Excluding childbirth, approximately 30 million people are admitted into US hospitals each year, a much higher fraction than in other countries. Many claim the results from this care does not meet world class standards, since the US is only 29th in longevity among other nations. However, this analysis ignores the effects of critical factors such as obesity, culture, and how child mortality is accounted for.

If we want to provide universal access to patients, we need to determine which of these admissions should pay more, and how much more will they need to pay. Of these admissions, over half originate in the emergency room (and only a portion of these are true emergencies).

By law, no person with an emergency condition can be refused care, even if they are unable to pay. There is also a well-documented tendency of low-risk individuals (for example, college students) avoiding insurance purchases, resulting in over 10% of the population not being covered by either private insurance or public programs (with little or no impact to mortality).  Today, the cost burden of these uninsured is borne by a combination of write-offs by medical providers and higher rates for those who can afford to pay (these costs are not borne by insurance companies or governments, through negotiations). So one of the key problems to provide 'universal' health care is to determine how to cover the costs of health care for this relatively healthy, young, and underinsured population. Should additional costs of coverage be borne by the healthy, the sick, or everyone, in order to achieve the best outcome for all over time?

The main problem with health care, though, is with the growth in costs to families, and the implications of that growth to the overall economy over time. According to the Robert Woods Johnson Foundation,

Americans who get health insurance for their families through their jobs have seen their premiums increase 10 times faster than their income... Nationwide, the amount employees pay for family coverage increased 30 percent from 2001 to 2005, while family policyholders’ income increased just 3 percent over the same period... Nationally, the average cost of family coverage increased nearly $2,500—from $8,281 in 2001 to $10,728 in 2005. The percentage of family premiums that employees pay held steady at about 24 percent. The amount that workers pay for family premiums, on average, increased $664, from $1,921 in 2001 to $2,585 in 2005. Meanwhile, the median income of people who hold family health insurance policies increased just $1,250 during the same period, from $40,818 in 2001 to $42,068 in 2005. The average cost that employers pay for their share of family coverage increased from $6,360 to $8,143, or 28 percent, during the period.

Thus, demand is up, and supply is decreasing, due to overregulation... so guess what happens to costs? Yet even this view is over-simplistic. As a nation, with an increasingly aging population, we choose to spend more on health care because we can. We are wealthy, and the costs of most other 'essentials' of life, as we age, decrease. Our homes are paid for. We eat less. We consume less technology. But when we have aches, pains, and wrinkles, we're willing to spend more than we did last year, trying just about anything, to protect our youth.

This exercise of 'following the money' demonstrates how easily a simple question about attempting to optimize one parameter (in this case, costs) can quickly result in the need to discuss other related parts of the system (like outcomes and the behaviors of selected stakeholders in interacting with the system). Understanding such interactions in a system is essential in order to determine how to improve it overall. According to a report by the Galen Institute, the value of uncompensated care in this country was 'only' about $41 billion in 2004, or less than three percent of our overall national health expenditures. The uninsured pay for only about 25% of the medical care they use. Federal and state taxpayers already pay 85% of the rest.

Part of the problem is that no nation, in the US or Europe, has solved moral hazard. As long as government, or our tax-subsidized employers, are pre-paying our healthcare, and we can leave your wallet at home and demand all the tests and treatments we are allowed, we are in trouble. It is a big Las Vegas buffet, and we are all high-rollers, pigging out and over-eating because the tab is on the house.

Prior, short-term cost control initiatives, in the absence of this systemic view, have often focused on reductions of obvious, tangible elements (physical assets, duplication of resources, etc), rather than more-important, higher-leverage items that could have longer-term and more sustainable payoffs. Such longer-term strategies typically tackle cultural issues and intellectual, rather than physical, work products. Such intellectual work products, and their associated decisions, are often difficult to observe and act upon within an organization, but tackling them typically will produce better results in the long term, since they are the basis of what drives underlying costs for stakeholders. You can't just inject $41B into the system, because the system will swallow it, and you'll be likely still down $41B. It's like Microsoft bidding on Yahoo; they bid (interestingly) $41B, and within a few days, their total stock valuation decreased by a corresponding amount. No one told anyone to sell, or how many; but the system adjusts; behavior emerges from the system.

Yet meaningful reform proposals, which attempt to do things like eliminate tax breaks for 'top shelf' health care plans (to 'level the playing field' around marginal costs), although well received, go no where, at least partially due to political gridlock. Longer-term strategies are thus relying on areas of concensus, such as on improving the flow and exchange of information through electronic health records. This will further enable understanding and exploiting patterns which exist in performing patient care, so that people can learn to do it more effectively, prepare to do it in a more orderly manner, and process the work more efficiently.

Until such long-term investments are implemented, and the results from those investments are realized, there are often limited opportunities of significance in the short term that can be done to reduce costs, other than reducing capacity, rationing care, or eliminating the very collection of the data that is necessary for meaningful, long-term improvements. It is difficult to have the time to implement improvements 'in real time' within any system, without a parallel effort (both in funding and resources) that can design the improvements themselves, and that is independent of the ongoing pressures that daily execution of health care involves. Such investments, unfortunately, can be too often the first thing to go when it's time to cut costs, within hospitals, and elsewhere.