Unbiased Analysis of Today's Healthcare Issues

Patient Reported Quality

Written By: Jason Shafrin - Oct• 18•15

Should doctors financial incentives be tied to patient perceptions of quality?  Some physicians are reluctant for this to happen.  Paula Chatterjee and co-authors, however, argue that patient satisfaction should play a role in how incentives tied to value-based care are measured.

Although I do not disagree with Chatterjee, it is not clear how patient perceptions of quality would not affect physician compensation.  Even if reimbursement rates were the same for all physicians, patients often find doctors they like by talking to friends or getting referrals from other physicians.  If patient satisfaction is low, your friends are not likely to refer you to that doctor; further, physicians don’t like referring patients to a specialist when the patient comes back to them unhappy.  Thus, even if patient satisfaction is not always directly tied to the reimbursement per visit or per procedure, patient satisfaction certainly affects patient volume and thus physician revenue and profitability.

It could be the case that capacity constrains mean that all physicians except the very worst can have a full load of patients since all the best physicians have long wait times for visits or procedures.  Nevertheless, at least on the margin, patient satisfaction must affect physician profitability and revenue.

One way to balance supply and demand is to allow physicians more ability to vary the prices they charge.  For Medicare patients, this would mean allowing balance billing.  Balance billing occurs when the insurance covers a flat rate, but physicians are allowed to charge a higher rate which the patient would owe out of pocket.  Physicians in this case would have an incentive to provide high quality in order to attract new patients and also to justify a high price above the minimum amount charged by insurance.

The current administratively-based measure of value is fairly crude and likely does not take into account quality.  Incorporating patient satisfaction is a good idea, but a better idea is to let patients themselves evaluate quality of care and to allow high-quality physicians to charge higher prices to make sure the market clears with fewer shortages.


ICD-10 Oddities

Written By: Jason Shafrin - Oct• 16•15

On October 1, Medicare switched from ICD-9 to ICD-10 diagnosis coding.  As a researcher, I appreciate that ICD-10 will give a much more granular representation of a disease.  However, the risk is that the level of granularity is so fine that the administrative costs on providers becomes high.  Some of the diagnose codes, in fact, seem downright odd.  Below are a list of some of the odder ICD-10 codes.

  1. R46.1: Bizarre personal appearance. Of all the crazy codes, this has to be one of our favorites. We’re not exactly sure what constitutes a person’s appearance as “bizarre.” That’s the doctor’s call. Most of Lady Gaga’s outfits are pretty bizarre, but that may not count as a medical condition (or maybe it does).
  2. V91.05XA: Burn due to canoe or kayak on fire, initial encounter. It’s difficult to imagine that a vehicle immersed in a body of water could catch on fire but, apparently, it happens.
  3. W16.222D: Fall in (into) bucket of water causing other injury, subsequent encounter. The picture that comes to mind is a daredevil high-diver plummeting from a 100-foot platform at a circus. It’s pretty much a given; some injury is bound to occur.
  4. W22.01XA: Walked into wall, initial encounter. Let’s face it. We all have those days when we just can’t find the door, right?
  5. W22.02XD: Walked into lamppost, subsequent encounter. Walking into a wall is one thing, but walking into a lamppost? And by “subsequent,” are we suggesting that this isn’t the first time it’s happened? (Yes, we know this means a subsequent encounter with a physician, but don’t spoil the fun, okay?)
  6. W53.21XD: Bitten by squirrel, subsequent encounter. Maybe the patient thought it was their cat? Prescription: get an eye exam.
  7. W55.42XA: Struck by pig, initial encounter. Heaven knows we don’t want our patients to suffer subsequent pig-striking encounters. One is bad enough.
  8. W59.29XS: Other contact with turtle, sequel. It’s a given that most contacts with turtles don’t result in injuries. But if the initial “other” type of contact was epic, certainly, it deserves a sequel.
  9. W56.29XD: Other contact with orca, subsequent encounter. We can’t imagine any contact with an orca would turn out well, much less a subsequent one.
  10. Y92.131: Mess hall on military base as the place of occurrence of the external cause. We all know high school and college cafeterias serve “mystery meat.” Perhaps military mess halls do as well.
  11. V95.42XD: Forced landing of spacecraft injuring occupant, subsequent encounter. NASA might have something to say about this. But, honestly, what doctor wouldn’t want to treat Buzz Aldrin?
  12. V80.731A: Occupant of animal-drawn vehicle injured in collision with streetcar, initial encounter. This could happen — were it the 1890s! Which makes us wonder, is there an ICD-10 code for “Uber-driver injured in collision with taxi?”
  13. Y92.834: Zoological garden (Zoo) as the place of occurrence of the external cause. Make certain not to stick your finger in the macaw cage. (And, yes, there’s a code for injury resulting from an encounter with a macaw: W61.11XD.)
  14. Y92.311: Squash court as the place of occurrence of the external cause. Better to play racquetball; the racquets are shorter. Tennis is even better. At least there players are separated by a net.
  15. Y92.146: Swimming pool of prison as the place of occurrence of the external cause. One has to imagine there are much more dangerous places where a prisoner is likely to get injured than a swimming pool.
  16. Y93.D1: Activity, knitting and crocheting. Anything involving needles can result in injury. And who has time to crochet anyway?
  17. W27.4XXD: Contact with kitchen utensil, subsequent encounter. Injury by spoon?
  18. V97.33XD: Sucked into jet engine, subsequent encounter. Birds getting sucked into a jet engine we understand, but a person? (The encounter itself definitely isn’t funny, but we’re talking about ICD-10 codes. And you have to admit, this one is.)
  19. W61.62XD: Struck by duck, subsequent encounter. Anyone who has ever taken a stroll by a lake in a public park can relate to this encounter, which is why you should have insurance from Aflac.
  20. Y92.72: Chicken coop as the place of occurrence of the external cause. This is what happens when you attempt to get an egg from under a hen with a bad attitude.

