CBA study that considered the effectiveness of installing Automatic External Defibrillators in planes. Found that it costed $35,500 /QALY recognized
Considered Medicaid subsidization of ACE inhibitors for heart attacks. $1,606 / beneficiary and 2.3 QALY's for the intervention.
Considered the 500 top live saving interventions. Found that Immunizations and Defibrillators costed the least, while endoscopies were the most.
Fogel and McKeown
The nutrition hypothesis. Rise in nutrition explains 100% of the gains until 1900, 50% of gains after 1900.
Cutler and Miller
Public Health. Filtration and Chlorination -- beginning late 1800s 74% decline in infant mortality, at $50/QALY Vaccination -- late 1800s, early 1900s Pasteurization, Public health measures to improve water quality and reduce typhoid
Neonatal care effectiveness. Regression discontinuity design with a birth weight cutoff at 1500 grams found that neonatal care increased pre-pubescent life expectancy.
Summarizing changes that increased life expectancy over time. Nutrition + Wealth (1750 - 1850) - Public Health (1850 - 1920) - Medical Care (1900s onwards)
Various mortality and life expectancy conclusions over time and space. Found that there is a difference between poor people across regions.
Considers a region where Medicare spends an inordinate amount, because there is a culture of administering a high level of treatment, but outcomes are comparably bad. Doctors are compensated for treatment and not quality, which is problematic.
Environment and lifestyle study that compared UT and NV. Found that NV had much higher excess death rates.
Pritchett and Summers
Instrumental variable study that tested health through wealth using exogenous variation in "good government," since there is some simultaneous causality between health and wealth. Found that a 3% increase in wealth leads to a 1% increase in infant mortality. Criticism was that "good government" might not actually be a good IV.
Cutler and Lleras-Muney
1.7 life expectancy for every year of schooling. When controlling for risky behaviors, effect reduced by 30%. When controlling for income and access to healthcare, effect reduced by 33%. Uses an instrumental variable strategy that compares differences in mandatory laws.
Link and Phelan
Fundamental causes of disease hypothesis, which posits that for certain diseases, such as cancer, income, wealth, or nothing else will impact your incidence of having the disease.
RAND Health Insurance Experiment
Experiment that randomized copayment rates. Found that increased access to health did not have an impact on objective physical outcomes.
Argues that the RAND results are misleading, because even a low level of insurance would encourage an individual to receive health insurance services.
A study of medicare to see if it actually saves lives. Uses a discontinuity design, because people only qualify for medicare after the age of 65. Found a 1% decrease in mortality because of medicare. The study measures the margin that more care is delivered when people are insured, but the issue that he specification is that the control group does not reflect a group that has no insurance.
Preyra and Pink
Canadian hospitals recognized large and unexploited gains due to consolidations.
Dranove and Lindrooth
Merged hospitals found significant savings 2-4 years after consolidation, but integration into systems did not lead to significant savings.
Asymmetric information with the lemons example
Author provides thought experiments regarding adverse selection in health insurance.
Cutler and Zekhauser
Considering the Harvard University death spiral. Both healthy and unhealthy people go into the poorer HMO plans relative to the generous PPO plans.
Zwanzinger and Meirowitz
HMOs prefer larger and non-profit hospitals. Generally, price is not a determinant of whether or not a HMO wants to partner
Framework for healthcare quality: Structure, Process, Outcome.
HMO enrolees have a lower hospitalization rate.
Length of stay is not different for HMO and FFS patients in a hospital.
This experiment randomly assigned people to a HMO or a FFS service. They found that HMO spending was 28% less than FFS spending.
Miller and Luft
HMO provide more preventative tests, but no changed outcomes
Public surveys during HMO backlash in the 90s.
Impact of HMO Backlash. Limited customer exit, limited plan switching.
Tennessee Employment Lock experiment -- people began searching for jobs when public insurance became unavailable.
Gruber and Madrian
Study considered the effects of COBRA. 65% of workers without any other source of insurance use COBRA. Increases likelihood of early retirement by 33% and reduces job lock by 10%.
Cutler and Scott Morton
Often a big metric is a C4 ratio -- the market share of the top four hospitals. Generally, this is about 80%. There has been significant consolidation in hospitals in the last few years. The benefit has been quality improvements. The harm is that there are higher prices, clinicians gain market power through consolidation and raise prices to payers (payers cannot afford to exclude large hospital systems that charge higher prices).
Melnick and Keller
Merged hospitals have higher prices, but not necessarily costs.
There are situations where hospital closures increase economic welfare. Considers the tradeoff between cost and quality improvements due to economies of scale, and the disutility of more travel for patients.
Uses instrumental variable strategy to understand casual effect of mergers on prices. There is a sizable, persistent and positive effect.
