MPOG QI - Quality Committee Meeting Notes – Monday, January 26th, 2026

Attendance:

Abess, Alex (Dartmouth)

Jacob, Seth (University of Kansas)

Stam, Benjamin (UMHS West)

Agerson, Ashley (UMH West)

Johnson, Rebecca (UMHS West)

Steadman, Randolph (Houston

Methodist)

Berndt, Brad (Bronson)

Kaper, Jon (Corewell Trenton)

Stumpf, Rachel (MPOG)

Bollini, Mara (WUSTL)

Karamchandani, Kunal (UT

Southwestern)

Tyler, Pam (Corewell Farmington

Hills)

Bow, Peter (Michigan)

Khan, Meraj (Henry Ford)

Vaughn, Shelley (MPOG)

Bowman-Young, Cathlin (ASA)

Kirke, Sarah (Nebraska)

Vitale, Katherine (Trinity Health)

Brennan, Alison (Maryland)

Kunkler, Bryan (Corewell West)

Wade, Meridith (MPOG)

Brown, Morgan (Boston Children’s)

Lacca, Tory (MPOG)

Wedeven, Chris (Holland)

Brown, Sheree (Trinity Health)

LaGorio, John (Trinity Health)

Weinberg, Aaron (Weill Cornell)

Buehler, Kate (MPOG)

Lalonde, Heather (Trinity Health)

Westfall, Christine (Sparrow)

Calabio, Mei (MPOG)

Liu, Bin (Michigan Medicine)

Wissler, Richard (University of

Rochester)

Cassidy, Ruth (MPOG)

Liu, Linda (UCSF)

Woody, Nathan (UNC)

Charette, Megan (MPOG)

Lewandowski, Kristyn (Corewell

Troy)

Yacoubian, Stephanie (B&W)

Chopra, Ketan (Henry Ford - Detroit)

Lopacki, Kayla (Mercy Health -

Muskegon)

Yeoh, Cindy (Moffitt)

Claybaugh, Deborah (MyMichigan)

Lu-Boettcher, Eva (Wisconsin)

Yuan, Yuan (MPOG)

Cohen, Bryan (Henry Ford - West

Bloomfield)

Ludwig, Kristen (Henry Ford)

Zhao, Xinyi (Sarah) (MPOG)

Coleman, Rob (MPOG)

Mathis, Mike (MPOG)

Zittleman, Andrew (MPOG)

Corpus, Charity (Corewell Royal Oak)

Mack, Patricia (Weill Cornell)

Cusick, Jordan (OHSU)

Malenfant, Tiffany (MPOG)

Cywinski, Jacek (Cleveland Clinic)

McComb, Joseph (Temple U)

Delhey, Leanna (MPOG)

McKinney, Mary (Corewell Dearborn

/ Taylor)

Denchev, Krassimir (St Joseph

Oakland)

Mentz, Graciela (MPOG)

Dewhirst, Bill (Dartmouth)

Milliken, Christopher (Sparrow)

Edelman, Tony (MPOG)

O’Conor, Katie (Johns Hopkins)

Ellison, Pavithra (WVU Medicine)

O’Dell, Diana (MPOG)

Esmail, Tariq (Toronto)

Ohlendorf, Brian (Duke)

Everett, Lucy (MGH)

Owens, Wendy (MyMichigan -

Midland)

Finch, Kim (Henry Ford Detroit)

Pace, Nathan (Utah)

Gedela, Radhika (University of

Vermont)

Pantis, Rebecca (MPOG)

Georgiadis, Paige (University of

Vermont)

Pardo, Nichole (Corewell Grosse

Pointe)

Gereboff, Avner (Cedar Sinai)

Paul, Jonathan (Columbia)

Gibbons, Miranda (Maryland)

Pimentel, Marc Phillip (B&W)

Goatley, Jackie (Michigan)

Poindexter, Amy (Holland)

Goldblatt, Josh (Henry Ford

Allegiance)

Qazi, Aisha (Corewell)

Greenblatt, Lorile (U Penn)

Rolfzen, Megan, MD (Michigan

Medicine)

Hall, Meredith (Bronson Battle Creek)

Russell, Michael (University of West

Virginia)

Harwood, Tim (Wake Forest)

Salamanca, Yuliana (Temple)

Heiter, Jerri (St. Joseph A2)

Schwerin, Denise (Bronson)

Herren, Melanie (MPOG)

Shah, Nirav (MPOG)

Horton, Brandy (Anes Associates)

Shettar, Shashank (OUHSC)

Huntington, Michelle (Corewell West)

Smith, Mason (MyMichigan)

Agenda & Notes

Opening, Attendance, and Minutes:

• Meeting Start: 10:03 am

• Roll Call: Via Zoom or contact Coordinating Center (support@mpog.zendesk.com) if you were

present but not listed on Zoom.

• Minutes from November 2025 Quality Committee Meeting

Upcoming Events:

2026 Meetings & Events

• MSQC + ASPIRE Combined Meeting: Friday, March 13 (Marriot, East Lansing).

• ASPIRE-only Meeting: Friday, July 17 (Weber’s Hotel, Ann Arbor).

• MPOG Retreat: Friday, October 16 (San Diego).

Sites outside Michigan are welcome at the Michigan in-state meetings—contact the

Coordinating Center for details.

Announcements

MPOG recognized Bethany Pennington, PharmD (Washington University) as the Featured Member of

the Month for January and February. The committee also welcomed Indiana University Health System as

MPOG’s newest participating site.

Two leadership updates were announced:

• Dr. Sharon Reale (Harvard Medical School / Brigham and Women’s Hospital) was introduced as

the new Obstetric Subcommittee Vice Chair.

• Dr. Eva Lu-Boettcher (University of Wisconsin) was announced as the new Pediatric

Subcommittee Vice Chair.

Both were acknowledged for their longstanding contributions to MPOG research and quality-

improvement efforts.

Glycemic Management Workgroup Update

Dr. Tony Edelman provided an update on the Glycemic Management Workgroup, which was formed

following committee agreement in September 2025 and met in December to explore outpatient

hyperglycemia metrics.

Key Discussion Points

• Patient inclusion: The group agreed that hyperglycemia itself, not a formal diabetes diagnosis,

should trigger inclusion. Non-diabetic patients with elevated glucose values were felt to

represent a clinically meaningful population.

• Treatment thresholds: There was extensive discussion about whether the threshold should be

180 mg/dL or 250 mg/dL, reflecting differences across existing guidelines (e.g., SAMBA vs. other

perioperative recommendations).

• Measure focus: The workgroup emphasized timely treatment and reassessment, rather than

insulin dosing precision.

• Data considerations: IV insulin was discussed but deprioritized due to limited use in ambulatory

settings. Capillary glucose monitor values were also questioned due to accuracy concerns.

Decisions

• MPOG will develop two outpatient hyperglycemia measures:

o GLU-15: Treatment of blood glucose >250 mg/dL within 60 minutes

o GLU-15b: Treatment of blood glucose >180 mg/dL within 60 minutes

Discussion

• Richard Wissler, MD, PhD (University of Florida): I had a question. For outpatients without a

diabetes diagnosis, how are people actually getting blood glucose measurements? Are we

proposing that everyone gets a blood glucose checked?

o Anthony Lewis Edelman (MPOG): No, not necessarily. There’s no expectation that every

patient walking in the door gets a glucose measurement. The intent is simply that if

hyperglycemia is identified, then treatment would be expected.

