Category Archives: economics

These Aren’t Mixed Results: Pioneer Accountable Care Organizations Worked

by Mike Miesen

After a slew of disheartening press releases from CMS about the Affordable Care Act’s growing pains, the Obama Administration is probably quite happy to see the initial results of one of the law’s most important provisions: Accountable Care Organizations (ACOs). You should be, too.

But the framing of the results may give the typical reader a false impression of what’s actually transpired. From an article by Melinda Beck in the Wall Street Journal titled “Mixed Results in Health Pilot Plan”:

All of the 32 health systems in the so-called Pioneer Accountable Care Organization program improved patient care on quality measures such as cancer screenings and controlling blood pressure, according to data to be released Tuesday by the Centers for Medicare and Medicaid Services. But only 18 of the 32 managed to lower costs for the Medicare patients they treated—a major goal of the effort. Two hospitals lost money on the program in the first year.

Does that sound mixed to you? Let’s restate it a bit, with two charts for emphasis:

All of the 32 health systems in the so-called Pioneer Accountable Care Organization (PACO) program improved patient care on quality measures such as cancer screenings and controlling blood pressure. 18 (56% of PACOs) managed to lower costs for the Medicare patients they treated, saving $140 million in costs; of that, they’ll receive $76 million, and $33 million will be returned to the Medicare Trust Fund. Only two of the 32 PACOs lost money in the first year.

Or, in charts:



Put that way, it sounds a bit more successful, right?

Partners Healthcare, the Boston-area hospital conglomerate, was one of the winners, and will get a check for $7 million from CMS. Of the two PACOs to lose money, one did so, its executive director said, because its cost baseline was artificially low; it will cut a check to CMS for $2 million. Even so, it isn’t one of the nine that are reported to be leaving the program; it’s sticking around.

Make no mistake: these aren’t mixed results. These are unambiguously positive results for one of the most important provisions in the Affordable Care Act.

Should we have expected all 32 PACOs to save money in the first year of a nascent program? Probably not. And they didn’t; 44% either failed to reduce costs or lost money on the program. But, as the director of CMS’s Innovation Center put it, we shouldn’t have expected all 32 to improve patient care, either: “It’s very rare that 100% of the participants outperform benchmarks.” And they did.

This is only the first year of a program that is reinventing the way the American health care system operates. And it worked. Full stop.

Three-quarters stop, actually: it’s important to keep in mind that the PACOs are, well, weird. For starters, they are Pioneer ACOs because they already had ACO-like qualities, so they have a head start on other hospital systems and are perhaps uniquely qualified to initially succeed. Then there’s the fact that they self-selected into a brand-new program that could cost them millions in missed Medicare reimbursements; their finance departments crunched the numbers and guessed that they could make money (or at least wouldn’t lose money). Most were right. And PACOs operate under slightly different rules than the average ACO will, starting in 2014; for the first two years, for example, PACOs are playing under the old fee-for-service rules, then will move to global payments, and only PACOs have two-sided risk – they can share in savings and losses, while typical Medicare ACOs will only share in savings.

In other words, PACOs aren’t “average” health systems, and they aren’t “average” ACOs; it’s obviously premature to say that all ACOs will work.

Still, that doesn’t invalidate these results, and it doesn’t invalidate the hope that ACOs are one policy innovation that can help bend the cost curve in a real, significant way.

More information is set to be released later today, so we’ll have a better handle on the magnitude of the savings compared to total spend, implementation costs, reporting costs, etc. We’ll link the report here after it’s released.


Mike is a healthcare consultant turned aid worker turned traveler (currently: East Africa) and freelance journalist. Follow him on Twitter @MikeMiesen or subscribe to the blog.


Blood Pressure Monitoring, Telemedicine, and Automated Hovering: A Future Model for Disease Management?

by Mike Miesen

Use of an at-home telemonitoring blood pressure device significantly reduced out-of-control high blood pressure, according to a recent study in the Journal of the American Medical Association. It’s another data point showing the potential of telemedicine to have a profound effect on American medicine, by positively modifying health behaviors, providing real-time data to clinicians through “automated hovering,” and helping Americans get and stay healthy – all of which holds the promise of bending the cost curve.

