Consensus Statement on Use of Continuous Subcutaneous Insulin Infusion Therapy in the Hospital

Abstract

In May 2017, the Diabetes Technology Society convened a panel of US experts on inpatient diabetes management to discuss the current and potential role of continuous subcutaneous insulin infusion (CSII) therapy in the hospital. The panel (1) discussed evidence for current use of CSII in the hospital, (2) recommended contraindications for use in the hospital, and (3) recommended guidelines to maximize chances for safe use of CSII in the hospital. Panel members agreed that larger, prospective, randomized studies are needed to evaluate safety and efficacy of CSII use in the hospital. As CSII technology becomes more complex and its use increases, it is imperative that institutional protocols be in place to ensure safe use of this technology and safe transitions across care areas. Providers need to be cognizant that not all patients currently using CSII as an outpatient are appropriate candidates for continued use in the hospital. This consensus statement provides guidelines for practitioners who may encounter patients using this technology in the inpatient setting.

In May 2017, the Diabetes Technology Society convened a panel of US experts in endocrinology in Atlanta, Georgia to discuss the current and future use of continuous subcutaneous insulin infusion (CSII) in the inpatient setting. Panelists reviewed the current evidence for CSII use in the hospital, identified knowledge gaps, made recommendations to ensure safe inpatient use of CSII, and discussed future concerns regarding CSII use in the inpatient setting. Three cochairs, Dr. Bithika Thompson, Dr. Curtiss Cook, and Dr. Mary Korytkowski, each served as a moderator for discussion of (1) current evidence for inpatient use of CSII, (2) recommended contraindications for inpatient use of CSII, and (3) best practices to ensure safe transitions of CSII during an inpatient stay. The focus of the discussion was to review the available evidence for CSII use, discuss potential benefits as well as contraindications for continued CSII use in the inpatient setting, propose guidelines for use of CSII in the hospital, and identify potential considerations for future use of this technology. The scope of these guidelines pertain to adult inpatients only, and do not include either perioperative or labor and delivery settings. We recommend that separate guidelines be considered for these scenarios.

Background

Continuous subcutaneous insulin infusion (CSII) therapy, or insulin pump therapy, is being used with increasing frequency for treating patients with insulin requiring diabetes mellitus. Currently, CSII therapy is FDA approved for use in the outpatient setting in people with both type 1 and type 2 diabetes. According to the latest market analysis, over 5 million people in the US are estimated to be using insulin pumps.¹ Recent data from the Type 1 Diabetes Exchange reported that of the 16,000 participants surveyed, 60% were using an insulin pump in 2014. Insulin pump use among people with Type 2 diabetes is also increasing.² Currently, it is not known how many patients on CSII therapy require hospitalization; however, as use of CSII increases, health care professionals can anticipate encountering patients who use this technology in the inpatient setting.

The currently available insulin pump devices approved for use in the U.S. are presented in Table 1. Insulin pumps are programmed to infuse a rapid (lispro, aspart, glulisine) or short-acting (regular) insulin preparation at a continuous (basal) rate with allowance for intermittent boluses that are triggered by patient instructions to the pump before meals and snacks (prandial dose), or for blood glucose excursions outside the desired or target range (correction dose). Older models of insulin pumps have a set basal rate (amount of insulin to be infused each hour), and individualized insulin to carbohydrate ratio and sensitivity factors programmed into the device. Insulin pumps are often now used in conjunction with a continuous glucose monitoring system (CGMS) that provide information regarding real-time blood glucose data to the patient. Some CGMS are linked wirelessly with the insulin pump. A recent consensus conference³ discussed use of CGMS in the hospital and concluded that further studies are needed to demonstrate evidence for safety or benefit of these devices in that setting. The most recent insulin pump and CGMS system is the MiniMed 670G, the only hybrid system that is currently commercially available at the time of this writing.⁴ This system differs from others in that when operating in “Auto” mode there is no set basal rate. Instead the system relies on continuous glucose data to auto-adjust the basal rate settings. Patients still administer bolus insulin according to individualized insulin to carbohydrate ratios and preset correction doses. As the complexity of insulin pump technology increases, there are new challenges for health care professionals who are faced with managing CSII therapy in hospitalized patients. It is imperative that health care professionals are familiar with this technology and that proper protocols are in place to ensure successful and safe management of these patients in the hospital.

Table 1.

Insulin Pumps Available in the United States.

