Consensus Considerations and Good Practice Points for Use of Continuous Glucose Monitoring Systems in Hospital Settings

Most continuous glucose monitoring (CGM) systems track glucose concentrations in the interstitial fluid (ISF) every 1–15 min with sensors that are wearable for up to 15 days, offering a detailed view of glucose fluctuations over time. These systems, which include a sensor inserted in the subcutaneous tissue of the arm, abdomen, or lower back, are divided into intermittently scanned CGM and real-time CGM (1). Using algorithms, CGM systems can predict impending changes in glucose concentrations and notify users of potential hypoglycemia or hyperglycemia episodes. While ISF glucose concentrations generally mirror blood glucose concentrations, discrepancies can occur based on the physiological conditions at the time of measurement, such as after a period of exercise or postprandially.

The enhanced accuracy and reliability of CGM systems have broadened their appeal for managing insulin dosing in hospitalized individuals (2,3). Hyperglycemia and diabetes are estimated to be present in as many as 25% of medical and surgical patients (4). Initially CGM was approved for outpatient diabetes management in various regions, including the U.S., Canada, and the European Union (EU), and then the pandemic saw a temporary authorization for hospital CGM use in Canada and a U.S. Food and Drug Administration (FDA) enforcement discretion for hospital CGM use in the U.S. This shift was aimed at reducing health care worker exposure to coronavirus disease 2019 (COVID-19) and conserving personal protective equipment (5,6). Supplementing traditional point-of-care (POC) glucose monitoring, CGMs have found their place in both intensive care unit (ICU) and non-ICU settings, assisting in glucose management for hyperglycemia and for people with diabetes on insulin therapy.

In the postpandemic era, there is growing interest among individuals and health care professionals in using CGM systems for hospital-based glucose monitoring instead of conventional POC methods. Many people with diabetes, accustomed to CGM systems at home, often prefer CGM over frequent finger stick blood glucose monitoring. This presents challenges for health care professionals, including laboratorians, to integrate CGM use safely within the hospital setting (7).

The aims of this consensus guidance document are to provide an update on the application of CGM in hospital patients and to provide good practice points for consideration, leveraging insights and opinions from laboratory medicine, endocrinology, and nursing. The CGM devices referred to include patient owned and hospital owned. This consensus report serves as a reference to foster best practices and develop a regulatory framework for policy and practice guidance. The document reviews evidence on hospital CGM use and covers the following topics.

• 1. Potential benefits of CGM systems for inpatient populations

• 2. Existing guidance for use of CGM systems in hospital settings

• 3. Analytical and clinical evaluation of CGM systems performance

• 4. Factors to consider for safe use of CGM systems in hospitals

• 5. Staff training, clinical workflow, and hospital policies

• 6. Current knowledge of CGM system performance in hospital settings, from findings of clinical studies

• 7. Management of CGM systems from a quality assurance perspective

• 8. Regulations and accreditation standards for CGM system use

• 9. Integration of CGM data in the electronic health record (EHR)

• 10. Cost considerations for hospital CGM system use

• 11. Barriers to CGM system implementation in hospitals

The scope of the document was determined by the consensus of the authors. Subgroups were formed to research each of the sections to be included and prepare a summary and key points.

MEDLINE was used to identify studies for analysis. The search parameters used were, “continuous glucose monitoring” or “flash glucose monitoring” AND “hospital” or “inpatient.” The search was limited to English-language studies published in the period 2012–2023.

Clinical studies have shown that CGM systems offer significant benefits in managing glucose measurements for ambulatory diabetes care, including improved glycemic outcomes, increased time with glucose in target range, fewer hypoglycemic episodes, and enhanced well-being (8–10). Hyperglycemia in diabetes can lead to severe complications and poor clinical outcomes. CGM systems are gaining traction in hospitals due to improved sensor accuracy (11–14).

