Aaron E. Carroll's Blog, page 139

Healthcare Triage: The Trump Budget and Meals on Wheels

I was on vacation and off the grid last week. So I missed posting on some stuff. Like this great HCT on Meals on Wheels. Go watch it!

This episode was adapted from a column I wrote at the Upshot. Links to further reading and references can be found there.

@aaronecarroll

Block grant funding of public health insurance: the Canadian example

Speaker Paul Ryan wants to reform Medicaid by “block granting” the program, that is,

by capping federal funding and turning control of the program over to states. The aim of such reforms is to reduce federal funding over the long term, while preserving a safety net for needy, low-income Americans. An additional valuable aim of this effort has been to advance federalism by reducing the federal government’s role and giving states and governors more freedom and flexibility in managing their Medicaid programs and helping people in their states.

What are the likely consequences of block granting? Benjamin Sommers and David Naylor write in JAMA about how Canada’s joint federal/provincial funding of health care provides lessons about the likely consequences of block granting.

Canada is a single payer health care system. However, there isn’t a Canadian single payer. Rather, there is a single payer for each province: I am covered by the Ontario Health Insurance Plan (OHIP). These plans are primarily funded by provincial taxes. However, provinces also receive a health transfer from the Canadian federal government, i. e., a block grant. The provincial health insurance plans are run by provincial health ministers, not the federal minister in Ottawa.

So, does provincial autonomy facilitate experimentation and tailoring by the provinces? Sommers and Naylor think not.

there is little evidence that the alleged advantages of block grants have materialized in Canada. Advocates argue that with greater flexibility and proper incentives, states can reduce costs by improving the efficiency of care. In Canada, however, the provinces’ primary means of coping with budget pressures under block grants has been to reduce funding to hospitals and bargain harder with provincial medical associations. Ironically, then, if this scenario plays out in the United States, it would exacerbate one of the chief Republican criticisms of Medicaid — that it pays clinicians such low rates that they have reduced incentives to care for low-income patients.

Indeed, physician refusal to take Medicaid patients is one of Speaker Ryan’s central criticisms of Medicaid.

What about the effects of a block grant system on federal funding of health care?

Once block funding was initiated in 1977, health care funding became a line item in the federal budget that could be arbitrarily cut or capped for fiscal or political reasons, as opposed to a level of spending pegged to the needs and health care use of the population. Importantly, these cuts occurred under both conservative and liberal federal governments.

When the Canadian health transfer began, the federal government paid 50% of provincial costs. However, the transfer has steadily declined, until it is now about 20%. Sommers and Naylor predict that US federal block grants would also decline, and this is clearly one of Speaker Ryan’s goals.

However, Canadian health care spending per capita has not declined.

Graph from the Fraser Institute.

As the cost of providing care has risen, but the federal health transfer has stayed fixed or declined, the provinces have taxed more and the federal government has taxed less. The provincial governments hate this, because they would rather have the federal government make the unpopular choice to raise taxes. But it’s not clear whether block granting has made a big difference in the health care received by Canadians.

American states could similarly increase taxes in response to a declining federal Medicaid block grant, but would they? The key difference between Canadian public health insurance and Medicaid is that the former is universal, while the latter is means-tested. Ontarians prefer lower taxes, but if Ontario decreases funding for OHIP, every Ontarian will experience longer waits for care. But American states can cut Medicaid — and reduce taxes — without affecting the health care of better off and able-bodied citizens.

The affluent and able-bodied are also the citizens most likely to vote. American states determine their own voting procedures. Block granting gives states an incentive to manage voting so as to reduce the participation of the marginalized communities who are most in need of public health insurance. Block granting is likely to undermine the health care for the poor and disabled, and it could reinforce the post-Shelby County v. Holder efforts to restrict voting.

@Bill_Gardner

Too little, too late

In their latest amendment to the American Health Care Act, House Republicans have created something called an “invisible risk sharing program.” The amendment is befuddling. The invisible program is a minor tweak that won’t improve the AHCA’s dismal coverage numbers. It’s not even really a program. If there’s any prospect at all of salvaging Republican-style repeal and replace, this newest amendment isn’t it.

The statutory text is spare. It appropriates $15 billion over nine years—or $1.67 billion each year—and tells the Secretary of Health and Human Services to use the money “to provide payments to health insurers with respect to claims for eligible individuals for the purpose of lowering premiums for health insurance coverage offered in the individual market.” The Secretary can supplement that funding with any money from the AHCA’s high-risk pools that states don’t find a way to use.

