What’s Past Is Prologue:
Medicare, Radiology, and the Strange Migration from Film to Digital
##
The CMS-RVU migration from film to digital radiology was not simply a technology upgrade; it became a revealing Medicare payment episode.
Around 2013–2017, CMS and the AMA/RUC translated the disappearance of film, processors, view boxes, and related supplies into the mechanics of practice expense RVUs. Film-era inputs were removed from hundreds of imaging codes, PACS workstations were added, and CMS estimated about $240 million in annual budget-neutral redistribution—small relative to the whole Physician Fee Schedule, but roughly 5% of radiology’s Medicare allowed charges.
The episode had a distinctive mood: inevitability, suspicion, and invoice-driven trench warfare. Radiology accepted that film was obsolete but argued that CMS was underpricing the digital replacement by treating PACS like a desktop computer.
The history is not a perfect parallel to digital pathology or AI, but it is a useful prologue. It shows how Medicare recognizes technology transitions: slowly, mechanically, and only when costs can be named, priced, and mapped to codes.
##
The Medicare payment system did not wake up one morning and say, “Radiology is digital now.”
But it did something more peculiar, and more Medicare-like.
It took one of the largest technology transitions in modern medicine—the disappearance of film, processors, view boxes, film jackets, and film-handling infrastructure—and translated it into the narrow grammar of the Medicare Physician Fee Schedule: practice expense inputs, equipment minutes, invoices, direct costs, indirect costs, and budget neutrality.
That history is worth revisiting. Not because radiology’s transition from film to PACS is a perfect analogy to digital pathology, whole-slide imaging, or AI-enabled diagnostics. It is not. In some ways, the analogy is attenuated almost to the vanishing point. Radiology had an obsolete physical input—film—that could be removed. Digital pathology is often trying to add a new digital layer where the old analog workflow still exists, or where the new costs are less easily reduced to a named piece of equipment. But “what’s past is prologue.” The better we understand the film-to-digital episode, the better our toolbox for thinking about digital pathology, AI software, and the economics of technological migration under Medicare.
The basic episode occurred around 2013–2017. The AMA/RUC Practice Expense Subcommittee’s Migration from Film to Digital Imaging Workgroup recommended that CMS remove film-related supplies and equipment from hundreds of imaging codes and replace them with PACS-related equipment. The AMA later summarized the workgroup’s recommendation as removing 21 supply items and nine equipment inputs from 604 imaging CPT codes and replacing those film-era inputs with PACS equipment. CMS accepted the basic concept, although it also went beyond the RUC recommendation in some respects, removing film-related inputs from additional codes that the RUC had intentionally excluded. [1]
The central Medicare question was not whether PACS existed. By then, that argument was over. The question was whether the Resource-Based Relative Value Scale could recognize the new digital cost structure without simply stripping away the old film costs and adding back a thin, underpriced substitute. In other words, was the new digital environment a true replacement cost, a lower-cost substitute, an indirect infrastructure cost, or some combination of all three?
CMS initially took a cautious and somewhat mechanical approach. For CY2015, CMS removed film-era practice expense inputs and used a desktop computer, equipment item ED021, as a proxy for a PACS workstation because it lacked adequate invoices for a more accurate PACS input. In CY2016, after receiving invoices, CMS raised the PACS workstation price from $2,501 to $5,557, creating a more specific technical PACS workstation input, ED050. [2] (Federal Register)
That first stage explains much of the radiology community’s irritation. CMS was not paying a bonus for digital adoption. It was saying: film is no longer typical, so film should come out. PACS is now typical, but PACS must be priced through the ordinary PE methodology. In practice, that meant subtraction first and replacement second. And in 2015, the replacement looked suspiciously like a desktop computer.
The mood of the time can be captured in the phrase “hidden reductions.” One radiology practice-management commentary from early 2015 described CMS as recognizing that most practices had moved from film to digital PACS, but removing costs associated with film production and storage while replacing them with only $2,501 for what CMS treated as a typical digital workstation. The article complained that this failed to recognize central PACS hardware and software, and cited ACR examples where the film-based practice expense fell sharply while the desktop-computer proxy added back very little. [3]
That was the radiology community’s basic point: yes, film is gone, but PACS is not a Dell desktop from Staples.
