Hospital Diabetes Meeting 2022

April 1-2, 2022; Virtual; Full Report – Draft

Executive Highlights

  • The Diabetes Technology Society’s Hospital Diabetes Meeting 2022 was as insightful as it was engaging and hats off to DTS President Dr. David Klonoff for orchestrating such an instructive gathering! As we understand it, sessions will become available online to those registered on April 11, 2022, and we highly suggest going back to watch an esteemed collection of faculty muse on the future of inpatient diabetes management.
  • In tech, the clear standout was a day #1 session on CGM-EHR integration. Dr. Juan Espinoza (Children’s Hospital Los Angeles) introduced attendees to the iCoDE Project (Integration of Continuous Glucose Monitor Data into the Electronic Health Record), which is working to develop technical standards and operational best practices for the integration of CGM data into EHRs. The iCoDE Project aims to support CGM data integration into EHRs by providing health systems with a set of replicable practices and guidelines to ensure data integration is safe and secure and is set up to help drive improved clinical outcomes for patients. Specifically, iCoDE includes six working groups that participated in an initial meeting in January 2022. As Dr. Espinoza explained, the working groups continue to meet regularly, and will reconvene for two additional meetings in May/June 2022 and October/November 2022 to report back on the standards they’ve developed and work to create a centralized resource on best practices for CGM integration into EHRs. We are very excited to follow the progress of this group as EHR integration remains one of the largest barriers to CGM adoption and use in clinical and inpatient care.
  • There were many engaging presentations on inpatient insulin management. We loved hearing Dr. Rima Bouajram’s (UCSF) talk showing that pharmacist co-management of insulin therapy enhanced glycemic management overall in critically ill patients. We also saw plenty on Glytec, including a pre-post analysis (n=178 type 1s and 2s) showing that Glucommander improved the percent of readings in range (70 – 180 mg/dL) from 60% to 84% across nine US hospitals, and a sponsored session from Chief Medical Officer Dr. Jordan Messler.
  • Elsewhere, we continued to see calls for more personalized in-hospital glycemic management. Highlights include Dr. James Krinsley’s (Stamford Hospital/Columbia University) presentation on how inpatient glycemic management isn’t “one size fits all,” and also Ms. Rebecca Longo’s (Lahey Hospital) SAWS framework for personal tech use in the hospital.

The Hospital Diabetes Meeting 2022 took place from April 1, 2022 to April 2, 2022, in a virtual format. As noted in our resource hub, you can check out the conference website, register, see our preview, and see the full agenda. Plus, in this full report, see our three themes to get a sense of our top-line takeaways.

Table of Contents 

Themes & Top Takeaways

Accelerated momentum toward widespread in-hospital CGM use; EHR integration remains the “holy grail”

The faculty at the Hospital Diabetes Meeting made one thing clear: CGM has the potential to be a transformative tool in the hospital if used effectively and integrated directly into the EHR. We could sense a palpable excitement about the future of in-hospital CGM, especially after Dexcom received FDA Breakthrough Designation for an in-hospital system in March 2022, potentially signaling FDA buy-in on the value of CGM to transform inpatient care. We most appreciated Dr. Juan Espinoza’s (Children’s Hospital Los Angeles) presentation on CGM-EHR integration and the work that he and Dr. Klonoff are doing with the iCoDE project to develop technical standards and operational best practices for the integration of CGM data into EHRs. While we’re far off from having widespread EHR-CGM integration, we’ve retained optimism following the International Diabetes Center’s successful integration of LibreView into Epic, presented at ADA 2021. And, as we learned from Dr. Athena Philis-Tsimikas (Scripps Whittier Diabetes Institute, San Diego, CA) and Dr. Elias Spanakis (University of Maryland), there’s still so much more to do beyond the EHR to optimize CGM use in the hospital, especially when it comes to securing support from hospital leadership, training nursing leadership, and optimizing glycemic outcomes in response to changing clinical conditions.

“It takes a village”: inpatient insulin management made possible by interdisciplinary care teams; is software the future?

Dr. Rima Bouajram (UCSF) gave a rousing presentation reminding us all that it takes a village to successfully manage inpatient glycemia. We were wowed by UCSF’s insulin management quality improvement initiative, which dramatically reduced episodes of severe hyperglycemia/hypoglycemia by making pharmacists co-manage subcutaneous insulin regimens with ICU providers. Elsewhere, we learned to not underestimate the potential of software to reduce provider burden and streamline workflows – just look at the outcomes Glucommander has been able to drive: a pre-post analysis (n=178 type 1s and 2s) showed that Glytec’s inpatient insulin control algorithm improved the percent of readings in range (70 – 180 mg/dL) from 60% to 84% across nine US hospitals. These findings greatly reinforce the utility of inpatient insulin management software to reducing long-term mortality and morbidity outcomes, which we’ve known since Professor Greet Van den Berghe's now landmark paper in NEJM in 2001. As Glytec’s Chief Medical Officer Dr. Jordan Messler put it during the company’s sponsored session, inpatient management software can close the “implementation gap” between what we know about inpatient insulin management, and what we actually do to manage glycemia in the hospital.

The horizon: adaptive glucose targets & personal technology use

A salient theme of the Hospital Diabetes Meeting was recognizing that as more tech is used in the hospital, we must think more about transitions of care, personalized glucose targets, and use of personal devices. This sentiment was made clear by Dr. James Krinsley (Stamford Hospital/Columbia University), who argued that diabetes management in the ICU isn’t “one size fits all” and that variable glucose targets should be used to meet the diverse needs of patients, as well as by Ms. Rebecca Longo (Lahey Hospital), who touched on: (i) the “SAWS (Supplies, Action, Want, Specialist)” framework for assessing safety of personal technology use in the hospital; and (ii) guidelines for transitioning from personal to hospital-based insulin delivery. As a reminder, the 2022 ADA Standards of Care include an updated clause stating that “patients who are in a position to safely use diabetes devices should be allowed to continue using them in an inpatient setting or during outpatient procedures when proper supervision is available.” We’ll be all ears for more thoughts on personalizing technology for the patient at upcoming conferences and gatherings, as we know this idea, while not particularly novel in principle, is rather nascent when it comes to implementation.

Inpatient Glucose Monitoring Highlights

Dr. Juan Espinoza on pathway to improved EHR integration for CGM data; iCoDE Project to develop technical standards and best practices for EHR integration currently underway

Dr. Juan Espinoza (Children’s Hospital Los Angeles) outlined six levels of possible CGM data integration into EHRs recognizing that CGM data currently exists outside of inpatient and clinical workflows. Thus, accessing CGM data requires providers to move back and forth between patient EHRs and CGM data sources (e.g., Clarity, LibreView, Tidepool, Glooko, etc.) adding burden and making it more challenging for providers to utilize CGM data in clinical decision-making. While there are limitations to current CGM data integration in EHRs, Dr. Espinoza discussed an “adapted interoperability framework for CGM data” detailing a path across six levels to increased interoperability in the future. Specifically, the six levels include: (i) transcription where there is no data exchanged between the CGM data portal and EHRs and providers must manually input information; (ii) static documents where PDFs of CGM data can be uploaded into EHRs, but not edited or interacted with; (iii) variable documents where CGM data source and EHR can exchange static documents, but providers can select components of documents and customize data visualization within reports; (iv) discrete structured data where CGM summary statistics can be pulled directly into charts in the EHR; (v) continuous structured data where CGM data is pulled continuously into the EHR; and (vi) device or app metadata where information on device utilization is available in addition to continuously streamed clinical data. Across these six levels, Dr. Espinoza suggested that most providers and healthcare systems operate between level one and level two with either static PDF documents or manual transcription of data by providers into EHRs. Dr. Espinoza advocated that improving the interoperability of CGM data into EHR systems along these six levels can also help drive increased CGM adoption as providers and hospital systems are more easily able to interact with and act on patient CGM data.