▪ Richard Wissler, MD, PhD (University of Florida): I’m just wondering how often

that’s realistically going to happen in the outpatient world without a diabetes

diagnosis. I’m supportive of the concept—I just don’t know how frequently we’ll

capture those cases.

o Nirav J. Shah (MPOG): I think that’s a very fair point. Where the workgroup landed was

that if, for whatever reason, a glucose is checked—whether it’s part of a basic metabolic

panel or another reason—and hyperglycemia is noted, then even without a diabetes

diagnosis, those patients should still be treated.

• Josh Goldblatt (Henry Ford Allegiance) (via chat): Henry Ford tests all ambulatory patients

except simple endoscopy and very simple procedures (like cataracts).

o Anthony Lewis Edelman (MPOG): Again, we’re not expecting that practice universally,

but it’s helpful context.

• Tariq Esmail (University Health Network) (via chat): Can you please add these units to the

measure proposal:

250 mg/dL → 13.9 mmol/L

180 mg/dL → 10.0 mmol/L

AKI Complication Phenotype Update

Dr. Nirav Shah reviewed recent updates to the AKI complication phenotype, prompted by its use as an

outcome in the Intraop-Ox Trial.

• Baseline creatinine logic: Baseline is now defined using the most recent preoperative creatinine,

rather than the highest preoperative value.

• KDIGO alignment: The logic gap related to detecting a ≥0.3 mg/dL rise within 48 hours was

corrected.

• Exclusion logic: The highest preoperative creatinine is still used to determine renal failure

exclusions, ensuring appropriate exclusion of recently dialyzed patients.

Impact

• AKI-01 and AKI-02-C scores will increase by approximately 2–3 percentage points across MPOG.

• Effects will be more pronounced for cardiac AKI measures due to smaller denominators.

• Historical data will be retrofitted to ensure apples-to-apples comparisons.

Discussion

• Michael Mathis, MD (University of Michigan): I can expand briefly on the first point. While we

are now using the most recent preoperative creatinine to establish baseline for AKI

determination, we still use the highest preoperative creatinine value when determining the

renal failure exclusion. This ensures that patients who recently required dialysis or had severe

renal dysfunction are still appropriately excluded, even if their creatinine appears normal

immediately before surgery.

o Jonathan Paul, MD (Columbia University Medical Center): I’m assuming that historical

data will be retrofitted to follow the new logic so that we’re comparing apples to apples.

▪ Nirav J. Shah (MPOG): Yes, absolutely. Thanks for bringing that up. Historical

data will be recalculated using the updated logic.

Measure Review — AKI-01 - Mike Mathis Review 1.26.26 - Google Docs

Dr. Michael Mathis presented a comprehensive review of AKI-01, reaffirming its appropriateness as a

quality measure.

Key Points

• AKI remains a clinically significant complication influenced by anesthetic management, including

hypotension avoidance, volume optimization, and medication decisions.

• Recent evidence does not support adding balanced crystalloids, dexmedetomidine, or other

medication-based interventions to the measure at this time.

• AKI-01 continues to rely on creatinine-based KDIGO criteria due to data reliability constraints.

• Several ongoing MPOG studies (e.g., VEGA-2, IV fluid composition studies) may inform future

revisions.

Recommendation

• No changes to AKI-01 at this time.

Discussion

Nirav J. Shah (MPOG): I just had a couple things to add from the Coordinating Center side. We review

measures for technical changes. We’ve transitioned to using kidney, lung, and liver transplant

phenotypes rather than CPT codes for exclusions, which are more accurate. We’re also using the organ

procurement phenotype instead of CPT codes.

We found some cases where Foley catheter CPT codes were being sent to MPOG along with procedure

CPT codes, and because that’s sort of kidney-related, those cases were being excluded. We removed

those CPT codes from the exclusion list so that if you got a Foley catheter, you wouldn’t be excluded

from this measure. We also updated the specification to reflect that the measure timeframe extends to

seven days after anesthesia end. The code already did this, but the specification now matches

Decision: Continue as Is

Anonymized Benchmarking — 2026 Planning Discussion

The committee discussed a proposal to expand anonymized benchmarking within the QI Reporting Tool.

Key Themes

• Strong interest in benchmarking by case volume, ASC vs. inpatient, and site characteristics

• Support for self-selection of categories (e.g., academic vs. community) with guardrails

• Recognition that sites near category thresholds need flexibility

• Emphasis on allowing all sites to view all anonymized segments, not just their own

• Discussion of risk adjustment vs. process measures, with acknowledgment that some outcomes

(e.g., AKI) require more nuance

Preliminary Direction

• Free-standing ASCs identified as a logical first category

• Further exploration of quartiles, severity-of-illness indices, and observed-to-expected

comparisons

Discussion

• Josh Goldblatt (Henry Ford Health – Allegiance) (via chat): Love the grouping. Can you add self-

selection criteria with the metric selection?

o Nirav J. Shah (MPOG): Josh, would you mind unmuting and expanding on that?

• Josh Goldblatt (Henry Ford Health – Allegiance): Sure. You already have site-level customization

built into the dashboard, and you are already controlling settings that are individualized to each

site. Instead of trying to centrally collate and validate all of that information, it might make

sense to allow sites to self-select their categories. For example, sites could self-identify as small,

medium, or large based on case volume, hospital size, or whether they are a hospital versus an

ambulatory surgery center. Some of those, like hospital versus ASC, could potentially be

automated, but others could reasonably be self-selected. This is especially relevant for the

academic versus community distinction. When you are part of a large health system, there is

often a lot of blending between those categories. Being able to toggle between different

categorizations and see how performance compares across them would be helpful.

o Xan Abess, MD (Dartmouth Health) (via chat): I agree with Josh. There is a lot of

blending between “academic” and “community,” and allowing sites to self-select seems

like a good idea, with guidelines.

• Kunal Karamchandani (UT Southwestern) (via chat): I think the two factors that matter most are

case volume and the severity of illness of patients undergoing surgery. Can we add quartiles

based on these?

o Anthony Lewis Edelman (MPOG): That could be a useful approach, since it allows

comparison to similarly sized sites within a reasonable range, even though edge cases

will still exist.

• Nirav J. Shah (MPOG): Thanks for that feedback. As we move into distinctions like academic

versus community hospitals and differences in payer mix, there is clearly a lot of nuance. Our

initial thought, as Tony mentioned, was to use standardized definitions such as those from the

American Hospital Association. We may also be able to pull information from external datasets,

but not everything will be available that way. That creates an opportunity to think carefully

about how self-selection could work. It would not make sense for a large hospital to classify

itself as a small hospital, but there may be ways to allow sites to provide information when we

do not already have it, while also building logic to prevent extreme mismatches. This is

something we can explore further, and Josh, we may reach out to you again as we begin

designing this.

• Mason Smith, MD (MyMichigan Medical Center – Midland) (via chat): I do not think it should be

anonymized. I believe it is helpful to compare your program to others inside and outside your

health system.

• Morgan Brown, MD (Boston Children’s Hospital) (via chat): I am interested in patient volume as

a category, but it would make more sense to define volume based on the number of patients

eligible for a specific metric rather than overall hospital size.

