Bundle in the Jungle

Bradley Flansbaum

Posted 12/09/11 on The Hospitalist Leader

First, before I discuss bundling as a method of payment for healthcare services, I wish to declare my neutrality on the subject.  The merits are appreciable if providers approach the implementation properly, and conversely, the detriments are equally clear if they (we) bungle it.  Enactment and downstream effects are so murky however, that prognostications are premature.

Second, bundling may not be the preferred payment method for all procedures or diagnoses.  Beyond CMS demonstrations and a scant sampling of commercial side ventures—all procedural—little to no chronic care evidence exists (COPD, CHF, sepsis).

Finally, I hear bundling referenced frequently, often by boosters distant from the realities of practice.  My desire is to invoke the tale, “Emperor’s New Clothes” at almost every utterance.  Shout it from the rooftops: “Bundling is not a panacea!”

With those brief comments, here are my observations on the subject and why we must avoid assenting to it reflexively:

1.   If you want to assemble a bundle, it entails estimates of provider totals in a patient’s care episode (or if by time, over a year).  Granted, some patients—the low hanging fruit—might visit only their primary care physician, an ophthalmologist, and an orthopod for routine problems.  Estimating utilization and costs, and parsing up the funds in these cases may not be problematic.  However, the costlier patients, by no fault of their own, who make doctor visits a vocational skill, present greater challenges. Look at the median number of providers engaging with chronically ill patients:

Additionally, note the major providers rendering care during a unique acute stay:

Despite the optimism of those sitting in dark rooms, the easy fix does not reside on a spreadsheet or a SAS program.  The efforts to streamline care and target the right providers for payment and measurement—by number and specialty, will fall in when the system is ripe.  This will not happen by 2014; who receives the “credit” for saving the COPD’er with severe pneumonia for example—the pulmonologist or hospitalist—is a coin toss.  This is a sliver of what confronts us.

2.   How about the evidence on bundling?  Does it decrease spending, reduce waste, improve the patient experience, or is it even operationally feasible?  Review this superb analysis by RAND to glean more, but suffice it to say, based on what we know, if bundling was a pill, you might have second thoughts before popping it.  Additionally, some of you might be familiar with PROMETHEUS—bundling with best evidence—and the work that emanates from their shop–mostly surgical and ambulatory in nature.  This Health Affairs release in November 2011 forecasts a tougher slog, even with their early dedication to this form of remuneration.

3. Given lower per capita spending internationally, it makes sense that capitated payments, as opposed to fee for service (for populations, diseases, or episodes of care), is de rigueur.  Acknowledging that cost growth is variable overseas, and no country has the right formula to clamp down costs, it is also notable that bundling is not prevalent in many high functioning health systems.  France, Germany, Switzerland, and others, still operate in a fee for service world—in more regulated environments—and the sky is not crashing downwards.  That is not to say they will not experiment or adapt, but my point is, cost growth moderation does not rely solely on a bundled platform.

4.   For all the upsides, and I cannot deny the conceivable benefits (better care at lower costs), we must simultaneously review the challenges before we beckon bundling utopia.  Here are is a brief list of impediments:

  • Splitting the bundle ($)
  • Tracking Claims (QI)
  • Setting the Price: all payer, negotiate, market (“you pick”), bidding
  • Quality: metrics; sophistication
  • Fair payment: case-mix, outliers, weighting–patients, sites, providers, payers, ? cost shifting
  • Multi-payer: ERISA, MCO, Public
  • Attribution: Individual versus group, sites
  • Defining an episode of care:
  • Setting, providers
  • Single vs. multi condition
  • Time (PPACA = -3, +30 days)
  • Heterogeneity within similar episodes
  • Legal Challenges
  • Mandatory vs Voluntary

These are system wide challenges, not just for bundling, but also for PCMH’s, ACO’, etc., but it gives you some perspective on the number of hurdles to overcome before bundling prospers—assuming it does.

5. Most policy experts are optimistic that bundling will lower costs.  While I mentioned above that evidence is lacking, predictions and modeling are in reasonable supply.  A recent Urban Institute brief examined the utilization of bundled payments by all payers (public and private), and their potential effects on lowering systemic costs.  The investigators mention that their assessment underestimates impact, as they excluded chronic care opportunities—presumably due to a paucity of data.  Nonetheless, based on current knowledge,  ten years savings is minimal (0.3%):

Even if we forecast the rosiest of scenarios, and the system “learns” quick, rapid answers are afoot— contrary to the ardent promoters of this form of payment.

6.   Finally, I will give the last word to Richard J. Gilfillan, the head of CMS’s Innovation Center, answering a question posed by Zeke Emmanuel at this forum on the subject.  Go to the 1’13″ mark to hear Gilfillan’s response to Emmanuel’s question regarding where chronic care bundling stands in 2011. It is worth a listen:


CMS is accepting applications for bundled payment proposals.  Good thing.  However, don’t bet on any go live projects involving complicated medical patients now or in the near term (2-3 years).  It is baby steps only, with predictable procedure-based interventions and simple, definable illnesses.  Anything more ambitious will hail from integrated systems that bear little resemblance to most of our practice environments today.

Bundling is coming.  However, it is not the only answer, and may not be appropriate for many of the DRG’s we bill.  The value payers and patients seek for those diagnoses may still reside in FFS, however, only in the context of other payment system and measurement changes our healthcare environment will undergo.  They will embody both supply and demand side interventions.

The emperor must keep shopping.

Brad Flansbaum MD is a hospitalist working in New York City and writing at the Hospitalist Leader.

One thought on “Bundle in the Jungle

  1. Bundling is a corrupted (certainly easily corruptible) form of disruptive innovation by committee, and Dr. Flansbaum correctly questions the likelihood of its success.

    However, he elsewhere recommends we attend to bell-curve distributions of clinician usage by chronic care patients. Our salient health reform issues reside at the extremes of power law distributions. Imagining central tendencies should be our first priority only contributes to confusion.

    And, on average, we don’t need to be doing that. Let’s look first instead at work being done in Camden & Atlantic City and other places where the most extreme examples of serial rescue care are being moderated, and copy that into more health care cultures.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s