Monday, June 22, 2026

CMS Coverage History: How "CAC" Meetings Evolved (2004, 2015, 2019)

 Medicare's contractor advisory committees have evolved greatly from 2005, to 2015, to the present (2019 forward).   I've put the Medicare LCD Manual from each period into one cloud zip file (here).

I asked Claude Opus to evaluate those three files, and compare CAC meetings under each era.

Afterward, I ask Chat GPT to put the CAC evolution 2004-2019-today, into a broader perspective reaching back to the state-level structure of Medicare in the 1960s.

###

From Carrier to Contractor: How Medicare Rebuilt the CAC Between 2004 and 2019

A reading of three Program Integrity Manual layers — and why today's Contractor Advisory Committee looks nothing like the standing state roster of 2005.

If you have ever sat on a Carrier Advisory Committee, the modern CAC can be disorienting. The thing many of us remember from 2005 was a standing roster of specialists — one seat per specialty society, often fifteen or twenty members, meeting quarterly in a hotel conference room, receiving each draft Local Coverage Determination as it went out the door to public comment. The thing we attend in 2026 is something else entirely: a half-dozen experts, frequently national rather than in-state, convened for a single afternoon to work through ten or twelve structured questions about the evidence on a topic that does not yet have a draft LCD attached to it — the whole session recorded and posted as a transcript.

That transformation is fully documented in three sedimentary layers of one manual chapter. Reading them in sequence shows that the change was not a gradual drift. It was a single, discontinuous response to a statute, bolted onto more than a decade of deferred housekeeping.

Read the revision stamps first

The three documents are not freestanding policies. They are the same Program Integrity Manual section at three points in time, and the revision stamps tell most of the story before you read a word of substance.

The 2004 text carries Rev. 63 (01-23-04) and sits in the old PIM numbering as §8.1. The 2015 text is the same manual, renumbered to §13.8.1, but its bones are still Rev. 71 (04-09-04), with the structure-and-process subsection bumped only to Rev. 473 (06-21-13). The 2019 text is a different creature altogether: Rev. 863, issued 02-12-19, effective 10-03-18, implementation January 8, 2019, relocated to §13.2.4.3.

That January 8, 2019 date is the tell. It is the go-live date of CMS's wholesale rewrite of the LCD process, issued under Change Request 10901 to carry out Section 4009 of the 21st Century Cures Act (P.L. 114-255, signed December 13, 2016), codified at Social Security Act §1862(l)(5)(D). So you are really comparing two eras, not three: a carrier-era artifact that barely moved between 2004 and 2015, and a Cures-driven reconstruction in 2019.

In a companion blog, I discuss a 2004 copy of "exhibit 3" about CAC membership.

Sidebar — The CAC across three manual layers

Feature 2004 (Rev. 63) 2015 (Rev. 71 / 473) 2019 (Rev. 863)
Name Carrier Advisory Committee Carrier Advisory Committee (text still says "Carriers") Contractor Advisory Committee; actor is the MAC
Geography One CAC per state, rigidly One CAC per state, rigidly Per state, per jurisdiction, or multi-jurisdictional / multi-MAC
Membership theory Specialty representation; one seat + alternate per specialty Same Evidence expertise; "more voices" (nurses, epidemiologists, etc.)
Core function Dissemination & liaison; carry drafts to societies Same Review the quality of the evidence
Link to comment clock 45-day clock starts on distribution to the committee Same Severed; comment period is a separate statutory step
Timing vs. draft CAC reviews the draft LCD as released Same May meet before a proposed LCD exists, on the topic's evidence
Cadence ≥ 3 meetings/year, ≤ 4 months apart Same No minimum; MAC sets frequency by LCD volume
Leadership CMD + elected physician co-chair; 2–3 yr tenure Same Co-chair / tenure language dropped; MAC approves members
Transparency Minutes to members & RO; web posting optional Same (+ PSC/GTL overlay) Recorded (audio/video), public, posted as LCD record
Compensation None; voluntary service None None (the one provision carried forward verbatim)

2004–2015: the standing state roster as a representation machine

In the first era the committee is the Carrier Advisory Committee, and almost every structural choice flows from that one word. One CAC per state; where a carrier spanned several states, each state got "a full committee and the opportunity to discuss draft LCDs." The committee was built around specialty representation: one member plus a designated alternate per specialty, with additional specialists pulled in when their expertise was implicated. In practice that produces exactly the body many of us remember — a standing roster running to fifteen or twenty seats, because the design goal was to have the relevant society's voice already in the room for whatever LCD surfaced.