Led by Karen Margolis, MD, MPH, a Senior Investigator at Health Partners Institute for Education and Research, the cluster-randomized study investigated whether using a cloud-connected, at-home blood pressure monitor paired with pharmacist and case manager support would lead to controlled blood pressure more than typical care, which involved check-ups with a physician.

Those using the telemonitoring device were 90% more likely to have controlled blood pressure at both the six and twelve-month checkups than the control group (57.2% and 30%, respectively), and had, on average, statistically significant lower systolic and diastolic readings.

Crucially, the effect persisted after the 12-month intervention. At 18 months – six months after the devices were taken back – those in the telemonitoring group were just as likely to have controlled blood pressure as at 12 months, and were 139% more likely to be in-control for the six, twelve, and eighteen-month checkups than the control group. The learned techniques and behavior changes brought about during the interventions were sticky.


The results are encouraging for the 67 million Americans with hypertension – one-third of all American adults. The disease is associated with 348,000 deaths each year, and a majority with it have trouble maintaining healthy blood pressure levels. It’s a disease of age – the older you get, the more likely you are to have it – so as the Boomers enter Medicare in droves, American taxpayers will increasingly foot the $47.5 billion (and growing) bill for direct medical expenses related to hypertension.


The intervention wasn’t cheap: spend was about $1,350 per participant. But, as a whole, it could be cost-neutral or cost-negative, if it led to fewer emergency department visits and inpatient stays. The study didn’t look at these savings, nor did it identify the cost implications of substituting physician labor with pharmacist labor, a limitation of the study. In an email interview, Dr. Margolis indicated that a more formal cost analysis will be released in the future.

Another limitation of the study is that the population was “generally well-educated with high income levels,” making it unclear if the results can be generalized to the entire American population. Future studies will likely need to enroll a broader cross-section of hypertensive adults.

Understanding why the intervention worked is critical. Dr. Margolis said that she believed “…the combination of frequent self-monitoring and pharmacist action based on the measurements creates a really powerful feedback loop. Both the patients and the pharmacists were getting positive reinforcement with this system.” As patients learn more about their condition and how to manage it, they become partners in care with providers.

Fortunately, interventions like this increasingly make financial sense for hospitals and primary care providers. Provisions in the Affordable Care Act have Medicare experimenting with bundled payments for hypertension, so clinicians will have the flexibility to treat patients optimally, rather than being forced to conform to ossified fee-for-service schemes that reward quantity over quality. Accountable Care Organizations, also a part of the Affordable Care Act, incent keeping patients healthy and, when they do get sick, treatment with effective, low-cost interventions.

Telemedicine allows what Professor Kevin Volpp refers to as “automated hovering:” clinicians can automatically monitor their patients’ conditions during “the other 5,000 hours” when a patient isn’t at the clinic, stepping in when the data shows a need. “I think this method of caring for chronic conditions is going to become much more common as the technology improves and goes down in cost,” Dr. Margolis said.

One concern with automated hovering is that, rather than having too little information about a patient, the clinician could have too much – like drinking water from a fire hose. Dr. Margolis said that providers have “concerns about being responsible for monitoring a massive influx of data from patients.”

But this is ultimately a design problem, not an inherent fatal flaw; data can be whittled down and summarized in a way that clinicians and patients find most useful. As Dr. Margolis put it, “we will have to really be careful to design smart systems to make it easy to interpret the data and to handle values that require urgent attention.”

There’s another reason to be optimistic about telemonitoring and automated hovering: this intervention is basically the simplest iteration of it, with one-way data transmitted to a clinician. It’s possible that an app like Withings’s Blood Pressure Monitor could improve adherence even further by quizzing the patient, offering algorithm-based suggestions, and providing a conduit for secure messaging between patient and clinician.

This space is nascent but growing, fueled by technological innovation, payment system reforms, and studies that show its efficacy. Used as one tool in America’s quest to get healthy and stay healthy, telemedicine may be a crucial component in helping bend the cost curve.


Mike is a healthcare consultant turned aid worker turned traveler (currently: East Africa) and freelance journalist. Follow him on Twitter @MikeMiesen or subscribe to the blog.