  • Dr. Espinoza introduced attendees to the iCoDE Project (Integration of Continuous Glucose Monitor Data into the Electronic Health Record), which is working to develop technical standards and operational best practices for the integration of CGM data into EHRs. The iCoDE initiative, developed over the last year by Dr. Espinoza and Dr. David Klonoff (Diabetes Technology Society), aims to support CGM data integration into EHRs by providing health systems with a set of replicable practices and guidelines to ensure data integration is safe and secure and can be done in a way to help drive improved clinical outcomes for patients. Specifically, iCoDE includes six working groups that participated in an initial meeting in January 2022. The working groups are divided across the six areas of (i) data standards; (ii) entity and identity resolution (confirming that data is correctly linked to patients); (iii) integration and interoperability; (iv) analytics and visualization; (v) clinical workflows: and (vi) partnerships and business models. As Dr. Espinoza explained, the working groups will reconvene for two additional meetings in May/June 2022 and October/November 2022 to report back on the standards they’ve developed and work to create a centralized resource on best practices for CGM integration into EHRs. We are very excited to follow the progress of this group as EHR integration remains one of the largest barriers to CGM adoption and use in clinical care in particular, to say nothing of inpatient care.
  • Looking to the future, Dr. Espinoza predicted that current EHR systems (e.g., Epic, Cerner, etc.) will become more like operating systems (e.g., iOS, Microsoft, etc.) with novel applications designed to meet the needs of providers and offer more personalized data management for individual patients. Specifically, Dr. Espinoza provided an example of a current computer application, such as one for notetaking, that is compatible with both iOS and Microsoft operating systems and expressed his hope that EHR systems could operate similarly in the future with disease or treatment specific applications that can be compatible across operating systems and allow for “more nimble and equitable” care by simplifying providers’ interactions with EHRs. In this discussion, Dr. Espinoza also noted concerns around the growing “digital divide” in access to internet and technologies that can exacerbate existing health inequities and urged attendees to remember that as providers transition to more digitally enabled care, underlying inequities must be addressed to ensure equitable access to and quality of care for all patients. It’s fascinating that Dr. David Feinberg, formerly of UCLA, then CEO of Geisinger, then exec at Google Health, is now CEO of Cerner – what’s going to happen there?
  • During Q&A, Dr. Nestoras Mathioudakis (Johns Hopkins University) raised the question of differences in EHR data integration for Abbott and Dexcom CGMs. Specifically, Dr. Mathioudikas shared his experience finding greater success working with Abbott to integrate FreeStyle Libre data into his hospital’s EHRs noting that Dexcom “didn’t seem to be in a place” where they could support EHR integration. In response, Dr. Espinoza asserted that Dexcom was actually the first CGM manufacturer to support an earlier version of EHR integration, but, due to necessary updates, has “sunset” that program and is now “entirely redesigning” its software to support greater options for EHR integration. As such, these features are not yet available, he said. Dexcom does maintain a very robust developer API to access device data for other applications (as a note, Dr. Espinoza clarified that he does not receive any financial support from Dexcom and was speaking from his own experience working to integrate Dexcom CGM data in clinical settings). This was a very interesting aspect of the conversation for us as we have long wondered what the role of CGM manufacturers will be in supporting EHR integration and we have not yet heard Dexcom discuss specifics of EHR integration, though software updates have been a common theme in the company’s recent updates.

Patients on glucose telemetry rt-CGM (n=57) spend +2.1 hrs/day between 70 and 250 mg/dL vs. usual care (73% vs. 63%, n=53); Dexcom G6 hospital intervention study to be read out at ADA 2022; new TIGHT study to assess telemetry systems with target of 90–130 mg/dL vs. usual care

Dr. Rodolfo Galindo (Emory University) and Dr. Elias Spanakis (University of Maryland) discussed the use of hospital telemetry systems for the prevention of hyperglycemia and hypoglycemia, respectively. As a reminder, glucose telemetry systems allow CGM readings to be directly transmitted from the hospital bedside to a central monitoring system in the nurse’s station, with the goal of preventing hypoglycemic events. Both Dr. Galindo and Dr. Spanakis presented data from glucose telemetry systems, showing their utility in managing patients’ glycemia within the hospital.

  • Dr. Galindo presented data from an RCT in which participants in a non-ICU hospital setting on telemetry monitored rt-CGM (n=57) spent +2.1 hours/day between 70 and 250 mg/dL compared to those on usual care (n=53) at 73% and 64% (p<0.05), respectively. As both cohorts spent 0% of time <70 mg/dL, Dr. Galindo argued that the increase in time between 70 and 250 mg/dL demonstrates that glucose telemetry systems can be useful in preventing severe hyperglycemia. Additionally, the mean glucose level of the rt-CGM cohort was significantly lower than the usual care cohort at 220 mg/dL vs. 238 mg/dL, respectively (p=0.03), although these mean glucose levels were still concerningly high and reflected very low levels of Time in Range across cohorts that were statistically equivalent (25% vs. 20%, respectively, p=0.15). The percentage of time <70 mg/dL in both cohorts was zero. While these study findings do illustrate the benefits of using telemetry monitored CGM compared to usual point-of-care testing in the hospital to reduce severe hyperglycemia, the participants’ glycemic outcomes are certainly far below what we would like to see from hospitalized patients in a non-ICU setting. Dr. Galindo concluded by urging attendees to stay tuned for ADA 2022, where investigators will read our results from the multicenter Dexcom G6 intervention study, which compares the glycemic control achieved through insulin adjustment by point-of-care testing vs. insulin adjustment from CGM readings in hospitalized type 1s and insulin intensive type 2s.
  • Dr. Spanakis followed with RCT data that he initially read out at ADA 2021 showing that telemetry monitored CGM was associated with lower rates of hypoglycemia and 20 fewer mins/day <70 mg/dl compared to point-of-care testing. Results from this investigation showed that the intervention was successful at reducing hypoglycemic events (<70 mg/dl) with an average of 1.7/patient in the control group compared to 0.7/patient in the CGM group (p=0.024). Additionally, time <70 mg/dl and <54 mg/dl were both significantly lower in the CGM arm at 0.4% and 0.05% compared to 1.9% and 0.8% in the control arm (p=0.002 and p=0.017, respectively). Notably, there was no significant increase in Time Above Range in the CGM arm compared to the controls indicating that prevention of hypoglycemia did not result in rebound hyperglycemia. Excitingly, a new study (NCT05135676) is underway to test whether in-hospital CGM with an intensive target of 90 – 130 mg/dL can reduce hyperglycemia without increasing hypoglycemia relative to standard of care. Aptly named the TIGHT (Time in Glucose Hospital Target) study, the multicenter investigation is set to conclude in September 2022 and will include 120 participants. While we are beyond certain that CGM telemetry systems hold great promise in simplifying inpatient glycemic management, we have been waiting for more action since NICE SUGAR – it has been such a long road. We look forward to seeing the results of the TIGHT study in the future.

