Written By Lisa-Jean Clifford, President, Gestalt
The length of time to install digital pathology equipment, interface, implement, train, and go live in 10 large laboratories ranged from under one year to about 18 months, and the time spent prior to that—evaluating options, gaining budget approval, making decisions, and completing the contract and negotiation process—took eight months to a year for seven of the 10. For two others, the evaluation, decision-making, and contracting process took about 18 months, and for one it took five years, largely owing to pandemic-related delay.
These are some of the findings that came from interviews with 10 of the largest laboratories that have adopted digital pathology in the United States, conducted by Lisa-Jean Clifford, president of Gestalt Diagnostics. She presented her findings at the Executive War College meeting in April.
Clifford offered the laboratories the choice to remain unnamed. Of the 10, five agreed to be named: Alverno Laboratories, Intermountain Health, Labcorp, Memorial Sloan Kettering, and Sonic Healthcare. Two laboratories started on their path to digital pathology in 2004 and 2007, and eight started between 2015 and 2019.
The aforementioned timelines the labs shared did not include the lengthy time spent before that deciding who among the laboratory staff and leaders would participate in the decision-making and seek approvals, determine what the organizational requirements were, develop the request for proposal, and determine which vendors would participate in the RFP.
Several factors explain the variation in the timelines, she said, among them the differing time frames, experience, and ability of the IMS vendors to interface; the availability of LIS, EHR, and other vendors and their resource availability to support the interfacing and implementations; and the number of the lab locations and complexity of the organizations’ workflows and whether the organizations went live with artificial intelligence.
The pandemic interrupted the process for two other labs, though they lost less time. Five of the laboratories accelerated their process during the pandemic. The two early adopters reevaluated their solutions and changed vendors when their existing contracts expired. “Both found that early technology was not as interoperable or adaptable to interfacing as they needed them to be,” Clifford said, calling it an artifact of the technology available at the time or attributing it to a lack of understanding of how the digital use cases would evolve once an organization deployed.
Most of the 10 labs started with a single type of use case or workflow and then expanded once that use case or workflow had been mastered internally, Clifford said. Noting that IMS vendors vary in the modules or workflows they support, she reported that all 10 labs were using their IMS for clinical workflows. Eight were using their IMS for primary diagnosis and tumor boards, four supported resident workflows, two were for molecular work, and five labs were also using it for research. “Only one was actively using their digital IMS vendor for consults,” Clifford said. “Ninety percent were performing consults manually or outside their LIS and IMS.”
The current state of digital pathology adoption in these 10 labs is not where any of them plan to remain, she said. Three of the labs did not disclose their percentage of adoption and case volumes; one lab disclosed that it is at 80 percent adoption but did not share its case volume. Of the six labs that disclosed case volumes and percentage of adoption, the three with the largest case volumes were at less than 10 percent of overall adoption. “The three laboratories with the overall lowest case volumes were at the highest rate of adoption,” Clifford said, “with one at 80 percent and two at 100 percent.”
Among the barriers to adoption the labs cited, IT was the biggest, with all 10 labs saying “they had issues in their IT processes for one reason or another,” Clifford said. Slide quality was another barrier, with the problems being the quality of the glass slides or of the scans. Cost was a third barrier, though it was not the cost of the IMS, she said, but of “the scanners and image storage for the organizations that wanted access to the stored images indefinitely.”
Two of the labs that cited pathologist resistance as a barrier said some of the resistance was a belief that the digital workflow was slowing pathologists down. One of the labs did an internal study in which they showed that by going digital, in three to six months (four months on average), “pathologists were faster on digital than on glass,” Clifford said.
All 10 labs said they realized benefits from adopting digital pathology, and the benefit-related categories were rated at high levels in terms of their realization. They are better patient care, turnaround time, specialty routing, no shipping of glass, workload balancing, and being able to instantly move cases to where the capacity is. AI is the only benefit category that was rated as a low realization, “not reflected in the perceived value of the benefit,” Clifford said, “but in the number of laboratories that had adopted AI,” which was half (with four using it for primary diagnosis). Of those that had not yet adopted AI, she said, “they understand this is the future, and they recognize that this is the way the industry is moving. They’re just not there yet.”