People are corporations, my friend. Or something like that.

by Adrianna McIntyre

There’s a ludicrous new meme emerging from Obamacare opposition: it’s not fair to enforce the individual mandate if the IRS is delaying the employer mandate—and Republicans want to level that playing field (a playing field I wasn’t aware existed). This was a central topic during a Ways and Means hearing on Wednesday morning. Rep. Eric Cantor tweeted about it. Legislative proposals are underway.

Just because “individual mandate” and “employer mandate” sound the same doesn’t mean that they are. Equating the two reaches an impressive new level of political theatre—and demonstrates willful ignorance of the motivations behind each mandate.

The individual mandate isn’t about fairness relative to employers. It’s about this chart. Pulled from the landmark paper on adverse selection by David Cutler and Richard Zeckhauser, this is the graphical representation of the dread insurance “death spiral”.


Adverse selection isn’t an especially difficult concept to grasp. With guaranteed issue (telling insurers they can’t discriminate based on pre-existing conditions), sick people are more likely to sign up for insurance. That makes the risk pool—the group of people an insurer covers—less healthy and more expensive. That causes premiums to rise and healthy people drop out, in a cycle that perpetuates until insurance is woefully out of financial reach. This is a market failure, and requiring individuals to purchase insurance works to correct that failure by keeping the young and healthy in the risk pool.

Sam Baker reports the problem with proposals offered by Republicans so far: they aim to delay the individual mandate, but don’t touch guaranteed issue. That’s exactly the recipe for adverse selection. We’ve seen it before—and we’ve seen it fixed by requiring individuals to obtain coverage. Guaranteed issue was a feature of the Massachusetts nongroup insurance market before the state installed their individual mandate in 2007.

Jonathan Gruber found that the average nongroup health insurance premium in the state fell from $8,537 in 2006 to $5,143 in 2009, a 40 percent reduction during a period in which premiums for similar plans rose by 14 percent nationally. (Source)

Time-limited by definition, a one-year delay to the individual mandate wouldn’t lead to a total unraveling of the market. But it also wouldn’t achieve anything other than higher rates for a year, exacerbating the rate shock phenomenon Obamacare critics seized on last month and creating a climate of pointless uncertainty for insurers across the nation.

Meanwhile, the employer mandate is just nifty accounting. I’m not going to spend a lot of time hashing this out, because a number of smart people have already done so (see here, here, and here). The employer mandate doesn’t serve a real policy purpose—not like how the individual mandate is essential to the solvency of health insurance markets.

The employer mandate is an accounting gimmick that contributes to the deficit-neutrality of the Affordable Care Act, writ large. By requiring large employers to offer affordable coverage to full-time employees, the feds shift costs to them and avoid shelling out subsidies. There are compelling arguments that this is bad policy and should be delayed indefinitely, but those arguments don’t translate to the individual mandate. People are not corporations, and a requirement to carry insurance is different from one to provide insurance. Earth-shattering, I know.

This “fairness” ploy to delay the individual mandate is just that: an opportunistic political maneuver intended to destabilize reform efforts. Perhaps there are other credible arguments for pushing it off until 2015. This is not one of them.


Adrianna is a graduate student in public policy & public health at the University of Michigan. Follow her on Twitter at @onceuponA, or subscribe to the blog. 

The “Micro-Voucherization” of Health Insurance: Can Reference Pricing Bend the Curve?

by Mike Miesen

Mitt Romney’s/Paul Ryan’s premium support/voucher plan was heavily derided during the dark days of Campaign 2012, but the devil was always more in the details than the theory. While the re-election of President Obama left premium support dead on the Medicare level, health insurers are increasingly turning to the ideas that drove it – choice, competition, and the power of a (carefully regulated) market – to address high costs on the procedural level. Call it the micro-voucherization of health insurance.

This is known by wonks as reference pricing, and its recent results in California are promising: the costs of hip and knee replacements fell by 19%, with no attendant decrease in quality. Using reference pricing is an assault on the status quo that holds the promise of “bending the curve” in a meaningful way, but it faces technical and political concerns that may consign it to the graveyard of promising-but-unfulfilled ideas.

Broadly-speaking, reference pricing is the act of offering a set amount of money for the purchase of a good, where the reference is an amount that can reasonably said to offer meaningful coverage for that good. Sometimes, reference pricing is focused on a given procedure – what I’ll refer to as “inputs-oriented reference pricing”; other times, a given outcome, or “outputs-based reference pricing.”