Improved glycemic management in the critically ill can reduce short-term mortality and ICU stays, though it may raise the risk of severe hypoglycemia, associated with higher mortality. Thus, frequent glucose monitoring is crucial, especially with intensive insulin therapy (14–18). Traditional capillary POC testing (POCT), often every 1–2 h in critical care and four times a day in noncritical care, is uncomfortable and increases nursing workload. CGM systems, offering glucose data every 1–5 min, could be a beneficial alternative for the critically ill requiring insulin therapy, especially those at risk of hypoglycemia (15,17–21).

During the COVID-19 pandemic, management of glucose in individuals with hyperglycemia and for those with diabetes in isolation presented challenges. CGM systems proved beneficial in reducing direct person-to-person contact and minimizing cross infection risks, thereby supporting safer, more frequent glucose monitoring (19–23).

In surgical settings, people with diabetes face higher risks of infection, and exhibit impaired wound healing, with extended hospital stays and increased mortality due to hypo- or hyperglycemia (24–26). CGM systems can help prevent dysglycemia during the perioperative period, though further study is needed on their reliability during procedures like coronary artery bypass surgery, due to the potential for data transmission interruptions (27–29).

Hyperglycemia and hypoglycemia are associated with higher mortality and morbidity in preterm or low–birth weight neonates, neonates born to mothers with diabetes, and those with other critical illnesses. Although CGM systems are not intended for use in neonates, preliminary studies show potential benefits of CGM use in allowing early detection of extreme glucose variation (30–35). Use of CGM during pregnancy has also been associated with improved maternal and neonatal outcomes, such as improved HbA_1c and reduced cesarean section rates (36,39).

The benefits of CGM systems in several inpatient populations, based on findings of recently published clinical studies, are summarized in Table 1.

Literature offers guidance on CGM use in acute care settings. Current clinical practice guidelines for noncritical care suggest using real-time CGMs along with bedside POC glucose monitoring for insulin dosing in adults with type 2 diabetes at risk of hypoglycemia (37). The increase in individuals wearing personal CGMs in hospitals is noted, and their continued use under hospital protocols is recommended. The American Diabetes Association “Standards of Care in Diabetes” continue to emphasize the role of CGM systems in noncritical settings with trained teams but advise against replacing POC monitoring for insulin dosing or hypoglycemia identification or management (38).

The 2020 hospital consensus guideline for inpatient use of CGM and automated insulin delivery systems highlights their potential to improve glycemic outcomes and reduce hypoglycemia risks. It acknowledges continued use of patient CGM devices in hospital and initiation of new CGM system use while in hospital. The guideline recommends individualized therapy, adequate health care professional training, and robust protocols for safe and effective implementation, including remote monitoring to integrate CGM data into EHRs, alarm management, and strategies for CGM accuracy in the first 48 h after sensor insertion (7).

The CLSI POCT05, 2nd edition, guideline focuses on the technical performance of CGM systems in clinical settings, emphasizing the need for accuracy, reproducibility, and ongoing clinical evaluation. It suggests that CGMs could benefit individuals with nocturnal hypoglycemia, calling for more extensive accuracy studies across different populations and settings (39).

The following recommendations are made for the design of analytical and clinical studies on CGM system performance and for the interpretation of data from CGM system manufacturer package inserts.

Studies should reflect situations and populations where CGM systems are intended for use. If the situations include certain procedures, administration of specific medications, or populations anticipated to have abnormal levels of endogenous interferents, impact on analytical performance should be assessed. If no data are provided by the manufacturer for the specific intended use, laboratory verification studies should be performed.

Table 2 lists and describes end points for clinical CGM performance evaluation. More details can be found in Supplementary Table 1.

Expected differences between whole blood capillary and circulating venous or arterial samples must be considered in selecting the comparator sample types used for evaluation of CGM systems that measure glucose concentrations in ISF.

The time from sample collection to glucose measurement will also impact the reported concentration. Comparison method measurements must occur within 15 min from sample collection, and the same tube/sample type must be used throughout the comparison study (40).

Evaluation of CGM system accuracy requires comparison of glucose concentrations with those measured with a reference method traceable to an established standard reference material and reference measurement procedure. The evaluation must include concentrations reflecting hypoglycemia, euglycemia, hyperglycemia, and values at the extremes of the reporting range, throughout sensor life and across sensor lots (39).