Six key drivers for implementing CGM in the hospital: importance of engaging hospital administration and training nursing leadership; does Dexcom’s FDA Breakthrough Designation signal accelerated momentum toward widespread CGM use in the hospital?

Dr. Athena Philis-Tsimikas (Scripps Whittier Diabetes Institute, San Diego, CA) kicked off an early morning session on CGM with a presentation on her experiences establishing a CGM program in the hospital. Specifically, Dr. Philis-Tsimikas laid out six “key drivers” for securing approval from important stakeholders for CGM use in the hospital: (i) accuracy; (ii) efficacy; (iii) safety; (iv) patient satisfaction; (v) efficiency; and (vi) cost. On the first three points, Dr. Philis-Tsimikas emphasized that physician and nursing leadership education was instrumental to ensuring that the hospital staff in charge of initiating patients on CGM were properly informed of how to manage patients on CGM and make the best use of the data. Historically, “champions” have been very helpful on these fronts. On the last three points, Dr. Philis-Tsimikas explained that conversations with hospital leadership were key to ensuring that leaders were aware of the financial and workflow benefits of inpatient CGM use. Dr. Philis-Tsimikas noted that when she worked to bring CGM to her diabetes clinic, she found it particularly important to get support from the Chief Medical Officer, as well as Corporate VPs and R&D leaders. Additionally, Dr. Philis-Tsimikas stressed that compliance and legal officers should be consulted, especially since the FDA’s allowance of in hospital CGM use continues to evolve.

  • Dr. Philis-Tsimikas laid out her clinic’s point-of-care testing guidelines for hospitalized patients with diabetes on CGM. Dr. Philis-Tsimikas explained that after patients are initiated on CGM, they are monitored via BGM at least once a day to ensure that the CGM values continue to be an accurate reflection of blood glucose levels (although presumably, the BGM isn’t being calibrated to YSI, we imagine that if they are generally far apart, particularly on day #1, then the BGM is considered more trustworthy). If there is a ≥20% variation (or ≥20 mg/dL variation for readings <100 mg/dL) between BGM and CGM readings, then hospital staff will revert the patient back to BGM until the readings fall back within the “20/20 guidelines.” If there is no alignment in 24 hours, Dr. Philis-Tsimikas explained that hospital staff will “consider removal or replacement” of the CGM and that for patients with persistent hypoglycemia (<80 mg/dL), after two treatments with glucose, CGM is immediately discontinued. Then, another CGM is placed.
  • Dr. Philis-Tsimikas very briefly noted that Dexcom received FDA breakthrough designation for an in-hospital CGM system. We were thrilled to hear Dr. Philis-Tsimikas mention this exciting development from last month. Most would expect that FDA took the opportunity to make this Breakthrough Designation for Dexcom to signal the Agency’s buy-in on the potential value of CGM inside the hospital – typically the designation doesn’t guarantee faster overall approval but it does make it more likely that FDA makes its regulatory leaders available during the time that sponsors are working on submissions – presumably this helps avoid bureaucratic delays. While neither the next step toward FDA approval nor the specific CGM system that would be submitted to the FDA were disclosed in Dexcom’s announcement, we expect that this news, coupled with the several trials evaluating inpatient use of Dexcom G6, signal a return to accelerated momentum toward widespread CGM use in the hospital. This designation comes just months after the ADA updated the 2022 ADA Standards of Care to address “consumer” inpatient technology use with a novel recommendation, stating that “patients who are in a position to safely use diabetes devices should be allowed to continue using them in an inpatient setting or during outpatient procedures when proper supervision is available.” From our view, while this was great to see, the breakthrough designation is working toward something very different from “consumer CGM” – it’s working toward something more similar to “the sixth vital sign,” where inpatient use of CGM in the ICU and eventually elsewhere would become standard. That would be far better than just not prohibiting use of traditional CGM. We’re staying very tuned … the wait in this arena has been unbelievable.
  • Aside from the ADA’s direct guidelines on inpatient technology use in its 2022 Standards of Care, Dr. Philis-Tsimikas emphasized that the updated recommendations for CGM use in basal-only type 2s will likely mean that “more patients will come [into the hospital] understanding CGM.” This certainly makes sense given that while nearly 30% of people with diabetes are on some kind of insulin, per CDC, and that long-acting insulin is the most common, and that basal insulin is now available at a cheaper cost. Although these same people have typically been diagnosed longer, dQ&A data on CGM in people with T2D from August, 2020 shows that many of them, even with coverage, haven’t yet been offered CGM – so we are hopeful but not necessarily optimistic that a majority will know CGM. But, we hope this changes over time. While some believe that much work lies ahead on the reimbursement front to provide CGM to those on basal-only therapy, we believe they’ll get it if they have education around it – from our view, a greater problem is that many clinicians who are not specialists still seem to perceive CGM as “high hassle” or “a lot of work” or technology that won’t be covered or learned easily. All this said, we certainly hope that Dr. Philis-Tsimikas’s prediction comes to fruition as CGM penetration rises further.
  • During a separate but closely related presentation on CGM, Dr. Eileen Faulds (Ohio State University) stressed the importance of nursing leadership in successfully implementing CGM within the hospital. While she previously thought that “having a robust inpatient diabetes management service” would be the most important contributor to successfully implementing CGM within the hospital, Dr. Faulds noted her surprise at how the implementation science required more investment than the protocol development aspects. We found Dr. Faulds’ comments to be a helpful complement to Dr. Philis-Tsimikas’s six “key drivers” for securing approval to use CGM in the hospital.