Brand	Models	Bolus calculator	CGM integration	Insulin suspend for hypoglycemia	Special features
Medtronic	1. 670G 2. 630G 3. 530G 4. Paradigm 5. Revel	Yes	1. Yes—Guardian 2. Yes—Enlite 3. Yes—Enlite 4. No 5. No	1. Yes 2. Yes 3. Yes 4. No 5. No	670G only hybrid closed-loop system on the market when operating in Auto mode
Tandem	1. T:slim 2. T:slim G4 3. T:slim X2 4. T:flex	Yes	1. No 2. Yes—Dexcom G4 3. Yes—Dexcom G5 4.  No	No	Touchscreen T:flex allows for larger volume insulin capacity
Insulet	Omnipod	Yes	No	No	Only detached “tubeless” pump on the market
Roche	Accu-Chek Spirit	No	No	No	Not applicable
Animas	1. OneTouch Ping 2. Animas Vibe	Yes	1. No 2. Yes	No	No longer selling pumps in the US and Canada
Sooil	Dana Diabecare IIS	Yes	No	No	Menu uses icons instead of words

What Is the Evidence for Use of CSII in the Inpatient Setting?

Moderator: Bithika Thompson, MD

The percentage of patients with diabetes who are admitted to the hospital using CSII is unknown. The majority of these patients are likely managing their diabetes well in the outpatient setting and are hospitalized for reasons other than diabetes. One report of 253 patients using inpatient CSII therapy indicated that these patients had a reasonable degree of outpatient glycemic control with a mean hemoglobin A1c (HbA1c) of 7.5%.⁵ Reports by others have found similar data.⁶ Thus, given that patients are self-managing their diabetes effectively with CSII in the outpatient setting, it is reasonable to assume that they would wish to continue to maintain their current level of glycemic control with CSII therapy while hospitalized. Patients often request to continue use of CSII in the hospital and report high patient satisfaction (86%) when allowed to do so.⁶

Asking a patient to remove their insulin pump device in the hospital could potentially contribute to decreased patient satisfaction given the perceived loss of control over self-management that is likely to occur. Patients may become frustrated if asked to transfer their diabetes management to hospital staff who may not be as knowledgeable about diabetes. However, some patients who are both competent and capable in managing CSII therapy in the outpatient setting may not be aware of potential situations in the hospital that can adversely impact their glycemic management and insulin requirements. Acute illness, infections, steroid therapy and other medications, changes in kidney function, and abrupt dietary changes can all result in variability in insulin sensitivity over time.

Inpatient diabetes self-management using CSII is a process by which patients, in collaboration with hospital staff are allowed to remain on insulin pump therapy. Table 2 summarizes studies of inpatient CSII use in patients already using this technology as an outpatient. Although data on safety and efficacy of these devices in the inpatient setting is limited, these studies suggest that with proper patient selection and standardized protocols, select patients can successfully transition CSII therapy from the outpatient to inpatient setting.^5-9

Table 2.

Studies of CSII in the Hospital Already on CSII as an Outpatient.

Study	Diabetes type	Population	Sample size	Type of study	Outcome measurement
Cook et al5	Both	General inpatient	136 patients/253 hospitalizations	Retrospective review	Protocol adherence Safety Mean bedside CBG Frequency of hyperglycemic and hypoglycemic events
Noschese et al6	Both	General inpatient	50 patients	Retrospective review	Mean bedside CBG Hypoglycemia frequency Safety Patient satisfaction
Bailon et al7	Both	General inpatient	35 patients/50 hospitalizations	Retrospective review	Protocol adherence Mean bedside CBG Hypoglycemic events Safety
Leonhardi et al8	Both	General inpatient	21 patients/25 hospitalizations	Retrospective review	Protocol adherence Safety
Nassar et al9	Both	General inpatient	65 patients/125 hospitalizations	Retrospective review	Protocol adherence Mean bedside CBG Safety
Kannan et al10	Both	General inpatient	50 patients/51 hospitalizations	Retrospective review	Mean bedside CBG Frequency of hyperglycemic and hypoglycemic events Safety
Faulds et al11	Both	General inpatient	2 patients	Case report	Safety
Cook12	Type 1	General inpatient	2 patients	Case report	Institutional guidelines

CBG: Capillary blood glucose.