AAPM’s 2014 comment letter made the same argument in a more technical register. AAPM agreed that many film-era supplies and equipment no longer represented typical digital practice. But it opposed CMS’s use of the desktop computer ED021 as a proxy for a PACS workstation and recommended delaying the migration until CMS could obtain accurate PACS pricing. [4] (AAPM)
There is something almost charmingly Medicare-ish about the dispute. Everyone agreed about the technological direction. No one was arguing that Medicare should preserve film inputs forever. The fight was over the proxy. Was a PACS workstation a desktop computer, a technical workstation, a professional workstation, an enterprise system, or an entire digital architecture? The answer, of course, was “yes,” but Medicare payment systems do not like “yes.” They like a price, a useful life, a utilization assumption, and equipment minutes.
By mid-2015, the radiology mood improved somewhat. CMS proposed increasing the PACS workstation price from $2,501 to $5,557 based on submitted invoices. Imaging Technology News framed this as CMS acting on ACR imaging-policy recommendations and said the change should improve reimbursement for many radiology exams and procedures. [5] AuntMinnie quoted ACR’s Cynthia Moran saying the College was “very pleased” that CMS had updated the workstation value, emphasizing the work ACR physicians and staff had done to obtain invoices supporting the higher number. [6] Radiology Today likewise treated the 2016 fee schedule as “mostly positive,” noting that CMS had moved from the $2,501 proxy to the $5,557 invoice-based price. [7] (AuntMinnie)
But the relief was partial. The technical PACS workstation did not settle the professional side of the imaging workflow. Radiologists argued that the interpreting physician’s workstation was a real, high-cost, direct resource replacing film alternators, processors, and view boxes historically embedded in practice expense inputs. CMS acknowledged the concept but struggled to fit it into the architecture of the PFS. Physician interpretation lives in the professional component and work RVUs; direct PE inputs are more naturally attached to the technical or global component. Nevertheless, CMS eventually accepted a professional PACS workstation input.
For CY2017, CMS created ED053, a professional PACS workstation priced at $14,616.93. The price was built from submitted invoices for a PC tower, two 3-megapixel monitors, keyboard/mouse, UPS backup, and a PACS monitor/workstation switch. CMS did not simply accept everything commenters wanted. It rejected some proposed additions, such as extra monitors, speech recognition equipment, and microphone items, as indirect costs or as not sufficiently analogous to the film-era inputs being replaced. [8] (Federal Register)
CMS also had to decide how many minutes of the professional PACS workstation to assign. For diagnostic codes, CMS assigned equipment time equal to half the preservice physician work time plus the full intraservice physician work time. For diagnostic codes without a detailed service-period time breakdown, CMS used half the total physician work time. For therapeutic codes, CMS used half the preservice plus half the postservice physician work time, reasoning that the workstation would generally not be used during the intraservice portion of many therapeutic procedures. [8]
That is a marvelous Medicare sentence: digital transformation translated into fractions of preservice, intraservice, and postservice physician time.
The financial scale was not trivial. The AMA text notes that CMS reported the migration from film to digital technology accounted for about $240 million in annual redistribution within the Medicare Physician Fee Schedule. That should not be read as CMS discovering that digital imaging cost $240 million more than film, nor as a literal film-savings account that PACS released. It was a budget-neutral redistribution caused by changing the relative values of many services. Film-related inputs came out; PACS inputs came in; the net effect shifted RVUs and dollars within the PFS.
The scale is useful. CMS’s corrected CY2016 impact table showed about $89.020 billion in total PFS allowed charges and about $4.494 billion for the radiology specialty line. Thus, $240 million was about 0.27% of the entire PFS, or roughly 1 part in 371. But relative to radiology alone, it was about 5.3%. If one uses a broader imaging-like denominator—radiology, IDTFs, interventional radiology, nuclear medicine, and portable X-ray—the denominator is around $5.67 billion, making $240 million about 4.2%. [9] (Centers for Medicare & Medicaid Services)
So, at the level of the whole Physician Fee Schedule, it was a small ripple. At the level of imaging economics, it was not small at all.