• Michael Mathis, MD (University of Michigan): I do not think sites necessarily need to assert what

they are. Objective, consistent definitions are important, and where possible, we should use

them. At the same time, it would be valuable to always allow sites to see their own hospital and

how it compares to others. For example, you could always see your hospital’s performance, but

then filter to compare yourself against small community hospitals or large academic medical

centers. That way, regardless of how a hospital is classified, it can still compare itself to different

types of institutions and understand performance across various phenotypes.

o Anthony Lewis Edelman (MPOG): That is an interesting point. Thinking about the

extremes, is there meaningful value in comparing a large inpatient academic facility to a

group of freestanding ambulatory surgery centers? Is that a question people want

answered? I am trying to think through where the utility is highest.

• Josh Goldblatt (Henry Ford Health – Allegiance) (via chat): Another issue is when a hospital falls

very near a threshold.

o Michael Mathis, MD (University of Michigan): That is exactly the type of situation where

self-definition could be helpful. Josh, could you expand on that?

o Josh Goldblatt (Henry Ford Health – Allegiance): We are a medium-sized hospital within

a large academic health system. For certain metrics, we may want to compare ourselves

to other small or medium hospitals, because many of the academic-center issues play

out differently at that scale. The threshold issue is important. For example, if a small

hospital is defined as under 500 cases per month, medium is 500 to 1,000, and large is

1,000 to 1,500, we might be right at 1,150 or 1,200 cases per month. Depending on the

context, leadership might want us treated as a medium hospital or as a large hospital.

Allowing self-selection would make those comparisons more intentional and

meaningful.

▪ Morgan Brown, MD (Boston Children’s Hospital): This gets to the issue of risk

adjustment. A benchmark value is just a benchmark value, and what we struggle

with is understanding whether we are performing well with a higher-risk

population or poorly with a lower-risk population. Because the benchmark is the

same regardless, the solution may not be a simple filter, but rather the ability to

adjust the benchmark for different scenarios. That would allow a more apples-

to-apples comparison of performance.

• Tariq Esmail, MD (University Health Network) (via chat): What Morgan is

describing is why NSQIP is so valuable: observed-to-expected rates.

• Xan Abess, MD (Dartmouth Health) (via chat): A reasonable starting point for the Coordinating

Center might be to separate free-standing ambulatory surgery centers, which are fairly

objectively defined.

• Xan Abess, MD (Dartmouth Health): I agree with Josh’s points about self-selection, particularly

for sites near thresholds and for distinctions like academic versus medical-school-affiliated

hospitals. As I mentioned in the chat, we have discussed free-standing ambulatory surgery

centers extensively, and there is clear interest in developing an ambulatory-focused dashboard

rather than lumping all sites together. I support allowing self-selection based on size and site

type.

o Anthony Lewis Edelman (MPOG): I agree. An ambulatory surgery center–specific

dashboard seems like a relatively straightforward and reasonable first step.

o Nirav J. Shah (MPOG): With our updated location mapping, we already have fairly good

data on which sites are freestanding ASCs. That is something we can definitely build as

one of the initial iterations.

• Kunal Karamchandani (UT Southwestern): The reason I raised this point is that raw numbers

alone are not very helpful for analysis. It is more informative to look at index values and case

mix. Many institutions rely heavily on Vizient data, including case mix index and comparisons

across peer groups. One challenge in perioperative quality is that we lack a robust way to

capture severity of illness. Unless we build something like a Johnson-style index, we are limited.

Vizient incorporates case mix and academic status, which is why observed-to-expected metrics

are so powerful. ASA status alone is too variable to be reliable. If we could capture severity of

illness in a meaningful way and combine it with volume-based quartiles, that would allow much

more accurate, apples-to-apples institutional comparisons.

• Tariq Esmail, MD (University Health Network) (via chat): It would be interesting to compare a

freestanding ASC to the ambulatory population within a hospital that does both, perhaps by

filtering to day-surgery patients. I realize that may be beyond the current scope.

o Joseph McComb, MD (Temple University Hospital): I agree that ASCs are fundamentally

different and should be examined separately. One thing I value about MPOG, especially

working in an underserved academic safety-net hospital, is that it does not allow us to

lower expectations simply by saying our patients are sicker. I often hear arguments that

poorer outcomes are acceptable because of patient complexity, and I worry that this

lowers the bar. I appreciate that MPOG allows me to show leadership that we are

performing as well as leading institutions, even with a sicker population. I support

thoughtful risk adjustment, but I do not want it to become a justification for accepting

poorer performance. I like being able to compare ourselves to the best in the country

and demonstrate that high standards are achievable.

• Josh Goldblatt (Henry Ford Health – Allegiance) (via chat): Risk adjustment has a different

meaning for outcome metrics than for process metrics.

o Nirav J. Shah (MPOG): I agree. Over the years, we have had to make strategic decisions

about where to invest time and resources. In some cases, we have relied on inclusion

and exclusion criteria rather than formal risk-adjustment models, which are resource-

intensive to build and maintain. In some areas, we may have gone too far in that

direction, and it may be time to reinvest in risk adjustment and severity-of-illness

modeling in selected, high-impact areas. The points raised by Joe and Kunal highlight

where that may be most appropriate. One final point is that as we build anonymized

benchmarking, it is important that all sites have access to all views. Small hospitals

should be able to see how medium and large hospitals are performing, and large

hospitals should be able to see smaller sites as well, even if anonymized. We want to

avoid a situation where sites only compare themselves within a narrowly defined peer

group. While de-anonymization can occur during in-person meetings, the reporting tool

itself should allow broad visibility across segments. As Tony mentioned, your site would

appear as the blue bar within its category, but you would still be able to view

anonymized performance across other categories. That is how we are envisioning

proceeding.

o Michael Mathis, MD (University of Michigan): I agree with all of that. Some measures,

like AKI, require substantial risk adjustment. Other measures are more straightforward

and focus on adherence to clearly endorsed practices. In those cases, the emphasis is

less on risk adjustment and more on consistently doing the right thing.

Action Items

• Coordinating Center to draft anonymized benchmarking framework

• Begin with ASC-specific benchmarking views

Best Practices Exchange Proposal

The committee reviewed a proposal to launch a Best Practices Exchange, aimed at sharing successful

workflows and QI strategies.

Proposal Highlights

• The Best Practices Exchange proposal was driven by repeated requests from sites to be

connected with peers who are performing well on specific quality measures or improvement

initiatives.

• The proposal includes two complementary approaches: identifying high-performing sites during

measure review and offering a brief, optional “how we do it” spotlight to share practical

insights.

• Presentations are intended to focus on transferable, real-world lessons such as culture,

workflow, and specific quality-improvement efforts rather than polished or theoretical content.

• The exchange is designed to be low burden, with short sessions, minimal preparation, optional

slides, and limited frequency to encourage participation without overtaxing sites.

• The initiative is meant to support both newer sites and experienced sites that are seeking

improvement but have not yet achieved desired results.

Discussion:

• C. Bowman-Young (American Society of Anesthesiologists) via chat: Could you link to an online

community or bulletin board to facilitate connection outside of specific meetings? That could be

used to encourage sharing of best practices in between meetings, and information could be

mined for education.

o Michael Mathis, MD (University of Michigan): We do have the MPOG Basecamp

community. That could be a place where people post questions, share documents, or

follow up after these kinds of discussions. MPOG Quality Forum

o Nirav J. Shah (MPOG): I agree. I think this fits well with how we already use Basecamp,

and it gives people a way to connect without requiring additional meetings.

Action Items

Rather than committing to a permanent structure immediately, the group agreed to:

• Pilot the Best Practices Exchange over the next few Quality Committee meetings

• Evaluate interest, feasibility, and value before expanding further

This incremental approach was intended to allow refinement based on real-world experience.