Cook and colleagues reported their institutional experience with transitioning patients on CSII from the outpatient to inpatient setting in sequential publications.^5,7-9 In the largest series to date reporting one hospital’s experience with 253 hospitalizations among 136 patients between 2006 and 2011, three different cohorts were identified. This included “pump on” patients who continued CSII therapy throughout the entire hospital stay (65% of hospitalizations), and “pump off” patients (15%) who did not meet criteria to remain on inpatient CSII therapy as described below. These patients were transitioned to alternate therapy with basal bolus insulin therapy on admission. In the “pump off” subset, CSII therapy was not restarted at time of discharge as these patients were not deemed to be good candidates for continued use. The third subset, “intermittent pump” cases (20%), was comprised of patients who may have stopped and restarted CSII during their hospital course as clinical circumstances changed. There was no difference in mean point-of-care glucose values between the 3 CSII patient subsets (P > .1). However, the percentage of hyperglycemic measurements was significantly lower in “pump on” versus “pump off” cases, and in “pump on” versus “intermittent pump cases” (P < .02). In addition, “pump on” patients had significantly fewer glucose measurements <40 mg/dl compared to “pump off” patients with no difference compared to “intermittent pump cases.” The authors reported that adherence to recommended documentation procedures as described in their protocol for insulin pump therapy improved over time (eg, completion of an endocrinology consult and patient agreement forms, placement of necessary orders).

Kannan et al reported on 51 patients previously using CSII who were admitted to a general inpatient unit.¹⁰ The purpose of this study was to investigate the relationship between patient knowledge regarding insulin pump therapy and glucose control during the inpatient period. Three groups of patients were identified: those who did not require any inpatient diabetes education and continued CSII (group A), those who received diabetes education once knowledge gaps were identified and continued CSII (group B), and those for whom continued CSII use was deemed inappropriate and were transitioned to alternate therapy (group C). There were no statistically significant differences in mean blood glucose or frequency of hyperglycemic and hypoglycemic events between groups A and B. There were no reported adverse events related to continued CSII use. These results suggest that even in patients with identifiable knowledge gaps, continuation of CSII can safely be allowed in the inpatient setting if education is provided.

Noschese and colleagues examined glycemic control in 50 CSII treated patients admitted to a general medicine ward between 2004 and 2006.⁶ The goal of this study was to determine safety and efficacy of CSII therapy following initiation of an inpatient insulin pump protocol (IIPP). Patients were separated into three groups: IIPP with an inpatient endocrine consultation (group 1), IIPP alone (group 2), and usual care (group 3). There were no differences in mean capillary blood glucose (CBG) or frequency of hypoglycemia (CBG <70 mg/dL) between the three groups although there were more patient days with CBG values >300 mg/dL in group 3 in which no components of the protocol were followed (P = .2). The types of insulin pumps used by patients in these earlier reports varied. Because the frequency with which nursing personnel encountered these was infrequent, it was felt that patients needed to be assessed for their ability to manage their own pump therapy and provide their own pump supplies to be considered candidates for continued use in the hospital.

In general, this review of the current literature indicates that CSII therapy can be safely continued in the inpatient setting in selected patients. Based on published experience, glucose control on CSII in the hospital is not inferior to subcutaneous insulin injections. Insulin pump malfunctions in the hospital while uncommon, have been reported. Cook and colleagues observed that there were no pump site infections, mechanical pump failures, or episodes of diabetic ketoacidosis (DKA). In their entire case series, there was only one adverse event of a kinked infusion catheter which resulted in hyperglycemia that was recognized early and corrected.⁵ Noschese reported only 2 minor events (1 pump malfunction and 1 catheter infusion site problem) in 50 consecutive patients treated with CSII in the hospital.⁶ Two cases of ‘runaway pump” phenomenon have been reported in which patients received an unsolicited bolus of insulin. In both of these cases, the patients developed severe hypoglycemia.¹¹

There is limited data for safety and efficacy regarding initiation of CSII therapy in hospitalized patients who were not using these devices as outpatients (Table 3). In one randomized controlled study of 60 patients admitted to the MICU, Bodur et al compared CSII to continuous intravenous insulin infusion (CIII) for obtaining glycemic control with a target of 74-110 mg/dL in patients admitted with hyperglycemia.¹³ Insulin doses were titrated by medical staff based on frequent monitoring of blood glucose values. Mean blood glucose was lower in the CSII group (P = .00055). There were no differences in hypoglycemic events or mortality between the two groups and no adverse events were reported in the CSII group. The authors concluded that strict glucose control was achieved more efficiently in the CSII group without an increased risk of hypoglycemia.

Table 3.

Studies of Patients Started on CSII in the Hospital.