Comparator methods should fulfill “desirable” analytical performance specifications based on biological variation (40,41). External quality assessment programs should be used for comparator methods to ensure against significant bias. Comparator methods could include measurement of glucose in plasma on a central laboratory chemistry analyzer or glucose measurement in whole blood on a blood gas analyzer.

• 1) If health care professionals want to use CGM systems beyond their intended use (e.g., to replace or reduce POC glucose measurements), analytical and clinical performance should be assessed.

• 2) If device labeling does not contain data about the specific use, verification studies should be performed, especially if no literature is available for the specific use of the CGM system or for the specific CGM system brand.

• 3) CLSI POCT05 (39) provides helpful guidance. Special care should be taken to use adequate comparator samples and methods.

Interruption of CGM system measurement and connectivity increases the risk of missed hypoglycemia or hyperglycemia events (42). Malfunction of CGM systems can result in data loss, and recalibration or replacement of the sensor can be required (43,44). Compression artifacts should be considered when CGM glucose readings are unexpectedly low (31). In one study, electrical interference with CGM systems in the emergency department was reported, with individuals who needed ventilation support (45).

Studies on CGM system interferences are limited. Commonly reported interferences include acetaminophen, hydroxyurea, mannitol, tetracycline, salicylic acid, and ascorbic acid (46–51). The effect of interferences may vary between manufacturers or CGM models and by glucose concentration (46,50). An option for mitigating interference risk is to prevent the use of CGM systems with known interferences in individuals receiving certain treatments. For example, in one clinical study, with use of the Dexcom G6 CGM system, use of the CGM system was not permitted in any individual receiving hydroxyurea, a known interferent (52). An alternative option is to use CGM systems while keeping a log of potential interferences and with reliance on POC or laboratory-based glucose measurements for treatment decisions, provided the POC and laboratory-based methods are not affected by the same interference (53).

Awareness of sensor “warm-up” or equilibration periods is key to ensuring accurate readings (31). CGM systems must equilibrate following insertion and can be less accurate in the first 24–48 h of use (39). It is important to develop protocols for guidance on use of POC or laboratory glucose measurements until warm-up periods are over and CGM system readings are deemed acceptable (53,54).

Hospital protocols for CGM should provide direction on the following:

• 1) Sensor equilibration time requirements, during which confirmatory POC glucose measurements should be performed until differences between CGM and POC are <20% for at least three consecutive time points.

• 2) Potential interferences that preclude patients from being eligible for CGM. Clinical staff must be aware that CGM cannot be used for clinical decision-making in these patients. Monitoring must be by POC or laboratory glucose measurements in these patients.

• 3) Glucose thresholds above and below which confirmatory glucose testing should be performed with use of another glucose measurement method. This will help mitigate the risks associated with sensor compression and ISF lag time.

• 4) Expected response to CGM alarms, including documentation and communication processes.

Before implementing CGM systems in the hospital, key stakeholders should be identified and hospital policies developed. Development of hospital policies and clinical workflow should include input from hospital leadership, diabetes specialist, prescriber, nursing, pharmacy, laboratory, information technology (IT), risk management, and legal teams (38). Written policy should include information on CGM systems and how they work and include clear guidance for appropriate individual selection for CGM use versus conventional POC glucose monitoring. The policy should indicate which personal and/or hospital CGM systems are permitted for use during hospitalization as well as when CGM use is to be discontinued. Details about onboarding and maintaining a CGM system, a data collection plan, and guidance for pattern identification, documentation, and protocols for responding to alerts and alarms for hypo- and hyperglycemic events should be included.

Education for clinical staff on CGM technology is critical for successful and appropriate use of CGM systems in hospitals. The curriculum should include the information listed in Table 3.

Incorporating trend arrows into insulin dosing algorithms in the outpatient setting improves average glucose, time with glucose in target range, and rates of hypoglycemia (55–58). Research is needed on the development and use of trend arrows in the inpatient setting.