Kessler and McClellan
Examined effects of hospital competition on the costs and outcomes for medicare benificiaries.Competition improved outcomes in some cases, but also raised costs. After 1990, there were substantial decreases in costs and substantial improvements in outcomes
Zwanzinger and Mooney
HMOs in NY which until 1996 regulated rates private insurers were required to pay for inpatient care: Result→ After 1996 reforms, HMOs were able to negotiate lower prices with hospitals that were located in more competitive markets.
Cooper and Gaynor
More competition reduced mortality in England (since no cost to patients, quality is the remaining dimension of competition).
Various considerations about transaction prices for hospitals and health outcomes when it comes to negotiated prices. Find that prices vary by a factor of 8 nationally, 3 within a HHR. Finds that negotiated transaction prices are recognized for 27% of the employer-sponsored population. And lastly, hospitals with less competitors are more likely to take discounted prices as opposed to DRGs.
Longitudinal specification to investigate the effect of malpractice settlements on state-level medicare spending. Find that 10% higher per-physician malpractice settlements lead to 1% higher physician Medicare spending, 2.2% more spending on imaging tests. At the end of the day, doctors cover malpractice risk with an extra MRI, not another treatment.
Since physicians both diagnose and treat, they crease their own demand.
"Target Income Hypothesis" -- physicians adjust workload until they hit target income
Suggests limits on surgeon training slots and certificate of need legislation, using inducement of C-sections.
Dranove and Wehner
Sarcastic perspective for inducement of childbirths instead of C-Sections.
Gruber and Owings
Use differential decrease in fertility across US states as identifying variation. Birth by C Section means that the doctor gets more money. They find that induemtn explains 2.9% of 9.1% of increase in C Sections
Top 1% of people take up 20% of medicare expense. Top 50% of people take up 47.5% of medicare expense. It is possible to identify where people go the ER most often as this is often a sign of subpar medical care and costs can be reduced by coordinating outside of hospital care for the most expensive patients.
60% of pharma drugs would not have been developed without patent protection
Acemoglu and Linn
1% increase in the market size leads to 4-6% increase in new drug approvals. Use differences in market size growth as a representative instrumental.
Does the market incentive lead to the "right" kind of discoveries or "me-too" drugs?
Copayments lead to reductions in drug overuse, and an increase in generic use
Big decreases in utilization when the most common drug classes after a doubling of copayments
"Offsets" -- higher drug spending might lead to lower spending for other types of care. Patients with high cost-sharing for drugs are more likely to be hospitalized (stent implantation and anti-platlet drugs)
Medicaid comprises four public insurance programs into one: First: Coverage of most medical expenses for low-income families. Second: Not medicare for low income elderly Third: Low income disabled Fourth: nursing home expenses for the institutionalized elderly
Medicaid granted through lottery -- randomized natural experiments Medicaid coverage reduced cost of care More use of outpatient care, no differences between inpatient care Issues: patients would have recieved this care even if uninsured, and some enrollees drop private insurance when eligible for medicaid.
Reduction in medical debt of between 600 and 1000 when they were enrolled in Medicaid through the ACA Expansion
Gross and Notowidigdo
Using differential changes in Medicaid eligibility criteria -- 10% increase in Medicaid eligibility leads to 8% decrease in personal bankruptcy
Examines the long term impact of medicaid spending. Using a 3% discount rate (and $872 per person annual cost of additional Medicaid eligibility), government recoups 56 cents of every dollar spent on childhood Medicaid by the time the individual reaches age 60
Regression discontinuity design compares cohorts of people that had vs. did not have Medicaid access between the ages of 8 and 13 - results are grouped for 'all races' and 'blacks.' Amongst black children 4.8% of increased eligibility in early childhood leads to 8-12% fewer hospitalizations and ER visits by age 25.
Health Care Quadrilemma. The four factors leading to spending growth are: tech growth, demand for healthcare, insurance coverage, healthcare R&D.
Part D insurance coverage for drugs increases R&D in drugs most likely to be demanded by this population.
Chandra and Skinner
Technology growth and expenditure growth in healthcare
Cutler and Lye
Provider prices reflect higher admin cost -- 40% of cost differences with Canada is due to admin expenses
Patient centered medical homes -- no reduction in costs, utilization or no improvement on 10/11 quality measures.
Modest savings for medicare ACOs after the ACA. 1.4% for first wave, 0% for the second wave
Pay for performance reduced readmissions from 21.5% to 17.8%
Joynt and Jha
There are many factors that influence the readmission rate outside the hospital's region, and P4P simply further penalizes regions with significant disadvantages.
Impact of Increased Deductibles. When considering a very tech saavy workforce and an increase in deductibles, spending down, mostly because of a reduction of care.
Figuring out what determines demographic variation in medicare spending -- practice variations or demand differences? Considered patients who move between different locations and if their spending adjusted to the new location average. Find that 40-50% of variation attributable to demand factors, rest to provider practice differences.