Misleading VA Health Information System Numbers

Health Affairs published an article regarding the performance of the Veterans Administration’s health information system, known as VistA, in its latest issue which largely dealt with health IT, and specifically with electronic medical records.  (Health Affairs Article) The article lauds VistA and claims that it has or will save the VA over $3 billion.  That number was picked up in the popular press and repeated in connection with the Administration’s claim that all the money being spent on electronic health records will yield both quality and cost benefits.  This article, however, is largely nothing more than speculation and has no methodological rigor.  The authors fail to mention the long history of VA health IT projects being over-budget and behind schedule; a history which led to a suspension of those projects last year because an audit found such poor project management. The authors also fail to note the significant body of research, some of which is found in the same issue of Health Affairs, finding that EHRs have a limited, if any, cost payback and that their impact on quality is uncertain.

The authors of the article benchmarked the adoption, cost and quality impacts of the VA system compared to the public sector.  They then also created a model for estimating the cost impacts of VistA.  In regard to the benchmarking, it is true that the VA is further along than most private medical systems in adopting electronic medical records.  The researchers also note that VA devotes a higher percent of its total spending to health IT than does the private sector, in particular, a higher level of operations and maintenance spending.  This is not necessarily a good sign; it likely reflects inefficiency and the poor management found in the audit.  A good system should not need high levels of operations and maintenance expense; when that exists it usually indicates extensive time being spent creating work-arounds and fixing bugs.

But it is the cost model that is almost laughable.  The authors did not make any attempt to collect real data within the VA to show savings.  Instead they talked to experts about kinds of savings might be possible as a result of implementation of EHRs.  Included among the buckets of  estimated “savings”, the largest is for prevention of unnecessary care from averting medication errors or avoiding unnecessary outpatient visits.  Since VA utilization and spending continues to rise at a pretty good clip, it doesn’t appear that in the real world the VistA has had much of an impact on those costs.  It is hard not to suspect that if the real data had showed savings, it would have been used, instead of a theoretical model.  EHRs are a good idea, but it doesn’t help to have proponents creating unrealistic expectations about quality or cost benefits.  EHRs are difficult to implement, cause workflow redesign issues, don’t always have all the functionality needed and take many years to provide the benefits they are intended to provide.