Use of glucose alarms and alerts in the hospital is recommended to prevent glycemic excursions (7,8), with consideration of frequency to avoid alarm fatigue, and thresholds should be provided at which interventions are necessary for safety (59). The thresholds should be included in the CGM hospital protocols. Although some alarms may be customized, some alarms, such as those indicating urgent low glucose, cannot be turned off or modified for safety reasons.

No CGM devices are currently formally approved for use in the hospital setting. With guidance for use in a hospital policy, a personal CGM can be worn for the individual's own information. Individuals wearing a personal CGM should be notified that the hospital POC blood glucose meter may still be used for treatment decisions, such as insulin dosing and hypoglycemia assessment and treatment, and documentation in the EHR. Individuals are encouraged to contact the nurse for confirmation with the POC meter when CGM trend arrows are indicating real-time or impending hypo- or hyperglycemia.

• 1 A CGM system and/or inpatient glycemic management committee should oversee the development and implementation of hospital-approved policies and procedures for use of CGM systems in hospital. This committee should have representation from nursing leadership, physician leadership (e.g., endocrinology, internal medicine, hospitalists), laboratory POCT leadership, and information services, hospital administration, pharmacy, and risk management/legal.

• 2) Policies for patient-owned and hospital-owned devices should be developed. A list of patient-owned CGM devices approved for use must be included.

• 3) Where CGM sensor calibration is required, protocols must be developed that include indication of acceptable agreement between POC or laboratory glucose and CGM values (<20% difference).

• 4) Protocols and training for sensor placement must be developed and specify, for example, sensor location and any deviations from manufacturer recommendations that are required for inpatient use or during surgical procedures.

• 5) Protocols and training must include information regarding patient populations who require or may require CGM device removal and/or procedures for which CGM devices may need to be or must be removed (e.g., during specific medical imaging procedures including MRI, computed tomography [CT] scan, nuclear medicine imaging, and bone density scan).

With expanded use of CGM in hospital during the COVID-19 pandemic, data from pilot studies on CGM performance in hospitals have been published. These studies have included assessment of the potential clinical risks of using CGM results for clinical decision-making for both adult and pediatric patients (18,20,22,52,60–63). In most studies investigators used hybrid approaches for CGM use with confirmatory POC monitoring required for certain, defined situations. Error grid analysis generally demonstrated CGM results to be clinically viable, showing greater ability of CGM to detect hypoglycemic events compared with POC monitoring but lower accuracy in the hypoglycemic range (18,22,52). In some studies, a switch to POC blood glucose monitoring occurred once CGM hypoglycemia or hyperglycemia alarms were triggered (20,22,52). Negative bias between CGM and POC measurements was also noted during patient hypoperfusion and when hypothermia protocols were in place (20). Complete loss of CGM signal occurred with patient cardiac arrest and with defibrillator use (20).

These studies have provided strategies for managing sensor warm-up time. Variance thresholds between CGM and comparison POC values have been suggested as confirmatory of sensor validation, with 20% used in one study (20) and an absolute value of 35 mg/dL (2.0 mmol/L) (52). A summary of the findings from these studies can be found in Table 4.

Use of CGM systems is associated with a reduction in the number of capillary POC glucose tests performed in ICUs and general medicine and surgical units (20,22,53,64), which has been shown to improve patient experience (20,64). Some CGM systems require periodic POC blood glucose calibration throughout the life of the sensor to maintain accurate correlation between ISF and blood glucose. At the time of writing, newer models such as the Dexcom G6/7 and Abbott FreeStyle Libre 2/3 have incorporated factory calibration (65–67), which enables greater user convenience. Of note, calibration of a sensor during a period of rapidly changing glucose concentration can introduce inaccuracy (68). For CGM systems with optional calibration, careful manual calibration at stable glucose concentrations may have a relevant impact on the accuracy and reliability of CGM readings.