SIDEBAR: The Anti-Film Modifiers FX, FY
The later modifier story added a second, more coercive layer. The practice-expense migration changed the base RVU architecture. The FX and FY modifiers created claim-level payment penalties for lagging technology. Beginning in 2017, the FX modifier applied to X-rays taken using film, reducing the technical component payment by 20%. CMS emphasized that this was a statutory payment reduction, created by the Consolidated Appropriations Act, 2016, and exempt from ordinary PFS budget neutrality. [8]
Then, beginning in 2018, CMS implemented the FY modifier for X-rays taken using computed radiography, meaning cassette-based imaging using an imaging plate. That reduction was 7% from 2018 through 2022 and 10% beginning in 2023, again applied to the technical component. [10]
This creates a useful distinction. The RUC/CMS practice-expense migration said: digital is now typical, so the base payment system should be rebased. The FX/FY statutory modifiers said: if you still use older technology, your payment will be cut. One was an RVU rebasing exercise. The other was a penalty regime.
For digital pathology, that contrast matters. There is currently no obvious analogue to an FX modifier saying, “If you continue glass-slide-only pathology, your technical component will be cut by 20%.”
Nor is there a sweeping CMS/RUC migration that says, “Whole-slide imaging is now typical across hundreds of pathology codes, so analog inputs should be removed and digital scanner, storage, and workstation inputs added.” (And recall, for imaging there was an AMA subcommittee.)
Instead, the field is still fighting at the margins: Category III whole-slide imaging add-on codes, software-enabled service taxonomy, AI coding debates, PLA and MAAA boundary problems, and the persistent question of whether digital infrastructure is direct PE, indirect PE, separately billable software, or simply bundled overhead.
The radiology episode was easier in one crucial way: the obsolete cost was visible. Film was a named supply. Film processors, view boxes, alternators, and film storage were physical things. CMS could remove them. PACS workstations were also physical enough to be invoiced, priced, assigned useful lives, and allocated minutes.
Digital pathology and AI are less cooperative. The relevant costs include whole-slide scanners, image management systems, storage, cybersecurity, interfaces, pathologist workstations, algorithm licensing, model monitoring, validation, version control, cloud compute, and enterprise IT. Some of these look like direct costs. Many look like indirect costs. Some are capital assets. Some are SaaS. Some support a particular patient encounter; others support the entire enterprise. Medicare PE methodology is more comfortable with boxes than with ecosystems.
That may be the most important lesson. When is it a doctor's workstation, when is it a server? CMS accepted the PACS workstation when radiology could define the box. CMS resisted the broader PACS ecosystem when the costs looked like infrastructure, software, training, maintenance, storage, or general IT. The same pattern is likely to recur in digital pathology. A scanner, monitor, or workstation may be easier to discuss in PE terms than cloud storage, AI model governance, cybersecurity, or platform licensing. The latter may be real costs, but “real” and “direct PE input” are not the same thing.
The film-to-digital episode also shows the value of invoice-driven trench warfare. The winning strategy was not simply to argue that digital imaging was important. It was to collect invoices, distinguish technical and professional workstations, contest proxy pricing, map equipment time to CPT codes, and show what was typical. That is not glamorous. But in Medicare PE policy, glamour rarely gets you paid.
The mood of the time, then, had three layers. First, inevitability: everyone knew film was disappearing. Second, suspicion: radiology feared CMS was using the disappearance of film to mine imaging for budget-neutral savings. Third, pragmatism: the field eventually improved its position by producing invoices and forcing CMS to recognize more accurate PACS inputs.
For digital pathology, the precedent should be used carefully. It does not prove that CMS will broadly fund the digital pathology transition. It may even explain why the transition is so difficult to fund through ordinary PFS mechanics. Radiology had a legacy cost that could be removed and a replacement workstation that could be priced. Digital pathology often has additive costs layered onto a still-functional analog architecture. AI adds another level of abstraction, because the resource may not be a box at all, but a software-enabled interpretive function.