That’s pretty vague, so let’s use the colonoscopy procedure (which has recently received a lot of attention thanks to an informative New York Times article) to help color this in. The inputs-oriented approach would see the payer asking: given the choice to have a colonoscopy – a procedure which varies wildly in cost without varying wildly in quality – what’s a reasonable price to pay? It would decide this based on some combination of price, quality, and geography, and would inform consumers of its spending cap.

Say it finds that most of its insured population can reasonably access a high-quality colonoscopy for $10,000; if a consumer choose provider that charges $15,000, he or she would pay the $5,000 difference out of pocket. Choice is preserved, but at a cost. The simple chart below shows how this may work:


But, if you read the colonoscopy article, you may be asking a separate question: why pay for a colonoscopy at all? A fecal occult blood test (FOBT), for example, is just as effective as a colonoscopy for colon cancer screening, but it’s cheaper to perform. The outputs-oriented approach is procedure agnostic, and identifies only what is most cost-effective; if a high-quality colonoscopy in a geographic area is $20,000 and a FOBT is $12,000, the insurer would pay for $12,000 of the total cost; chart below [1]. Again, the consumer has freedom to choose what procedure he or she would like, but would pay the difference out of pocket.


Stylized examples aside, reference pricing is catching on in America. The 19% decrease referenced above came from the California Public Employees’ Retirement System’s (Calpers) use of reference pricing for hip and knee replacements. To accomplish this, Calpers went to individual hospitals and made an agreement: charge no more than $30,000, and the hospital will be included in the health plan. Those that didn’t agree weren’t included in the plan [2].

Safeway, the self-insured grocery chain, has also used reference pricing in limited circumstances, including for colonoscopies and lab tests.

And it’s very common in the formularies of pharmaceutical reimbursement; generics are often covered 100% but brand-name drugs will cost the insured extra. A variety of systematic reviews have found that reference pricing typically leads to reductions in pharmaceutical expenditures without an attendant decrease in health outcomes or increase in physician office visits.

So, if scaled to many procedures, reference pricing looks a bit like an insurance plan that contains a bundle of vouchers: one for hip replacement, one for FOBT, et cetera. It’s the micro-voucherization of health insurance [3].


1. Theoretically, the insurer would first decide a) which procedure is more cost-efficient, then b) use the inputs-oriented approach to define the appropriate reference price for that procedure.

2. This isn’t pure reference pricing, but it provides one path for large insurers to exert downward influence on the price of health services. A purer form would be to simply set the cap – to provide the micro-voucher for a procedure – and let the market drive down costs, as health care consumers choose.

3. To a point, of course; some procedures don’t seem a good fit – emergency surgery comes to mind. It’s not mean to be all-encompassing


Mike is a healthcare consultant turned aid worker turned traveler (currently: East Africa) and freelance journalist. Follow him on Twitter @MikeMiesen or subscribe to the blog.

Summer Journal Club, Week 3: What happens when you force employers to offer insurance?

by Allan Joseph

Welcome to Week 3 of Project Millennial’s 2013 summer journal club. Previous posts on why people want health insurance and why employers provide it can be found here. This week, we’ll study what happens when policymakers force employers to provide health insurance to their employees, using “How do Employers React to a Pay-or-Play Mandate?”, by Carrie Colla et al., and “After the Mandates” by Jon Gabel et al (which cites other research we’ll discuss). Unfortunately, these papers are gated — if someone knows of an ungated version, please let us know and we’ll update.

Last week, we learned that employers who offer health insurance do so because it’s tax-exempt and their employees demand it — and have the bargaining power to get it. That approach, we learned, leaves millions of Americans uninsured while making employer-sponsored insurance (ESI) the core of the American insurance system. That’s why the Affordable Care Act (ACA) used an “employer mandate” as one of its tools for expanding coverage.

The ACA requires any company with at least 50 employees to offer health insurance to any employee who works at least 30 hours a week (though a bill may been introduced to change that to 40). There are two ways a company can be penalized for not offering coverage starting in 2014:

  •  If the company does not offer insurance to at least 95% of employees, and just one employee receives a subsidy (this is the “trigger” mechanism), the company must pay a fine of $2,000 per employee—all employees (technically all employees minus 30), whether or not they sought a subsidy.
  • If the company does offer insurance to 95% of employees but an employee still seeks insurance on the exchange—either because they’re in the small fraction of individuals who was not offered coverage or the plan offered by the employer is skimpy enough that individuals qualify for a subsidy—the employer must pay $3,000 per employee receiving a subsidy.