CGM systems have been reported to function well in the operating room, as well as in the preoperative, interoperative, and postoperative periods, providing there are no sensor errors (69). Blood gas analyzers are commonly available for POCT in operating rooms and can be used to confirm CGM results, if required. As fingers are not always accessible for capillary POC glucose monitoring in the operating room due to positioning, the functionality of CGM in this setting was noted as a unique benefit (69).

The Association for Diagnostics & Laboratory Medicine Academy (ADLM Academy, formerly known as the National Academy of Clinical Biochemistry) in the U.S. has issued a guidance document focused specifically on management of POCT (70). The Canadian Society of Clinical Chemists has also issued position statements touching on essential aspects of a POCT program and recommended quality assurance practices (71,72). Transferability of these POCT recommendations to CGM system use in hospitals is complicated and requires careful consideration, as described in Supplementary Table 2. Successful implementation of a robust CGM system quality assurance program in a hospital setting has not yet been reported, but we anticipate that close collaboration will be required among the laboratory, clinical stakeholders, and organizational leadership and people with diabetes. Input from relevant accreditation agencies and regulatory bodies may also be needed.

CGM systems are approved by the FDA as noninvasive, remote monitoring devices that can provide information on trends in glycemic patterns and facilitate therapy adjustment in the home by lay users. As discussed above, the FDA issued an enforcement discretion policy for noninvasive remote monitoring devices used to support glucose monitoring during the COVID-19 public health emergency. This policy ended on 7 November 2023, meaning hospital CGM use is now technically off-label (5,73).

The Clinical Laboratory Improvement Amendments (CLIA) regulations establish quality standards for laboratory testing performed on specimens from humans for the purpose of diagnosis, prevention, and treatment of disease or assessment of health (CDC CLIA Law Regulations) (74). The Centers for Medicare & Medicaid Services implements the CLIA program, and each state agency enforces requirements equal to or more stringent than CLIA requirements. Some confusion or controversy exists as to whether CLIA regulations apply to CGM systems. CGM systems perform glucose measurements of a biological sample (ISF), and results can be used for therapeutic decisions, suggesting that CLIA regulations are applicable. However, CGM systems are approved by the FDA as noninvasive remote monitoring devices without assignment of CLIA test complexity, suggesting that CLIA regulations do not apply. Moreover, CGM systems are not listed in FDA CLIA Database (75). For clarification, we made inquiries with each of 50 state agencies as to whether the state enforces laboratory regulatory requirements for CGM systems. Of 50 state agencies, 29 responded. Most responses for the states indicated that CLIA regulations are followed and there are no state-specific laboratory requirements for CGM systems. A few responses alluded to in vivo specimen testing not subject to CLIA regulations, and others indicated that even though CLIA does not apply, the requirement remains to follow the manufacturer’s instructions for use. Interestingly, the state of Nevada agency indicated that CGM system measurement meets the definition of a medical laboratory procedure under the Nevada Revised Statutes (NRS 652.060) (76) and would require a CLIA certificate based on the FDA categorization of test complexity. Yet, test complexity has not been assigned to CGM systems by the FDA.

Clinical laboratories in the U.S. operating under a CLIA certificate may select from seven accrediting agencies with deemed status according to the CLIA application for certification (Form CMS-116) (77). One of these is The Joint Commission (TJC). In response to our inquiry about potential CGM system–related accreditation standards, TJC indicated that CGM systems are approved by the FDA as monitoring devices and are not considered laboratory tests; therefore, they are not regulated by the TJC-specific laboratory standards. Moreover, for monitoring devices the minimum requirement is to follow the manufacturer’s instructions for use. Per the College of American Pathologists (CAP), current CGM systems are not FDA approved for use in diagnosis or treatment of individuals in a hospital setting and CAP standards do not apply.

CGM systems currently available in Canada have been approved by Health Canada for the self-management of diabetes. During the COVID-19 pandemic, Health Canada temporarily authorized the expanded use of select CGM systems in all hospital and professional health care settings (6).