Beyond that, however, the statute tells us next to nothing about how the program is supposed to work. Hilariously, a section of the statute titled “Details of Program” contains no details. It says, for example, that the program should include “[a] definition for eligible individuals,” but leaves the defining up to HHS. So too with “[t]he identification of health conditions” that, if an eligible person has them, would qualify her insurer for extra payments.

Oh, and the program is supposed to be in place in time for the 2018 plan year.

Read generously, this newest amendment tells HHS to create a kind of reinsurance program for insurers who enroll high-cost individuals. The statute doesn’t use the word “reinsurance,” maybe because Republicans have spent years railing against the risk corridor and reinsurance programs as insurer bailouts. But if those were bailouts, then this is too.

Judging from the title, the program is supposed to look something like the proposal pioneered by Maine and described in this Health Affairs post. But Republicans are delusional to think that the Secretary can establish and implement a complex reinsurance-style program in time for the 2018 plan year. Insurers that want to participate on the exchanges have to submit bids to HHS by June 21. Even if the AHCA passed tomorrow—which it won’t—there’s no chance that Secretary Price could ramp it up in time.

Nor does the amendment explain how the new program is supposed to interact with the ACA’s risk adjustment program, which the AHCA leaves in place. The point of risk adjustment is to equalize risk across insurers: those with healthier-than-average enrollees have to pay into a central kitty, and those with sicker-than-average enrollees get some of that money. But if insurers get “invisible” risk sharing money for high-cost individuals, should they get less in risk adjustment money? The amendment doesn’t say.

In any event, the money is too insubstantial to make much of a difference. Sure, $1.67 billion per year sounds like a lot of money. But $1.67 billion is chump change compared to the subsidy reductions that are contemplated under the AHCA. It’s like using a band-aid to treat a gunshot wound.

@nicholas_bagley

JAMA Forum: Safe injection facilities

Last winter, the mayor of Ithaca, New York, Svante Myrick, proposed to provide a safe and legal space in which people could inject heroin. It may sound like a radical and desperate way to reduce the harms of drug use. But its effectiveness—and cost-effectiveness—is well supported by research.

So begins my latest post on the JAMA Forum. It’s probably got more links to studies than anything I’ve ever posted there. Take a look!

 

AcademyHealth: Referral patterns and hospital-owned physician practices

According to a recent study, when hospitals employ physicians, patients may not benefit. More in my new AcademyHealth post.

@afrakt

 

Diversion of stimulants prescribed to adolescents

Too many Americans are dying from opioid overdoses and too many of those opioids were pharmaceuticals that had been diverted to non-medical use. According to National Institute on Drug Abuse Director Nora Volkow,

opioid analgesics are widely diverted and improperly used, and the widespread use of the drugs has resulted in a national epidemic of opioid overdose deaths and addictions. More than a third (37%) of the 44,000 drug-overdose deaths that were reported in 2013 (the most recent year for which estimates are available) were attributable to pharmaceutical opioids; heroin accounted for an additional 19%.

However, opioids are not the only controlled substances that are commonly diverted: it’s also a large problem with benzodiazepines, sleep medications, and stimulants. We have known about the diversion of stimulants prescribed for attention-deficit/hyperactivity disorder [ADHD] for a long time. But I had no idea just how big the problem is.

Sean McCabe and colleagues have a recent article on stimulant diversion in the Journal of the American Academy of Child and Adolescent Psychiatry.

Objective

To assess the prospective 17-year relationship between the medical and nonmedical use of prescription stimulants during adolescence (age 18 years) and educational attainment and substance use disorder (SUD) symptoms in adulthood (age 35 years).

Method

A survey was self-administered by nationally representative probability samples of US high school seniors from the Monitoring the Future study; 8,362 of these individuals were followed longitudinally from adolescence (age 18, high school senior years 1976−1996) to adulthood (age 35, 1993−2013).

In these samples, 21% of kids reported using a stimulant medication at least once, including both usage as prescribed and diverted (non-medical) uses. But among those kids who used a stimulant, the kids who used them only as prescribed were a small minority.

It’s striking that there were almost three times as many kids who only used diverted stimulants as there were kids who only used them as prescribed. Keep in mind that this is not a clinical or convenience sample: the authors used nationally representative probability samples. This is as good a description of stimulant use in the US adolescent population as we are likely to get.

