In May 2016, Uber announced a partnership with the Southeastern Pennsylvania Transportation Authority (SEPTA)1 to provide discounted ride-sharing services to “bridge the first and last mile gap” and encourage people to ride the regional rail system. It was a potential win for all — increased ridership for Uber and SEPTA, decreased traffic and pollution. The partnership was lauded for testing an innovative way to advance social goals.
Contrast this partnership with one that might be arranged in health care. For instance, a partnership between a health system and a ride-sharing service to provide free rides for patients with transportation barriers might help elderly patients with disabilities or those with limited transportation options get needed care.2 However, it might be illegal.
Two federal laws prevent health care providers from using inducements to increase demand for care or encourage selection of one provider over another. Under the Anti-Kickback Statute, no provider or institution receiving federal dollars can offer anything of financial value that may increase referrals for either their publicly or privately insured patients. Violators risk criminal penalties and substantial fines per kickback under the Civil Money Penalty Law. That law allows some incentives for care, a “nominal value exception” of no more than $15 per item or $75 per year per patient. Triggers for investigating fraud have a low threshold: increasing referrals doesn’t have to be the primary reason for providing the service or good — it just needs to be one possible reason or consequence.
But two recent changes in health care invite new thinking. First, these laws were enacted when health care financing largely involved patients who receive care, physicians and hospitals who provide care, and insurance companies (and the government or employers behind them) who pay for care. The same stakeholders exist today, but rearrangements in how the money flows have changed who is at financial risk for what. For example, as the financial risk for care is redistributed toward providers with bundled payment and readmission penalties, it makes less sense to retain harsh penalties for inducing patients to seek care.
Second, along with these shifts, health systems have had financial reason to develop new approaches to improving outcomes. Services such as providing low-salt food parcels for patients with heart failure and safe housing for patients with addiction — services that would never have been considered under traditional payment models — are now seen as potential ways to avert readmissions and associated penalties.3 A recent large, randomized trial revealed that financial incentives shared by patients and physicians can lead to substantial improvement in lipid management in patients with high cardiovascular risk.4 Such incentives are precisely what the relevant statutes were designed to prevent, because they may be seen as inducements to seek services. And yet if it makes sense to pay for the statins that patients need, it may also make sense to offer patients financial incentives to take them. And if those incentives are acceptable because they help achieve the patient outcomes we want, should it matter who pays them?
It might. Even as we recognize that poor medication adherence greatly limits the management of chronic disease, we might worry if pharmaceutical companies began paying patients to take their drugs. Indeed, coupons from pharmaceutical companies that reduce patients’ out-of-pocket expenses are prohibited in government insurance programs. What makes coupons unacceptable (even if their use is legal in commercial markets) is that they reduce both patients’ incentives to seek value and companies’ incentives to make price concessions.5 We might worry less about such consequences if insurance companies were the ones paying patients to take their medications — because, presumably, they would make such payments only under circumstances of high value.
The same considerations seem relevant for health care providers. If a health system gives free rides to patients for surgical treatment of varicose veins, a payer or a competitor may cry foul because vein stripping is profitable and free rides may induce demand or divert clients. Rather than using profit or unfair competition as the primary metrics, a more socially constructive distinction might be whether the service is high value. If the procedure is indicated and the price is right, as it might be for a screening colonoscopy, what’s the problem with sending a private jet? If, however, we are providing colonoscopies at exorbitant costs or to people who don’t need them, then offering a transportation inducement seems problematic.
The Department of Health and Human Services adopted a new safe-harbor provision in December that was intended to clarify the rules and permit health care providers to pay for certain forms of ride-sharing services. Though the provision is intended to ease restrictions, it focuses on the cost of the ride, who is eligible for it, and the types of cars that can be sent. It also prohibits marketing the services. What drives the new approach is still concern over inducing demand without distinguishing between high- and low-value care.
Some health systems have avoided the perception of inducement with some logistic gymnastics. The website of Medstar Health in Maryland provides a link to advertise the option to use ride sharing for appointments but probably escapes the anti-kickback statutes by not paying for the rides. Hackensack University Medical Center in New Jersey does pay for rides — but only the rides home, perhaps because a ride to the medical center for specific services might look like an inducement, but a ride away for patients in general might not. In collaboration with Lyft, we are studying the impact of ride-share–based medical transportation on attendance at primary care appointments. Our lawyers advised us that Lyft Plus and Premier are luxury vehicles and therefore prohibited inducements.
Uncertain and overlapping motivations make it hard to judge these programs. Some services and incentives help patients receive high-value care by overcoming barriers they couldn’t otherwise surmount. Others may unleash demand for low-value care that generates high profit margins for providers.
So perhaps we should instead consider their ability to achieve what we want to achieve. We believe that if inducements support the receipt of high-value services, they shouldn’t be viewed negatively. And if an institution provides transportation, thereby encouraging participation, raising satisfaction, and wooing patients from competitors, that’s positive, too — as long as we know that the services being encouraged provide benefit at reasonable and competitive prices. Under these circumstances, we might be applauding inducements, just as people applauded the deal made between Uber and SEPTA. It’s only the fact that we often can’t agree on what is high value that makes it easier just to condemn all these strategies as forms of graft. But in doing so, we also limit our ability to test them.
What to do? We could increase the dollar limits, hoping for more high-value inducements and fewer low-value ones. We could judge inducements by who pays for them, reasoning that graft is less likely when inducements come from parties with greater risk sharing. Or we could judge inducements by the kind of care they support — not at the level of each individual service, which might be impossibly burdensome, but by modifying the safe harbors to include broadly categorized high-value services, such as recognized prevention. Any of these approaches seems better than what we have. After all, sweeping prohibitions against patient inducements never really made sense, in our view, because sometimes it’s good to get patients to seek care.
Our improved understanding of the forces influencing patient behavior helps us reimagine currently prohibited inducements as tools for driving high-value care, not just engines of fraud or value-empty demand. New considerations could free health care institutions to provide incentives or services with the purpose of improving overall health, even if it means inducing patients to seek care. Applying value-based criteria to inducements is challenging. But so much of health care financing is moving toward value-based assessments — we might as well bring inducements along for the ride.
From the Robert Wood Johnson Foundation Clinical Scholars Program, University of Pennsylvania (K.H.C., D.A.A., D.T.G.), and the Cpl. Michael J. Crescent Veterans Affairs Medical Center (K.H.C., D.A.A.) — both in Philadelphia.
1. SEPTA and Uber announce transit partnership. Philadelphia: Southeastern Pennsylvania Transportation Authority, May 25, 2016 (http://www.septa.org/media/releases/2016/05-25-16a.html).
2. Powers BW, Rinefort S, Jain SH. Nonemergency medical transportation: delivering care in the era of Lyft and Uber. JAMA 2016;316:921-922. CrossRef | Medline
3. Asch DA, Pauly MV, Muller RW. Asymmetric thinking about return on investment. N Engl J Med 2016;374:606-608. Free Full Text | Medline
4. Asch DA, Troxel AB, Stewart WF, et al. Effect of financial incentives to physicians, patients, or both on lipid levels: a randomized clinical trial. JAMA 2015;314:1926-1935. CrossRef | Medline
5. Dafny LS, Ody CJ, Schmitt MA. Undermining value-based purchasing — lessons from the pharmaceutical industry. N Engl J Med 2016;375:2013-2015.
Free Full Text | Medline
This Perspective article originally appeared in The New England Journal of Medicine.