Study	Diabetes type	Population	Pump vendor	Sample size	Type of study	Outcome measurement
Bodur et al13	Both	MICU	Deltec Cozmo	60 patients; randomized to CSII or IV insulin infusion	Randomized Controlled Study	Mean blood glucose level Hypoglycemic events
Lee et al14	Type 2	General inpatient	Medtronic	27 patients- (8 CSII, 10 CIII, 9 conventional)	Prospective randomized study	Mean glucose levels Fasting glucose levels Safety Patient satisfaction
Kannampilly15	Type 2	General inpatient	Medtronic	4 patients	Case Series	Fasting and postprandial CBG A1c Safety
Levitt et al16	Type 2	General inpatient	OmniPod	Total 16 patients in 3 groups according to CSII use and CGMS assignment	Randomized prospective; subjects in two groups blinded to CGMS data	Mean capillary glucose Hypoglycemia frequency Time in target.

MICU: Medical intensive care unit. CIII: Continuous intravenous insulin infusion. CSII: Continuous Subcutaneous insulin infusion. CBG: Capillary blood glucose. Continuous glucose monitoring system.

Lee et al randomized 27 hospitalized patients with type 2 diabetes to use of CSII, CIII, or subcutaneous basal-bolus insulin therapy.¹⁴ The goal of the study was to examine the effect of instituting inpatient CSII on glycemic control in the hospital. Fasting glucose levels on day 2, 3, and 4 were all lower in the CSII group whereas no consistent significant change was seen in the CIII or conventional therapy group. No symptomatic hypoglycemia was observed in the CSII group. Patient satisfaction scores were much higher with CSII than CIII or conventional treatment (P < 0.001). A separate “psychological well-being” score indicated that patients using CSII felt better about themselves and their treatment method (P = .008).

Kannampilly presented a case series of four patients with poorly controlled type 2 diabetes and chronic cirrhosis admitted to the general inpatient ward.¹⁵ All 4 patients were poorly controlled at baseline on conventional insulin therapy. The goal of the study was to retrospectively examine the glycemic effect of instituting CSII therapy in the inpatient setting and then allowing for continued outpatient use of this technology at discharge. In the inpatient setting, the authors observed immediate and significant improvements in fasting and postprandial CBG, despite the fact that 3 of 4 patients were able to decrease their total daily insulin doses. There were again no reported adverse events with use of CSII in the inpatient setting

A recent study by Levitt et al examined initiation of insulin pump therapy combined with CGMS in hospitalized type 2 DM patients who were new to CSII therapy. Subjects were randomized to one of 3 groups: group 1 received basal bolus insulin therapy + blinded CGMS (n = 5); group 2 received insulin pump + blinded CGMS (n = 6), and group 3 received insulin pump + nonblinded CGMS (n = 5). Blinded CGMS referred to whether subjects had access to the real time glucose data. Use of CSII did not offer any superior glucose control, however, samples sizes were small. In addition, as the authors noted, initiation of CSII in the hospital setting was labor intensive.¹⁶

One limitation of the above reports is that different CSII vendors were employed (Table 3). In addition, it was not clear who assumed the primary responsibility for management of the technology once applied. These differences make comparisons about efficacy difficult to interpret.

Consensus Reached by the Panelists

Larger, prospective studies of CSII use in the hospital are needed to guide recommendations. There is currently no large prospective, randomized study investigating the safety and efficacy of inpatient use of CSII and it is not clear if this will occur in the future. Most studies reporting on CSII use in the inpatient setting have focused on adherence to institutional protocols and documenting safety of continued use of these devices in the hospital among patients already using these devices at time of admission. There is very little data comparing CSII to conventional basal-bolus insulin therapy in hospitalized patients. Although CSII use has been demonstrated to be noninferior to standard therapy with regard to glycemic control, examination of other outcomes (eg, length of hospital stay) is lacking. Panelists agreed that randomized controlled trials would be challenging to design because of the inherent difficulty of blinding caregivers as to which subjects receive CSII compared to those who do not. Studies to date have been single-site rather than multicenter. In addition, funding opportunities for future studies may be limited given that continued CSII use in the inpatient setting has become standard of care at many institutions, primarily among patients who are using these devices prior to admission.

Panelists agreed that patients successfully using this technology in the outpatient setting should be allowed to continue use of CSII as an inpatient provided they have been evaluated as being competent to do so and no contraindications (see below) to continued use are identified. There is evidence that patients who are allowed to continue CSII in the hospital setting feel more involved in their care, and as a result report higher levels of satisfaction with their care. Panelists also pointed out that evaluation of pump use in the inpatient setting is a unique opportunity for the physician to examine current settings for basal rates, insulin to carbohydrate ratios, and correction factors on a day to day basis in a structured setting. This is an opportunity for practitioners to make adjustments to CSII settings which can benefit patients as they transition back to the outpatient setting.