Is non-medical use of stimulants by teenagers a major problem? Unlike diverted opioids, few people die from using diverted stimulant medications. This study does not have a design that can identify a causal effect of non-medical stimulant use. Nevertheless, the descriptive findings of this study are not encouraging.

Among past-year adolescent nonmedical users of prescription stimulants, 97.3% had used at least one other substance during the past year.

Moreover, the Monitoring the Future data allowed McCabe et al. to follow the high school seniors to age 35. The kids who used diverted stimulants as adolescents were doing less well as adults.

Medical users of prescription stimulants without any history of nonmedical use during adolescence did not differ significantly from population controls (i.e., non−ADHD and non−stimulant-medicated ADHD during adolescence) in educational attainment and SUD symptoms in adulthood. In contrast, adolescent nonmedical users of prescription stimulants (with or without medical use) had lower educational attainment and more SUD [substance use disorder] symptoms in adulthood, compared to population controls and medical users of prescription stimulants without nonmedical use during adolescence.

The takeaways from this study are, first, that from the mid-seventies to the mid-nineties about 1 in 5 US high schoolers used stimulants, medically or non-medically. Second, many more kids used diverted stimulants than used prescribed ones. Lastly, the kids who used diverted stimulants didn’t get as far in school and had more adult SUD problems. We can’t say that use of diverted stimulants caused these problems. Among other reasons, essentially all of kids using diverted stimulants also used other drugs. We can say that using diverted stimulants is a good marker of a kid who may be in trouble.

The finding that far more kids use stimulants non-medically than medically — coupled with the high rates of opioid diversion — means that our system for supplying pharmaceuticals is broken. In the future, I am going to have difficulty saying “`controlled’ substances” without sardonic quotes around ‘controlled’.

@Bill_Gardner

Expanded coverage has pushed health services employment up by roughly 240,000 jobs

This post was coauthored by Charles Roehrig, Ani Turner, and Katherine Hempstead. Charles Roehrig is the Altarum Institute Fellow and Founding Director of the Institute’s Center for Sustainable Health Spending. Ani Turner is the Co-Director of the Altarum Institute Center for Sustainable Health Spending. Katherine Hempstead is Senior Adviser to the Executive Vice President at the Robert Wood Johnson Foundation.

In a November 2015 Health Affairs blog post, we argued that expanded coverage was responsible for much of the increase in health job growth that we observed from the middle of 2014 through the middle of 2015. In this post, we carry our analysis forward to December 2016 and provide our best estimate of the total number of additional health jobs that are attributable to the expansion in health insurance coverage that began in 2014 under the Affordable Care Act (ACA). While for now, the American Health Care Act (AHCA) is dead, should it be resurrected and a portion of these coverage gains be reversed as projected by the Congressional Budget Office (CBO) in their scoring of the bill, we would expect a proportional reversal in these job gains.

Coverage and Health Job Growth: National Data

The percentage of the US population with health insurance increased slowly from 2010 to 2014 and then rose sharply in 2014 and 2015, coinciding with the expanded coverage provisions of the ACA that were introduced in 2014 (Figure 1). As expected, coverage began to level off in 2016 with a relatively small increase in the percent insured. Health job growth was stable through 2012, but dropped in 2013.  Despite the jump in coverage, it dropped further in 2014, but then more than doubled in 2015.

Figure 1:  Percent Insured and Health Job Growth (National Data for 2010 – 2016)

Source:  Altarum Center for Sustainable Health Spending.  Health job growth is estimated using U.S. Bureau of Labor Statistics (BLS) data while the percent insured is taken from the American Community Survey combined with CSHS analysis of Gallup poll data (for 2016). The shaded area represents coverage expansion under the Affordable Care Act.

What we expected.  When coverage expands, the newly insured increase their overall utilization of health care and reduce their levels of uncompensated care. As a result, health care providers experience a growth in demand for care and in revenues. This should lead to increased hiring as they have both the need for additional staff and the means to pay. As expanded coverage levels off, these effects recede and hiring should return to normal growth rates.

What we saw.  The patterns in Figure 1 are generally consistent with these expectations and reveal a lag between coverage and hiring. The greatest increase in coverage occurred in 2014 but the jump in the rate of hiring did not occur until 2015. This lag is not surprising, given the timing of coverage expansion (much of the expansion in 2014 occurred part-way through the year) and the time it takes for the newly insured to find a provider and for providers to increase hiring in response.