Inpatient Insulin Management Highlights

Retrospective study of medical-surgical UCSF ICU patients (n=210) over 18 months shows 33% and 3% had one severe hyper/hypo episode, respectively; post-hoc analysis shows pharmacist co-management of subcutaneous insulin regimens decreases severe hyper/hypo episodes to 9% and 0% of cohort, respectively; UCSF EHR uses visual cues for dysglycemia

Dr. Rima Bouajram (UCSF) shared data indicating that pharmacist co-management of insulin therapy enhanced glycemic management overall in critically ill patients. In a retrospective study of medical-surgical ICU patients (n=210) from January 2017 to June 2018, Dr. Bouajram noted that 188 patients (~90% of cohort) experienced ≥one hyperglycemic event, 69 (~33%) had ≥one severe hyperglycemic event, 36 (~17%) had ≥one hypoglycemic event, and 7 (~3%) had ≥one severe hypoglycemic event. Day-by-day, the data indicated that poor glycemic management persisted over five days in a high majority of patients, with only mild improvement. As such, UCSF initiated a new quality improvement project to improve glycemic management and workflow efficiency, under which pharmacists would co-manage subcutaneous insulin regimens with providers during the work week. Dr. Bouajram then presented results from a post-hoc analysis, showing the same endpoints from the retrospective study she first presented at six weeks and six months following the implementation of the new quality improvement protocols. Impressively, all four endpoints decreased: ≥one hyperglycemic event from 90% to 72% of cohort; ≥one severe hyperglycemic event from 33% to 9% (that’s incredibly impressive); ≥one hypoglycemic event from 17% to 2%; and ≥one severe hypoglycemic event from 3% to 0% – wow! These results are simply outstanding and reflect the power of multidisciplinary care teams to accomplish so much synergistically for their patients by leveraging a collective body of expertise and far more attention to TIR and TITR. Dr. Bouajram noted that one of the biggest barriers to optimizing glycemic management in the ICU is the lack of understanding or experience that inpatient providers often have managing glycemia, and so we commend UCSF for investing in this important quality improvement program leveraging the expertise of pharmacists that so clearly has improved patient outcomes and can hopefully serve as a model for other hospitals to follow.


Preliminary Study (Jan 2017 – June 2018)

QA after six weeks (Jan 2020 – Feb 2020)

QA after six months (Jan 2020 – June 2020)


≥1 Hyperglycemic (BG > 180 mg/dL) event




≥1 Severe Hyperglycemic (BG > 300 mg/dL) event




≥1 Hypoglycemic (BG ≤70 mg/dL) event




≥1 Severe Hypoglycemic (BG < 54 mg/dL) event




  • Dr. Bouajram explained that there was a “bleed-over effect” of pharmacist-provider collaboration from the weekdays to the weekends, when only providers were caring for patients. On weekends when pharmacists were not responsible for co-managing subcutaneous insulin regimens, the data from the post-hoc analysis indicated that providers were delivering outcomes that did not significantly differ from the outcomes delivered by both pharmacists and providers during the week suggesting that ICU providers may have learned more about insulin therapy while collaborating with pharmacists that they were then able to translate into their stand-alone practice. Additionally, these outcomes were still markedly better than before the co-management pilot was initiated.

  • Dr. Bouajram illustrated how the UCSF EHR system incorporates screening tools and dashboards to manage glycemia in the hospital. Inside the UCSF ICU patient list, Dr. Bouajram showed how each patient’s “Rx Glucose Monitoring” status was strategically located next to the Room/Bed column, with a small but prominent visual flag to draw attention to severe hypoglycemia and hyperglycemia episodes. Inside each patients’ chart, Dr. Bouajram demonstrated that HCPs can see a more detailed trace of patient BGM values over the course of a day, in line with medication doses and other metrics of interest. Dr. Bouajram argued that these simple but effective visual cues, which improve the accessibility of patient reports, “may support improved glycemic control,” although this has yet to have been demonstrated in a scientific study.

Glucommander pre-post analysis (n=178 type 1s and 2s) improves percent of readings in range (70 – 180 mg/dL) from 60% to 84% across nine US hospitals; reinforces utility of inpatient insulin management software

Dr. Joseph Aloi (Wake Forest University) presented a pre-post analysis (n=178 type 1s and 2s) demonstrating that Glytec’s Glucommander inpatient insulin dosing software increased the percentage of recorded glucose values in range (70 – 180 mg/dL) while also reducing hypoglycemia. As a reminder, Glucommander is Glytec’s FDA cleared algorithm and provides insulin dosing support for inpatients and their caregivers via the company’s eGlycemic Management System (eGMS). Glytec’s eGMS is currently implemented in over 60 health systems, representing 300 hospitals in the US. We also learned earlier this year that Glucommander will be integrated with Roche’s cobas pulse inpatient BGM upon FDA approval. Turning to the study, Dr. Aloi explained that this pre-post study was conducted in the cardiovascular surgery and medical ICU units across nine Sentara Healthcare hospitals, comparing glycemic management across the 178 participants 30 days before implementing Glucommander and 30 days after. Notably, the percentage of recorded blood glucose levels that were in range rose from 60% using an IV EHR calculator to 84% using Glucommander’s insulin management software. Glucommander was also successful at reducing the percentage of recorded glucose readings <70 mg/dL, from 2.1% using an IV EHR calculator to only 0.9% using Glucommander. These results are certainly impressive and reinforce the potential of inpatient insulin management software to improve glycemic management in the hospital, especially given that many ICU staff may not specialize in diabetes management. Less than 10% of hospitals use glycemic management software as of ATTD 2020, and so we hope that these data and others will continue to drive adoption of Glytec’s eGMS and other similar solutions, as we’ve known since Professor Greet Van den Berghe's now landmark paper in NEJM in 2001 showed that inpatient glycemic management is highly valuable in reducing mortality and morbidity.

  • Dr. Aloi also noted that EndoTool, a comparable solution to Glucommander, has also improved glycemic management among his patient population. Dr. Aloi’s remarks, similar to the success of Glucommander as documented above, reinforce that inpatient insulin management systems can tighten glycemic management while reducing hypoglycemia.
  • Dr. Matthew Bouchonville (University of New Mexico) followed Dr. Aloi by sharing benefits and challenges of inpatient glycemic management software. He included an illuminating table, which we’ve repurposed below:



Improves glycemic metrics:

  • Reduces hypoglycemia & hyperglycemia
  • Regulates transition from IV to SQ insulin
  • Attains & sustains glycemic goals more effectively (reduces glycemic variability)

Change: loss of familiar, mistrust of new, greater time commitment in learning new procedures/skills

Standardization of practices

Effects on medical education: fear more automation = less critical thinking and may limit learning opportunities

Ease of practice for providers, decrease cognitive burden

Nursing burden: alert fatigue, transcribe orders, more software prompts/alarm reminders to take BGs on time and record dosages; learning to trust mechanized decision processes

Decreases length of stays/costs & utilization of ICU beds

Variation: if allowing different treatment options in the same facility

Glytec-sponsored session: Chief Medical Officer Dr. Jordan Messler presents roadmap to glycemic management success; plans to integrate inpatient CGM within company’s eGlycemic Management System

In a highly anticipated concluding session to the 2022 Hospital Diabetes Meeting, we heard from Glytec’s Chief Medical Officer Dr. Jordan Messler on “uniting people, process, and technology to achieve optimal glycemic management.” We first met Dr. Messler at DTM 2020, during which he spoke on Glytec’s EMR-integrated eGlycemic Management System (eGMS) and explained how it can streamline hospital workflows and efficiently integrate insulin dosing data. We also wrote about Dr. Messler when he was appointed to spearhead continuous improvement initiatives for Glytec’s clinical strategy and product development as the company’s new Chief Medical Officer in July 2021. Dr. Messler gave an outstanding presentation, centered on Glytec’s eGMS and how it can close the “implementation gap” between, as he put it, what we know about inpatient insulin management, and what we do to manage glycemia in the hospital. To overcome the implementation gap, Dr. Messler presented a “roadmap to glycemic management success,” with two coexisting frameworks that position hospitals to deliver the best glycemic outcomes to their patients: (i) merging the best people with the best processes and most advanced, personalized technology; and (ii) driving optimal utilization, preventing hypoglycemia, and treating hyperglycemia. We commend Glytec for the very important work that it is doing. The inpatient insulin management field is heavily underpenetrated, with not even 10% of hospitals using glycemic management software as of ATTD 2020. It’s a notoriously challenging arena – while there is so much more room in this field for development to improve the lives and outcomes of hospitalized people with diabetes, this arena is not (at all) for the faint of heart. The graphic below certainly is vivid ...