It is imperative that patients and practitioners understand that the hospital environment differs from the outpatient setting and that the patient’s clinical course can change rapidly during periods of acute illness. The studies reviewed here suggest that on initial assessment there are patients for whom continued CSII use is not appropriate based on clinical status or lack of patient ability to demonstrate competent use of this technology. Patients in the hospital need to be reassessed on a regular basis to determine if continued CSII use remains appropriate.

Development and implementation of institutional protocols for inpatient use of CSII can help ensure compliance with desirable behaviors (eg, obtaining an endocrinology consult, documenting patient agreement forms, entering necessary orders). Insulin pump use in the hospital can cause confusion among hospital staff who are unfamiliar with this technology; increasing risk for medication errors and adverse glycemic events. Initiating established protocols upon admission ensures that hospital personnel are aware of patient CSII use and decreases the potential for adverse patient events and outcomes. The available literature suggests that with supervision by a diabetes or endocrinology team or practitioner familiar with CSII use, patients can be continued on this technology or transitioned on and off this technology in a safe manner.

Data on safety and efficacy of initiating CSII therapy in patients previously naïve to this form of insulin delivery is limited. While studies suggest there may be some glycemic benefit to use of CSII in this group of patients, these findings need to be explored in larger, multicenter, prospective trials as there is currently insufficient evidence to support this practice. It is important to point out that in this population of patients, CSII is being administered and monitored by hospital staff who also perform all dose titrations. Panelists point out that to initiate CSII therapy in patients not familiar with its use will involve having a sufficient number of adequately trained nursing and medical personnel. In addition, CSII is an expensive and intensive therapy. Panelists agreed that extensive clinical data supporting its effectiveness would need to be present to justify the cost of initiating such therapy.

What are Recommended Contraindications to Use of CSII in the Inpatient Setting?

Moderator: Curtiss B. Cook, MD

Consensus Reached by the Panelists

Panelists agreed that not all patients currently using CSII in the outpatient setting are good candidates for continued use in the hospital. Upon admission, patients should be assessed for their ability to safely continue CSII. Table 4 summarizes recommended contraindications for CSII self-management in the hospital. Special consideration should be given to the reason for hospitalization (eg, monitoring, surgery, treatment), the acuity of the condition for which the patient is hospitalized, as well as the willingness and ability of the patient to operate the pump during the hospitalization.

Table 4.

Recommended Contraindications for CSII Self-Management in the Hospital.

Patient with altered state of consciousness or cognitive impairment

Metabolically unstable patient (diabetic ketoacidosis [DKA]/hyperglycemic hyperosmolar state [HHS] at admission)

Critically ill patient requiring intensive care Patients who are unable to use their hands or arms due to injury or trauma

Patients with suicidal ideation

Patients unwilling to comply with hospital policies of CSII self-management

Patients unable to demonstrate competency with using CSII

Patients without CSII supplies

Patient preference

Physician judgement

CSII use should be discontinued in patients who are critically ill or metabolically unstable (including those admitted with DKA or hyperglycemic hyperosmolar state [HHS]). In this subset of patients, CIII should be initiated according to established protocols and titrated to achieve adequate glycemic control and stabilization of metabolic abnormalities. Once stabilized, these patients can be transitioned back to CSII in the hospital if deemed appropriate. Patients on outpatient CSII therapy admitted with DKA or HHS should have their insulin pump examined prior to resuming this therapy to ensure that pump malfunction was not the cause of the DKA/HHS. If there is any concern that the insulin pump may be malfunctioning, patients should be transitioned to basal bolus therapy in the hospital and at discharge, with follow-up plans established for ordering a new device if the patient wishes to resume CSII in the outpatient setting.

In specific cases, practitioners may proactively consider transitioning a patient to basal-bolus insulin therapy. This may be reasonable to do in the setting of upcoming major surgical procedures. In this situation, practitioners will want to consider the length of the surgery, possible dietary changes following surgery, and need for narcotic medications postoperatively which may impair a patient’s ability to safely operate the insulin pump.

