Healthcare is expensive.
Its current trajectory in America – one that yields an ever-accelerating rise in both cost and complexity – is clearly unsustainable. This reality, combined with the outsized role government plays in regulating and paying for healthcare, presents Washington with an enormous policy challenge.
Healthcare’s web of interconnectedness means any change in one stakeholder’s business model – be it a hospital, physician practice, skilled nursing facility, device maker, pharmaceutical company, insurance company, etc. – is much closer than Six Degrees from every other business model across the delivery system. Kevin Bacon can eat his heart out.
Under the guise of “bending the cost curve” policymakers may be inclined to impose across-the-board cuts (like sequestration) in an effort to “finally get things under control.” Instead of imposing blunt cuts that tell providers “just do less” – and potentially cut the exact kinds of care what we want to incentivize in a population health world – we need to use pay-for-value programs to drive a shift in the kinds of things we do.
CMS has experimented with this concept in both the Value Based Purchasing and ACO models (among others), using cost metrics to facilitate the move away from fee-for-service. Now isn’t the time to give up on cost/utilization metrics – its time to double down and get real.
Here’s my unsolicited advice for policymakers on how to keep driving the shift to population health *while simultaneously bending the cost curve in the short and long-term* through the thoughtful incorporation of cost/utilization metrics into their pay-for-value programs:
1. Use metrics that get at real under-utilization
Deming was right – driving quality decreases cost. As our experience with “Core Measures” showed we have not and do not consistently perform evidence-based, low-cost preventative care 100% of the time. These are basic things like prescribing aspirin at discharge for heart attack patients, *never* electively delivering a baby before 39 weeks, and universally screening for cervical and colorectal cancer.
Measuring and tying payment to these metrics drives performance, and providers have demonstrated the ability to rapidly top-out process measures at >99% achievement. The temptation is then to “move on” to measuring something else… and this is where we risk missing the long-term cost impact. We cannot lose sight of the tight relationship among ensuring the routine execution of evidence-based process, better outcomes, and lower lifetime cost. However, since the cost impacts of many process measures (think high blood pressure control) are generally felt over multiple years (if not decades) we must maintain focus on inappropriate under-utilization to avoid “drift” once the Eye of Sauron has moved on to a new target.
2. Use metrics that get at real over-utilization
There is real waste on healthcare – that is, money we spend paying for diagnostic and therapeutic services that are either simply not indicated or where a different, less costly option would be equivalent. Gawande wrote about this at great length (and with much greater eloquence than I can) in the New Yorker’s “America’s Epidemic of Unnecessary Care” so I’ll point you there for the full treatise. Suffice it to say that we frequently “do too much” and we do so because of choices made at every level of the healthcare decision chain (including patients, providers, and payers).
So how to get at over-utilization within a metric program? Rather than go after coverage (e.g., universally refusing to pay for certain services) we must begin measuring and reporting some of the examples discussed at Choosing Wisely – including:
- The routine prescription of antibiotics in the setting of newly-diagnosed, uncomplicated ear infection in otherwise healthy 2-12 year-old children whose parents are able to maintain contact with the practice for 48-72 hours
- The routine use of lower-back imaging in the ED for adults with low back pain and no history of trauma or signs of nerve / severe injury
- The routine use of head CTs in the pediatric ED setting for children with minor head injuries who do not meet the PECARN criteria
There are literally hundreds of examples, each recommended by the relevant clinical specialty society (e.g. the American Academy of Pediatrics for the ear infection guidelines, the American College of Emergency Medicine for the ED guidelines, etc). Yet we still do many of these things.
It’s time we measure, report, and expect explanations for over-utilization.
To be clear, I’m not saying payers should make either individual or universal COVERAGE decisions based on clinical algorithms… Choosing Wisely’s website itself actually states:
…recommendations should not be used to establish coverage decisions or exclusions. Rather, they are meant to spur conversation about what is appropriate and necessary treatment
Nor am I saying that practitioners must follow these guidelines 100% of the time (note the word routine in each of the examples above). Instead, we need to be transparent with utilization metrics that are tied to real evidence, discuss them with our patients and incorporate their reporting into pay-for-value programs.
While there will always be exceptions to any clinical algorithm, we must begin asking clinicians to document (in an easy-to-use electronic workflow) their medical decision making that indicates antibiotic use for an uncomplicated ear infection or imaging for low back pain in the ED. This in and of itself will be a critical first step towards decreasing real over-utilization and have an immediate impact on both short and long-term cost.
3. Use metrics that actually drive change in care delivery… while resourcing the creativity, innovation, and investment required to achieve success
Driving accountability to the provider level for population health outcomes – e.g., preventable hospitalizations for ambulatory sensitive conditions like asthma or cellulitis – will require innovation, experimentation and investment. It also presumes Americans are universally ready and able to accept it (see previous post for a longer discussion on this point). While bundled service payments can simultaneously drive this innovation and fund it through reasonable target costs and bonus payments, we must drive deeper to meaningfully change our delivery system.
Cross-cutting metrics within broader pay-for-value programs can support the innovations needed to weather the transference of risk onto providers… but policymakers must be cautious. There are very few if any providers prepared for the level of accountability that comes with metrics like those included in the Institute for Healthcare Improvement’s Whole System Measures 2.0 – including “unmet healthcare needs” and “disparities in high school graduation rate” in addition to preventable hospitalizations. Incorporating metrics like these into pay-for-value programs transparently pushes the responsibility onto providers to chart the path to population health, and doing so must be coupled with the necessary resources to innovate, iterate, and ultimately broadcast the best interventions nationwide.
In the end, we can and will meaningfully impact the cost curve in American healthcare. We simply have no other choice if we want to avoid bankrupting the country. How we get there remains to be fully determined, and our window is small to drive policymakers towards programs that facilitate real changes in care delivery instead of wielding the sledgehammer of global cuts.