Meeting Concluded: 1103

Appendix A — Full Transcript

ASPIRE Quality Committee — January 26, 2025

(Original wording preserved, but sentence structure, punctuation, and flow corrected for clarity.)

Opening Announcements

10:03:34 — Anthony Lewis Edelman (MPOG): We have some new announcements to start with. First is

our Featured Member of the Month for January and February. This is Bethany Pennington, PharmD,

from Washington University. Her profile will be featured on the website, so please feel free to take a

look and read more about her work if you’re interested.

Our newest MPOG site is officially up and running. Indiana University Health System has joined MPOG,

so welcome to the group. I don’t know if we have any members from IU Health on the call today, but

we’re very excited to have you.

We also have a couple of new subcommittee Vice Chairs to announce. First, congratulations to Dr.

Sharon Reale from Harvard Medical School and Brigham and Women’s Hospital. She serves as the

Obstetric Anesthesia Fellowship Program Director at Brigham and Women’s and is the MPOG Research

Lead. Dr. Reale continues to publish impactful obstetric anesthesia studies using MPOG data, including

work on difficult intubation and maternal cardiac arrest.

We are also pleased to welcome our new Pediatric Subcommittee Vice Chair, Dr. Eva Lu-Boettcher. She

is a pediatric anesthesiologist at the University of Wisconsin, where she serves as Associate Vice Chair

for Quality and Safety. She has been the site PI and MPOG Quality Champion since 2021, leading local

quality-improvement efforts using MPOG data. Congratulations to both of you, and welcome to the

team.

Glycemic Management Workgroup Update

10:05:05 — Anthony Lewis Edelman (MPOG): I’ll move on to some key operational updates. Back in

September 2025, this committee agreed to begin discussions around a new glycemic management

workgroup. That group met in December to discuss the feasibility of developing clinically meaningful

metrics addressing hyperglycemia thresholds and treatment in ambulatory surgery.

A few key themes emerged from that meeting. First, there was strong consensus that patients without a

formal diagnosis of diabetes should still be included. The threshold here is hyperglycemia itself, not

whether a patient carries a diabetes diagnosis. We believe there is a meaningful subset of non-diabetic

patients who present with elevated glucose values.

There was extensive discussion about treatment thresholds, specifically whether that threshold should

be 180 mg/dL or 250 mg/dL. Different guidelines recommend different values—SAMBA guidelines

reference 250, while other guidelines recommend 180. The discussion focused on balancing ideal clinical

practice with operational feasibility and the ability to gain clinician buy-in at ambulatory sites.

The group also emphasized that the measure should focus on timely treatment and reassessment,

rather than insulin dosing accuracy. There was discussion about excluding IV insulin given how rarely it’s

used in ambulatory settings, as well as considering exclusion of capillary glucose monitor values and

focusing instead on finger-stick or arterial values due to accuracy concerns.

Ultimately, the decision was made to build two new outpatient hyperglycemia treatment measures.

GLU-15 will measure treatment of blood glucose greater than 250 mg/dL within 60 minutes. A sub-

measure, GLU-15b, will look at treatment of blood glucose greater than 180 mg/dL within the same 60-

minute window. We’re meeting again tomorrow to continue refining the assessment and glucose-

checking logic. I’ll pause there—are there any questions, concerns, or discussion around these new

glucose measures?

Discussion — Glycemic Management

Richard Wissler, MD, PhD (University of Florida)

I had a question. For outpatients without a diabetes diagnosis, how are people actually getting blood

glucose measurements? Are we proposing that everyone gets a blood glucose checked?

Anthony Lewis Edelman (MPOG): No, not necessarily. There’s no expectation that every patient

walking in the door gets a glucose measurement. The intent is simply that if hyperglycemia is

identified, then treatment would be expected.

Richard Wissler, MD, PhD (University of Florida)

I’m just wondering how often that’s realistically going to happen in the outpatient world

without a diabetes diagnosis. I’m supportive of the concept—I just don’t know how frequently

we’ll capture those cases.

Nirav J. Shah (MPOG): I think that’s a very fair point. Where the workgroup landed was that if,

for whatever reason, a glucose is checked—whether it’s part of a basic metabolic panel or

another reason—and hyperglycemia is noted, then even without a diabetes diagnosis, those

patients should still be treated.

Josh Goldblatt (Henry Ford Allegiance) (via chat): Henry Ford tests all ambulatory patients except simple

endoscopy and very simple procedures (like cataracts).

Anthony Lewis Edelman (MPOG): Again, we’re not expecting that practice universally, but it’s

helpful context.

Tariq Esmail (University Health Network) (via chat): Can you please add these units to the measure

proposal:

250 mg/dL → 13.9 mmol/L

180 mg/dL → 10.0 mmol/L

MPOG Phenotype — AKI Complication Update

10:09:36 — Anthony Lewis Edelman (MPOG): I’ll turn things over now to Nirav to give an update on the

AKI complication phenotype.

Nirav J. Shah (MPOG): Great, thank you. Most of you are aware that we started measuring acute kidney

injury many years ago at MPOG—probably seven, eight, maybe even nine years ago—and we’ve done

that in two different ways. One approach has been through our AKI performance measures, AKI-01 and

others, focused on specific case types. The second has been through the AKI complication phenotype. At

different points, those have been used separately and at other times merged together.

Recently, as part of one of our new MPOG pragmatic clinical trials—the Intraop-Ox trial, which is

examining different intraoperative FiO₂ strategies—we took a closer look at the AKI phenotype because

AKI is a planned outcome for that trial. During that review, we identified several issues in the AKI logic

that needed to be addressed.

Those issues have now been fixed, and I’ll spend a few minutes reviewing the changes. These updates

will impact AKI-01 and AKI-02 scores, and they will also affect the AKI-01 risk adjustment model. Once

these updates are implemented, we’ll revisit and update the AKI-01 risk adjustment model accordingly.

I want to share a summary of the changes, and we’re very happy to walk through any of this offline if

people have questions. Mike Mathis is also on the call and has been reviewing AKI-01, so he can chime

in as well.

Previously, when determining baseline creatinine, we referenced the highest preoperative creatinine

value. We’ve updated that logic to instead reference the most recent preoperative creatinine. For

example, if a patient had AKI several weeks before surgery and their creatinine was elevated at that

time, but then improved prior to surgery, we will now use the most recent value as the baseline rather

than the earlier peak.

The second major change relates to how we identify postoperative creatinine increases. To align with

KDIGO criteria, we assess AKI in two ways: either a rise of at least 0.3 mg/dL within 48 hours, or a 1.5-

times increase from baseline within seven days. Previously, there was a gap in the logic where we did

not always check for the 0.3 mg/dL rise within the first 48 hours if a later creatinine was higher. That gap

has now been corrected.

10:13:27 — Michael Mathis, MD (University of Michigan) (via chat): Note, to determine preop

renal failure exclusion, we still reference the highest preop creatinine, so as to not miss renal

failure patients who were recently dialyzed immediately before procedure.

I want to specifically thank our Vanderbilt colleagues for identifying this issue. It turned out to be

frequent enough to meaningfully impact results. As a consequence of capturing these early creatinine

rises more consistently, sites will see an increase in AKI-01 rates—on the order of about two to three

percentage points across MPOG, with some site-level variability.