Canada does not have an equivalent to the U.S. federal CLIA regulations for diagnostic laboratory testing. The quality of testing is ensured through provincial/territorial acts and regulations and/or through accreditation standards of nongovernmental organizations and professional societies. The acts and regulations identified did not provide specific details on quality requirements for CGM systems, but most required that diagnostic laboratories operate in compliance with the standards of an accreditation program and/or the province’s College of Physicians and Surgeons.

Diagnostic laboratories in Canada are accredited by a variety of different agencies. In a 2017 environmental scan, the Canadian Agency for Drugs and Technologies in Health (78) found that many of the Canadian agencies base their requirements for POCT quality and competency on ISO 22870:2016: Point-of-Care Testing (POCT) — Requirements for Quality and Competence (79), or on an adaption of this document, adapted to the Canadian setting (80). Note that ISO 22870:2016 has been replaced by ISO 15819:2022.

Of the six Canadian agencies identified, none were found to have explicit standards for CGM systems. None of the agencies were found to consider their POCT standards to be applicable to CGM systems.

A summary of international regulations can be found in Supplementary Table 3. In the EU, medical devices must be used in accordance with their intended use. Otherwise, with “off-label” use, the full responsibility of a manufacturer is transferred to the operator of the device. This is a concern for health care professionals and hospitals, as most CGM systems are intended for home use by lay users.

Many EU member state guidelines also require that measurement systems in the health care context be quality controlled. For traditional analyses in the hospital setting, these measurements are in vitro diagnostic measurements and the respective analyzers can be used with quality control and proficiency testing materials. The same cannot be said for CGM systems, which perform in vivo measurements. In some EU member states, this difference limits the coverage by health care insurers for costs associated with CGM systems measurements.

A comprehensive review of CGM systems and laboratory regulations beyond the U.S., Canada, and the EU is outside of the scope of this document. Interested readers should refer to resources available from national regulatory agencies for more information.

In 2022, the International Organization for Standardization (ISO) transferred its POCT standards to ISO 15189: Medical Laboratories — Requirements for Quality and Competence (79). Annex A of ISO 15189:2022 contains a POCT applicability statement that excludes individual self-testing altogether, though also a qualifying statement that elements of the document may be applicable. Personal communication with the Standards Council of Canada, a national ISO member, revealed that there have not been specific or granular discussions regarding the oversight or operation of CGM systems.

• 1) Given the rapidly growing interest in CGM system use in professional health care settings, clear regulations and accreditation standards should be developed by provincial, state, national, and international agencies for this application of CGM use.

• 2) CGM devices sample a biological fluid for measurements, which may be used for clinical decision-making. Use of CGM devices in hospital settings, for clinical decision-making, should fall under CLIA regulations in the U.S.

• 3) Outside the U.S., use of CGM devices in hospital settings, for clinical decision-making, should be considered as POCT and should be regulated/accredited as such.

The Integration of CGM system Data into the Electronic Health Record (iCoDE) report (81) provides detailed and practical recommendations for health care institutions that wish to interface CGM system data with their EHR.

The typical flow of data from a CGM system to the EHR is outlined in Fig. 1. This is adapted and simplified from what was included in the iCoDE report (81) depicting the relationships among individuals, CGM manufacturers, and health care organizations/professionals for patient-owned CGM systems.

iCoDE recommends a “data pull” from the CGM system portal into the EHR via computerized provider order entry requests, as preferrable to a “data push,” where CGM data automatically transmit to the EHR without a specific request. For a successful data pull and linkage, enough patient identifiers must match between the CGM system portal and EHR.

iCoDE recommends adoption of the Institute of Electrical and Electronics Engineers (IEEE) P1752 family of standards for mobile health data (82). These include standards for time stamps, which must include the date and time as well as the time zone. Time stamps are critical for data collection by the CGM sensor, for transmission to and from the manufacturer cloud, when data are transmitted from the manufacturer cloud to the CGM system portal and eventually when data are transmitted to the patient EHR. A standardized date format is recommended, according to the ISO 8601 standard (83).