Those penalties add up quickly, but that’s only part of the story. According to the Kaiser Family Foundation, the average employer health insurance plan in 2012 cost $5,600 for a single person and over $15,000 for a family — far more than the fines for not offering health insurance.

In addition, the money raised from the fines goes to subsidize the cost of insurance for people who buy insurance individually through the ACA’s new “exchanges,” which are highly regulated and allow anyone to purchase insurance. Thus, there’s a worry that employers could drop their health insurance coverage and save a lot of money, while their employers would still have access to coverage — just not the one they have now. Whether or not this phenomenon, termed “dumping,” is actually a problem is a matter of some debate, but it is a concern for many people, especially those who want to minimize disruption in the healthcare sector.

Luckily for us, we actually have a couple good examples of how this would work on a smaller scale, which is what we’ll take a look at today. In 2006, both the city of San Francisco and the state of Massachusetts passed healthcare reform that included a “pay or play” employer mandate: employers of a certain size had to either provide health insurance (play) or pay a fine (pay). Colla et al. studied the San Francisco mandate, while Gabel et al. studied the Massachusetts mandate — and the results may surprise you.

It’s worth mentioning that San Francisco’s mandate was slightly more stringent than the ACA’s version, though it also covers a high-income city with a strong safety net. Massachusetts’ mandate was actually far more worrisome to those who worry about dumping: It covers any firm with at least 20 employees (the ACA starts at 50), and has a fine of less than $300 per employee (as opposed to the ACA’s $2,000). However, the structure of the Massachusetts market and Massachusetts reforms were much like that of the Affordable Care Act. Together, the two experiences should give us a decent window into what the ACA might bring.

In San Francisco’s case, only 25% of employers were in line with the law’s requirements before it took effect, but only 1 in 5 employers took the option to pay, while the rest played, suggesting most employers were willing to offer qualifying health insurance to their employees in order to comply with the mandate. In addition, employers were largely supportive of the mandate and didn’t identify it as a serious source of problems.

In Massachusetts, the experience was even more positive. As this figure from the paper shows, when the mandate took effect in 2008, more firms (even ones that weren’t affected) offered health insurance:


In addition, there wasn’t any noticeable increase in the number of firms who said they planned to drop coverage, corroborating evidence from other studies cited in the paper, which showed that the mandate had no effect on how many employers dropped coverage or, for the most part its generosity. The majority of employers in Massachusetts were also generally happy with the reform package.

These aren’t perfect analogies by any means (the penalties are different, and so are the options for employees whose employers don’t offer health insurance), but they do suggest that there has to be a pretty good reason employers are “playing” when it’s so much cheaper for them to simply pay. It’s hard to say what that reason is for sure, but we can speculate.

One of the reasons may be that dumping employees and paying the fine would cause a lot of discontent — people generally like what they have, and they don’t want to change their health insurance. Another big reason is that when firms competing for the same workers offer health insurance, one firm’s unilateral decision to pay instead of play will make it much harder to recruit workers, who will expect to receive ESI. Finally, Colla et al. suggest an intriguing possibility. The fines companies pay for not carrying health insurance are essentially invisible to the worker, who will then demand a corresponding increase in cash. As we learned last week, employers generally pay for health insurance by reducing wages, but workers who aren’t getting ESI won’t accept the lower wages. Thus, the employer can’t shift the cost of insurance as easily to the employee.

Regardless of the reason these reforms haven’t created as much of a stir as expected, it remains to be seen how the ACA’s employer mandate affects the market — but given the experiences in San Francisco and Massachusetts, it’s probably a good bet that it won’t be as disruptive as it might appear at first.

Next week: Now that we’ve learned quite a bit about health insurance in America, we move to perhaps a more fundamental question: Does health insurance even work?


Allan Joseph is a senior at the University of Notre Dame studying economics and pre-medical studies. You can follow him on Twitter @allanmjoseph.

Tagged , , , , ,