...I think that it's possible that Abbot has the following scenario in mind as a long-term goal: sell to clinical labs, as an integrated package, analytic instruments, reagents, and test result management supported by a cloud-based LIS/LIMS. Pricing would be on a per-click or taxi-meter basis. This would be the first PaaS offering for the clinical lab industry.
For the sake of this discussion, I will refer to the integrated environment provided by an IVD vendor (lab testing plus a web-based LIS) as a cloud-LIS. One of the key questions pertaining to the practicality of such an environment is whether the reports generated by such a system are "integratable" into the workflow and database of a classic LIS running locally in that same lab. This new model will receive little attention in the market absent such functionality. To get an answer for this question, I turned to my colleague Dr. Ulysses Balis who is Director of Pathology Informatics in the Department of Pathology, University of Michigan Medical School. Here is his response:
Firstly and looking at the internal Starlims relational architecture, it is clear that the architects have elected to employ modern, best-practice programming and structural approaches including the use of XML, constrained name spaces (probably utilizing ISO-11179), and service oriented architecture (SOAs) to deliver a highly interoperable LIMS platform. Such a starting point represents an optimal minimal set construct for leveraging the significant asset-tracking requirements of the growing menu of molecular diagnostics tests. Specifically, features present in this environment and not present in the typical LIS environment include the need for expanded and real-time tracking for the entire range of test kits available in the market in addition to simply tracking work-center results. LIMS architectures are generally well-suited to this task and provide, additionally, workflow automation, bar code tracking, chain of custody, and domain-specific decision support tools for enhanced reporting. Consequently, limitations on the integration of LIS-cloud-generated test results into a hospital LIS data architecture are largely mitigated, if not completely removed, in the setting of several specific but common XML approaches.
The use of a constrained XML schema, in association with a LIMS-provided .XSLT file and a constrained name space, allow the purchaser to map the result data stream from the LIMS application into a legacy hospital-based LIS without resorting to the traditional HL7 multi-step approach. This latter method is typically associated with multiple developmental iterations and changes. It would thus appear, at a strategic level, that Abbott’s acquisition of Starlims may represent a practical strategy for rapid adoption of sophisticated molecular reporting capability without the requirement of extensive hospital LIS modifications. Moreover, it can also be expected that the final integration of the two disparate architectures will yield reports that are indistinguishable from those of a hospital LIS designed as a seamless solution. For this reason, we should expect the field of laboratory information systems to increasingly depend on XML, service oriented architectures, and the interoperability features that these technologies provide beyond currently available traditional HL7-based tools.
Given that test results generated in an IVD-vendor cloud LIS can be easily and rapidly integrated into the database and reporting capabilities of a traditional hospital-based LIS, the next question is why such a "package" would be appealing to a clinical lab director and IVD vendor alike. I will return to this question in future notes, but here's a brief answer to this question. In molecular, genomic, and proteomic testing, it is important that testing not be constrained by any limitations of hospital-based LISs. With a cloud-LIS, both lab test workflow and interpretation algorithms (i.e., rules) can be changed quickly and on-the-fly. This is one of the fundamental advantages of such a specialized web-based system. In essence, the "tyranny of the outdated or inadequate LIS" ceases to be a limiting factor for lab managers. This approach moves the goal line for molecular diagnostic testing from the quality of individual test results to the quality of the diagnoses rendered in the cloud and reported to clinicians.