The cost of switching to ICD-10-CM has been underestimated: Proof.

The Department of Veterans Affairs (VA) is spending $211 million on software and services for the switch to ICD-10-CM. And this expenditure is on a third-party alone, let alone the time its employees are devoting to managing the switch.

So, how does this dollar figure comport with the estimates that various entities made of the costs of the switch to ICD-10-CM?

Well, as it happens, it is well out of bounds of any of the estimates.

First, the RAND report (warning: PDF) that estimated the costs of the switch failed to account for any costs to the VA for the switch. Oops. The RAND report estimated a cost of $425–1,150 million for the switch, so this $211 million expenditure on the part of the VA inflates RAND's estimate by 19-50%.

Therefore, RAND's estimate of the overall cost of the switch is significantly low.

Next, the Department of Health and Human Services (HHS), in its final regulatory rule (warning: PDF) mandating the switch, did account for VA software expenditures to manage the switch. HHS estimated that the VA would spend a total of $24-31.35 million on software and systems, not counting training and planning. Even if we add training and planning (assuming that the contract includes those functions), HHS estimated no more than $113.8 million in spending on software/systems by the VA.

Thus, the $211 million figure is nearly double the amount that HHS expected, at a minimum.

HHS pegged total costs of the switch at ~$2.3-2.6 billion. So this overrun of its estimate for the VA represents 3.7% of the highest estimate by HHS for the total, national cost of the switch.

What other overruns are government agencies, providers, and payers experiencing. In other words, where else, and by how much, are the estimates too low?

Stay tuned.

We have always maintained that the estimates were unrealistically low, and now we have presented proof.