We estimate that roughly 240 thousand health jobs have been added by expanded coverage between 2014 and 2016.  Health jobs grew at an average rate of 2.5% in 2015 and 2016. As a rough estimate of what the growth rate would have been without coverage expansion, we use 1.7%, which is the average rate for the years 2010 through 2013.  Using this rate, we estimate that of the 745 thousand health jobs added between 2014 and 2016, about 240 thousand were due to expanded coverage. This is, of course, a rough estimate based upon the assumption that health jobs would have grown at 1.7% in 2015 and 2016 in the absence of expanded coverage and that the faster growth was due solely to expanded coverage.

To examine the hypothesis that the acceleration in health job growth in 2015 and 2016 was due to expanded coverage, we analyzed state-level data to see if states with higher rates of expanded coverage showed higher acceleration in health job growth.

Coverage and Job Health Growth:  State-Level Data

Figure 2 displays the relationship between state-level coverage expansion and health job growth. The horizontal axis shows the change, from 2013 to 2015, in the percent of the state population covered by any form of health insurance. The vertical axis displays the change in the health job growth rate following coverage expansion. Specifically, it is the difference between health job growth during the expanded coverage period (June 2014 through December 2016) and the pre-expanded coverage period (June 2012 through June 2014). Forty states are included in the analysis due to lack of data for 10 states (listed in the footnote to Figure 2).

What we expected.  We expected the change in health job growth rates to be greater for states that experience the largest increases in coverage. If expanded coverage accounts for all of the job growth change, we would expect states with very little expanded coverage to show very little change in health job growth rates.

What we saw.  The data show the expected positive relationship between coverage expansion and change in health job growth.  A weighted regression (using the square root of the number of health care sector jobs as weights) shows that this relationship is statistically significant (Figure 3). The regression intercept (0.2%) suggests that only 0.2 percentage points of job growth change would have occurred in the absence of expanded coverage. Thus the state level data are consistent with the hypothesis that expanded coverage accounted for most of the change in health job growth in 2015 and 2016.

Figure 2:  Increase in Health Job Growth Rate Vs Increase in Percent Insured

Source:  Altarum Center for Sustainable Health Spending. The percentage point increase in the insured population refers to 2013 to 2015 and is derived from American Community Survey data. The percentage point change in health jobs is the difference in job growth between June 2014 and December 2016 and June 2012 and June 2014 (annualized rates). Health jobs data are from the BLS Current Employment Statistics and are available for 40 states (the 10 missing states are AK, IA, KS, MS, NV, NH, NM, SC, SD and WV).

Figure 3:  Weighted Regression Statistics

Source:  Computed from data in Figure 2 using the square root of health jobs in 2015 as weights.

Implications for 2017 and Beyond  

As projected by the CBO, the AHCA, if it were to become law, would reduce the number of people with health insurance by 14 million in 2018, through reductions in tax credits and Medicaid funding. This would largely reverse the previous gains in coverage (18 million) from 2013 through 2016, and therefore, based on our analysis, largely reverse the growth in health jobs that resulted from expanded coverage.

With AHCA tabled, but ACA repeal or modification not fully abandoned, the health sector will continue to operate under uncertainty. Early 2017 jobs data suggest that providers and health systems may have already begun to slow hiring, whether due to this uncertainty or a gradual return to the pace of growth seen prior to expanded coverage. During January and February, monthly health sector job growth averaged only 19 thousand, compared to the monthly average of 32 thousand for 2015 and 2016. Should monthly gains continue at this rate, the health job growth rates will drop below 2% by mid-year and stabilize near the 1.7% average growth seen in 2010 through 2013.

Notes:

Health jobs refers to jobs in health care services such as hospitals, physician offices and clinics, nursing homes, home health, and dentist offices.

The drop in the rate of hiring in 2013 and 2014 was driven mainly by hospitals and a likely contributor was the Sequester which reduced Medicare payments to hospitals by 2% starting in the middle of 2013.

For evidence of such lags, see exhibit 2 in Donohue, et. al, showing how prescription drug use of new enrollees increases incrementally over many months rather than all at once.

This represents a 1.6% increase in jobs that we are attributing to the 18 million individuals gaining coverage between 2013 and 2016.  It is roughly in line with previous research suggesting that, for every 10 million people who gain coverage, health care utilization and workforce requirements would increase by about 0.7%, with some added short term increases due to double payments for previously uncompensated care. One recent study estimated a loss of 912 thousand health jobs by 2019 under ACA repeal while another estimated roughly 400 thousand.

The actual number of health jobs increased from 14.7 million in 2014 to 15.4 million in 2016.

