Unbiased Analysis of Today's Healthcare Issues

Leadership Rules

Written By: Jason Shafrin - Oct• 14•14

From a book on economics, complexity theory and economics, below are some good guidelines for managers and leaders developed by the Prussian army:

  1. A commanding officer should always give an order for an outcome, never for an action.  This leads the person receiving the order to reflect on and interpret it with all his prior knowledge and in its relevant context.
  2. The subordinate officer receiving the order was expected to report back to the commanding officer how it was going to be executed, institutionalizing a feedback loop and ensuring that there would be no misunderstandings or any missing information from the top.
  3. Officers were fully expected to deviate from orders from time to time, when they saw that this would be a more sensible course of action.  After all why have highly educated officers, if they are not allowed to exercise their own judgment? “Obedience is a principle, but the man stands above the principle” said Von Moltke in 1870, the Prussian general who designed the system.

The 2014 Nobel Prize in Economics goes to…

Written By: Jason Shafrin - Oct• 13•14

Jean Tirole. The full scientific background is here.  Marginal Revolution has a number of posts or you can check out Wonk Blog.  Digitopoly discusses how Tirole is like Louis Pasteur.  Vox has a nice overview of Tirole’s work as well and an example from newspapers…

Newspapers, for example, are one area the Nobel Committee points to that illustrates this. They often give information away at a discount as a loss-leader to gain market share and increase advertising revenue. While a government might consider that illegal “predatory pricing” in some other context, such an approach doesn’t make sense in a press context.

Vox also has another example on platform markets, including a description why Google’s search technology is free.   An excerpt from the press release is below.

Many industries are dominated by a small number of large firms or a single monopoly. Left unregulated, such markets often produce socially undesirable results – prices higher than those motivated by costs, or unproductive firms that survive by blocking the entry of new and more productive ones.

From the mid-1980s and onwards, Jean Tirole has breathed new life into research on such market failures. His analysis of firms with market power provides a unified theory with a strong bearing on central policy questions: how should the government deal with mergers or cartels, and how should it regulate monopolies?

Before Tirole, researchers and policymakers sought general principles for all industries. They advocated simple policy rules, such as capping prices for monopolists and prohibiting cooperation between competitors, while permitting cooperation between firms with different positions in the value chain. Tirole showed theoretically that such rules may work well in certain conditions, but do more harm than good in others. Price caps can provide dominant firms with strong motives to reduce costs – a good thing for society – but may also permit excessive profits – a bad thing for society. Cooperation on price setting within a market is usually harmful, but cooperation regarding patent pools can benefit everyone. The merger of a firm and its supplier may encourage innovation, but may also distort competition.

The best regulation or competition policy should therefore be carefully adapted to every industry’s specific conditions. In a series of articles and books, Jean Tirole has presented a general framework for designing such policies and applied it to a number of industries, ranging from telecommunications to banking. Drawing on these new insights, governments can better encourage powerful firms to become more productive and, at the same time, prevent them from harming competitors and customers.

Residency Hour Restrictions and Patient Outcomes

Written By: Jason Shafrin - Oct• 12•14

Residency has been a right of passage for physicians. Long 36-hour shifts mix with the excitement of new learning. However, recent regulations have now capped residency weekly hours at 80. Although patient medical care in the short-run may improve (because physicians are less tired), in the long run, are physicians exposed to as wide a variety of cases as before? Consider the following sentiment from Chris Sonnenday, a former fourth year resident:

You might expect that, as residents, we’d stand up and rejoice that these regulations have been passed. But I’ll tell you, if you’re the chief resident on the GI service and a case comes up that you may have one or two opportunities during your entire residency to do … well, many of us have to be dragged kicking and screaming out of the hospital.

More important than the doctor’s feelings about residency requirements, how did the hours restrictions affect long-run patient quality of care? This is the question Jena, Schoemaker and Bhattacharya (2014) attempt to answer in their recent Health Affairs paper.