This effect is even more pronounced for the cardiac AKI measure, AKI-02-C, due to smaller denominators

in open cardiac cases. We’ll be discussing this in more detail at the Cardiac Subcommittee meeting as

well.Are there any questions about these updates, which we’re planning to implement over the next

couple of weeks?

Discussion — AKI Complication Update

10:15:32 — Michael Mathis, MD (University of Michigan): I can expand briefly on the first point. While

we are now using the most recent preoperative creatinine to establish baseline for AKI determination,

we still use the highest preoperative creatinine value when determining the renal failure exclusion. This

ensures that patients who recently required dialysis or had severe renal dysfunction are still

appropriately excluded, even if their creatinine appears normal immediately before surgery.

Jonathan Paul, MD (Columbia University Medical Center): I’m assuming that historical data will

be retrofitted to follow the new logic so that we’re comparing apples to apples.

Nirav J. Shah (MPOG): Yes, absolutely. Thanks for bringing that up. Historical data will be

recalculated using the updated logic.

Measure Review — AKI-01 (Acute Kidney Injury)

Anthony Lewis Edelman (MPOG): And then now we’ve got Dr. Mathis with his formal review of AKI-01.

Mike, do you want to share, or would you like me to share your screen?

10:17:02 — Michael Mathis, MD (University of Michigan): Yeah, maybe I can share my screen so I can

scroll through it as we’re chatting. We can put the link in the chat, or I think it’s already on the invite.

Some of this review is unchanged from three years ago when I did the review as well, but I think it’s still

important to rehash some of that to give context for what direction we’re headed in. One of the

questions is whether this is an appropriate quality measure to hold an anesthesiologist accountable to. I

still think yes.

This is an important complication that has associations with other comorbidities and other

complications, and it can lead to increased healthcare expenditures. It’s a significant problem, and

anesthesiologists can do things to help mitigate the risk of AKI. We can avoid hypotension when it’s

indicative of hypoperfusion. It’s not always the case, but to the extent that it commonly is, avoiding

hypotension and renal hypoperfusion is part of the anesthesiologist’s responsibility.

We can optimize volume status and cardiac output through potentially goal-directed therapies. We can

potentially use balanced crystalloids. I’ll get into this a little bit more later. We can mitigate the risk of

potentially nephrotoxic agents and weigh the relative risks versus benefits of certain medications that

could increase AKI risk. We also talked earlier about hyperglycemia, and certainly glyco-oxidative injury

can contribute to AKI.

While we recognize that not all AKI is avoidable, and that the anesthesiologist is not solely responsible

for mitigating AKI, I still think this measure is appropriate to the extent that we can uphold our part in

contributing to AKI risk reduction through these practices and potentially participating in broader

hospital-level initiatives that could impact AKI, like glycemic management or enhanced recovery

protocols.

In terms of evidence, I did look for any updated guidelines since the last review, from nephrologists,

anesthesiologists, or surgeons. There really hasn’t been anything since the 2021 consensus statement

that went into as much detail as that article did.

That was a joint consensus statement across nephrologists, anesthesiologists, and other health

professionals. They offered a number of consensus statements, many of which are not capturable within

MPOG. We can’t reliably track a lot of them. But the two that we could track were hypotension and

balanced crystalloids.

For hypotension, we already have attribution within the AKI measure where the attributable provider

could be one who was signed into the case during periods of prolonged hypotension. That was a

recommendation supported by grade C evidence, but a recommendation nonetheless.

The other recommendation was the use of balanced crystalloids rather than normal saline to reduce the

risk of postoperative AKI. Digging deeper into that, for otherwise healthier perioperative populations—

like orthopedic or colorectal patients—the SOLAR trial published in Anesthesiology in 2020 randomized

patients to normal saline versus lactated Ringer’s and did not see a difference in acute kidney injury.

Presumably, the effect of balanced crystalloids versus normal saline plays out more strongly in sicker,

critically ill ICU populations. That’s been shown in trials like SALT-ED and SMART. But in the

perioperative population, with relatively healthy patients and the brevity of the intraoperative

encounter, there was not a significant difference. Because of that, I don’t think balanced crystalloids

should be included in this quality measure at this time.

There are a few newer medication-related studies. Acetaminophen has been studied for its antioxidant

effects. Observational studies show an association with reduced AKI, but there haven’t been strong

randomized controlled trials in perioperative populations.

Dexmedetomidine has also been studied. There have been some randomized trials, but they’ve had

conflicting conclusions, with some supporting benefit and others refuting it. The evidence isn’t strong

enough to include dexmedetomidine in a quality measure for AKI.

There was a sub-analysis of the RELIEF trial looking specifically at AKI. That trial compared restrictive

versus liberal fluid strategies for major abdominal surgery. One secondary finding was that more

restrictive strategies were associated with increased AKI, but because it was a secondary outcome, no

strong conclusions were made. A subsequent sub-analysis published in 2024 suggested that NSAIDs or

COX-2 inhibitors could increase the odds of AKI, likely through prostaglandin inhibition and renal

vasoconstriction. Again, the evidence isn’t strong enough to be practice-changing.

There was also a multicenter randomized trial in Europe published in Anesthesiology in 2025 evaluating

hypotension prediction index-guided hemodynamic therapy versus standard care. The primary outcome

was stage 2 or 3 AKI, and they found no difference between groups.

Looking ahead, there are several MPOG-sponsored AKI studies on the horizon. There are several in the

cardiac literature that will likely be discussed in cardiac subcommittee meetings. There’s also the VEGA-

2 trial comparing phenylephrine versus norepinephrine as first-line vasopressors. We demonstrated

feasibility, and full results are anticipated around 2027 or 2028.

There are also observational studies examining IV fluid amount and composition. One study is leveraging

the IV fluid supply disruption from Hurricane Helene as a natural experiment to look at AKI outcomes.

Another study led by one of our MPOG research fellows is comparing saline versus balanced crystalloids

in elective major non-cardiac surgery.

In terms of inclusion and exclusion criteria, we continue to exclude minor procedures where

postoperative AKI is more likely related to other factors. We also exclude procedures with complex

pathophysiology where it’s not an apples-to-apples comparison. We rely on the creatinine component

of the KDIGO criteria because other elements, like oliguria or dialysis, are not reliably captured in MPOG.

Overall, while there is evolving evidence, I think the quality measure should remain as is. My

recommendation is no changes.

Discussion — AKI-01 Review

10:30:29 — Anthony Lewis Edelman (MPOG): Great. Thanks, Mike. Really appreciate the in-depth

review. Any questions or comments?

Nirav J. Shah (MPOG): I just had a couple things to add from the Coordinating Center side. We

review measures for technical changes. We’ve transitioned to using kidney, lung, and liver

transplant phenotypes rather than CPT codes for exclusions, which are more accurate. We’re

also using the organ procurement phenotype instead of CPT codes.

We found some cases where Foley catheter CPT codes were being sent to MPOG along with procedure

CPT codes, and because that’s sort of kidney-related, those cases were being excluded. We removed

those CPT codes from the exclusion list so that if you got a Foley catheter, you wouldn’t be excluded

from this measure. We also updated the specification to reflect that the measure timeframe extends to

seven days after anesthesia end. The code already did this, but the specification now matches.

Anonymized Benchmarking – 2026 Plan (Part 1)

10:35:41 — Anthony Lewis Edelman (MPOG): Uh, so another thing we wanted to discuss today is a plan

for anonymized benchmarking. Uh, going forward. And this is an idea that came out of this group. Um,

and the thought here is to, um… build more anonymized benchmarking into the QI Reporting Tool, and

this can be done in a number of different ways. It can be done based on case volume at sites, whether or

not it’s ambulatory or inpatient centers, health system, uh, health system-wide, um, metrics, academic

versus community hospitals.