Income and weight gain

Written By: Jason Shafrin - Oct• 14•15

Cross sectional analysis finds that individuals with lower income are more likely to be overweight or obese. Does this imply that increased income causes weight loss?

A paper by Au and Johnston (2015) find the opposite result.

In this paper, we use nationally representative panel data and exogenous wealth shocks (primarily inheritances and lottery wins) to shed light on this issue. Our estimates show that wealth improvements increase weight for women, but not men. This effect differs by initial wealth and weight—an average-sized wealth shock received by initially poor and obese women is estimated to increase weight by almost 10 lb.

Why are poor people likely to be overweight but the authors find that increased income causes weight gain?

I have two explanations.  First, although the wealth shocks the authors identify are generally exogenous, they may have an adverse affect on activity or diet.  Inheritances or lottery wins increase the likelihood of removing oneself from the labor force, which may have adverse effects on diet and exercise.  A change in income (e.g., a 10% raise) would likely not have a similar effect; since it is a flow rather than a stock payment, pay raises are only realizable if one does not remove oneself form the labor force.

Second, weight is likely affected by one’s peers.  Previous studies have shown that your peer’s weight affects your weight.  Higher income households may have more social pressure to maintain their weight than lower income household.  This is a statement about our current time and place as in other cultures or time periods wealthy households may value overweight individuals more.  Inheritances or lottery wins do improve income, but likely have a smaller effect on one’s peers.  Inheritances in particular are likely to be received by older individuals who are less likely to change their peer group.

Thus, I agree with the authors that the marginal effect of income on weight is positive, but the driving force of an individual’s weight—besides genetic factors—is likely the preferences of their peers.


Innovation: The good and the bad

Written By: Jason Shafrin - Oct• 13•15

Our data, for example, show that a third of Medicare’s spending in physician or outpatient settings in 2012 reflects technology that did not exist a decade earlier…When it comes to technology development, the central challenge is to encourage high-value innovation while discouraging innovation that drives up costs without much improving health.

Angus Deaton wins nobel prize in economics

Written By: Jason Shafrin - Oct• 12•15

You can find the Nobel website press release here:

The work for which Deaton is now being honored revolves around three central questions:
How do consumers distribute their spending among different goods?Answering this question is not only necessary for explaining and forecasting actual consumption patterns, but also crucial in evaluating how policy reforms, like changes in consumption taxes, affect the welfare of different groups. In his early work around 1980, Deaton developed the Almost Ideal Demand System – a flexible, yet simple, way of estimating how the demand for each good depends on the prices of all goods and on individual incomes. His approach and its later modifications are now standard tools, both in academia and in practical policy evaluation.

How much of society’s income is spent and how much is saved?
To explain capital formation and the magnitudes of business cycles, it is necessary to understand the interplay between income and consumption over time. In a few papers around 1990, Deaton showed that the prevailing consumption theory could not explain the actual relationships if the starting point was aggregate income and consumption. Instead, one should sum up how individuals adapt their own consumption to their individual income, which fluctuates in a very different way to aggregate income. This research clearly demonstrated why the analysis of individual data is key to untangling the patterns we see in aggregate data, an approach that has since become widely adopted in modern macroeconomics.

How do we best measure and analyze welfare and poverty? In his more recent research, Deaton highlights how reliable measures of individual household consumption levels can be used to discern mechanisms behind economic development. His research has uncovered important pitfalls when comparing the extent of poverty across time and place. It has also exemplified how the clever use of household data may shed light on such issues as the relationships between income and calorie intake, and the extent of gender discrimination within the family. Deaton’s focus on household surveys has helped transform development economics from a theoretical field based on aggregate data to an empirical field based on detailed individual data.

N.Y. Times and Washington Post coverage provide additional thoughts.
Marginal Revolution has more coverage as well.

“Best price”

Written By: Jason Shafrin - Oct• 11•15

What price does Medicaid pay for drugs? The answer is, the “best price”.  What is “best price?”