A frequently encountered situation in the hospital setting is the initiation of high doses of glucocorticoid therapy which are known to cause significant elevations in blood glucose levels in the majority of patients.^17,18 In these cases, insulin management can become very challenging because of highly variable glucose levels in the setting of an increase in insulin resistance. Because patients are often not equipped to adjust insulin aggressively, it may be reasonable to transfer the insulin dosing and administration from the patient to the inpatient practitioners.

Insulin pumps must be removed for various radiographic studies, such as MRI procedures. Temporary suspension and disconnection of the pump for short radiographic procedures may be warranted if the device is going to be in the field of ionizing radiation. A process should be in place for transitions that occur within the hospital (eg, for procedures; see below). Practitioners need to be proactive in determining when a pump should be removed and how long it is expected to stay off. It may be appropriate to transition the patient to CIII or subcutaneous injections of insulin therapy temporarily to ensure patient safety and avoid untoward glycemic excursions if it is anticipated that the pump will remain off for an extended period of time.

Panelists agreed that patients should be assessed for competency in operating the pump in some manner upon admission. There is no consensus on the method for assessing patient CSII competency in the inpatient environment nor any agreement on the key competencies that should be tested while in the hospital. Many panelists indicated that some method of assessing competency was built into their institutional protocols and assessed on admission when writing initial pump orders. Kannan et al reported that at their institution patients are assessed for their ability to demonstrate important pump skills.¹⁰ These skills include the ability to suspend the pump, use the built-in pump calculator, program a temporary basal rate, and deliver a correction bolus. Patients are also questioned regarding knowledge on hypoglycemia correction, carbohydrate counting, and use of back up insulin in case of pump malfunction. While many institutions may have some system in which they measure competency, there is no evidence to suggest which competencies best predict a patient’s ability to effectively and safely use the insulin pump in the hospital. Panelists recommended that at a minimum patients be assessed for their ability to suspend insulin delivery from the insulin pump, and that patients demonstrate how they would navigate the pump menu to administer an insulin bolus. In doing this, practitioners have the opportunity to observe whether the patient has the dexterity necessary to manipulate the pump which is important for patient self-management in the hospital. Further studies are needed to identify key competencies that predict safe and effective patient use of this technology in the hospital setting.

Patients who are not capable of operating the pump safely in the hospital or unwilling to abide by institutional pump policies should be transitioned to basal bolus insulin therapy. This includes patients with altered levels of consciousness or with suicidal ideation. Because patients are being asked to self-manage insulin pump therapy in the hospital and are often more knowledgeable about insulin pump use than the nursing staff caring for them, successful inpatient use of CSII demands a conscious, involved patient to operate the technology. Removing a pump from an angry or unwilling patient can be difficult for the provider and distressing to the patient. Care should be taken to explain to the patient that this is being done in the patient’s best interest and that the pump can be restarted when the clinical course stabilizes. In difficult cases, the physician should not hesitate to involve risk management or other members of the hospital team to assist with making these decisions to transition off CSII which are in the best interest of both the patient and the institution.

Some patients may choose to voluntarily transition to basal-bolus therapy because the thought of self-managing their insulin pump during an acute illness can be overwhelming. In other cases, patients must be transitioned to basal-bolus therapy because they do not have the necessary pump supplies with them in the hospital to allow continued use. Given the number of available insulin pumps, hospitals cannot be expected to maintain supplies of all the specific insertion devices, catheters, and tubing specific for each of these devices.

A special situation occurs in the setting of a patient admitted to the hospital who uses an insulin pump that is managed by another person in the family. This scenario is best dealt with on an individualized basis. Panelists agreed that the insulin pump can be continued safely if the caregiver commits to staying in the hospital 24 hours a day with the patient. If this is not possible, the patient should be transitioned to alternate therapy.

In summary, it is important to assess every patient using CSII who is admitted to the hospital to determine candidacy for continued use in the inpatient setting. In many cases insulin pumps should be removed for patient safety. As the clinical course changes, this decision should be constantly reassessed to determine the need to remove the pump or the ability to restart the pump. Providers must work closely with patients to decide on the best course of action.

What Can Be Done to Ensure Safe Transitions of Outpatient CSII to the Inpatient Setting?

Moderator: Mary Korytkowski, MD

Consensus Reached by the Panelists

The literature suggests that development of institutional protocols for CSII use in the inpatient setting can increase compliance with desirable behaviors as well as ensure safe transition of insulin pumps across care areas in the hospital. It is important to recognize that the patient is usually far more knowledgeable regarding this technology than inpatient providers. In a busy hospital setting with acutely ill patients and rotating medical staff it is a very common occurrence that the nurse and primary medical team caring for the patient may never have encountered a patient on CSII previously. A successful outpatient to inpatient transition of CSII therapy requires, above all, a collaborative relationship between the patient and hospital staff.

