Assessment of nursing perceptions of three insulin protocols for blood glucose control in critically ill patients

Conflict of interest: none. ABSTRACT Objective: To evaluate nurses’ perception regarding three different blood glucose control protocols for critically ill patients. Methods: As part of a randomized trial comparing three blood glucose control protocols in critically ill patients (Computer-Assisted Insulin Protocol, Leuven Protocol, and conventional treatment), all nurses participating in the study were asked to fill in a questionnaire to assess their perceptions of efficacy, complexity, feasibility, and safety (as to the occurrence of hypoglycemic episodes), an to indicate which protocol they would like to see adopted as the standard one in the Intensive Care Unit they worked. Results: Sixty nurses answered the questionnaires. Computer-Assisted Insulin Protocol was considered the most efficient protocol to maintain blood glucose levels within the target range by 58% of the nurses, compared to 22% for Leuven Protocol (p<0.001) and 40% for conventional treatment (p=0.04). Computer-Assisted Insulin Protocol was considered easier to use than Leuven Protocol (p<0.001) and as easy as conventional treatment (p=0.78). Out of the nurses, 37% considered Computer-Assisted Insulin Protocol more feasible than Leuven Protocol and conventional treatment. A total of 51% of nurses chose Leuven Protocol as the protocol more often associated with hypoglycemia, while 27% chose Computer-Assisted Insulin Protocol and 8% conventional treatment. Finally, 56% of the nurses selected Computer-Assisted Insulin Protocol as the protocol they would like to see adopted as the standard one in the Intensive Care Unit they were based, as compared to 22% that selected Leuven Protocol and 15% that selected conventional treatment. Conclusion: Computer-Assisted Insulin Protocol was considered more efficacious, easier to use and safer than Leuven Protocol by nurses. The complexity and feasibility of Computer-Assisted Insulin Protocol were considered similar to conventional treatment. Most nurses chose of Computer-Assisted Insulin Protocol as the protocol they would like to see adopted in their Intensive Care Units.


INTRODUCTION
Tight blood glucose control with continuous intravenous insulin administration may reduce morbidity and mortality in critically ill patients (1)(2)(3) .Since the publication of the reference randomized controlled trial of Van den Berghe et al. (1) , intensive insulin therapy strategy has been recommended based on guidelines from several medical associations (4,5) .
Different results regarding efficiency and safety have been published related to the implementation of various insulin protocols in Intensive Care Units (ICUs) (1,(6)(7)(8) .The best results were found with nursedriven protocols, which quickly led to more effective blood glucose control in critically ill patients, compared to physician-led management (9,10) .Therefore, highly involved and motivated ICU nurses are essential for successful implementation of an efficient and safe protocol (10)(11)(12) .One of the most important obstacles in the implementation of a tight glucose control algorithm is the increased nursing workload imposed by frequent blood glucose determinations and insulin adjustments (11,12) .Nursing staff concerns regarding hypoglycemia risk and a lack of knowledge about normal blood glucose maintenance benefits may also hinder protocol acceptance (12,13) .Some studies were performed to evaluate the workload, time, and costs related to algorithms for tight glycemic control.Only a few studies examined the perception of nurses regarding intensive insulin therapy (12,(14)(15)(16) .However, we are not aware of any studies that have compared nurses' views on three different protocols implemented simultaneously as part of a randomized controlled trial.

OBJECTIVE
Our aim was to evaluate nurses' perception of the efficacy, complexity, feasibility, and safety (measured by the occurrence of hypoglycemic episodes), as well as which protocol they would you like to see adopted as the standard in their ICU, among three different blood glucose control protocols administered to critically ill patients as part of a prospective, randomized, controlled, multicenter trial (17) .The three insulin protocols evaluated in the randomized controlled trial were Computer-Assisted Insulin Protocol (CAIP) (17) , the Leuven Protocol (LP) (1,2) , and conventional treatment (CT).