State health job growth rates are also impacted by factors such as population growth and aging.  Our analysis assumes that these factors did not change significantly during our study period.

Figure 1 shows coverage increasing by 5.7 percentage points which, applied to the 2016 US population of 324 million, gives coverage gains of about 18 million.

We report year-over-year growth.  For monthly data, this means comparing health jobs in the current month with health jobs in the same month of the preceding year.

Healthcare Triage: Health Care Reform, Tradeoffs, and the Need for Negotiation

We did an episode on the AHCA. Then we did an episode on the CBO report on the AHCA. Then… everything seemed to fall apart. What happened? It turns out, nobody is interested in negotiating about healthcare laws. Which makes it really hard to pass a new healthcare law.



@aaronecarroll

Good news for health-services research

Without fanfare, a rule restoring full research access to Medicare and Medicaid data has gone into effect. The rule had been adopted by the Substance Abuse and Mental Health Services Agency (SAMHSA) in the waning days of the Obama administration, but was placed on hold when President Trump took office. It apparently passed muster with the new administration, and a SAMHSA spokesperson has confirmed that the rule took effect on March 21.

This is good news for the research community, and a welcome end to a two-year battle to restore access to unbiased government data. The rule isn’t perfect: as I’ve discussed before, it appears to leave all-payer claims databases with no mechanism to receive data containing records about substance use disorders. But it’s a big step forward.

For those of you who have received scrubbed data from prior years, it’s not clear yet how quickly those data can be re-released in unscrubbed form. If ResDAC—the contractor that manages the research distribution of CMS data—has the unscrubbed data on hand, the rule now allows their release. If not, ResDAC will have to go back to CMS to secure the complete data sets.

@nicholas_bagley

The complications of House v. Price

I’ve got a piece at Vox discussing what happens next with House v. Price (formerly House v. Burwell), the litigation over whether Congress has appropriated the money to make cost-sharing payments.

To bring you quickly up to speed: a district court in Washington, D.C., concluded last year that the Obama administration was breaking the law in making the cost-sharing payments. The court entered an injunction prohibiting the payments; the injunction was then stayed pending appeal. The question for the Trump administration is what to do with that appeal.

I wanted to explain in non-legalese what the case is about and why it has insurers so worried. I also wanted to address the common assumption that the House of Representatives and the Trump administration could cut some kind of deal to keep the cost-sharing payments flowing. Even assuming that’s how both parties want to play it, the lawsuit will be a big headache.

The most straightforward fix would be for Congress to appropriate the damn money. That’s probably a nonstarter, given how many Freedom Caucus members would cry foul at funding Obamacare. Nor is it clear whether the Republican leadership is willing or able to broker a deal with Democrats.

If Congress won’t appropriate the money, the House and the Trump administration could try to bury the hatchet and settle the case. They might say, in effect, “We’ve agreed between ourselves to drop the lawsuit and that we’re better off without the district court’s injunction.” Now that the case is on appeal, however, it’s not so easy as that. The Supreme Court has said that appeals courts can’t overturn district court orders when parties settle their cases, even if both parties ask nicely.

So to get out from under the district court’s injunction, the parties may have to go back to the district court. But the court can modify its prior order only if there’s been a “significant change either in factual conditions or in law.”

Does Trump’s election qualify as such a “significant change … in factual conditions”? Perhaps. Certainly it would be strange to keep an injunction in place when no one on either side of the legal fight wanted it anymore. Judges don’t usually ask too many questions when opposing parties agree about something.

But still, there’s something fishy about the asking the court to vacate the injunction — and allowing the payments to proceed. Both the district court and the House of Representatives still believe (correctly, in my view) that it’s unconstitutional for the executive branch to keep making the cost-sharing payments. The Trump administration’s lawyers likely share that assessment. (Until recently, those same lawyers were raging about Obama’s lawlessness.)

The only reason to vacate the injunction, then, is because it’d be awfully convenient to keep making the cost-sharing payments — even though the judiciary, the executive, and the legislature all think those payments are unconstitutional. The judge might well balk. Indeed, she might be offended at the effort to enlist the federal courts in an unconstitutional scheme.

If I’m right about this, the best course of action for Trump administration and the House of Representatives might be to ask the D.C. Circuit to put the case on hold. But that will drive insurers crazy: the president could dismiss the appeal at any point, which would lead to chaos on the exchanges. Will insurers really stick around if the exchanges are that vulnerable?

@nicholas_bagley