  • Looking to the future, Dr. Messler explained that Glytec will aim to forge additional partnerships revolving around inpatient CGM, and that the company is “actively working in that space.” Dr. Messler presented this information alongside a picture of the Roche logo, which had us wondering if there are plans for Roche and Glytec to develop an inpatient CGM system, or whether this was just an allusion to Glytec’s continued partnership with Roche’s cobas pulse inpatient BGM system. Of course, as Dr. Messler was musing about the “incredible work done in [the inpatient CGM] space,” we couldn’t help but think of Dexcom and how it recently received FDA Breakthrough Designation for an in-hospital CGM system. According to Dr. Messler, Glytec hopes to make CGM integrate “seamlessly” with Glucommander, which we find very exciting, as this could eventually lead to more widespread use of insulin pumps and AID systems within the hospital in a broad swath of populations, although this is likely far off in the future and represents speculation at this point.

  • Dr. Messler explained that Glytec’s business model revolves around its hospital and health system partners, as well as around its strategic business partners. Dr. Messler shared that Glytec currently serves “300+ hospitals and health system partners,” aligning with our last correspondence with the company in January 2022 (to be exact, we then heard 60 health systems, representing 300 hospitals). In terms of strategic business partners, Dr. Messler reiterated Glytec’s announcement from January 2022 that it has partnered with Roche Diagnostics USA to integrate its Glucommander inpatient insulin dosing software with cobas pulse when the device becomes available, pending FDA approval. Dr. Messler also highlighted Glytec’s partnership with Premier GPO, under which Glytec is the “sole supplier of insulin management software” in Premier’s Patient Safety Solutions category.
  • Glytec’s “main focus,” as Dr. Messler put it, remains on its inpatient solutions – Glucommander IV and Glucommander SubQ. While the company has an outpatient solution that supports adaptive, long-term insulin dosing for patients after they are discharged from the hospital, Dr. Messler explained that the company is more focused on its in-hospital solutions, that aim to optimize the transition from IV to SubQ insulin, as well as make insulin dosing recommendations at the time of hospital discharge. Dr. Messler also shared a mockup picture of a Cerner EHR, showing how the Glucommander dashboard integrates directly into the EHR via the company’s “SmartClick” single-sign-on button.

  • Dr. Messler’s title of “Chief Medical Officer” was quite visible and we were glad to hear him focus on both clinical outcomes as well as demonstrating economic value. On the clinical side, Dr. Messler presented data from a three-year retrospective review showing that IV insulin therapy managed with Glytec’s eGMS led to a marked reduction in the percentage of glucose readings below 40 mg/dL (the experiment’s sample size was not mentioned). On health economics, Dr. Messler presented a cost savings case study of the “Kaweah Delta Health Care District,” showing that a 71% reduction in patients with hypoglycemic, a 33% reduction in patients with hyperglycemic, and a 24% reduction in average length of stay helped save “over $9 million annually.” We are very eager to follow up on the methodology of this cost savings analysis.
  • We appreciated that Dr. Messler’s presentation was highly interactive. Kicking off his presentation, Dr. Messler used an audience polling software to ask attendees who they thought would emerge victorious from the final four of the men’s basketball teams in March Madness. An outspoken Duke fan, Dr. Messler continued to ask questions throughout the presentation, geared toward inpatient providers, and we’ve included the questions and responses below. And, boy, was this instructive from our view, to see that pieces that we see as on the more basic side, like setting targets and setting doses and using correct timing, were a challenge. Some things, like even the mention of Sliding Scale Insulin, to say nothing of “calculating starting dose” will be challenging for some to understand at all. Other things like “counting carbs” and matching carbs to prandial insulin, will be very well understood, particularly given the high number of carbs in many hospital meals and other major challenges that can be found in many hospitals. Obviously, these responses were anecdotal, but very important to hear and certainly imply that even more advocacy is needed to increase use of CGM and automated insulin delivery in the hospital, to say nothing of standardizing them far more than they are now.   


Sample Size


“What are your institutional challenges with IV insulin management?”


“Change management”

“Setting glycemic targets”

“Rn time. Following algorithms.”


“What are the main challenges in your institution when transitioning from IV to SubQ insulin?”


“Correct dosage of basal and correction”


“To make sure that basal insulin is ordered and not just SSI”

“Not over lapping [sic] long acting”

“No basal dose overlap”

“Nutrition changes occurring at the same time.”

“What are your challenges with SubQ insulin?”


“matching prandial insulin to carb intake”


“it being held without orders”


“Determining a starting dose”

Big Picture Hospital Highlights

Diabetes management in the ICU isn’t “one size fits all”; Dr. James Krinsley advocates that insulin therapy, A1c at admission, and ICU mean glucose are all important factors to consider when developing ICU glucose targets

Dr. James Krinsley (Stamford Hospital/Columbia University) presented data from two publications on the impact of inpatient hyperglycemia as well as prior insulin therapy on ICU outcomes for people with diabetes. Across both studies, Dr. Krinsley highlighted the importance of individualized insulin therapy, even in the hospital, taking into account patient treatment histories and diabetes management.

  • Investigating the relationship between A1c at hospital admission, inpatient glycemic management in the ICU, and mortality, Dr. Krinsley found that inpatient hyperglycemia had a differential impact on mortality based on A1c at admission (n=5,567). Specifically, across the eight-year duration of this study, for patients with an A1c <6.5% (n=4,406) at hospital admission, a mean ICU glucose of >180 mg/dL was strongly associated with higher mortality rate (p<0.0001) whereas, among patients with an A1c >8% (n=405) at hospital admission, a mean ICU glucose of >180 mg/dL was associated with a lower mortality rate (p< 0.0027). As Dr. Krinsley explained, this data “flew in the face” of the idea that one glucose target or target range can apply to all patients. That said, there has certainly been research demonstrating that tight glycemic control is associated with poorer outcomes for critically ill patients with diabetes such as NICE-SUGAR. However, data demonstrating that for patients with admission A1cs >8%, a mean glucose value >180 mg/dL while hospitalized is actually associated with the lowest mortality rate at 5.2% compared to 13.2% for patients with a mean ICU glucose between 140-180 mg/dL and 15.7% for patients with a mean glucose of 80-140 mg/dL is surprising. Ultimately, Dr. Krinsley advocated for properly designed RCTs to test these findings. He shared that his own medical-surgical ICU employs two different glycemic targets: 80-140 mg/dL for A1c <7% and 120-180 mg/dL (recently revised from 110-160 mg/dL) for A1c >7%, and noted that a more multi-tiered approach awaits CGM. 