Panelists agreed that all institutions should have an inpatient policy for CSII use.^12,19 Ideally, initiation of this policy should begin with selection of a computerized order set which then propagates other essential components of the protocol. A computerized order entry system which dictates directions to nursing and propagates required forms and order sets that direct entry of basal rates, insulin to carbohydrate ratios, sensitivity factors, and active insulin times can help educate staff and ensure safe use of this technology.

Panelists agreed that all patients admitted with an insulin pump should have consultation by an endocrinology service or by a certified diabetes nurse educator (CDNE) who is knowledgeable and comfortable with insulin pump therapy. With a computerized order system, selecting a medication order for an insulin pump would automatically generate an option for consulting endocrinology or the CDNE. In institutions that are not using computerized entry or do not have these order sets in place, education should be provided to admitting physicians and nurses regarding the need for endocrine consultation in this subset of patients. As insulin pump technology becomes more complicated; specifically with the recent release of the first hybrid closed loop system, it becomes imperative that medical personnel familiar with the complexities of this technology be consulted to assess the patient and the device and determine if continued use is appropriate.

Panelists discussed the optimal protocol for patients admitted to small hospitals that do not have access to endocrinologists or CDNE. Suggestions for these institutions included use of a “virtual consult” in which the provider could remotely consult with an endocrinologist at another institution, transitioning from CSII to conventional basal bolus insulin therapy, or transfer of these patients to larger hospitals with endocrinologists on staff. The best course of action in this scenario is not known. Further investigation into how best to manage these patients in smaller, remote hospital systems is necessary to ensure patient safety.

Once the endocrinology service or CDNE is consulted, the endocrinologist, nurse practitioner, CDNE, or other trained personnel should assess the patient and ensure that continued CSII use in the inpatient setting is appropriate. Patients should be tested for competency in using the insulin pump with consideration of contraindications for continued CSII use (see earlier discussion for suggested components of the competency assessment). If continued use is considered appropriate and the patient demonstrates competency, the institutional pump policy should be discussed with the patient in detail. If the patient is willing to comply with the pump policy, the patient is asked to sign an agreement form that becomes part of their medical record.

Table 5 summarizes the recommended components of a CSII patient agreement form. Patients must allow hospital staff to inspect the pump and review pump settings including basal rates, carbohydrate ratios, sensitivity factors, active insulin times, and blood glucose targets. Patients must agree to change their infusion site every 48-72 hours or as needed. Most institutions require that patients obtain insulin from the inpatient pharmacy when the pump needs to be refilled. While hospitalized, it is felt that using insulin supplied from the hospital is safer than using insulin brought in from home. Patients must provide their own insulin pump supplies (infusion sets and reservoirs) as these are not stored in hospital pharmacies. Although patients are self-managing their diabetes and administering boluses and corrections on their own, they must record all insulin bolus doses administered on a bedside record sheet so these amounts can be entered into the medical record by nursing staff for documentation purposes. It is preferred that personnel who are familiar with insulin pump therapy query the pump for accuracy of recorded basal and bolus insulin doses on at least a daily basis.

Table 5.

Recommended Components of a Hospital Based CSII Patient Agreement Form.

Allow staff to inspect the pump on admission to record basal rates, carbohydrate ratios, sensitivity factors, insulin on board times, blood glucose targets, and type of insulin in the medical chart

Change the infusion set every 48-72 hours or as needed

Patients provide own pump supplies

Patients record basal rates and bolus doses and provide this information to nursing and medical personnel

Patients agree to point-of-care glucose testing by hospital staff using the hospital’s glucometer

Patients report any pump problems

Understanding that a family member assisting with the operation of the insulin pump must remain in the hospital during the entire hospital stay

The pump may be discontinued and a different insulin delivery given for any of the following:

Physician order

Changes in judgment

Changes in patient’s level of awareness or consciousness

Any X-ray or other procedure (eg, MRI) that requires temporary discontinuation of CSII