The functional theory was dissemination and liaison, not adjudication of evidence. The "Role of CAC Members" section is explicit: members existed to carry proposed LCDs back out to their state and specialty societies to solicit comment, to carry Medicare program information back in, and to flag inconsistent or conflicting medical-review policy. They were conduits. The contractor medical director co-chaired with one physician elected by the committee, and the co-chairs set the agenda and pushed background material out at least fourteen days ahead.

Two features matter most for understanding what later broke. First, the CAC was wired directly into the comment clock: the 45-day comment period "starts when the proposed LCD is distributed to the committee members," and the co-chairs "present all proposed LCDs to the CAC for discussion." The CAC meeting was the public-facing event for a draft LCD — which is precisely the co-distribution-to-CAC-and-public simultaneity of the 2005 model. Second, the cadence was fixed and frequent: a minimum of three meetings a year, no more than four months apart. That is a standing quarterly body by design, because a standing body is what you need when the committee is the routing point for every draft in the jurisdiction.

What is striking is how little changed by 2015. The Medicare Modernization Act had already replaced carriers with A/B MACs years earlier, yet the 2015 text still says "Carriers shall establish," still says one CAC per state, still prescribes three meetings a year, still triggers the comment clock on distribution to the committee, still names an elected physician co-chair, still suggests two- to three-year tenures. The only real accretion is a maintenance overlay reflecting the program-integrity contractor apparatus — the parenthetical "(for PSCs, the GTL, Co-GTL, and SME)" threaded through the directory, waiver, and recordkeeping provisions — plus the 2013 (Rev. 473) refresh of the structure section. Between 2004 and 2015, CMS updated the plumbing references and left the committee's constitutional theory entirely intact. The 2015 document is a carrier-era document wearing a MAC-era nametag.



2019: the Cures rupture

The 2019 rewrite changes the committee's purpose, not merely its logistics, and nearly everything else follows from that.

Carrier becomes Contractor. The committee is renamed, the actor throughout is "MACs," and the carrier vocabulary is finally retired a decade and a half late.

The geographic unit is loosened. Where the old rule was rigidly one-CAC-per-state, 2019 permits one CAC per state, one per jurisdiction, or a multi-jurisdictional CAC with representation from each state — and CMS's companion LCD Process Q&A confirms that multi-MAC CACs are also on the table. This is the textual hinge that permits the modern reality. Once you may convene a single panel for an eleven-state jurisdiction, the chair of pathology at one academic center and the chair at another become natural picks, because you are no longer staffing for in-state representation. You are staffing for the best available evidence reviewers.

Representation becomes evidence review. This is the conceptual core. The old purpose — a formal mechanism for physicians in the state to be informed of and participate in LCD development, plus a liaison forum — is replaced by a purpose centered on the evidence: members "review the quality of the evidence used in the development of an LCD." CMS said the quiet part plainly in the Q&A: because Cures now requires LCDs to include a summary of the evidence considered and a rationale, the role of CAC members should evolve to reviewing the quality of that evidence instead of only representing a constituency. The 2019 language even imports the deliberative register of the national MEDCAC — advice is "most useful when it results from a process of full scientific inquiry and thoughtful discussion with careful framing of recommendations," and the panel exists to ensure an "unbiased and contemporary consideration of 'state of the art' technology and science." That is consciously MEDCAC-style framing dropped onto a local body.

The comment-clock coupling is severed. Note what disappears: the entire mechanism by which distributing a draft LCD to the committee started the 45-day clock, and the instruction that co-chairs present "all proposed LCDs" to the CAC. This is the most consequential deletion. Cures created, by statute, a separate and mandatory public process — the MAC must post the determination in its entirety, identify where and when it was first made public, and supply a rationale, at least forty-five days before the effective date. Once the open public-comment period and the proposed-LCD posting are independent statutory steps, the CAC no longer needs to be the comment vehicle. It is freed to move upstream: it now convenes around a contemplated topic and its evidence base. The Q&A makes the new optionality explicit — a MAC may host the CAC before posting a proposed LCD, to inform the upfront analysis, or after. That is exactly the shift from the 2005 picture (draft presented to the roster, co-distributed to the public, clock running) to the 2026 picture (a future topic, structured evidence questions, no draft on the table yet).