And the idea is that you’ll be able to segment and review within the specific sub-area that your

concerns, hospital or site, uh, fit. And, um, you know, so if your concern is, you know, an outpatient

freestanding ambulatory surgery center, and it’s really being reviewed against other freestanding

ambulatory surgery centers, you would still see your blue bar against the gray background of the other

sites, so still anonymized.

Um, but a couple of questions and discussion areas we would love thoughts about from the group is,

you know, coming up with some of the definitions for these things. Um, so we’re able to use the

American Hospital Association definitions for some standardization. And then we have some

phenotypes that we’re able to use, like medical-school associated or not. Phenotypes, but there will be

some need to discuss more intricate definitions depending on how we sub-segment the sites.

Then our initial thought was to allow all sites to see these visualizations and filter by those, um, sub-

segments. And if you, again, have a site that meets criteria for that segment, you’d be able to see it with

that blue bar. If you don’t, all you would see is the anonymized gray bars throughout. And then finally,

you know, what should those groupings look like based on hospital size, small hospital versus large

hospital?

Uh, could that be based on beds, based on case volume, based on inpatient status? Um, so this is really,

you know, an opportunity, everyone, to give some feedback, their thoughts, some ideas. Um, any initial

thoughts that anyone might have? Is this something that people are interested in seeing?

10:37:38 — Josh Goldblatt (Henry Ford Health – Allegiance) (via chat): Love the grouping! Can

you add self-selection criteria with the metric selection?

10:38:08 — Nirav J. Shah (MPOG): Josh, do you mind being able to unmute and kind of expand upon

that a little more?

Josh Goldblatt (Henry Ford Health – Allegiance): Yeah, I was just thinking, uh, you already have

us on the dashboard, and you’re already controlling some of the settings that we have

individualized to a site. Um, so instead of trying to collate the information and maybe validate it

in a particular way, that you would just let sites self-select and say, “Okay, I’m small, medium,

large in case volume; I’m small, medium, large on hospital size; hospital versus ASC,” and just…

Well, maybe that one can be… can be automatic, but, uh, but some of those others, um, we can

kind of, um, self-select and pick our category.

Um, especially the academic versus community. There’s a lot of blending when you’re part of a

large health system, um, so it might be helpful to flip the switch one way and then flip the

switch another way, and see how we’re comparing in different categories.

10:38:17 — Kunal Karamchandani (UT Southwestern) (via chat)

I think the two factors that matter most are the case volume and the severity of illness of

patients undergoing a surgical procedure. Can we add quartiles based on these?

10:39:16 — Nirav J. Shah (MPOG): Yeah, yeah, thanks for that feedback. Um, yeah, I think there’s, um… I

think especially as we get into kind of academic versus community and, you know, government payer

mix, and some of those things, there’s going to be a lot more nuance there.

Um, so, you know, one of the things that we were trying to do was, as Anthony mentioned, use some of

the standard definitions via the American Hospital Association. And we might be able to get some of

that information from other external datasets, but not everything. So I think there’s going to be an

opportunity for feedback, how some of this… the self-selection would work.

It’d be weird if a large hospital were to select themselves as, like, a small hospital. But maybe there are

still some things we can do to actually… where, if we don’t have the information, you know, we can give

the site an opportunity to, um… to share that with us. And maybe have some logic to prevent extreme

mismatches. So, um, it’s something that… that I think… I think we can look at. So, Josh, maybe we’ll be

reaching out to you again when we start building it, to get some more thoughts on it.

10:39:33 — Mason Smith, MD (MyMichigan Medical Center – Midland) (via chat)

I don’t think it should be anonymized. I believe it is helpful to be able to compare your program

to others inside and outside of your health system.

10:40:01 — Morgan Brown, MD (Boston Children’s Hospital) (via chat)

Yes, interested in patient volume as a category. But I think it would make sense as the volume of

patients your site has eligible for a particular metric, rather than overall hospital size.

10:40:02 — Xan Abess, MD (Dartmouth Health) (via chat)

Agree with Josh — I think there is a lot of blending with “academic” vs “community,” and

allowing sites to self-select seems a great idea (with guidelines).

Michael Mathis, MD (University of Michigan): I was just gonna say, yeah, I don’t know if you… you may

not need to assert what you are, you know, yourself. I mean, again, yeah, there is this issue of just, like,

you know, you want to have some objective, consistent definitions, and to the extent that you can, you

should. But, um, to the extent that you could, we should, like, have filters to, like… you should maybe

always be able to see your hospital and how you stack up against other hospitals.

Sure, like, you could always see your hospital, but then filter and compare your hospital to, um, you

know, small community hospitals, or, you know, your hospital to larger academic medical centers. You

know, agnostic to whatever your hospital is, but you can always see your hospital. Maybe that’s what

we want to do. You don’t want to assert what you are, but you… it’d be nice to be able to compare your,

uh, performance to, um, you know, uh, different phenotypes of different kinds of hospitals.

Anthony Lewis Edelman (MPOG): But I think that’s an interesting thought — just thinking, say, in the

extremes, is there meaningful data or meaningful presentation looking at, say, a large inpatient, you

know, academic facility as compared to, you know, a group of freestanding ambulatory centers? Is that a

question that people would want to answer? You know, I’m just trying to think about, in my head, um,

you know, where the utility is.

10:41:36 — Josh Goldblatt (Henry Ford Health – Allegiance) (via chat): Other issue is if a hospital

falls very near the threshold.

Michael Mathis, MD (University of Michigan): Like I said, maybe I’ll turn that over to Josh, you know, to

say a little bit more about that. Because, you know, I think that’s exactly the kind of thing where, if you

could self-define yourself — yeah, define yourself, you know, compare yourself to other small hospitals

versus other large academic medical centers — to, you know… is there, uh, value there?

10:42:16 — Josh Goldblatt (Henry Ford Health – Allegiance): Yeah, so, I mean, we’re a medium-sized

hospital, or a smaller-sized hospital in a large academic health system. And so, I think, you know, we

would probably want the ability to say, “Okay, for this metric, we want to compare ourselves to other

small or medium-sized hospitals,” because a lot of the academic issues come to play at a smaller scale.

Um, I think the other thing that I just threw in the chat would be, you know, if your hospital is near the

threshold. So, you know, if you said, you know, a small hospital has case volume under 500 cases a

month, and then a medium hospital is 500 to 1,000, and a large hospital is 1,000–1,500. You know,

we’re at, like, 1,150 to 1,200 cases a month… 1,220, you know, depending. Do we fall in the medium, or

do we fall in the large? We might be really near that threshold. And for our comparison purposes, my

leadership might say, “Treat us like a medium,” or they might say, “Treat us like a large.” Um, so that’s

where it gets a little… a little clearer and more intentional if we get to self-select.

10:43:37 — Anthony Lewis Edelman (MPOG): Yeah, and there was mention in the chat around quartiles.

I think that’s an interesting idea, too, because then, really, you’re looking at other similar volume-sized

[sites] within a range that seems reasonable, although you’ll still have edge cases within those quartiles.

Uh… Morgan?