  • Investigating the relationship between insulin use prior to hospitalization, inpatient mean glucose, and mortality, Dr. Krinsley found that, among people with diabetes treated with insulin, increased mean inpatient glucose was associated with lower mortality. Of note, these results did vary across A1c at admission. Specifically, across the eight years of this study, mortality rates for patients with diabetes on prior insulin therapy (n=538), patients with diabetes on oral medications only (n=986), and patients without diabetes (n=3,721) hospitalized in the ICU who A1c and blood glucose values were compared. For patients with an admission A1c >8%, mean glucose in the ICU >180 mg/dL was associated with lower mortality regardless of treatment with insulin or oral agents only (p<0.0001). However, for patients with an admission A1c <6.5%, mortality rates differed based on diabetes status. For patients with admission A1c <6.5% on insulin therapy, a mean ICU glucose between 80-140 mg/dL was associated with the highest mortality while a mean glucose ≥180 mg/dL was associated with lower mortality (p=0.0058). However, tor patients with an admission A1c <6.5% on oral agents only or without diabetes, a mean ICU glucose of ≥180 mg/dL was associated with the highest rate of mortality (p<0.0001). Collectively, Dr. Krinsley argued that these data demonstrate the importance of individualized insulin therapy for critically ill patients and called for the development of RCTs to test these findings, as these retrospective analyses have indicated that prior glycemic management may significantly impact optimal glycemic targets with implications for overall mortality risk. Additionally, Dr. Krinsley advocated for the use of CGM in the design of such RCTs to accurate capture glycemic data -- we certainly hope to see greater CGM use in clinical trials.

Ms. Rebecca Longo highlights “SAWS (Supplies, Action, Want, Specialist)” framework for assessing safety of personal technology use in the hospital; guidelines for transitioning from personal to hospital-based insulin delivery

Ms. Rebecca Longo (Lahey Hospital and Medical Center) discussed the use of personal diabetes technology among hospitalized patients outlining best practice guidelines on the use of personal insulin pumps in the hospital. As a reminder, the 2022 ADA Standards of Care include an updated clause stating that “patients who are in a position to safely use diabetes devices should be allowed to continue using them in an inpatient setting or during outpatient procedures when proper supervision is available.” Recognizing that hospitalization can be a stressful time for people with diabetes and that many may want to continue using personal insulin pumps rather than transitioning to hospital-based therapy, Ms. Longo highlighted the importance of an initial assessment upon hospital admission to establish whether a patient can safely continue to use their personal insulin pump. Specifically, Ms. Longo listed four questions providers must assess prior to supporting a patient using a personal pump in the hospital: (i) was the patient able to safely use the technology before hospitalization; (ii) is the patient able to safety use technology during hospitalization; (iii) has something changed in the patient’s cognition, clinical condition, or emotional state that may prevent them from safely using their technology; and (iv) does the patient want to continue pump therapy? To assess these four questions, Ms. Longo introduced attendees to the SAWS (Supplies, Action, Want, Specialist) framework. Starting with Supplies, Ms. Longo highlighted the importance of visualizing the extra pump supplies and skills patients have including (i) two additional infusion sets or pods; (ii) two reservoirs (if required in the pump system); (iii) rapid-acting insulin (based on hospital protocol); (iv) pump charger/batteries; and (v) patient’s ability to change their pump/pod independently. Of note, Ms. Longo emphasized that providers must specifically see that patients have these resources and skills to ensure that they are able to safely manage pump therapy should their current pump system fail. Moving to Action, Ms. Longo explained that providers should ask patients to demonstrate how they would (i) deliver a bolus, (ii) correct for a given blood glucose reading, (iii) suspend insulin delivery via their pump, (iv) set a temporary basal rate, and (v) treat a low blood glucose to confirm that the patient is able to safely manage their glycemic levels throughout their time in the hospital. While many of these scenarios occur frequently in diabetes management, Ms. Longo discussed results from a 2014 study of insulin pump skills and knowledge among hospitalized people with diabetes that found as many as a quarter of patients surveyed struggled to demonstrate pump settings, set a temporary basal rate, or suspend insulin delivery on their pump systems reinforcing the necessity of assessing these skills upon hospital admission when considering the use of personal systems in the hospital.  For patients can demonstrate these skills, Ms. Longo reinforced that in order to continue personal pump use in the hospital patients must Want to do so and have access to a Specialist who can support the patient throughout their hospital stay rounding out the last two components of the SAWS framework. With this framework in mind, Mr. Longo emphasized that inpatient personal pump use does still differ from home-based personal pump use as patients should confirm any hypoglycemia events via a point-of-care blood glucose measurement and providers may want to consider adjusting glucose targets in-line with hospital practices (140 mg/dL – 180 md/dL or 110 mg/dL – 140 mg/dL).

  • For patients who are unable to continue personal pump therapy in the hospital, Ms. Longo outlined best practices for transitioning on and off of insulin pump therapy. When transitioning off of insulin pump therapy to hospital-based insulin infusion Ms. Longo encouraged providers to look at patient insulin pump settings in the context of the patient’s historical total daily insulin dose, weight-based dosing guidelines, and clinical condition to determine the most appropriate basal insulin dose. For bolus insulin, Ms. Longo encouraged providers to use existing patient carb ratios in conjunction with weight-based dosing, and prescribed doses for estimated meals if not carb counting, as well as considering the patient’s clinical context and how this may change their bolus insulin needs. Importantly, when discontinuing insulin pump use, Ms. Longo emphasized that this should not be done until two hours after long-acting basal insulin has been administered to ensure patients have received the correct basal does. As patients are transitioning from hospital-based insulin delivery back to personal pump use, Ms. Longo highlighted the importance of reconciling any differences in therapy between hospital-initiated insulin delivery and personal pump settings. Additionally, Ms. Longo reiterated the importance of not discontinuing insulin infusion until two hours after long-acting basal insulin is administered and also not restarting pump-based basal rate until 22-24 hours after the last hospital-based basal dose to ensure patients are not stacking basal insulin.

  • Ms. Longo also discussed inpatient use of personal CGM and AID, though neither has been approved by the FDA. Despite this, Ms. Longo explained that CGMs can be especially beneficial by providing trend and alert information that can then be verified by point-of-care blood glucose testing. However, Ms. Longo noted that personal CGM should not be used for insulin dosing decisions and could instead help inform nursing staff of when additional point-of-care tests should be done that could then inform insulin dosing. Similarly, as AID systems rely on CGM data for insulin dosing, Ms. Longo emphasized that if patients continue to use personal AID systems in the hospital that they must be overseen and co-managed by a clinical expert with experience using AID systems. Back in 2020, the Continuous Glucose Monitoring and Automated Insulin Dosing Systems in the Hospital Consensus Guidelines Panel published a report that largely called for additional research on the use of personal CGM and AID devices among hospitalized patients and most TGC (tight glycemic control) experts would certainly agree that more data is needed in this field to better understand the benefits and drawbacks of personal technology use. On a different note, earlier this year in March, the Dexcom’s “hospital CGM system” received FDA breakthrough designation and we are eagerly awaiting more data on this front across Dexcom’s in-hospital CGM trials as this would presumably represent a significant step forward for inpatient glucose management should this system get FDA approval.

