Part of the theory of ACO value creation is trading off more primary care (resulting in better care coordination, fewer missed time bombs, and use of lower cost care options) against reduced use of specialists, ERs and hospitals (few stays, shorter stays). Early results seem to describe substantial promise (although not for everyone who tries the model). Let’s assume this promise will be realized in broader roll-out for the purposes of this post.
One fear that is ACOs will drain volume away from unaffiliated medical specialists and hospitals, leading to a large capacity overhang and sharply weaker economics. Indeed, if ACOs were rolled out all at once and they perform as promised, the shock to specialist and hospital utilization would be considerable, forcing higher end medical services to reallocate outsized overhead (cost sharing) or exit.
This scenario assume a sudden switch in the shape of healthcare demand across medical services. However, ACO roll-out will likely be gradual and other demand growth drivers (the reform coverage expansion jolt and ripples, overall population growth, increased average age from baby-boomers, technology) will help make up for reduced demand from the existing population. Couple these perspectives with a view that some 3-5% of physicians retire each year (assuming 25-30 year career) and ongoing renovation of the hospital infrastructure allows for optimization over time.
Key will be figuring out the right proportion of the delivery system to be operating as an ACO over time. Carefully “manage” that growth rate, and ACOs can become a mechanism to optimize current capacity against growing, demographically driven need.
Here’s a straw man scenario:
Suppose ACOs learn to optimize care for their patients as follows (aggressive assumptions premised on continued maturation of reported positive results):
A few additional assumptions to the scenario:
- Overall population grows at 1% per year
- Healthcare utilization (across all service categories) increases an additional 1% a year as a result of demographic drivers (aging baby boomers)
- Capacity across provider services remains constant (so assuming retirees are backfilled consistently but no growth)
If the ACO roll-out is aggressive (for illustrative purposes: incorporating 10% of patients at the beginning and growing to over 50% of patients within 5 years), there will be a large chunk of unused capacity (up to 10-15% across specialists and 5-10% of hospital beds), requiring “cost shifting” to shore up the economics: 
However, if the roll-out of ACOs is more measured (say, starting at 7% of patient volume — the inverse of the AMGA survey result suggesting that 93% of their membership would not participate in ACOs as designed in the draft rules — and growing to 38% by year 5 – remember we are assuming success here), the utilization rates of advanced services can be kept closely in line with current levels without reducing (or growing) capacity. Under these assumptions the following scenario of ACO roll-out does the trick:
One issue is that the demand for PCPs exceeds supply in both scenarios. This is as expected, given the assumed ACO trade-off across medical services. Of course, a lot of this additional care can be managed through higher leverage team models (using more RNs and NPs) and outsourced services provided by a population health vendor such as HWAY. Even so, primary care will certainly be a choke point under the new system.
Key implications:If this analysis is correct, a relatively small share of initial ACO participation in the CMS program is desirable, not a sign of failure; too fast an uptake can disrupt the economics of the rest of the provider system, encouraging undesirable capacity exits (given looming baby-boomer demand) or, worse, provoking crippling provider resistance to the overall program. In this context, the rather laborious rules CMS has offered may have had an incidental – even semi-intentional – benefit: if the rules are too easy, too many will want to participate, provoking stronger resistance from those parts of the system left out. If this reasoning is right, do not expect the final version of the rules to be a lot more user friendly than the first draft.In addition, burdensome CMS rules may also help dissuade “leading systems” from playing (e.g. they have easier ways short-term to win share through traditional reputation based strategies), focusing the real commitment from second tier systems who have no choice but to go the extra mile (think UPMC vs. West Penn Allegheny). This will help ensure a more balanced distribution of demand across each local delivery system (rather than skewing more volume to the leading systems). If true, the lack of interest from Mayo, Geisinger and Cleveland clinic int he Pioneer program should not cause too many headaches at CMS. That said, commercial plays with ACOs may disrupt a gradual roll-out built on Medicare reimbursement by unleashing a number of competitive dynamics: first, to the extent commercial ACOs use growth in share of patients (rather than share of savings) as an incentive (as seems to be happening with the BSC pilots), growth becomes imperative and therefore more threatening to other providers. Second, primary care is a clear choke point for the ACO and markets where payers and providers are aligning; an arms race is underway (think California with multiple payer acquisitions and ACO initiatives). Also, successful ACOs will leave a rump of underutilized advanced services capacity whose overhead costs will need to get paid for…by competing health plans (in the form of cost-shifting and higher rates). These competitive threats may lead to a much faster push for ACOs on the commercial side (especially leading into the 2014 super cost-sensitive Exchange environment).Fortunately, commercial volume of any one payer is usually a fraction of overall volume (especially compared with Medicare), so the stakes should be lower and the risk of provoking provider resistance smaller. Also, the growth of commercial ACOs will help support the program in Medicare: One can imagine as the members become eligible for Medicare, they will either (1) roll into their current plan’s Medicare Advantage offering or (2) go into Medicare FFS. Either way, the patients will be trained to stick within an ACO operating world. As a result, the Medicare share of ACO lives will grow simply as a result of the passage of time (and at a non-threatening gradual rate).