Still, the old radiology story gives us a vocabulary: typicality, direct versus indirect PE, invoices, equipment minutes, specialty impact, budget neutrality, and statutory modifiers. Those are not the whole toolbox for digital pathology. But they are tools worth carrying.
#####
SIDEBAR:
The AMA "Migration from Film to Digital Imaging Workgroup"
The Migration from Film to Digital Imaging Workgroup was an AMA/RUC practice-expense workgroup created to solve a deceptively simple problem: Medicare’s radiology practice-expense inputs still contained the old world of film, processors, view boxes, VHS tapes, envelopes, developer, fixer, and film alternators, while radiology practice had largely moved to PACS.
The workgroup sat under the Practice Expense Subcommittee of the AMA/Specialty Society RVS Update Committee, or RUC. In April 2012, the PE Subcommittee recommended forming a multispecialty workgroup, including ACR and other specialty societies, to develop PACS practice-expense inputs to replace film-related inputs for appropriate codes. By October 2012, the group had met twice by conference call, developed a draft list of codes, solicited specialty-society comment, and identified 605 CPT codes for possible migration, along with 21 film-related supply items and nine film-related equipment items to be replaced. (American Medical Association)
By January 2013, the workgroup was wrestling with the problem in a more granular way. Its stated task was to identify the typical PACS system that a physician or group of physicians would own in a nonfacility physician practice. It planned to group inputs by broad modality buckets, such as plain film, CT, angiography, and echocardiography. It also specifically asked whether storage costs varied enough by service to justify different PACS storage inputs across different imaging code families. (American Medical Association)
By April 2013, the workgroup had completed its work. The final RUC minutes say that 604 imaging CPT codes had been identified for transition to digital equipment. The workgroup recommended removing the specified film supplies and equipment from those codes, replacing them with recommended PACS equipment, providing invoices from specialties, and making no immediate clinical labor modifications for existing codes. For new or newly reviewed imaging codes, however, the workgroup recommended more detailed digital-era clinical labor activities, such as confirming availability of prior images, technologist QC of images in PACS, scanning exam documents into PACS, completing the exam in the RIS, and populating the radiologist work queue. (American Medical Association)
CMS later summarized the RUC recommendation in the CY2015 Physician Fee Schedule rule. CMS said the RUC recommended removing film-related supply and equipment items because they were no longer typical, and adding PACS equipment because PACS was now typically used in furnishing imaging services. CMS also immediately exposed the methodological fault line: some PACS-related items looked like direct PE inputs, while others looked like indirect PE costs, and CMS would not treat broad infrastructure costs as direct costs unless they were individually allocable to a particular patient and procedure. (Federal Register)
The workgroup’s importance is that it created the formal bridge from an obvious technological fact—film was disappearing—to the machinery of the Medicare fee schedule. It gave CMS a crosswalk: remove these film inputs, consider these PACS inputs, apply them across hundreds of CPT codes, and revise digital-era labor descriptions prospectively as codes come up for review. The AMA’s later practice-expense guide summarized the final arc: after two years of effort, the workgroup’s recommendations led to removal of 21 supply items and nine equipment inputs from 604 imaging CPT codes, replacement with PACS equipment, CMS’s initial use of the ED021 desktop-computer proxy, later use of ED050 PACS workstation pricing, and eventually the ED053 professional PACS workstation. (American Medical Association)
For your article, the sidebar point is this: the workgroup was not a glamour committee about “the future of imaging.” It was more like a forensic cleanup crew for obsolete practice-expense inputs. Its legacy is that it showed how a technology migration becomes real in Medicare payment policy: not through rhetoric about transformation, but through code lists, invoices, equipment names, storage assumptions, labor tasks, and the eternal fight over whether a cost is direct or indirect. some PACS-related items looked like direct PE inputs, while others looked like indirect PE costs, and CMS would not treat broad infrastructure costs as direct costs unless they were individually allocable to a particular patient and procedure. (Federal Register)
The workgroup’s importance is that it created the formal bridge from an obvious technological fact—film was disappearing—to the machinery of the Medicare fee schedule. It gave CMS a crosswalk: remove these film inputs, consider these PACS inputs, apply them across hundreds of CPT codes, and revise digital-era labor descriptions prospectively as codes come up for review. The AMA’s later practice-expense guide summarized the final arc: after two years of effort, the workgroup’s recommendations led to removal of 21 supply items and nine equipment inputs from 604 imaging CPT codes, replacement with PACS equipment, CMS’s initial use of the ED021 desktop-computer proxy, later use of ED050 PACS workstation pricing, and eventually the ED053 professional PACS workstation. (American Medical Association)
The sidebar point is this: the workgroup was not a glamour committee about “the future of imaging.” It was more like a forensic cleanup crew for obsolete practice-expense inputs. Its legacy is that it showed how a technology migration becomes real in Medicare payment policy: not through rhetoric about transformation, but through code lists, invoices, equipment names, storage assumptions, labor tasks, and the eternal fight over whether a cost is direct or indirect.