Prof. Ketan Dhatariya discusses controversies in the diagnosis and management of DKA, calling for updated ADA guidelines that recognize euglycemic DKA and the importance of ketone monitoring

Professor Ketan Dhatariya (Norfolk and Norwich University Hospitals, UK) compared guidelines from the ADA and the Joint British Diabetes Societies for Inpatient Care (JBDS-IP)’s on DKA management, arguing that the US guidelines are outdated due to their failure to acknowledge euglycemic DKA. Prof. Dhatariya explained that DKA diagnosis should be based on the three core components of this hyperglycemic crisis: (i) the “D” component, or glucose concentration; (ii) the “K” component, or the presence of ketones; and (iii) the “A” component, which requires confirmation of an acidosis. On the first point, Prof. Dhatariya noted that the ADA criteria – published over ten years ago, in 2009 (!) – require a glucose level >250 mg/dL, whereas the JBDS guidelines require a glucose level >200 mg/dL or a previous history of diabetes. Prof. Dhatariya argued that the ADA’s requirement for glucose >250 mg/dL is no longer appropriate given the increasing use of SGLT-2s and the increased risk of euglycemic DKA in patients taking these newer agents. Overall, Prof. Dhatariya emphasized that there is no need for a lower glucose limit to diagnose DKA, provided the patient has a past diagnosis of diabetes. On the second point, Prof. Dhatariya noted that ketone measurements are a crucial component of diagnosis and lamented that many HCPs diagnose DKA without measuring ketones despite the widespread availability of ketone meters. Indeed, an article published in Diabetes Care in February 2022 came to this same conclusion, stating that ketone measurement is necessary for diagnosis and management of DKA. Finally, on acidosis, Prof. Dhatariya noted that the differences in classification reflect the different healthcare systems in the US and UK. The US classifies degrees of acidosis as “mild,” “moderate,” and “severe,” categories which are linked to different billing codes and presumably, different treatment costs. Meanwhile, the UK does not make such distinctions between levels of acidosis, as the NHS provides free healthcare to the public, regardless of insurance status or treatment cost. Overall, Prof. Dhatariya called for an update to the ADA consensus guidelines on DKA management, echoing statements in a Lancet article that he co-authored with Emory’s Dr. Guillermo Umipierrez in 2017, titled “Guidelines for management of diabetic ketoacidosis: time to revise?”. In closing, Prof. Dhatariya noted that all JBDS guidelines, including those on DKA management (recently updated in February 2022), are available for free download online at the Joint British Diabetes Societies (JBDS) for Inpatient Care Group website.

Dr. Jean Charles Preiser pushes for adaptive blood glucose targets in the ICU; recent CONTROLING RCT studying adaptive ICU glucose targets stopped early due to high hypo risk; Accu-Chek Inform II BGM unlikely to cause insulin dosing errors in critically ill patients

Dr. Jean-Charles Preiser (Erasme University Hospital, Brussels) championed the use of individualized and adaptive blood glucose targets in critically ill hospitalized patients. In what he called a “reasonable and appealing hypothesis,” Dr. Preiser stressed that insulin resistance is a component of the body’s innate stress response, reflecting its metabolic state and ability to use macronutrients (Bellomo et al., 2013) in a state of “adaptive survival” (Soeters et al., 2012). As such, Dr. Preiser stressed that the “optimal” blood glucose level for a critically ill ICU patient varies from the early acute phase of healing through the end of recovery. Dr. Preiser highlighted his research from 2018 in Critical Care Medicine, in which hospitalized patients (n=316) with a glycemic ratio (admission blood glucose to average blood glucose over stay) closest to one had the lowest ICU mortality. In contrast, patients with a glycemic ratio below or above one had a higher chance of ICU mortality (see the graph below with the red mortality curve). Dr. Preiser explained that the need for an improved blood glucose target calculation protocol for ICU patients led to the multicenter CONTROLING RCT (2021) that aimed to individualize blood glucose targets based on patients’ pre-admission usual glycemia. However, that was a balancing act without a positive end. Notably, the study was stopped early due to the possibility of complications from severe hypoglycemia. At the conclusion of the study, both groups had different glycemic profiles, but equal rates of mortality. Most concerningly, 3.9% of patients in the adaptive-ICU-target group recorded severe hypoglycemia episodes below 40 mg/dL, compared to only 2.5% of patients receiving standard of care (p=0.09). Dr. Preiser used the CONTROLING study to emphasize the need for further research into how blood glucose targets can be optimized for critically ill patients in the hospital to ensure that they are achieving the best glycemic outcomes. We are eager to hear more impressions since ending a study early for safety reasons certainly is not confidence-inspiring. Yet and still, Dr. Preiser’s presentation was very insightful, in both how it highlighted how much more can be done to improve outcomes for people with diabetes in the hospital.

  • In a closely related session following Dr. Preiser, Dr. James Nichols (Vanderbilt University) presented a recent multicenter evaluation Roche’s Accu-Chek Inform II BGM in critically ill patients. Dr. Nichols noted that most current BGM systems do not meet the FDA guidance criteria for accuracy in all hospitalized patients (as they are very stringent). However, turning to his recent study (2021) published in the Journal of Applied Laboratory Medicine, Dr. Nichols demonstrated across 10 US hospitals that patients in the ICU (or “critical” patients outside the ICU) had a low risk of potential insulin-dosing errors when using the Accu-Chek Inform II BGM, despite only meeting the FDA’s accuracy guidelines in certain patients. Dr. Nichols reinforced that critically ill patients “challenge the analytical performance of BGM,” and particularly pointed to those with poor peripheral perfusion as cases where HCPs should consider arterial or venous samples instead of capillary fingersticks to “better reflect physiologic central circulation.”

Poster Hall Highlights

Post-hoc analysis of RAPIDS RCT (n=238) shows proactive inpatient care drives 0.6% A1c decrease at median five months post discharge and 3.2x increased odds of treatment intensification vs. usual care; reinforces importance of inpatient diabetes care

Dr. Mervyn Kyi (Royal Melbourne Hospital, Victoria, Australia) presented a post-hoc analysis from the RAPIDS RCT (n=238 with follow-up A1c) associating an early intervention model of inpatient care for type 2s with greater treatment intensification and greater improvement in A1c following discharge. As a reminder, the RAPIDS study was originally presented at ADA 2017 and later published in Diabetes Care in 2019, showing that an early intervention model of bedside diabetes management (within 24 hours of admission) was associated with decreased hyperglycemia and lower rates of hospital-acquired infection. As motivation for the post-hoc analysis, Dr. Kyi noted that widespread clinical inertia often means that the vast majority of patients with suboptimal A1c at admission may not receive the treatment intensification that they need to see improved outcomes, and also cited several barriers complicating this process including: (i) the treatment team’s lack of knowledge or confidence; (ii) the lack of prioritization for diabetes care; and (iii) the lack of continuity of care following discharge. As such, the RAPIDS study created a new proactive model of diabetes care, aiming to provide early bedside consultations to all inpatients with diabetes (see picture below).