…we studied whether hospital mortality and patients’ length-of-stay varied according to the number of years a physician was exposed to the 2003 duty-hour regulations during his or her residency. We examined this database of practicing Florida physicians, using a difference-in-differences analysis that compared trends in outcomes of junior physicians (those with one-year post-residency experience) pre- and post-2003 to a control group of senior physicians (those with ten or more years of post-residency experience) who were not exposed to these reforms during their residency. We found that the duty-hour reforms did not adversely affect hospital mortality and length-of-stay of patients cared for by new attending physicians who were partly or fully exposed to reduced duty hours during their own residency.



Written By: Jason Shafrin - Oct• 09•14

Joe Paduda has posted this week’s Health Wonk Review and keeping with the overall mission, he says that there’s “lots of health policy stuff, all right here!”

How important is having a good nurse?

Written By: Jason Shafrin - Oct• 08•14

Perhaps unfairly, doctors get all the credit. When patients have better health outcomes, physicians are rewarded; when they have worse health outcomes, the physician is blamed. However, the quality of nurses also likely affects the quality of care that patients receive. A recent paper by Yakusheva, Lindrooth and Weiss (2014) finds that nurse quality does affect outcomes. Using an EHR-based measure of patient health (the Rothman Index), the authors find the following:

Nurse effects were jointly significant and explained 7.9 percent of variance in patient clinical condition change during hospitalization. NVA was positively associated with having a baccalaureate degree or higher (0.55, p = .04) and expertise level (0.66, p = .03). NVA contributed to patient outcomes of shorter length of stay and lower costs.

One issue not addressed is nurse assignment to patients. The empirical approach basically assumes that nurse assignment to patients is uncorrelated with nurse quality. In reality, the highest quality nurses may take the most difficult cases in some hospitals. One might believe that the highest quality nurses would be assigned to case where they would have the largest improvement in outcomes. However, management’s utility function may not be linear in patient outcomes. For instance, management may assign high-quality nurses to critically ill patients and these high-quality nurses may move the patients from critically ill to sick; a feat that may not be feasible for low-quality nurses. Low quality nurses could move moderately ill patients to excellent health, but this feat may be easier than moving patients from critically ill to sick. Thus, if risk adjustment is inadquate, high quality nurses may be more valueable even then the figures from this study state.

One conclusions that the authors incorrectly draw is that their results “…strengthen the body of evidence in support of strategic federal, state, and health system initiatives to build toward a baccalaureate-educated workforce.” Although it is true that nurses with a BA are more likely to be high-performing nurses, there is no evidence that getting a BA has a causal effect on outcomes; nurses who were motivated to get a BA were likely going to be the best nurses regardless of the education received.

Although these results are not surprising–high-quality nursing should clearly result in better patient outcomes than low-quality–this study does make the important point that physicians are not the only healthcare providers who play an important role in hospitalized patients’ health outcomes.

Do specialty drugs provide good value?

Written By: Jason Shafrin - Oct• 07•14

Specialty drugs have greatly improved quality and duration of life for many patients.  Specialty drugs have been developed for a variety of illnesses, such as cancer, hepatitis C, rheumatoid arthritis, and multiple sclerosis.  However, these drugs are typically very expensive.  Are they worth the cost?

A paper by Chambers et al. (2014) find the following:

…specialty drugs offered greater QALY gains (0.183 versus 0.002 QALYs) but were associated with greater additional costs ($12,238 versus $784), compared to traditional drugs. The two types of drugs had comparable cost-effectiveness. However, the distributions across the two types differed, with 26 percent of specialty drugs—but only 9 percent of traditional drugs—associated with incremental cost-effectiveness ratios of greater than $150,000 per QALY. Our study suggests that although specialty drugs often have higher costs than traditional drugs, they also tend to confer greater benefits and hence may still offer reasonable value for the money.


Pharmaceutical Cost Offsets

Written By: Jason Shafrin - Oct• 06•14

Does increasing use of prescription drugs decrease cost?  There is evidence on both sides of the aisle.  However, a recnet paper by Deb, Trevedi and Zimmer (2014) does find evidence of cost offsets using a sophisticated copula-based bivariated dynamic hurdle model.  This model, models drug and non-drugs spending and for each dimensions models the probability of $0 spending using a probit and then models log spending levels conditional on positive spending using a gamma distribution.  Using Medical Expenditure Panel Survey data, the authors find:

There is evidence of greater than dollar-for-dollar cost-offsets of expenditures on prescribed drugs for relatively low levels of spending on drugs and less than dollar-for-dollar cost-offsets at higher levels of drug expenditures.