#####
This article was entirely researched and written by Chat GPT 5.5.
References
[1] American Medical Association. Practice Expense Component / AMA practice expense manual excerpt discussing “Migration from Film to Digital Practice Expense Inputs.”
https://www.ama-assn.org/system/files/practice-expense-component.pdf
[2] Centers for Medicare & Medicaid Services. CY2016 Physician Fee Schedule final rule. Discussion of PACS workstation pricing, ED021 proxy, ED050, and film-to-digital PE changes. Federal Register, November 16, 2015.
https://www.federalregister.gov/documents/2015/11/16/2015-28005/medicare-program-revisions-to-payment-policies-under-the-physician-fee-schedule-and-other-revisions
[3] HAP / Healthcare Administrative Partners. “Hidden Reductions in the Medicare 2015 Physician Fee Schedule.”
https://info.hapusa.com/blog-0/hidden-reductions-in-the-medicare-2015-physician-fee-schedule
[4] American Association of Physicists in Medicine. “Physician Fee Schedule 2015 Proposed Rule Comments.” August 2014.
https://www.aapm.org/government_affairs/CMS/documents/PhysicianFeeSchedule2015ProposedRuleCommentsFINAL08-14.pdf
[5] Imaging Technology News. “CMS Proposes Actions on ACR Imaging Policy Recommendations.”
https://www.itnonline.com/content/cms-proposes-actions-acr-imaging-policy-recommendations
[6] AuntMinnie. “Radiology averts cuts in proposed 2016 MPFS.”
https://www.auntminnie.com/practice-management/article/15612682/radiology-averts-cuts-in-proposed-2016-mpfs
[7] Radiology Today. “Medicare Fee Schedule: Mostly Positive for Radiologists in 2016.”
https://www.radiologytoday.net/archive/rt0216p32.shtml
[8] Centers for Medicare & Medicaid Services. CY2017 Physician Fee Schedule final rule. Discussion of ED053 professional PACS workstation and FX modifier. Federal Register, November 15, 2016.
https://www.federalregister.gov/documents/2016/11/15/2016-26668/medicare-program-revisions-to-payment-policies-under-the-physician-fee-schedule-and-other-revisions
[9] Centers for Medicare & Medicaid Services. CY2016 Physician Fee Schedule corrections notice, including corrected impact table showing total allowed charges and radiology allowed charges. Federal Register, March 8, 2016.
https://www.federalregister.gov/documents/2016/03/08/2016-05054/medicare-program-revisions-to-payment-policies-under-the-physician-fee-schedule-and-other-revisions
[10] Centers for Medicare & Medicaid Services. CY2018 Physician Fee Schedule final rule. Discussion of FY modifier for computed radiography payment reduction. Federal Register, November 15, 2017.
https://www.federalregister.gov/documents/2017/11/15/2017-23953/medicare-program-revisions-to-payment-policies-under-the-physician-fee-schedule-and-other-revisions
