The discipline of diagnostic pathology has benefited over the years from invention of new technologies such as immunohistochemistry and fluorescence microscopy and the introduction of genomics. The use of this information to make clinical decisions has heralded some of the most significant advances in the field.1 Through the advances of imaging and information analytics of recent years, we have witnessed the evolution of pathology from an ancillary clinical service to a key driver of the use of technology to enhance patient care and the patient experience.
No advancements have had as much potential to transform the work of the pathologist to the extent of digital pathology. As the projected number of pathologists entering the field is expected to decline, it becomes increasingly important to empower pathologists through the incorporation of transformative technologies into the pathology workflow.2 Digitization of tissue slide images could address numerous challenges faced in medicine thereby enabling telepathology and increasing the accuracy and precision of diagnostic/prognostic determinations. Efficient digitization and analysis of images has advanced by improving the tracking and organization of information, facilitating greater efficiency of the pathology laboratory, expediting identification and interpretation of unique data (such as for rare diseases), and using computational pathology.
Analysis of digital images is data intensive and requires large data sets to power the underlying statistical analysis. The introduction of artificial intelligence (AI)-based decision support tools leveraging cloud-based services, big data sets, and computational pathology into routine clinical care holds great promise to deliver even more efficient and accurate care. While the specifics of these underlying technologies may not be transparent to users, clinicians and healthcare providers will see performance benefits in parameters such as turn-around time, data quality, and more comprehensive patient reports. In this paper we report on the adoption of these technologies and how they might impact clinical innovation, payment reform, value-based care, and coverage and reimbursement decisions.
Current Landscape and What We Know Today
The integration of this transformative technology into the daily workflow has proceeded slowly. Barriers to implementing digital pathology are numerous: the technology is expensive, there are large upfront costs for hardware as well as for integration in the current infrastructure, and significant resources are required to initiate change management. Additionally, current reimbursement for pathology services is insufficient to cover the cost of technology acquisition, the pathologist work involved in working with digital pathology tools, and new systems to incorporate the technology effectively.
When pathologists and lab directors advocate for technology adoption within their health systems they must be able to describe and justify reimbursement benefits to articulate the return on investment of these tools and incentivize investment. At the same time, payment mechanisms in the US are transitioning from fee-for-service mechanisms and fee schedules to value- based care models, necessitating innovative approaches to care delivery and new concepts to define and demonstrate value. Although the current environment presents complexities, it also provides the potential to establish new and improved methods to achieve reimbursement for these technologies.
Similar to other innovative diagnostic technologies, digital pathology faces the dilemma of how best to demonstrate its essential value to stakeholders. This article is intended to provide recommendations to the broader digital pathology community to create momentum for optimizing reimbursement for these technologies in the United States. The path to improved reimbursement begins by defining the most valuable use cases for these technologies and aligning on how to support the demonstration of value through evidence generation. It also requires clarifying the value for various stakeholders and developing the evidence to incentivize reimbursement and uptake.
Reimbursement (coding and coverage pathways)
Three primary approaches implemented by digital technologies that each individually or collectively offer distinct value to the healthcare system justifying reimbursement. In practice, the advancement and integration of these approaches concurrently may be necessary to realize optimal benefit to patients and stakeholders.
The first key category of digital pathology includes technologies that incorporate advancements to improve the pre-analytics workflow, i.e., sample process, accuracy, and efficiency in comparison to existing approaches. Novel alternative CPT codes for digital techniques that enable the use of these incrementally innovative tools and distinguish their approaches are needed to ensure appropriate reimbursement and realize benefit to patients. It must be noted that current codes do not account for the full cost and value of digital pathology technologies, including both the technical (TC) and professional components (PC), neither in isolation nor where there is a blending of TC and PC (e.g., clinical interpretation of unclear TC output). While this distinction is highly specific, it is vital to ensuring the appropriate allocation of payment back to various individual contributors.
