The US spends far more on healthcare than other advanced countries, but we have worse health outcomes. Ideally, we could slow the growth of healthcare spending and improve outcomes by investing in prevention, creating incentives for providers to give high-value care, and eliminating care that’s unnecessary or harmful. While many of the efforts to achieve these goals involve arrangements between payers and providers, some also target consumers. However, as two recent pieces by Vox’s Sarah Kliff make clear, it’s hard to turn healthcare consumers into the kind of savvy shoppers who can contain costs while getting the care they need to get or stay healthy.

Earlier this month, Kliff wrote about findings from a new study by Zaret Brot-Goldberg of UC Berkeley and colleagues, who took advantage of a natural experiment to study how consumers respond to being shifted to an employer-sponsored high-deductible health plan (HDHP). These plans typically have lower premiums, but require individuals to cover the full cost of most care until they’ve paid each year’s deductible. In 2014, employer-sponsored HDHPs required single people to pay an average of $2,099 and families to spend $4,332 before insurers would start paying for care. Employers who offer HDHPs often contribute to health savings accounts that are paired with the high-deductible plans, and the money in these HSAs can accumulate tax free. (Unlike flexible spending accounts, HSA funds can be carried over from one year to the next.)

In theory, HDHPs can work out well for both employers and workers. Employers can pay lower premiums and contribute some (or all) of the savings to workers’ HSAs, and workers who use healthcare judiciously can watch their HSA funds accumulate. The thinking is that workers with generous traditional coverage have few incentives to shop around for high-value care, and that the relatively low co-payments associated with tests, drugs, and doctor visits let workers consume more of these than they really need. Supporters of HDHPs contend that if consumers have to pay the full cost for some of this care, they’ll be more likely to shop around and think about which expenditures are really necessary.

But will consumers be good judges of which healthcare they need and which they can skip without harming their health? The RAND Health Insurance Experiment addressed this question when, from 1971 – 1982, researchers randomly assigned 2,750 families to five different health insurance plans with different levels of cost sharing. They found:

Participants in cost sharing plans spent less on health care; this savings came from using fewer services rather than finding lower prices. Those with 25 percent coinsurance spent 20 percent less than participants with free care, and those with 95 percent coinsurance spent about 30 percent less.

… The analysis found that cost sharing reduced the use of effective and less-effective care across the board (see the table). For hospitalizations and prescription drug use, cost sharing likewise reduced more-effective and less-effective care in roughly equal amounts for all participants. The proportion of inappropriate hospitalizations was the same (23 percent) for cost-sharing and free-plan participants, as was the inappropriate use of antibiotics.

In other words, requiring consumers to spend more money on their healthcare can induce them to get less care, but not to become smarter about which care they get. Will these trends necessarily hold more than three decades later, though, now that the internet gives us access to so much information and so many of us are accustomed to thinking about premiums, deductibles, and co-payments?

The Brot-Goldberg et al. study takes advantage of a natural experiment, in which more than 50,000 employees of a single company and their dependents were switched from a cost-sharing-free health plan to an HDHP with a $3,750 deductible. The company contributed $3,750 into an HSA, so the only workers who would truly pay anything out of pocket would be those who needed so much healthcare that they met the deductible and then had to pay coinsurance (10% in network, 30% out of network) for additional care. The company also provided an online tool to let workers compare healthcare prices. Kliff describes the study’s findings:

Workers’ health spending dropped, and did so quickly. Average per-patient spending fell from $5,222.60 in 2012 to $4,446.08 in 2013. That’s about a 15 percent decline in a single year — and it held true across all types of health services. Between 2012 and 2014, there was a 25 percent drop in emergency room spending, an 18 percent decline in physician office visits, and a 6 percent decrease in mental health services.

In one sense, then, the high-deductible plan did accomplish a key goal: lower health spending. But when the researchers looked at why spending dropped, they found it had nothing to do with smarter shopping. The average price of a doctor visit wasn’t dropping.

Instead, under the high-deductible plan, workers just went to the doctor way less. The paper finds that “spending reductions are entirely due to outright reductions in quantity.” Workers did use less “potentially wasteful care,” like imaging services, but they also cut back on “potentially valuable care,” like preventive visits.

Even more striking: The sickest workers were those who were most likely to reduce their use of care while still under the deductible — even though this is the group that needs lots of care and is most likely to blow through the deductible by the end of the year. Once these sick workers actually exceeded their deductible, though, use of medical services rebounded.

It’s also noteworthy that 62% of the employees earned $125,000 or more per year, so these are people with substantial resources – the ones we might expect to be able to shop around for the best healthcare value. Perhaps the only group we might expect to be savvier about shopping for healthcare is people whose jobs involve delivering or studying healthcare. That group includes Sarah Kliff, who’s been reporting on the intricacies of US healthcare policy for several years. But, as Kliff reported in another Vox article published five days after the one on the HDHP study, her own recent efforts to be a cost-conscious healthcare consumer didn’t go as smoothly as she’d have liked.

Kliff needed an MRI for a foot stress fracture that hadn’t healed despite treatment, and her orthopedic specialist referred her to an academic medical center for an MRI. She made the appointment as suggested, but then got a call from her insurer, who asked her if she’d be willing to go to a different facility that charged the insurer less. Being a healthcare wonk, Kliff agreed enthusiastically, and got the cheaper MRI. But when she returned to her specialist’s office, the imaging center hadn’t sent the results to the doctor’s office. And when it did arrive, the doctor wasn’t impressed:

Then there was the image itself. It did show a stress fracture that hadn’t healed, but it was blurry and a little harder for the doctor to make out what exactly was going on. The academic center he refers patients to, he told me, typically sends back much clearer images.

I had actually expected this, thanks to a conversation I had this weekend with a radiologist who’s engaged to one of my good friends. I told him excitedly about my cost shopping experiment (because I am, obviously, a very fun person to hang out with at parties). He was skeptical and mentioned how bad some of the images he looks at are. It’s just like cameras, he said: Some people have expensive DSLRs, and others are working with a point-and-shoot.

Kliff reflects on the experience:

Knowing what I know now, I’m way more torn on whether I would choose the lower-cost MRI facility.

On the one hand, I did save my insurance plan about $400. If everyone at Vox shopped like I did, then that might do a lot of work to reduce our premiums. That would probably make us all pretty happy! My slightly blurry scan was still enough for my doctor to diagnose the fracture, although I still don’t know if he’d have seen something else on a clearer image.

At the same time, my own experience convinces me I’ve downplayed the trade-offs inherent in shopping for even basic health services. If I went back and did it again, I probably would have gone to the academic medical center for the scan. Because when it comes to health care, I’m selfish — I think most of us are. I want the best shot at fixing my foot, even if that means adding a few pennies to my co-workers’ premiums (sorry!).

Kliff’s insurer was able to tell her how much two different imaging centers would charge her, but didn’t answer the real question most of us are asking: who will give me the best care? Being sick or injured can be scary, and it’s rational to worry that we can end up with a worse outcome if we go to a doctor who has less experience with our condition (or who doesn’t take our symptoms seriously), a practice where office staff don’t always follow up the way they should, or a facility with worse equipment.

Healthcare consumers don’t want doctor visits; we want health. But, as old and new research has demonstrated, we don’t necessarily know which healthcare services will help us achieve our goal. I’m glad that researchers are experimenting with a wide variety of strategies for improving the value we get for healthcare spending, because asking consumers to be savvy healthcare shoppers doesn’t seem to be achieving all its supporters hoped it would.