This is a perfectly sensible conclusion. Effective drugs do reduce medical spending. Likely, cost-offset are smaller when a drug is branded.  For instance, consider an effective drug that reduces hospitalization costs by $10,000 but costs $12,000. This is a less than dollar-for-dollar offset. Now assume that after a few years the drug becomes generic and costs only $500. In this case, the reduction in hospilization cost is still $10,000 so there is a more than dollar for dollar offset. Some people may interpret this finding as that generic drugs are more cost effective than branded drugs. While this may be drug in some static sense, without the higher prices for brands, innovative new drugs would not come to market. It would be interesting to see whether over the life of a drug (say 20-30 years) are the cost offsets more or less than dollar for dollar when taking into account both the time the drug spent on and off patent.

Regardless, the Deb paper clearly shows that many drugs do produce cost offsets.


What does an ACO look like?

Written By: Jason Shafrin - Oct• 05•14

ACOs are all the rage, but what does a typical ACO look like ? A study by Shortell et al. (2014) attempts to answer this question using data from the National Survey of ACOs.

This survey includes ACOs participating in Medicare, Medicaid, or commercial payer programs.  The authors consider eight attributes for creating an ACO taxonomy including: “ACO’s size, number of different types of participating provider organizations within the ACO (including nursing or postacute care facilities), the scope of services offered, whether the ACO belongs to an integrated delivery system (IDS), the percent of primary care clinicians, their institutional leadership model, the performance management system used for accountability, and the ACO’s prior experience with payment models other than fee-for-service.”

They classify ACOs into three separate types:

  • Larger integrated delivery systems (40%). Largely self-identifying as an integrated delivery system, most of these ACOs are large (mean of 566 FTE physicians) and offer a broad scope of services.  About 40 percent of these organizations are physician-led.  Further, about 30 percent of ACOs in this category have postacute facilities. “These ACOs have the most experience with payment reform but are relatively lower on their use of performance management/accountability mechanisms.”
  • Smaller, physician-led ACOs (34%). These ACOs are physician-lead and are not associated with an IDS. They offer a narrow scope of services, with fewer doctors (mean of 181 FTE physicians)  and few include nursing facilities.  Most have many PCPs, but little experience with payment reform.   They have a high degree of performance management/accountability in place.
  • Hybrid ACOs (26%). A hybrid of the two categories above, these groups are of moderate size (mean of 351 FTE physicians) and a moderately large range of services, with about 3 in 10 including post acute services.  “They tend to be hospital-led, coalition-led, state/region/county-led, or some other arrangement…” and these entities have some experience with payment reform and relatively low scores on performance management/accountability.

In summary, one should not think of ACOs as a single simplistic entity as this study highlights some of the diversity–as well as some of the commonalities–across ACOs.

The authors also describe relative ACOs performance on cost and quality measures across these three groups.

Weekend Reading

Written By: Jason Shafrin - Oct• 04•14

Funding for Digital Medicine on the rise

Written By: Jason Shafrin - Oct• 02•14

What is digital health?  Although a single definition is hard to fine, digital health uses data collected from patients to improve care.  Mobile health, the use of electronic medical records, the “quantified self” movement and various forms of health information technology all count as digital health.

Is digital health just a fad or will it revolutionize health care?   If you look where the money is going, a good bet is on the latter. WonkBlog reports:

Companies in the Big Data and analytics space are the biggest fundraisers this year, with 77 deals totaling $1.17 billion, according to StartUp Health. That’s followed by companies that help navigate the health-care system (40 deals, $794 million) and practice management (26 deals, $598 million).

This isn’t just a San Francisco phenomenon either, though that city leads the country in digital health funding this year (62 deals, raising $948 million). It’s followed by New York ($584 million, 33 deals) and the D.C./Maryland area ($536 million, 14 deals).

The question seems now whether or not digital health will be the wave of the future, but rather exactly how it will affect the care provided to patients in the US and around the world.