The following reporting elements should be provided by the CGM system, as per the 2019 international consensus recommendations on CGM metrics for clinical care (84):

• No. of days sensor has been worn

• Percentage of time CGM system was active

• Mean glucose

• Glucose management indicator, if possible (requires minimum of 10 days’ sensor use)

• Glycemic variability (% coefficient of variation)

• Time with glucose above target range as percentage of readings and time >13.9 mmol/L (>250 mg/dL)

• Time with glucose above target range as percentage of readings and time 10.1–13.9 mmol/L (180–250 mg/dL)

• Time with glucose in target range as percentage of readings and time 3.9–10.0 mmol/L (70–180 mg/dL)

• Time with glucose below target range as percentage of readings and time 3.0–3.8 mmol/L (54–69 mg/dL)

• Time with glucose below target range as percentage of readings and time <3.0 mmol/L (<54 mg/dL)

• Ambulatory glucose profile CGM system report, with attention to days in and out of hospital and when different therapies were given.

In addition to the elements above, the ability to pull discrete time-stamped CGM measurement values into the EHR is an important consideration for hospital CGM use. Glucose telemetry systems have enabled real-time data transfer from individuals’ CGM systems to paired devices in the hospital, which is a more practical way for clinical staff to access data (14,20). One study highlighted that the Dexcom G4 CGM system could be paired to a receiver that would send glucose results to a bedside phone securely affixed to the table, with the results then sent by Wi-Fi to a nursing station tablet (16). Alternatively, smart phones or tablets can be fixed to the wall or cart outside patients’ rooms with a similar purpose (52). Of note, to date no CGM devices have been directly interfaced with an EHR.

• 1) Hospital information services, laboratory, and clinical stakeholders must work together to determine how CGM data will be integrated with the patient EHR.

• 2) Clinical access to CGM data in real time for decision-making is an important consideration. Clinical stakeholders must work with local information services groups to develop a process and advocate for the required infrastructure.

• 3) The iCoDE recommendations offer a guideline for CGM-EHR data integration.

Few studies have been focused on cost implications of CGM use in hospital settings. However, in one randomized controlled trial in a medical-surgical ICU, in the Netherlands, investigators did assess cost (85). Individuals were randomized to either a control group where CGM system data were not available to nurses or an intervention group where CGM data were visible for use by nurses for treatment decisions. Several costs were assessed as part of this trial, including nursing time, devices, consumables, and laboratory costs. Nursing time was divided into several categories, including the time it takes to perform a POC measurement, insert a sensor, calibrate a sensor, or read a CGM value and record it into the insulin dosing algorithm. The time to perform these tasks per 24-h period was determined as the average time it took to perform that task on 10 occasions across the trial period. A comparison of the nursing time required per ICU subject per day revealed that individuals equipped with CGM systems require only mean ± SD 17 ± 12 min vs. 36 ± 24 min of nursing time for individuals whose insulin was dosed based on POC glucose results. Furthermore, the investigators of this study concluded that the total daily cost, with consideration of nursing time, consumables, devices, and laboratory testing, was significantly different for the control group (decisions based on POC measurements: €53/day) versus the intervention group (decisions based on CGM data: €41/day).

Despite the emergency use authorizations granted during the COVID-19 pandemic, the lack of formal regulatory authorization for CGM use in hospital remains an obstacle to its implementation 86. In many institutions, insufficiency of no. of trained staff to take on the increased workload required for initiation and monitoring of new devices is also a barrier to CGM system implementation (15). Lack of knowledge among staff on CGM data interpretation and the use of cloud-based applications for sharing and monitoring CGM system data, especially hospitalists and nurses who are not diabetes experts, remains a challenge. Preliminary studies have suggested that successful implementation of CGM programs in hospitals requires an inpatient diabetes management team in place to help oversee the program and provide education to frontline staff. This may not be practical at smaller, nonacademic hospital sites (15,53). A technical support team, including the vendor, hospital IT, and the laboratory, may be required for troubleshooting issues with CGM systems (87) as they can be difficult for clinical teams to troubleshoot (88).