A second category of digital pathology technologies includes those that can perform analyses that were previously impossible within traditional pathology. An example of such an application would be an AI-driven software algorithm to provide unique patient risk stratification information. For these offerings, rather than relying on a series of codes for broad use of the technology (e.g., slide capture and storage), we would urge innovators to consider pursing coding for specific use cases and applications of the technology. This could be achieved by leveraging the new PLA coding system which allows for creation of codes specific to particular offerings/applications. These codes could be assigned varying values based on the clinical benefit or the complexity of diagnosing a particular disease.
Finally, a nascent category of technologies is emerging that will support improvements in clinical care at the health system or even the population level. For these tools, support for novel reimbursement must occur within the broader context of the shift of healthcare payment systems from fee-for-service to value-based reimbursement. The evidence needed to support reimbursement in a value-based system will be inherently different, requiring a shift from viewing digital pathology as solely a reduction in cost to a source of value.
In the broader context, pathologists should re-examine the value of pathology services beyond that of slide reading to an integration into complex services. Here the efficiencies provided by the three categories of technology advancements will resonate because they could provide equivalent or even better prediction as well as outcomes with more standardization, reduction of retesting or confirmatory testing, and less effort on a case-by-case basis. Achieving consensus on the value of these efficiencies should involve entities such as thought leading institutions and groups developing integrated delivery networks that are at the forefront of clinical innovation, payment reform, and value-based care.
“The integration of this transformative technology into the daily workflow has proceeded slowly”
To Whom It May Concern
Support for reimbursement for digital pathology is contingent upon identification and building programs that educate payer stakeholder about applications and technologies that provide value to practitioners. This value may be different to various stakeholders, but they must be aligned to tell a cohesive value story. Digital pathology will impact pathologists, clinicians, patients, hospitals, and payers differentially. For example, pathologists may value efficiencies in workflow and better-informed decision making whereas clinicians and patients benefit from greater accuracy and speed of test results. It is important to consider all the ways in which digital pathology may influence decision-making for each stakeholder and how those different values may build to something greater before pursuing evidence development.
Coverage decisions for novel technologies are influenced by numerous factors. Identifying key sources of information about these factors is complicated by differing priorities across the major healthcare payors in the US. Influential factors such as patient-reported outcomes, comparative effectiveness, and patient advocacy can all weigh into coverage decisions.3,4 Generally, evidence of reduced costs and improved clinical outcomes are highly valued in payer decision-making, however, what is most important is to demonstrate that these novel technologies have clinical utility. Evidence of clinical utility can take various forms, but the most direct argument is whether the use of a technology results in a change in clinical decision making, which can further translate to implementation and a positive impact on patient care. Widespread adoption and implementation can subsequently further increase the impact on coverage and payer decisions.
The annual operating costs for laboratories are highly affected by the purchase of new instruments, their implementation, validation, maintenance, down time, and effectiveness. Decisions to purchase and implement new equipment in the lab are highly dependent on the ROI analysis as well as benefit to patient care; consequently, the potential for reimbursement of a technical component can be highly influential in the decision-making process. In general, decisions for technology implementation may be based on business models developed years before that show ROI based on the arc of projected performance improvements of existing or legacy technology. Innovation such as digital pathology and artificial intelligence algorithms are so novel that they have yet to be considered for implementation in many laboratories, let alone baked into ROI business models. Hence, coverage and reimbursement determinations could strongly influence business decisions to adopt novel technology in laboratories.
Numerous benefits of digital pathology to pathologists have been documented in the many studies that have correlated its use with greater accuracy and precision in pathology results.5-7 Any variability in these studies tends to be not due to the skill level, or lack thereof, between pathologists, but rather the lack of good, comparable consensus standards. On the other hand, the reference for measuring accuracy and precision for digital pathology devices is often individual pathologists! Without a standard, there will always be large variation between individual pathologists and between pathology labs, making it difficult to compare diagnoses and determine appropriate treatment decisions. For example, significant variability exists between pathologist interpretation of Gleason scoring in prostate cancer, but some have been able to use digitization of tissue architecture to better classify lung and prostate cancers.8-11 To address this standardization issue, digitization of pathology images and introduction of objective decision support tools will create a harmonization of diagnosis and consistent treatment decision-making. The ability to read digitized images and use analytical tools for assessment will likely reduce the need for additional consultations, leading to improved workflow efficiency and, in some cases, reduction in the intra- and inter-pathology scoring/interpretation time that inherently exists for some diagnoses such as cancer. Furthermore, these tools may reduce the need for confirmatory testing or second opinions. Lastly, digital pathology facilitates synthesis of novel information for the detection of rare diseases. For example, the Thomas Fuchs Lab be used safely and effectively in a broader population, leading to a lack of confidence in the technology. This impression may be due to the need for better education, communication, or understanding of the application that may be met with training and/or widespread dissemination of the results and evidence. Others are fearful that the technology will obviate the need for trained pathologists. On the contrary, the aim of digitization and especially computational pathology is to maximize the impact of the pathologist, utilizing the decision support tools for the more readily automated work and leaving the complicated cases to the skilled pathologist (so-called augmented intelligence). Better articulation of the benefits using scientific evidence from digital pathology, informed by the pathologist priorities, are needed to implement a culture of change within the community and shift conversations away from the “replacement of the pathologist” towards improvement of care delivery.