Transparency is mandated, and it reshapes the format. The 2019 rule requires the MAC to record every CAC meeting — video, audio, or both — and to maintain that recording on the contractor website as part of the LCD record. All CAC meetings are open to the public to attend and observe; only the non-evidence portions (provider practice-trend reporting, fraud-and-abuse discussion) may be closed. The old regime asked only for minutes distributed to members and forwarded to the regional office within ten days, with web posting optional and FOIA-redacted. The published transcript of a modern CAC is the direct artifact of this recording-and-posting mandate, and the structured, ten-to-twelve-question, roughly two-hour format is the natural operational form of "review the quality of the evidence" plus "careful framing of recommendations." It is a local echo of the MEDCAC voting-question script, engineered to produce a clean, citable evidentiary record rather than a roundtable of society liaisons.

The standing body dissolves into an ad hoc panel. Several deletions point the same way. The fixed cadence is gone; the Q&A confirms MACs now set frequency "based on the appropriateness and volume of LCDs requiring CAC input," and a CAC is not required for every LCD. The two- to three-year tenure guidance is gone. The elected physician co-chair structure is gone; member approval now rests simply with the MAC. The dissemination-to-societies role is gone, because the statutory public-comment process now performs that solicitation directly and at scale. The reference to the Practicing Physicians Advisory Council is gone (PPAC itself had been sunset earlier in the decade). The substitution rule tightens and formalizes — a substitute may attend only if the MAC is notified and approves at least one week ahead — which is the housekeeping you would expect once a meeting is a recorded, public event of record. About the only provision carried forward verbatim is the honorarium prohibition: participation remains voluntary, expenses on the member or the society.

The net trajectory — and why 2015 is the instructive case

Put the two eras side by side and the arc is coherent. The carrier-era CAC was a representative standing committee — a per-state roster of society delegates, meeting quarterly, functioning as the routing point and comment trigger for draft LCDs, steered by the contractor medical director and an elected physician. The Cures-era CAC is an evidence-review panel — convened at jurisdictional or multi-jurisdictional scale, often stocked with national figures chosen for expertise rather than in-state constituency, addressing a future topic through structured questions in a recorded, public, transcribed session that becomes part of the LCD record, with the comment period and the proposed-LCD posting hived off into their own statutory channels.

The 2015 document is instructive precisely because it isn't a transition. It shows that as late as 2015 CMS was content to leave the 2004 design frozen, which means the 2019 change was not a slow evolution but a single discontinuous response to a statutory mandate. The "many other ad hoc changes" — the multi-jurisdictional option, the substitution mechanics, the recording medium, the closable fraud-and-abuse portions — are best read as CMS using the Cures rewrite as the occasion to clear a decade of deferred modernization in one transmittal.

One caveat for readers who want to follow the functions rather than the committee: these provisions capture the CAC specifically. The full 2019 Chapter 13 rewrite also created or relocated the informal pre-LCD meeting, the formal open meeting, the new-LCD request process, and the comment-and-response (RTC) machinery. Those companion provisions are what make the upstream repositioning of the CAC coherent. The dissemination and comment functions the CAC used to carry did not vanish in 2019; they were redistributed into a statutory public process that the committee no longer needs to host.

Key documents


###

###

I ask Chag GPT to put the CAC evolution 2004-2019-today, into a broader perspective reaching back to the state-level structure of Medicare in the 1960s.

###

###

Yes — that is exactly the deeper historical frame. The 2004 CAC language was not just a technocratic CMS instruction. It was also a fossil record of Medicare’s original state-based administrative architecture, especially the world of state Blue Cross and Blue Shield plans, state medical societies, local carriers, local intermediaries, and local physician advisory networks.

