In response to a recent note (see: Why Healthcare IT Companies Have a Finite Lifespan), Jeff Hawley of Impact Advisors submitted a comment an excerpt from which I list below:
Some of the observations Dr. Friedman espouses are substantiated by looking at the pedigree of where the large EHR systems came from -- Cerner, McKesson, Allscripts, etc. with one notable exception: Epic. Epic's lineage derives from a single source with no attempt to merge other commercial systems into the suite to provide functionality....
...[T]he trend over the past 10 years is the migration from best-of-breed to integrated solutions relative to the EHR. The primary driver of this migration is the need to define single sources of truth for data relevant and important to patient care such as medications. Medication reconciliation is very difficult with disparate systems (ED, acute EHR, separate ambulatory systems) and their separate definitions, standards, formularies, etc....
The portion of his comment that I found the most interesting is the following: ...[T]he trend over the past 10 years is the migration from best-of-breed to integrated solutions relative to the EHR. The primary driver of this migration is the need to define single sources of truth for data relevant and important to patient care such as medications.
I have always been a strong advocate for best-of-breed LISs that is based on the idea that LISs should optimally function as the "single source of truth" for lab data (see: Single Source of Truth). This means that lab test results are stored in the LIS and only in the LIS -- no where else. In other words, the LIS database is the single source of truth for test results in the hospital. This strict definition is impossible to achieve in most hospitals because physicians and nurses normally access test results in the EHR and not in the LIS. For this reason, LISs are interfaced to the EHR and send copies of test results to this system. Because an SST status is impossible for LISs currently, I usually refer to the LIS as the "source of truth" regarding lab data.
All of this means that the LIS is the reference database of record and always the "correct" source as opposed to any copies of lab data from other systems. Such a statement is based on two truths: (1) only the professionals that generate some body of medical information understand it in sufficient depth to QC and maintain it; and (2) copying that information to any other system (e.g., from the LIS to the EHR) inevitably introduces errors in it such as typographical, interpretive, or formatting. Therefore, only the LIS can be the source of truth for lab data and, similarly, the RIS for any imaging reports. Similarly, the EHR is the source of truth for clinical notes by physicians and nurses because the EHR is the system in which such notes are entered and maintained. In short, there are usually various sources of truth in a hospital and never only one.
I can't argue with Jeff that there has been a trend in healthcare computing from best-of-breed systems to integrated EHR solutions, which is to say that the EHR is the key system to which physicians and nurses turn to generate and review patient records. However, I also think that this EHR dominance has been a major error on the part of the healthcare industry and its IT vendors and will be corrected in the near future. The vendors of EHRs aspire to function as the "single source of truth" in hospitals but this is an impossible goal because EHRs are so large, complex, and lack adequate functionality in comparison to specialized, best-of-breed, departmental systems. In other words, I believe that seeking an "enterprise wide solution" from a single vendor is a chimera that will ultimately be demonstrated as impractical.