METHODS Participants
A randomized controlled trial was conducted to compare three insulin administration protocols in intensive care patients (17) .When this study was complete, a questionnaire was given to all nurses to assess their perceptions about efficacy, complexity, feasibility, and safety (measured by the occurrence of hypoglycemic episodes), as well as which protocol they would you like to see adopted as the standard in their ICU (Supplemental Digital Content / Methods).
The study was conducted in five ICUs from different Brazilian organizations: Hospital Estadual Mário Covas, a 32-bed closed teaching ICU in a 321-bed hospital; Hospital Israelita Albert Einstein (HIAE), a 30-bed open teaching ICU in a 450-bed hospital; Hospital Municipal São José, a 8-bed, teaching closed ICU in a 200-bed hospital; Hospital Dona Helena, a 7-bed, non-teaching, closed ICU in a 120-bed hospital; and Centro Hospitalar UNIMED, an 8-bed, non-teaching, closed ICU in a 140-bed hospital.
Adult medical patients admitted to the ICU were eligible for the study if they had at least one blood glucose measurement ≥150 mg/dL plus one of the following: 1) mechanical ventilation for an acute process, with an expected duration ≥24 hours; 2) trauma; 3) burns; and/ or 4) systemic inflammatory response syndrome (SIRS, modified criteria), with at least three of the following: a) a core temperature ≥38°C or ≤36°C; b) heart rate of ≥90 beats per minute, except in patients with a medical condition or who were receiving a medication known to prevent tachycardia; c) respiratory rate ≥20 breaths per minute or a partial pressure os carbonic dioxide (PaCO 2 ) ≤32mmHg; and/or d) white blood cell count ≥12,000/mm 3 or ≤4,000/mm 3 , or >10% immature neutrophils.
Patients were excluded if they were: younger than 21 years old, surgical patients, admitted because of diabetic ketoacidosis or a non-ketotic hyperosmolar state, or in a state in which death was perceived as einstein.2012;10(3):347-53 imminent.The study protocol and consent form were approved by the ethics review board of each institution.The study was performed in accordance with ethical standards stated in the Declaration of Helsinki.Written consent was obtained from every patient or the next of kin when the patient was unable to give it.

Insulin protocols (Supplemental Digital Content/Methods)
The three insulin protocols evaluated in the randomized controlled trial were: CAIP, a protocol developed by the authors, based on continuous infusion of insulin with adjustments guided by a hand-held device or desktop software with target glucose levels between 100-130mg/dL (17) .LP is the standard strict blood glucose control protocol described in two large clinical trials (1,2) .It is a continuous insulin infusion protocol with glucose target levels between 80 and 110mg/ dL.CT involves intermittent subcutaneous insulin administration according to a sliding scale starting with levels above 150mg/dL, with glucose target levels for glucose levels between 80 and 150mg/dL.A computer-generated centralized, blinded allocation sequence, with blocks of six patients and stratified by center was performed.All nurses involved in the study were trained in all protocols before enrollment (17) .

Data collection
A waiver for informed consent was obtained from the hospital Institutional Review Board to distribute surveys to critical care nurses.
At the end of the randomized study, we sent the questionnaires to the research coordinator of each of the five centers involved in the study.Then, the questionnaires were delivered by a person not involved in the study to the nurses who were involved in the administration of all three protocols, who filled them out without the presence of the investigators at the nursing station or in the break room, during normal shift hours.To protect the subjects' anonymity, the survey did not ask for any personal identifiers (name, age, ICU, shift etc.), and nurses were instructed to complete the questionnaire without including any identifying information.This questionnaire, developed by the investigators before the randomized controlled trial began, was first tested by six critical care nurses not involved in this study, who provided feedback regarding its validity.
The questionnaire consisted of 10 multiple choice questions (Appendix 1).The first part of the questionnaire (questions A1-A4) asks about the efficacy of the glycemic control: questions A1-A3 allowed the nurses to evaluate the performance of the CAIP, LP, and CT protocols regarding their efficiency in maintaining glucose levels in the target range (% of time that blood glucose levels were under control), and question A4 asked which protocol (CAIP, LP, or CT) was the most efficient for the control of blood glucose levels.The second part (questions B1-B4) evaluated protocol complexity (time spent to execute the protocol tasks) and feasibility: questions B1-B3 asked the nurses to evaluate the complexity of CAIP, LP, and CT, and question B4 asked which protocol (CAIP, LP, or CT) was the most feasible.The third part, question C1, was about safety and asked which protocol had the most episodes of hypoglycemia (blood glucose ≤40mg/dL).The final section of the survey (question D1) asked which of the three protocols (CAIP, LP, and CT) the nurses would like to see adopted as the standard protocol in their ICUs.

Statistical analysis
Comparisons were made between CAIP versus LP and CAIP versus CT.Categorical variables were displayed as absolute and relative frequencies.Nominal variables (with more than two categories) were displayed as proportions with 95% confidence intervals (95%CI).Comparisons of proportions were made using a χ 2 test or Fisher's exact test, when appropriate.P-values <0.05 were considered statistically significant.The Statistical Package for Social Sciences (SPSS) version 18.0 (SPSS Inc ®; Chicago, IL, USA) was used for statistical analysis.