The pathologist is the “doctor’s doctor.” Since clinicians’ value accurate and timely pathology reports, improved communication between the referring physician and pathologist is a valued service that pathologists provide.13,14 Where digital pathology is employed to increase workflow efficiencies, the availability of pathology consultations for clinicians could increase and, in conjunction with harmonized and standardized pathology reports, clinician understanding of pathology reports and subsequent decision-making could be improved.15 The opportunity to share annotated, whole slide imaging at multidisciplinary tumor board meetings, for example, will lend to increased understanding and better implementation of interventions (such as surgery). This could result in more efficient interventions and, possibly, fewer patient complications. The ability of the pathologist to access and diagnose digitized images of a specimen at any time, especially remotely, can provide the clinician with the information needed to initiate an appropriate course of therapy sooner and thus provide a better outcome for the patient.
Uptake of digital pathology has the potential to reduce turn-around time and improve the accuracy of pathology reports, both of which are important for patients. It also has the potential to reduce the time from diagnosis to treatment.16 Patients can experience significant anxiety when awaiting a pathology report to determine diagnosis, prognosis, or treatment plan, and incorrect test results lead to significantly delayed treatment initiation.17 Further, the clinician’s recommendations are highly influential to patient decision-making and confidence in their medical team is essential.18 With the uptake of digital pathology, patients may benefit from a greater potential to receive second opinions and expert consultations for difficult to treat specimens, particularly for patients living in rural communities for example, who have less access to care at academic medical centers.19 Knowing that diagnostic results will influence treatment decisions, patients have clear benefits from receiving timely and appropriate therapy. This could result in better outcomes, including longer survival, and fewer complications.
A Framework for Evidence Generation
A key component of improving reimbursement for digital pathology will be not only defining the value it can provide but also producing evidence of its clinical value for payers and application to payment guidelines. When evaluating the various stakeholders to be impacted by digital pathology, both clinical and economic arguments of value should be considered. Use cases where value has been successfully demonstrated through evidence for other innovative technologies need to be explored to guide efforts for digital pathology. There are three key steps that leaders in the space will need to take to create evidence that will support reimbursement for digital pathology technologies.
“To address this standardization issue, digitization of pathology images and introduction of objective decision support tools will create a harmonization of diagnosis and consistent treatment decision-making”
1. Define key endpoints and outcomes
First, innovators in this space must outline the key outcomes relevant to demonstration of the value-add of the For payers this includes not only the changes in decision-making for patient care, but also the influence of those changes on clinical outcomes. In its fullest version, one could seek to evaluate changes in morbidity and mortality but, in the near-term, more intermediate endpoints such as improvement in response to therapy or delay in disease progression are useful. Additionally, we must consider benefit-risk assessments and development of cost-effectiveness data that can impact to relevant stakeholders including the hospital system and commercial payers.
2. Identify limitations of the current care paradigm
Before focusing on the value add of digital pathology services specifically, it will be important to create a foundational understanding of the limitations of the current care paradigm so that the unique benefits of the technology can be highlighted. This can most effectively be established via a retrospective review of existing care decision-making and treatment patterns via chart review or claims analysis. For completeness these evaluations should include both community and academic settings given that the care paradigm can vary considerably in different settings.