The 1990 Omnibus Reconciliation Act (OBRA) codified that pharmaceutical manufacturers must give Medicaid steep discounts in order to receive coverage by state Medicaid agency.  How big are the discounts?  The answer is the larger of a fixed percentage of the average manufactuerers price (AMP) or the difference between the “best price” and AMP.  The best price is defined as the is the lowest manufacturer price paid for a drug by any purchaser (defined by the Medicaid statute as “any wholesaler, retailer, provider, health maintenance organization (HMO), or nonprofit or government entity…).”  Some payers–such as the VA, Department of Defense, and Indian Health Service–are exempted from the list of entities considered in the best price calculation.

Best price sounds like a good idea.  It lowers the price of drugs for people who need it most.  But is it really?

One drawback is that it limits the price discounts that can be paid to large commercial insurers.  This may be less of an issue if society feels that Medicaid should get the best deal.

However, it also limits the ability for drug manufacturers to engage in risk sharing agreements and other creative contacting forms.  For instance, drug manufactures may want to charge a high price for patients who derive significant benefits from a drug and a lower price for those who receive moderate benefits.  Selling a drug at a lower price for a specific subpopulation in the private market, however, could trigger the “best price” provision which would lower the drug price received by all Medicaid patients, even if they were in the group that highlight benefited from the treatment.  Additionally, creative risk sharing contracts where drug manufacturers contract based on health outcomes become increasingly risky if poor performance for patients at a single payer could result in a low best price for all Medicaid patients.

Thus, what was motivated by good intentions–provide Medicaid patients with lower cost, affordable medications–has a number of unintended consequences.

HWR is up

Written By: Jason Shafrin - Oct• 09•15

Joe Paduda has posted Health Wonk Review’s overflowing cornucopia! at Managed Care Matters.  What’s in this cornucopia? ACA, FDA, ACO, GOP, CEO and much more.  Check it out!

Friday Links

Written By: Jason Shafrin - Oct• 08•15

The end of dose-based drug pricing?

Written By: Jason Shafrin - Oct• 07•15

A Health Affairs post by Dana Goldman and Darius Lakdawalla argues that dose-based pricing for pharmaceuticals is suboptimal.  They make a clear distinction between typical goods, where cost and benefits are roughly proportional to quantity consumed, and pharmaceuticals.

Buying two bunches of bananas naturally costs twice as much as one bunch. Twice as many bananas can feed twice as many people. However, why should a patient who responds best to a 100 mg injection pay twice as much as another who happens to need a 50 mg version?

The 100 mg patient rarely receives twice as much value as her 50 mg peer. Even worse, charging more to patients on higher doses discourages physicians from titrating dosage upward, even when it is clinically warranted. What’s more, PPD forces us into a copayment model where patients are penalized for better adherence.

What is the solution?  The authors recommend tying pricing to value.  Higher value treatments should have a higher price; lower value treatments should have a lower price.  The interesting feature of their proposal is that value would not just be measured across treatments but also by type of patient who receives the treatment.  They use new cholesterol lowering drugs–known as PCSK9 drugs as an example.

The best solution may be to reimburse pharmaceutical companies for PCSK9i therapies on the basis of heart disease risk—something cardiologists are already good at classifying—and to eliminate copayments per prescription regardless of patient risk. The highest risk group includes those with genetic disorders that elevate their cholesterol to dangerous levels, and who develop heart disease at a very early age. For these patients, the currently announced price of about $13,000/annually is a great deal.

But the answer is not to restrict the drugs just to this group. For other high-risk patients with less elevated cholesterol—e.g., atherosclerotic cardiovascular disease patients for whom statins lower LDL significantly, but not all the way to goal—a different, lower price should apply. With this differential pricing, payers would no longer have incentives to limit coverage. And, prescribing decisions would focus on the clinically optimal way to lower LDL, instead of on the least expensive dosing strategy.

A key barrier to this type of pricing strategy is assessing clinical risk and measuring outcomes.  Nevertheless, CMS is already planning to have 90% of payments tied to value by 2018.


Why research in health economics is particularly interesting

Written By: Jason Shafrin - Oct• 06•15

The Federal Reserve Bank of Minneapolis has a long but very interesting interview with Amy Finkelstein, one of the preeminent health economist of our time.    The interview covers a number of topics including the correlation between adverse selection and risk aversion, annuity markets, geographic variation in health care spending, the Oregon Health Insurance Experiment, and other. Despite these various topics, I wanted to highlight one particular excerpt giving an overview of why studying health economics is so interesting.

There are areas of economics that are incredibly important and the policy world has not caught up, but where the economists are mostly in agreement on what the optimal solution is. But what’s exciting to me about this work on health care delivery is, well, if you made me king of the world, I wouldn’t actually know what we should do.

The constraints in health care delivery aren’t just constraints of the political process; there are a lot of real intellectual constraints. There’s a lot we don’t yet know about how best to design these systems, and that makes it an extremely fun and exciting area to work in and to advise students in.

In fact, this rationale–that health economics has so many unsolved research questions–is one of the main reasons I decided to focus my research in this area.