Other reasons deemed necessary by medical staff

Panelists agree that use of hospital glucometers for dosing is necessary for safe use of CSII in the hospital. Because hospital glucometers must be calibrated frequently per regulation, hospital personnel can be assured that these point-of-care (POC) values are accurate. Accuracy is absolutely necessary as these values will be used to determine insulin doses when entered into the insulin pump. Many patients using CSII may also be using CGMS in conjunction with the insulin pump. In a previous consensus statement on inpatient use of CGMS in the hospital, Wallia et al³ recommended that hospital glucometer POC values be used to calibrate CGMS to ensure accuracy. This is particularly important in the case of the newest hybrid closed loop insulin pump system (Medtronic 670G) as it allows values generated by the CGMS to automatically change insulin infusion rates based on the blood glucose values. Patients must agree to use hospital glucometer POC values for insulin dosing decisions and calibration. Ensuring that the glucometers used to calibrate these systems are accurate will increase safe use of these technologies in the hospital. To date, there are no published data with use of the hybrid pump in the inpatient setting. With these considerations, and until more data is available on how patients are using this newest technology, converting the 670G from auto to manual control would be the best option if used in the hospital. Conversion to manual mode also facilitates entry of discrete basal data into the electronic medical record.

Panelists recommend that patients on CSII therapy be reassessed at least daily (and sometimes more frequently) for changes in clinical status that could affect current insulin doses. Infusion sites need to be examined on at least a daily basis for signs of inflammation or infection, and the insulin pump itself should be assessed daily to ensure there is enough insulin remaining in the pump. Pump site changes should be documented in the medical record every 48-72 hours. If the insulin pump is removed for any reason; it should be locked up in the room by nursing to ensure safe keeping of the device if there is no reliable family member available.

Special attention should be given to CSII use when transitioning care across the hospital. Inpatients are routinely taken off their usual nursing unit for radiographic studies or diagnostic procedures such as cardiac catheterization, endoscopy, colonoscopy, or bronchoscopy. For studies and procedures that require pump removal, the endocrinology team should be aware of the proposed time the pump will be off to allow a determination of a need for alternative insulin therapy. The pump should be removed and handed to the patient’s nurse so that it can be locked up in the patient room or nursing station to avoid risk of loss of the device. Ideally, a bar coded label with patient name can be affixed to the device. If the patient has been off the floor for longer than 1 hour, a blood glucose should be checked in the procedure area and the endocrinologist contacted if needed to determine whether an insulin bolus injection is required at that time. If the pump remains off for more than an hour glucose should be checked hourly and reported to the endocrinologist. In most cases, the patient can be managed with subcutaneous injections of bolus insulin until the pump can be reattached. If the pump is removed at any point during a transition of care, the nurse should be alerted so the pump can be safely secured and the endocrinologist can be contacted for further instructions.

Panelists agreed there are potential liabilities that need to be addressed at an institutional level for inpatient use of CSII that require an examination of the risks and benefits associated with either continuing or discontinuing this therapy. For institutions that lack ancillary support in the form of endocrinology services, nursing expertise, and CDNE, use of CSII may not be reasonable as there is no person with the expertise necessary to provide patient and staff support for the use of these devices. Requiring patients to sign agreement forms and demonstrate competency in using this technology as described above can help to decrease liability at an institutional level. A growing but significant concern regarding continued use of CSII devices in the inpatient setting is cybersecurity, or whether this glucose and insulin dosing data can be compromised. Hospitals may not feel safe allowing continued use of a device that has not met cybersecurity standards.

As CSII technology continues to advance, other currently unforeseen issues may arise regarding use of the devices in the hospital. For instance, the MiniMed 670G,⁴ the first hybrid system on the market, is the first integrated technology which allows for automated basal rate changes based on continuous glucose data. There are many potential concerns regarding use of these systems in the hospital, all of which need to be addressed through further discussion and data collection. Documentation of basal rates is more difficult as there is no set basal rate on these devices when operating in “Auto” mode. Instead, the basal rate is continually changing based on glucose monitoring data. A common framework is needed on how to document variable basal rates in the medical record. Patients may be switching between “Auto” and “Manual” modes which further complicates the ability to assess patient competency. In these cases, documentation that the patient is able to use this device in both modes may be necessary. As noted above, setting the device into manual mode would be the best option.

Future Directions

To date, published experience with CSII use in the hospital is limited to just a few institutions and primarily to those instances involving transitioning of the therapy from the outpatient to inpatient setting. It would be helpful if more institutions could report on their approaches to assessment and management of insulin pumps in the inpatient setting. In addition, improved data on rates and causes of inpatient stays for patients on CSII therapy would be helpful in defining the burden of hospitalizations for this subset of diabetes patients. Finally, this consensus did not focus on transitioning CSII therapy across the perioperative period, a topic which warrants additional discussion.