Staff education on acceptable agreement between CGM and capillary POC blood glucose measurements is important, especially for CGM systems that require calibration (89). Studies have demonstrated that some health care professionals prefer to use POCT over CGM systems because they do not trust CGM measurements and are more comfortable with POCT (87,90,91).

Wireless network functionality at an institution can operate as a barrier, as wireless technology is necessary for transmission of data from CGM systems (89). Unstable signals can result in data transmission errors (87). Hospital network security firewalls can also pose challenges if data from devices are sent to an intermediary cloud-based platform as part of the connectivity solution (90). Often, health care professionals must log in to a third-party software system to download and upload the required data to the EHR, which can be a challenge from an IT data privacy and security perspective (92). It can also be difficult for staff to troubleshoot third-party systems when required.

Lastly, there are practical limitations during inpatient care regarding CGM sensors themselves. Hospitals do not necessarily stock replacement sensors for patient-owned devices. Devices must also be removed during some imaging studies (87); however, there are some preliminary data demonstrating that removal of the devices before imaging procedures may not be necessary for regular X-rays, CT, and angiography (93).

• 1) Development of specific guidance around accessing CGM data in hospital settings for decision-making and guidance for documentation of CGM results in the EHR.

• 2) Extensive and detailed CGM interference studies to provide guidance on use of CGM in hospital settings for certain patient populations.

• 3) Reliability of CGM devices during specific surgical and other invasive procedures (e.g., bypass surgery).

• 4) Reliability of CGM devices during specific imaging procedures. Currently, devices must be removed during MRI, CT scan, nuclear medicine imaging, and bone density scan procedures and require lead shielding for X-ray, fluoroscopy, and electrophysiology procedures.

• 5) Reliability of CGM measurements in clinical situations or treatments where glucose concentrations are changing rapidly.

• 6) Procedures that are higher risk for loss of CGM signal, for risk mitigation (e.g., defibrillator use)

• 7) Sensor placement locations best suited for hospitalized patients to avoid compression loss-of-signal or hypoperfusion effects.

• 8) Performance of CGM in neonatal patients for hypoglycemia monitoring, given the challenges related to reliability of POC glucose measurements in this patient population.

• 9) Cost analysis studies for CGM that include comparison with POCT and laboratory and nursing costs and also include consideration of patient outcomes, such as surgical site infection rate, length of stay, and readmission rates.

The body of data from clinical studies demonstrating the value of CGM system usage for hospitals is growing. Clear guidance on the use of CGM systems in hospitals is needed. If CGM systems are used beyond a patient’s self-care within the intended use, appropriate steps needed for clinical and analytical verification, with potential studies performed prior to implementation, have to be defined.

Hospitals looking to implement CGM systems must ensure that all appropriate stakeholders are engaged and that policies and protocols are in place to ensure safe and effective use. Training of clinical staff on the use of CGM systems as well as their limitations is key to successful implementation. Key components for hospitals to address prior to implementation, with a clear plan, are timely responses to CGM alerts and integration of CGM data into the EHR.

Acknowledgments. The authors thank the International Federation of Clinical Chemistry and Laboratory Medicine Working Group on Continuous Glucose Monitoring (IFCC WG-CGM) for their contribution to reviewing this article.

Duality of Interest. G.F. is general manager and medical director of the Institute for Diabetes Technology (Institut für Diabetes-Technologie [IfDT]), which carries out clinical studies, e.g., with medical devices for diabetes therapy, on its own initiative and on behalf of various companies. G.F./IfDT have received speakers’ honoraria or consulting fees in the last 3 years from Abbott, Berlin-Chemie, BOYDSense, Dexcom, Lilly Deutschland, Novo Nordisk, Perfood, PharmaSens, Roche, Sinocare, Terumo, and Ypsomed. G.F. chairs the IFCC WG-CGM. S.P. is an employee of IfDT and is a member of the IFCC WG-CGM. No other potential conflicts of interest relevant to this article were reported.

Handling Editors. The journal editor responsible for overseeing the review of the manuscript was Steven E. Kahn.