The 1965 design: Medicare as a federal program administered through local private plans

When Medicare began, it was national in statute but deliberately decentralized in administration. The 1965 Medicare law created a national Part A hospital insurance benefit and a voluntary Part B medical insurance benefit, but the federal government did not build a single national claims-processing agency from scratch. Instead, Medicare relied on private health insurers to administer fee-for-service claims. CMS still summarizes this historical structure plainly: since Medicare’s inception in 1966, private health care insurers processed Medicare claims; the Part A entities were called fiscal intermediaries, and the Part B entities were called carriers. (Centers for Medicare & Medicaid Services)

That architecture reflected the insurance market Medicare inherited. Blue Cross plans had grown around hospital coverage; Blue Shield plans around physician services. In many states, they were local or state-level nonprofit plans. The Blue Cross Blue Shield Association still describes the system as an association of “independent, locally operated” companies, and its current list still shows many state-named plans: Blue Cross and Blue Shield of Vermont, Massachusetts, Rhode Island, Alabama, Kansas, Michigan, North Carolina, Tennessee, and so on. (Blue Cross Blue Shield Association)

So the early Medicare program did not begin with something like “Noridian Jurisdiction E” or “NGS Jurisdiction K.” It began in a world where the natural administrative unit was often the state or the local Blue plan service area.

Part A and Part B also mapped onto different local institutions

A further complication is that Medicare did not have one kind of contractor. It had fiscal intermediaries for institutional claims and carriers for physician/supplier claims. CMS’s 2006 MAC transition announcement described the old division clearly: fiscal intermediaries processed Part A claims for hospitals, skilled nursing facilities, and other institutional providers, while carriers processed Part B claims for physicians, laboratories, and suppliers. (Centers for Medicare & Medicaid Services)

That distinction matters for CAC history because LCDs, especially Part B medical-review policies, emerged heavily from the carrier/CMD world. Carrier medical directors were often working with state physician societies and local specialty leaders. Thus the CAC was not an abstraction; it was an organized relationship between a Medicare contractor and the state physician community.

In the 1960s and 1970s, that relationship made intuitive sense. A Vermont carrier or Vermont-associated Blue plan could talk to Vermont physicians. A Maine intermediary could know Maine hospitals. Massachusetts Blue Cross could work with Massachusetts institutions. The contractor was not merely processing claims; it was embedded in local provider networks, local professional societies, and local insurance history.

Consolidation of Blues plans made consolidation of CMS contractors inevitable.

Over time, that world consolidated. Blue Cross and Blue Shield plans merged, affiliated, converted, formed holding companies, or operated across multiple states. 

Health services researchers writing in the early 2000s noted that Blue Cross plans had incentives to merge across state lines, including economies of scale and the need to serve multistate employers. (PMC) A 2014 industry review similarly described ongoing consolidation among Blues plans and noted that several Blues organizations operated in multiple states. (Mark Farrah Associates)

By the early 2000s, Medicare contractor territories had become much more aggregated than the 1965 mental model. But the culture and legal-administrative memory of state-level Medicare administration remained. That is why the 2004 CAC rule has the flavor it does: even if one carrier served many states, each state still had to have a full CAC and an opportunity to discuss draft LCDs and issues presented in that state. A single multi-state CAC was allowed only with permission and only if CAC members across the jurisdiction agreed.

That is the “darn it, Vermont gets a Vermont CAC” principle. It is not just bureaucratic fussiness. It reflects the residue of a Medicare program that had been built through state-level carriers, state medical societies, state Blue plans, and state provider relationships.

2004 was a transition moment, not the beginning of the MAC world

The timing is important. The Medicare Modernization Act of 2003 directed CMS, through Section 911, to replace the old Part A fiscal intermediaries and Part B carriers with A/B Medicare Administrative Contractors selected under federal acquisition rules. (Centers for Medicare & Medicaid Services) But in 2004, CMS was still living inside the old contractor world. The 2004 CAC manual still uses “carrier” language, not the later MAC-centered language. It still assumes carrier medical directors, state CACs, physician co-chairs, specialty-society designees, regular meetings, and a state-oriented roster.

CMS did not begin the practical transition from legacy contractors to MACs until later. GAO summarized that CMS started transitioning Medicare claims administration from 51 legacy contractors to MACs in 2006. (GAO) CMS’s own 2006 announcement said the first MAC awards were part of contracting reform and that, when fully implemented, fiscal intermediaries and carriers would be replaced by MACs responsible for both Part A and Part B claims. (Centers for Medicare & Medicaid Services)

So the 2004 CAC text is perched at a hinge point. The MMA had just told CMS to move toward MACs, but the operating manual still preserved the carrier-era, state-based advisory structure.