10:43:57 — Morgan Brown, MD (Boston Children’s Hospital): Yeah, I think this gets to the question of,

um, risk adjustment, right? Because the benchmark value is just… the benchmark value. And I think what

the… what sometimes we’re struggling with is: do we have a riskier population that we’re actually doing

well with, or do we have a lower-risk population where we’re actually doing poorly?

Because it’s the same benchmark no matter what. And so, maybe the answer rather is, you know, in the

filters — maybe it’s not exactly a filter — but is being able to adjust that benchmark to different

scenarios that would then let you have more of exactly what you’re thinking is an apple-and-apple

comparison of your performance.

Anthony Lewis Edelman (MPOG): Yeah, I think that was the initial thought — was to be able to compare

similar, you know, like apples to like apples.

10:44:47 — Tariq Esmail, MD (University Health Network) (via chat)

What Morgan is talking about is why NSQIP is so valuable. Observed vs. Expected rate.

10:45:11 — Xan Abess, MD (Dartmouth Health) (via chat)

Perhaps a reasonable starting point for the Coordinating Center is at least to separate free-

standing ambulatory surgery centers — those are fairly objectively recognized (I think).

Xan Abess, MD (Dartmouth Health): Sure, thanks. Yeah, I mean, I think Josh’s points are great about the,

um, ability to self-select, especially on places that are near thresholds, and what’s an academic center,

what’s a medical-school-affiliated location or not.

I, you know, I put it in the chat — you know, perhaps a thing to think about — like, we’ve talked a lot

about free-standing ambulatory surgery centers, and there is definitely interest in creating this kind of,

like, ambulatory-type, you know, dashboard really, rather than just having everything lumped together.

So I do like that ability to kind of self-select based upon size, etc.

Anthony Lewis Edelman (MPOG): Yeah, I mean, I agree. I think the ambulatory surgery center-type

dashboard is a pretty low-hanging fruit and, you know, pretty reasonable dissection point. Awesome.

Nirav J. Shah (MPOG): Yeah, I don’t know if Kate or Meridith want to comment. I think we have pretty

good data on what’s a freestanding ASC right now with our updated location mapping. Would you agree

with that? Alright. Yeah, you would. Okay, so that is something that we can do as well. Cool, thank you.

So that’s something that we can definitely build, at least as one of the first cuts.

Kunal Karamchandani (UT Southwestern): Yeah, hi. So, you know, the reason I mentioned that was

because, you know, when I work with data, just looking at numbers is not helpful. I think it’s more

helpful to look at index numbers and case mix. A comparison makes analysis more accurate.

A lot of places — like, I know where I work — everybody’s very big on, you know, their Vizient data and

what others are doing, what’s their case mix index, and if you’re actually providing that quality of care

that you should. And some of the challenge in perioperative quality is we don’t have… unless we build in

a Johnson-[type] commodity index, or something like that, to capture the severity of illness.

And like Tariq mentioned, you know, Vizient looks at their CMI and, you know, academic status and

whatnot. So maybe it’s about time that, you know, we and ASPIRE, as for perioperative quality as well,

kind of concentrate on O:Es — not just the observed, but what we would expect. Because if you only

look at ASA status, given the variability and how ASA status is graded, I don’t think you’ll get anywhere.

But if we can build a measure where we capture their severity of illness in some shape or form, and then

use that plus the volume to create quartiles, that might be a more accurate way to compare institutions

and apples to apples. Just a thought.

10:48:29 — Tariq Esmail, MD (University Health Network) (via chat): @Xan Abess it would be interesting

to be able to compare an ASC to the ambulatory population at a center that does both. Perhaps a filter

within the overall organization to allow filtering by day-surgery patients (i.e., have ASC-like population).

(Although I know that’s not what we are talking about lol.)

10:49:22 — Joseph McComb, MD (Temple University Hospital): Yeah, thank you. Good discussion. I think

the ASC, I agree, is very different, you know — to be able to look at those separately would be nice. The

one thing I do like about MPOG, though, is that it doesn’t give you the ability to do what I hear in some

other venues in the house when I’m in an academic underserved safety-net hospital — that sometimes I

hear, “Well, it’s because our patients are sick,” or, “Our outcomes aren’t as good,” and, you know, we’ve

proven that our patients are sicker.

And so we’re concerned that it lowers our bar. And we don’t want to push it as hard, and, you know,

when we don’t tolerate that, we just accept that our outcomes aren’t as good as other hospitals. Um, I

like the idea of, though, trying to risk-adjust because certainly, yes, you know, we’re doing… a lot of our

CMI are like lower-income, government-payer patients.

But I’m worried — and I don’t have an answer to that — but the one thing I do like about MPOG is that

when I look at our data and I see us moving up the rankings and I can say, “No, we’re actually doing as

well as others, even though our patients are sicker.” And I present that to leadership.

So I just wanted to sort of voice that I like that this is one area where I present data that we’re not, um,

sort of washing it down. I like that I’m comparing myself against the leaders in the country; I don’t want

to be in a situation where we say, “Well, we’re risk-adjusted and so we’re doing okay for our

population,” and then accept poorer performance.

Maybe some of the answer is some of these markers we could subset — more easily select out patients

who are very high risk objectively, not high risk because “our patients are sicker.” Again, I don’t know if

that gives you any help, but I just wanted to say I like that I’m comparing myself against the leaders in

the country.

10:51:37 — Josh Goldblatt (Henry Ford Health – Allegiance) (via chat)

Application of risk in Outcome metrics has different meaning than for Process metrics.

Nirav J. Shah (MPOG): Totally. And, you know, I… you know… um, and, you know, over the years, as

we’ve had to make choices about, like, where to spend time and energy and resources, um, you know,

did we build, you know, a new measure where, you know, where we don’t necessarily have… we’re, um,

we’re not as worried about the, you know, severity of illness, where we build inclusion and exclusion

criteria instead of building a risk-adjustment model, which is very resource-intensive and maintenance-

intensive.

Um, that’s… Joe, that’s kind of where we’ve landed, you know, for the exact reasons that you’ve

mentioned, I think. Um… and I think one could make an argument that, in some ways, we’ve gone a little

bit too far, and now we need to get back to, um, focusing some resources on risk adjustment and

severity of illness in selected cases and in selected ways.

I think you mentioned one or two of those, and Kunal, you mentioned one or two as well, so I think

those are all… those are all good points. So thanks for making them.

Anthony Lewis Edelman (MPOG): Yeah, I mean, it’s a great discussion. I think there’s maybe more here

than we can realistically get into today, but I think there’s a lot to discuss and a lot to dissect. Um, so…

Nirav J. Shah (MPOG): Yeah. The one thing I will say — and Mike — is that I do think it is important that,

you know, as we’re building these anonymized graphs, that everybody has access to all of them. So, for

example, like, you know, Anthony was mentioning small hospitals, medium hospitals, large hospitals.

I think it’s important that the small hospitals should also be able to see how the medium and large

hospitals are doing, anonymized. And similarly, large hospitals should be able to see how small and

medium-sized hospitals are doing. I think that’s important, because, you know, Joe, to your point, we

don’t want people to just say, “Well, this is my group, I’m going to compare myself only to my group,

and that’s it.” So, um, I think, you know, we have these forums for identifying or de-anonymizing sites,

and those are typically at our in-person meetings. For, like, putting up the graphs and saying, you know,

“This is us,” or, “That’s you,” and having those conversations. But in the tool itself, I think it’s important

that everybody should have access to all of these graphs, even in anonymized fashion.

