In a report released Tuesday of this week the Congressional Budget Office concluded that "By itself, the adoption of more health IT is generally not sufficient to produce significant cost savings.” This conclusion is significant, as it is the CBO's official role to forecast the budgetary effects of proposed legislation. Their skepticism about health IT benefits may slow passage of legislative proposals that include funding for health IT. Because of this, the CBO report is sure to generate heated debate. Â

Another potential source of friction is the CBO's criticism of two well-known health IT benefit studies, from the RAND Corporation, and the Center for Information Technology Leadership (CITL). The RAND study comes in for especially harsh treatment, as many of its methods and conclusions are directly challenged by the CBO.

The CBO feels that RAND's benefit forecasts are overstated because:Â

1. RAND considered the "potential" and not the "likely" impact of health IT. RAND assumed changes in the health care system that would, for example, reward providers who cut costs instead of penalizing them for lower utilization. The CBO assumes that the likely results of health IT implementation would not involve these system changes.Â

2. The RAND study looked only at positive examples of health IT implementations, ignoring the available evidence of failed implementations.Â

3. The RAND study did not adjust for natural growth in health IT use over time "as the CBO would do in a cost estimate for a legislative proposal".

4. Separate from these general issues, the CBO feels that some of RAND's individual savings calculations "appear to be overstated" or are based on insufficient evidence.

So who's right? Well, the CBO is...at least in the context of forecasting national health IT benefits. However, their conclusions aren't necessarily relevant to the decisions of individual healthcare providers about whether to purchase health IT.

Taking the four points above one at a time:

1. The incentives for, and costs of, health IT adoption are out of alignment. That is, those who pay the costs (the providers) don't get a lot of the benefits. The CBO correctly states that this is a major barrier to the adoption of health IT. Individual providers must correctly identify the benefits that accrue to them, vs. those they are giving to their payers. Unless providers can negotiate a "share" of these payer benefits they mustn't be counted on the provider side of the ledger.

2. The CBO is right to point out that what is possible for an individual provider is not what is most likely to occur for all providers. It's certain that there will continue to be failed health IT implementations, and some that don't deliver all of the expected benefits. However, when making a specific health IT investment decision a provider must assume that they can perform or hire a successful implementation; otherwise they shouldn't make the investment.

3. RAND can't be faulted for not adjusting for ongoing growth in health IT adoption, since it wasn't their primary purpose to consider the impact of incremental government spending on adoption. The CBO, on the other hand, must evaluate the incremental impact of government spending. This difference between the two studies is just a matter of context.

4. It's true that some of RAND's benefit forecasts are based on very sketchy data. And that they use unrealistic assumptions about, for example, staff cost savings. I point out several of the same issues that the CBO raises in a new paper, Using Published Data to Estimate Hospital EMR Benefits, Healthcare Financial Management, in Press. Â

In short, the CBO's conclusions are appropriate for their role as an impartial analyst of the effects of government policy. This doesn't mean that there are not benefits from health IT, just that the possibility of failure must be considered, along with market forces and operational realities that limit benefits to providers.  Most individual providers can make a case for investment in health IT, irrespective of government policy.