Why the old "state CACs" mattered politically and professionally

A state CAC was not merely a meeting format. It embodied several claims:

First, local physician legitimacy. If a draft LCD affected gastroenterology in Vermont, Vermont physicians should have a Vermont gastroenterologist in the advisory chain. If it affected surgery, Vermont surgeons should not be represented only by someone from Boston or Hartford.

Second, state medical society mediation. CAC members were expected to disseminate proposed LCDs to colleagues and specialty societies, solicit comments, and bring those comments back. That made sense when state medical societies were important intermediaries between Medicare and practicing physicians.

Third, local utilization context. The 2004 manual expected meetings to include comparative utilization data related to proposed LCDs, with CAC members helping interpret the data and advise on overutilization. Local physicians might know whether an apparent spike reflected abuse, local referral patterns, new technology adoption, coding confusion, or one dominant academic center.

Fourth, state-level trust. The CAC reassured physicians that policy was not being imposed by a distant regional contractor without local input. This mattered because LCDs could have immediate consequences for coverage, claims payment, documentation, and medical practice.

The old contractor model was not always perfectly state-level

We should avoid saying “always” state-level in 1965. The early Medicare contractor map was heavily influenced by state and local Blue plan footprints, but it was not mechanically one contractor per state for every function and every period. There were differences between Part A and Part B, hospital associations and physician insurers, Blue Cross and Blue Shield entities, commercial carriers, and later specialty contractors. Some contractors served more than one state; some states had more than one relevant contractor relationship; and the Part A/Part B split meant a hospital’s Medicare intermediary and a physician’s Part B carrier were not necessarily the same entity.

But as a historical approximation, your point is right: the original administrative ecology was much more state/local-plan based than the modern MAC map. By 2004, aggregation was well underway; by the MAC transition after 2006, it became the organizing principle.

The 2004 CAC rule as institutional memory from 1970s, 1980s

Seen through this lens, the 2004 rule is almost a compromise document.

On one hand, CMS knew contractors were aggregating. The manual explicitly contemplates one carrier serving many states and even allows a single CAC for an entire jurisdiction with permission.

On the other hand, CMS insisted that each state retain a full committee and an opportunity to discuss draft LCDs. That is the institutional memory of the older carrier/intermediary world: state doctors, state hospitals, state Blues, state medical societies, and state-level Medicare relationships.

The manual’s insistence on state CACs can therefore be read as a guardrail against excessive regionalization. CMS was saying, in effect: regional administration may be efficient, but local Medicare policy still needs state-level professional input.

How this differs from the 2019/2026 model

By 2019, that historical memory had largely been displaced by a different legitimacy model. The legitimacy of the CAC no longer came mainly from state representation. It came from transparent evidence review. The 2019 manual allows one CAC per jurisdiction or multi-jurisdictional CAC with representation from each state, and it describes the CAC’s role as reviewing the quality of evidence used in LCD development.

That is a profound change. The 2004 question was: Has the local/state physician community had a chance to review and comment on this draft LCD? The 2019 question is closer to: Has the MAC exposed the evidence base to qualified experts, public observation, and a recorded process?

In 2004, Vermont mattered because it was Vermont. In 2019, Vermont matters if the MAC can say the multi-state process included representation and an opportunity for evidence discussion, but the central currency is no longer state medical identity. It is expertise, transparency, and an evidence record.

A useful way to phrase it

The evolution can be described this way:

Medicare began as a federal benefit administered through a patchwork of local and state private insurance institutions, many of them Blue Cross and Blue Shield plans. 

The CAC inherited that world. Its original function was not simply to “get expert input”; it was to preserve a state-level professional advisory channel between the Medicare carrier and the local physician community. By 2004, that state-based logic was already under pressure from consolidation and the imminent MAC reform, but CMS still preserved it in manual language requiring one CAC per state. 

After the MAC transition and especially after the 21st Century Cures-era LCD reforms, the CAC’s center of gravity shifted from local representation to evidence transparency. The modern CAC is less the descendant of the Vermont Blue plan’s physician advisory committee and more a public evidentiary workshop (like a CMS MedCAC)  for a regional Medicare contractor operating at multi-state, and sometimes functionally national, scale.