3. Define Clinical Utility of Digital Pathology Intervention
With an understanding of the historic care paradigm in hand, we are now well positioned to create evidence that highlights the specific value of digital pathology system interventions. In its simplest form this can be achieved through an observational study which compares the selected endpoints from a program using the digital pathology technology to one without. Several key questions will need to be answered/ evaluated including how one would know:
- the correct tests were selected and executed?
- the diagnosis was correct/incorrect?
- the patient received the wrong intervention, including treatments?
- the clinical impacts been improved and/or side effects could have been reduced if a different intervention was considered?
In its fullest form, a randomized controlled trial that shows the benefit of a system using digital pathology technologies versus those without would be very useful, but we also recognize that there are relatively few examples of medical technology innovators that have executed these types of studies versus those that have done these in the pharmaceutical development space.
Approach Moving Forward
This is a pivotal moment to demonstrate the value of digital pathology and maximize the impact of this technology to transforming medicine. A unified, collaborative coalition may be required by the digital pathology community to make the case for the timely uptake of digital pathology technology through an aligned effort outlined above involving the different stakeholder groups (see below), each group taking concrete steps to establish interoperable platforms. This requires each stakeholder thinking about the perspective of payers to understand their motivations and evidence needs.
1. Hospital/Laboratory Administrators
Hospital administrators and payers will play an important role in funding the adoption of digital pathology The same integrated delivery networks and thought leading institutions we outline above should be well positioned to facilitate data collection to demonstrate the impact and justify the use of technologies in practice. All too often after a decision to adopt has been made there is no evaluation of its benefits. Instead, we (the authors) recommend early adopters of these technologies to share their clinical and claims data with innovators so that the impact can be analyzed, and we can build on the early successes and learn from the failures.
Entities such as the Alliance for Digital Pathology and the Digital Pathology Association can coordinate efforts but, at its core, successful value demonstration will require the combined efforts of individual physician champions. This will need to be a combination not only of the pathology users but also the clinicians that use this information to inform care. They should also consider collaborative efforts to develop standards and harmonization
“These aligned efforts should involve harmonizing data collection and interoperable platforms, leading to a virtuous cycle to gather evidence”
efforts to streamline development of supportive clinical evidence. At the specialty society level, these groups should explore frameworks to develop information in post-market setting to contribute to the totality of evidence to support clinical utility.
Innovators must identify key use cases for the technology that can serve as models for the disruptive value that the technology can provide. They must also help define what evidence can support reimbursement and provide guidance/funding to generate it. Further, leaders should join together to reform coding, as well as developing new coding for a spectrum of locked- and continuous- learning algorithms and proprietary laboratory analyses (PLA) determinations. Finally, they should consider pursuing legislative changes or policy requirements for government-sponsored studies designed to support digital pathology.
Patient groups will need to be engaged first, to solicit input and support and then recruited to define how the clinical value propositions translate to improvements in their care. Particular components that are problematic for patients such as a long diagnostic odyssey or incorrect diagnoses that could have been averted or avoided can be highlighted and articulated to support an understanding of the impact of the technologies in practice.
Payers also have a responsibility to pay attention to the space and define the data they need to make decisions on reimbursements. CMS has often taken the lead in creating innovative programs and payment pathways that are then taken up by commercial payers (e.g., DRG system). Once technologies have been initially adopted, commercial payers have access to new claims data they can employ to create real-world evidence about the benefits/limitations of technology use in practice. Given that most payers have only limited detail in claims, ideally this data can be combined with EMR systems to track not only services/activity changes but also clinical impact and outcomes.
Long Term and Planning for the Future
This coalition-based approach can be very successful when there are clear goals and aligned incentives in place. Improvement to reimbursement for digital technology requires not only creation of novel CPT coding but also alignment with the growing trend towards value-based care. Getting this alignment right is critically important to establish a pathway for all the great digital pathology technologies that have yet to be invented for this space. These aligned efforts should involve harmonizing data collection and interoperable platforms, leading to a virtuous cycle to gather evidence; align on approaches to demonstrate clinical utility and comparative effectiveness; and provides data resources to justify coverage and reimbursement decisions.