  • Participants receiving proactive care saw a 0.6% larger decrease in A1c compared to those in the usual care arm with A1c decreases of 0.9% and 0.3%, respectively (p=0.03) at a median five months post discharge. As expected, in a subgroup analysis with participants whose baseline A1c was >8%, this difference was even higher at 0.9%. Specifically, those who received early intervention saw a 1.5% A1c decrease compared to those in usual care with a 0.6% A1c decrease (p=0.03). These results are hugely encouraging and reinforce the importance of inpatient diabetes care not just for in-hospital glycemic management, but also for outcomes post discharge.

  • The proactive care arm also saw a higher odds of treatment intensification at discharge compared to the usual care group (adjusted OR=3.2, p<0.001). Without adjusting for age, gender, and Charlson index, the percentage of patients receiving treatment escalation at discharge was 18% higher for those receiving proactive care compared to usual care at 37% and 19%, respectively (p<0.001). While there were no outcomes available to characterize SGLT-2 or GLP-1 usage among participants or medication adherence, we think these findings reinforce the importance of inpatient patient education and tailored diabetes care within the hospital to ensure that patients are receiving the care that they need and being discharged with an appropriate treatment regimen to ideally prevent future hospitalizations. We imagine that the increased treatment intensification was largely responsible for the A1c improvements witnessed post discharge.

Seven-month UCSD QI study (n=250): EHR optimization and provider education improves percentage of blood glucose values logged into “Glucose Management” tab from 56% to 79%

Dr. Nicholas Albino (UCSD) kicked off the 2022 Hospital Diabetes Meeting’s poster session by presenting data from a quality improvement project to standardize the logging of at-home BGM readings in the EHR. Dr. Albino explained that in his clinic (Rady Children’s Hospital), all patients with new onset diabetes receive training on how to use BGM independently at home, but the lack of standardized processes to document these readings in the EHR has complicated hospital workflows given that Rady Children’s Hospital has many outside rotators delivering care. Prior to the quality improvement project, there were three different ways to log home BGM values into the EHR, making it difficult to consistently retrieve blood glucose levels from the same place in a patient’s chart. Plus, Dr. Albino noted that home BGM values were not visualized within the EHR in the same way as in-clinic values, making it difficult for rotating providers to understand patients’ glycemic management out of the clinic, as well as to compare at-home BGM readings to in-hospital ones. Dr. Albino then turned to his QI study, a chart review of patients admitted with type 1 or type 2 diabetes (n=250) from June 2021 to January 2022 aiming to improve the percentage of blood glucose values documented in the EHR “Glucose Management” tab from a baseline of 56% to at least 70%. The interventions, Dr. Albino explained, included: (i) multidisciplinary engagement and feedback; (ii) creation of an order for home-use BGM within the EHR; (iii) updated “Glucose Management” tab in the EHR; and (iv) workflow education to nurses and providers.

  • The QI interventions increased the percentage of blood sugar records documented into the “Glucose Management” tab by 23%, from 56% (June 2021) to 79% (December 2021). These results, exceeding the study’s initial goal of at least 70% of glucose values logged, are certainly impressive and speak to the value of streamlined clinic workflows coupled with provider education. The synergistic value of workflow optimization and provider education is most visible in a graph that Dr. Albino shared (see below). As Dr. Albino was having a baby in the middle of the study, there was a slight time gap between the time of the EHR redesign and the clinic education. However, this time gap allows one to see two incremental steps that led to the 23% improvement in blood glucose logging: (i) the EHR modifications (~14%); and (ii) the nurse/provider education (~9%). We continue to hear from providers about the value of simplified glucose tracking, and so it’s very encouraging to see those sentiments fortified by this analysis.

  • Excitingly, Dr. Albino shared that upcoming QI initiatives at Rady Children’s Hospital will enable CGM-EHR integration for hospitalized patients. We continue to believe in the transformative value of CGM in the hospital when used effectively and integrated directly into the EHR, and so we look forward to hearing more about this project, potentially at a future Hospital Diabetes Meeting!

Inpatient EHR-based insulin dosing calculator developed to improve bolus doses taking into account glucose targets, hypoglycemia and hyperglycemia correction factors, insulin to carb ratios, current glucose levels, and carb intake

Dr. Celeste Thomas (University of Chicago) presented a new model for inpatient bolus insulin dosing via a dose calculator located in the EHR taking into account blood glucose levels, carbohydrate intake, insulin to carb ratio, and correction factors. As Dr. Thomas explained, prior to the design and implementation of a new insulin ordering and dosing component of the EHR, providers were limited in their ability to specify insulin doses, instead only have the option to select a fixed insulin dose or add more personalized dosing instruction as an EHR note that could unfortunately be easily missed. To address these challenges, Dr. Thomas’s group developed an insulin dosing calculator within the hospital’s Epic EHR with sections to enter: (i) patient glucose targets during the day and at night; (ii) hyperglycemia and hypoglycemia correction factors during the day and overnight; and (iii) carb ratios for breakfast, lunch, and dinner, as well as overnight. Then, when providers needed to dose insulin for hospitalized patients, as long as these parameters had been established, they would only need to enter which meal the patient was eating, the patient’s current glucose level, and the patient’s current carb intake. Given this information, the system could then calculate the correct bolus insulin dose for the patient. This insulin dosing calculator was launched at Comer Children’s Hospital in August 2021 and was recently launched in the University of Chicago adult hospitals in March 2022 following “hospital-wide nursing training.” While an official survey of provider attitudes toward this new system has not been conducted, Dr. Thomas described hearing “good things” from nursing staff who were “satisfied” with the new insulin dosing calculator.

  • Highlighting a case study with a patient on oral prednisone, which can affect insulin resistance, Dr. Thomas discussed how the insulin dose calculator provided more appropriate insulin dosing recommendations than a fixed insulin dose. Specifically, Dr. Thomas discussed two meals during which the patient ate 49 and 17 grams of carbohydrates respectively. Had the patient received a single fixed insulin dose that didn’t take into account their prednisone prescription and their adjusted insulin to carb ratio, Dr. Thomas argued that the patient would have been at risk of hyper or hypoglycemia, depending on what the fixed insulin dose had been. However, with the ability of nursing staff to use the EHR-based insulin dose calculator, bolus doses of 32.5 units and 11.5 units, respectively, were able to be delivered helping to keep the patient within target glucose ranges.
  • During Q&A, Dr. Thomas explained that this insulin dosing calculator can currently only be used when an endocrinologist signs off on the insulin dosing orders. As Dr. Thomas explained, this step has been implemented as a safe-guard to ensure that providers ordering insulin understand insulin to carb ratios and correction factors to ensure patients receive an appropriate amount of insulin. We certainly appreciate this safety-minded approach, but we are curious what other mechanisms may be able to be implemented to provide this level of personalized insulin dosing and safety without the need for an endocrinologist to sign off to potentially help alleviate burden on providers and expand access to care beyond those served by endocrinologists, even while hospitalized.


--by Armaan Nallicheri, April Hopcroft, Hanna Gutow, and Kelly Close