profit - Normally, bad time = ↓non-profit size - **Profit deviator model** - Non-profit hospitals **maximize [profits + quantity]** [ ] - **Non-profits = ↑volume/ quantity + ↓ price, but [unknown mechanism (unobserved quality) ]**(motive = altruism and "empire building") - All 3 models are not satisfactory to explain non-profit objectives - **Q2. Why do non-profit and for-profit hospitals coexist in same market?** - 1\. Non-profit and for-profit = different services - No, they produce similar services - 2\. Non-profit and for-profit = different organizational goals - Different objects ≠different behaviors, may behave similarly but difficult to explain switching between non-profit and for-profit - **3. Non-profit = solve median voter problem under-provision of services** (can't provide services to everyone who needs them) - Private / public goods - Public goods = goods not competing for same resources (i.e. park) - Private goods = excludable goods (i.e. private guards) - Median voter problem = in a majority-rule voting system, policies tend to **reflect the preferences of the median voter** -  - Healthcare = [public good] [high level of heterogeneity] (lots of diff goals), ↓gov, ↑non-profit, \~for profit - **4. Non-profits solve an asymmetric information problem** - Quality can't be measured **non-profit status = warranty = signal of good quality** - Non-profit = high quality + high price - For-profit = low quality + low price - Problem - 1\. Asymmetry = hospitals (no physicians!) vs patient - 2\. No evidence showing non-profit charging higher price + quality - 3\. Non-profit hospitals don't mention non-profit status nonprofit status = **quality signal in name** (i.e. Saint Francis = for profit, used to be non-profit) Lecture 13 - **Q3. Why do [non-profits] dominate [hospital] industry while [for-profits] dominate [nursing home] industry** - 1\. More **unobservable quality** in hospitals than in nursing home - 2\. Hospitals started out as charitable / **altruistic** organizations - 3\. Nursing home / outpatient = not as **heavily regulated** as inpatient / hospitals - 4\. Less **need to hold excess capacity** due to less demand uncertainty - Hospital need to have ppl/room ready for emergency not profitable - Non-profit = concentrated in [hospitals] and [inpatient facilities] - For-profit = concentrated in **[hospice]** and [outpatient] - Hospice shift from non-profit to for-profit b/c **patient in hospice shifted** from (almost dead, short life, no profit, i.e. cancer) to (unknown life expectancy, very profitable, i.e. AD) - **Q4. Do non-profit hospitals justify their tax exemption status?** - ↑ED visits, ↓ED departments b/c ↑[uninsured ppl] going to ED for [minor] issues - Non-profit exempt from property, sales, and income taxes deal = **hospital treat patients unable to pay (community benefit), gov grant tax exemption** - Gov ↑tax revenues to pressure private nonprofit hospitals to justify tax exemption - Community benefits = charity care, unprofitable units, lower prices, community programs - Intentional vague meaning to justify tax exemption - 1\. Should taxes paid by for-profit = community benefit? - A white text with black text Description automatically generated - Big problems - 1\. Assume nonprofit surplus = for-profit taxable profit - Nonprofit surplus = tax exemption, for-profit profit = after tax - 2\. Shrink taxes paid by for-profits = only impact healthcare - Set % of tax to beneficial to healthcare only is wrong, b/c \$ impacts beyond healthcare - 3\. **Can't compare non-profit total tax exemption vs for profit shrunk tax** - **2. Value of uncompensated care vs value of taxes** - No benchmark of uncompensated care (for-profit also provides uncompensated care) - Uncompensated care = calculated with listed price (theoretical), not actual cost (after insurance patient pay much less) - **When should non-profits be allowed to have market power?** - Nonprofits create [benefit] for society + Society [can't achieve that benefit] via market system - Problem: definition of "community benefit" is so vague doesn't really justify tax exemption status - [Inconsistency in law about nonprofit ] - **IRS 501(c)(3)** = exempt nonprofit from paying tax = **nonprofits are different differential treatments** - **Antitrust law** = prevent hospitals perform anti-competitively (M&A to avoid competition) **not distinguish nonprofit vs profit** - Indigent care pool = handle bills for low/middle income patients many nonprofit hospitals drew from indigent care pool but still went after patients - Charity care vs bad debt - **Charity care + bad debt = both uncompensated care** - Charity care = provide care without expectation of payment - Not be rigorous to collect debt - Bad debt = provide care but expect payment - Hospitals sell debt to collection agency to get debt back - **Uncompensated care: Non-profit provide \~ for profit (4%), gov provide most (13%)** - Bad b/c gov wants to finance care, not provide care - For profit provides uncompensated care b/c its nature of business - **[Charity care volume doesn't change] (no real activity), increase in \$ b/c inflation** - **Non-profit hospital expands into areas that have no uninsured (rich areas)** - Catchman area = area that the facility draws its patients or clients - [Increase travel time only effect privately insured], not uninsured - **Non-profit hospitals have same admission rate compared to for-profit hospitals (same patient access)** - **Financing uncompensated care at hospitals** - **Cost shifting / income redistribution (use \$ from rich patient to help poor patient) within organization, not society level** - Hospital CEO tasked with income redistribution of society (critical social mission) they are not trained, and too much pressure **\ ** **Nonprofit Hospitals (3) Main take-aways** - **Q3. Why do [non-profits] dominate [hospital] industry while [for-profits] dominate [nursing home] industry** - 1\. More **unobservable quality** in hospitals than in nursing home - 2\. Hospitals started out as charitable / **altruistic** organizations - 3\. Nursing home / outpatient = not as **heavily regulated** as inpatient / hospitals - 4\. Less **need to hold excess capacity** due to less demand uncertainty - Hospital need to have ppl/room ready for emergency not profitable - Non-profit = concentrated in [hospitals] and [inpatient facilities] - For-profit = concentrated in **[hospice]** and [outpatient] - Hospice shift from non-profit to for-profit b/c **patient in hospice shifted** from (almost dead, short life, no profit, i.e. cancer) to (unknown life expectancy, very profitable, i.e. AD) - **Q4. Do non-profit hospitals justify their tax exemption status?** - **No, because:** - **1. Non-profits provide [\~ uncompensated care volume] than for-profit** - Charity care + bad debt = both uncompensated care - [For profit provides uncompensated] care b/c its nature of business - [Charity care volume doesn't change] (no real activity), increase in \$ b/c inflation - **2. Non-profit hospital [expands into rich areas] that have no uninsured** - **3. Non-profit hospitals have [\~ admission rate] than for-profit hospitals (same patient access)** - Deal = hospital treat patients unable to pay (community benefit), gov grant tax exemption - **When should non-profits be allowed to have market power (tax exemption)?** - Nonprofits create [benefit] for society + Society [can't achieve that benefit] via market system - **Community benefits** - Definition = intentional vague to justify tax exemption - **1. Should taxes paid by for-profit = community benefit?** - Community benefit = % tax + uncompensated care - **\*\*\*Big problems\*\*\*** - [1. Assume nonprofit surplus = for-profit taxable profit ] - Nonprofit surplus = tax exemption, for-profit profit = after tax - [2. Shrink taxes paid by for-profits = only impact healthcare] - 3\. Can't compare non-profit [total tax exemption] vs for profit [shrunk tax] - **2. Can't compare value of uncompensated care vs value of taxes**