So, as Tony mentioned, you know, let’s say if we’re doing large hospitals versus small versus medium

hospitals, your blue bar will be in that group, but you still have access to look at the anonymized graphs

for the medium and large, so that you can, you know, make those comparisons across segments. And I

think that will be an important component of this. That’s… that’s kind of how we were planning on

proceeding with it. Um, Mike?

Michael Mathis, MD (University of Michigan): Yeah, yeah, I agree with all of that. I was going to add just

one additional thought — which is that I think there’s certain measures that need a lot of risk

adjustment. AKI is probably the best example of one that needs a lot of risk adjustment.

And there’s other measures that are just, like, “Don’t be lazy — just do this thing,” that is, you know,

clearly endorsed, um, by, you know, society guidelines and everything else. And, you know, in that case,

it’s less about risk adjustment and more about: “This is the thing, and you know, the thing can be

solvable, and it doesn’t need to have a lot of critical thinking — it’s just more, ‘Do the thing.’”

Nirav J. Shah (MPOG)

Yeah.

Best Practices Exchange Proposal

10:55:34 — Anthony Lewis Edelman (MPOG): Alright, so the next thing we wanted to talk about is

something we’re calling a Best Practices Exchange. This is still very much a proposal, and we wanted to

get feedback from the group before moving forward with anything.

The idea here is to identify high performers during measure review and then create a short, low-prep

opportunity for those sites to briefly share how they do it. This could be during the Quality Committee

meeting itself or potentially during an adjacent forum.

We’re envisioning this as a short spotlight — less than ten minutes — where a high-performing site or a

site that’s made a big improvement could talk through workflows, policies, culture, or specific

interventions that helped them succeed on a measure.

There are really two lenses we’re thinking about here. One is consistently high performers — sites that

are always at the top and likely have strong underlying culture and processes. The second is intentional

improvers — sites that weren’t doing well, recognized that, and then implemented specific quality-

improvement strategies like PDSA cycles, education, reminders, or workflow changes.

The goal is not to put anyone on the spot, but to make it easier for sites to learn from each other

without having to reinvent the wheel.

We think this could be especially useful for newer sites or for measures where traction has stalled. The

idea would be that interested sites could then connect offline for deeper dives if they want to.

In terms of logistics, we’re thinking something very lightweight. Slides would be optional, but some

material would probably be encouraged. The cadence would likely be two to three times per year,

aligned with Quality Committee meetings. Topics would be driven by upcoming measure reviews, site-

initiated ideas, or suggestions from the Coordinating Center.

The presenter could be the quality champion or a designee — whoever is most familiar with the work.

And again, this is meant to be low burden and high value.

So I’ll stop there. Thoughts? Reactions?

10:58:03 — C. Bowman-Young (American Society of Anesthesiologists) via chat

Could you link to an online community or bulletin board to facilitate connection outside of

specific meetings?

Anthony Lewis Edelman (MPOG): Yeah, that’s a great question. I think that ties in nicely

with this idea. We’ve talked before about ways to encourage more interaction between

meetings.

Michael Mathis, MD (University of Michigan): We do have the MPOG Basecamp community.

That could be a place where people post questions, share documents, or follow up after these

kinds of discussions.https://launchpad.37signals.com/basecamp/2754939/signin

C. Bowman-Young (American Society of Anesthesiologists)

That could be used to encourage sharing of best practices in between meetings, and information could

be mined for education.

Anthony Lewis Edelman (MPOG)

Yeah, exactly. That’s really the spirit of what we’re thinking. Having a space where people can continue

the conversation, share tools, or ask follow-up questions after hearing a Best Practices Exchange

presentation.

Nirav J. Shah (MPOG)

I agree. I think this fits well with how we already use Basecamp, and it gives people a way to connect

without requiring additional meetings.

Anthony Lewis Edelman (MPOG)

Great. So it sounds like there’s interest in exploring this further. We’ll take this feedback and come back

with something more concrete.

10:56:02 — Nirav J. Shah (MPOG)

Okay, alright, thanks. Yeah, so, you know, folks have, um, in the past, they’ve, um, reached out to us at

the Coordinating Center many times, said, “Hey, can you introduce us to someone who’s doing really

well in a given measure or a given quality improvement initiative?” And so.

Um, this concept is based on, you know, that, um, requests that we’ve had, you know, time and time

again, and at the Coordinating Center, we’ve made those connections in the past.

Uh, we’ve usually, you know, looked and asked the folks ahead of time, “Hey, do you mind if we make a

connection with someone who’s interested in learning more about the great work you’re doing at your

site?” And so, um, thought about it in two ways.

One is that during a measure review, for example, like Mike’s AKI review, we would identify who the

high performers are, if it made sense for that measure then put their contact information up so that

others could reach out to them. And to learn a little bit more about their processes and how they got to

where they were. was one component.

The second component is, to add, like, a short—not something that we want folks to spend a

tremendous amount of time on — but a low-prep, kind of “how we do it” spotlight. We’ve done versions

of this at our in-person ASPIRE meetings in the state of Michigan or at the MPOG retreat.

Call it a “QI story” in those scenarios, but, something like that where, for a few minutes, a high-

performing site, kind of shares a little bit about, like, what’s the culture, the workflow, The work that

they’ve done in a particular area. Through a couple of different lenses. One is, is it a culture issue, a

process that… that, you know, they’ve always been great at this, why is that? Or. Was there a specific

project that they initiated and were successful at, that others could learn from?

And then, sites that were interested in reaching out could connect with those folks offline if they wanted

To have a more in-depth conversation. And we think it’ll be useful, you know, not just for newer sites,

but for also sites that have been working on something for a while, and just have not seen the progress

that they’ve wanted.

We thought that, it could be brief, less than 10 minutes, as I mentioned. Low-prep, slides optional, but

maybe some material would be encouraged to help transmit the info maybe not with every Quality

Committee meeting, just a couple times a year to start, and we’ll see how successful it is or not and then

topics could come either from the measure review — so as we’re doing the measure reviews, we could

say, “Hey, is this a good topic for this, you know, 10-minute ‘how we do it’ thing?” Or, if a site was

interested in sharing the great work they’ve done, then we would create a forum for that, or if we at the

Coordinating Center realized, you know, through our own case review, that a site is doing amazing in a

certain area, we would… we could reach out to them as well.

And it could either be the Quality Champion or someone they designate as well.

Um, so we wanted to get some very, very early feedback. We pretty much only have time for the vote

right now about two things.

One is that, is it okay — if you feel comfortable — with identifying a high performer, not a low, but a

high performer, during our measure review?

And number two, should we create this new segment — you know, calling it Best Practice Exchange…

Meridith will come up with a better name — um, if folks are interested in it.

I’m just gonna share this as a vote, and then, just to kind of get, you know, early thoughts about what

people think. And then…

Anyone can vote here; it does not have to be one vote per site. We are simply looking to gather some

early feedback. Because this would involve additional effort for participating sites, there are two

considerations. First, for the initial question, a high-performing site would be identified. Second, for the

Best Practices Exchange itself, participating sites would need to invest a small amount of time and effort,

though we intend to keep that burden minimal.

We’ll give everyone just a few moments to respond. We are at time, so we’ll stop the poll shortly and

share a brief summary of the feedback.

Based on the responses, it appears that most participants feel this is a good idea, so we will continue to

explore it further. We plan to pilot the first Best Practices Exchange over the next few Quality

Committee meetings. Before identifying any high-performing sites during a measure review, we will

communicate that plan in advance, at least for the next few meetings.