RESULTS
From August to December 2006, 60 nurses from the 5 institutions involved in the study filled out the questionnaires.Out of total of 600 questions, only 2 were left unanswered for an overall response rate of 99.7%.
Nurses' perceptions of the efficacy of protocols are presented in table 1. CAIP was considered efficient for maintaining glucose levels in the target range more than 75% of the time by 35/60 (58%) nurses.A total of 13/60 (22%) and 24/60 (40%) nurses considered LP and CT, respectively, efficient in maintaining glucose levels under control (p<0.001 for CAIP versus LP and p=0.04 for CAIP versus CT).When considering the protocols' efficiency to maintain glucose levels in the target range more than 90% of the time, 25% (15/60) of the nurses chose CAIP as the most efficient, followed Corrêa TD, Almeida FP, Cavalcanti AB, Pereira AJ, Silva E by CT with 13% (8/60) and LP with 5% (3/60) of the nurses (p<0.001 for CAIP versus LP and p=0.16 for CAIP versus CT) (Table 1).

DISCUSSION
In 2001, a large randomized controlled trial by Leuven demonstrated that the normalization of blood glucose levels using an intensive insulin infusion protocol improved clinical outcomes in patients admitted to a surgical ICU (1) .Since this publication, intensive insulin therapy strategy has been recommended by guidelines from several medical associations (4,5) .
Although nurses agree with the need for glycemic control and believe that such intervention is beneficial for patient care, they also recognize the increased work effort associated with maintaining tight glycemic control (10) .Reaching blood glucose goals requires extensive efforts from nurses, including frequent bedside capillary glucose monitoring and the implementation of a variety of insulin infusion protocols, with different degrees of complexity.Moreover, the prevalent fear of hypoglycemia among hospital staff further hinders the widespread acceptance of intensive blood glucose control protocols administered to critically ill patients (10,11) .
In the present study, nurses considered CAIP more efficacious in maintaining blood glucose levels within the target range, easier to use, and safer than the LP.In fact, nurses' perceptions corresponded to the randomized controlled trial findings in which the mean of patients' median blood glucose was 125.0 (±17.7)mg/dL in CAIP and 127.1 (±32.2)mg/dL in LP (p=0.34).In the CAIP group, 21.4% of patients had at least one episode of hypoglycemia ≤40mg/dL, as compared to 41.4% in LP (p=0.02) (17).
When CAIP was compared to CT, a statistically significant difference regarding efficacy in maintain blood glucose levels in the target range for more than 75% of the time was found.Both CAIP and CT were considered equally easy to use and feasible.Finally, the majority of nurses chose CAIP as the protocol they  In answering the question, "which protocol do you believe is the most efficient?",34/59 (58%; 95%CI: 44-70) of the nurses chose CAIP as the most efficient, 9/59 (15%; 95%CI: 7-27) chose LP, 8/59 (14%; 95%CI: 6-25) chose CT, and 8/59 (14%; 95%CI: 6-25) believed no difference existed between them.
Out of the nurses, 51% (30/59; 95%CI: 37-64) chose LP as the one most often associated with hypoglycemia Assessment of nursing perceptions of three insulin protocols einstein.2012;10(3):347-53 in the setting of a randomized controlled trial in five different ICUs.This increases the confidence that differences in nurses' evaluations are due to the protocols' properties instead of an imbalance in patient characteristics (e.g., the sickest group of patients being treated with any specific protocol).All nurses involved in the administration of the protocols filled out the questionnaires, with a very low frequency of unanswered questions.Only few studies examined nurses' perceptions as to intensive insulin therapy (10,12,(14)(15)(16) .We are not aware of any study comparing nurses' views of three different protocols implemented simultaneously.Based on our results, it may be interesting to involve the nursing team in an insulin protocol implementation process, and it also may help to support future improvement actions.

CONCLUSIONS
The CAIP was more efficacious, easier to use, and safer than the LP according to evaluation by nurses.Compared to CT, the feasibility and safety of CAIP were considered similar.Most nurses chose CAIP as the protocol they would like to see adopted as the standard algorithm in their ICU.would like to be adopted as the standard algorithm in their ICUs.We believe that CAIP was considered easier to use than LP because, in CAIP, adjustments are quick and easily completed using the computer software, while, with LP, nurses must consult a written and relatively long protocol (1,2,17) .
The reason why LP was considered less safe than the other protocols may be due to the fact that nurses observed a higher incidence of hypoglycemia with this algorithm.However, the fact that, with LP, decisions regarding insulin adjustments are often left to the discretion of the nurses may have contributed to the feeling of less safety.
Blood glucose control in intensive care should be a nurse-driven task (10)(11)(12)15,16) . Nurse-riven protocols achieve the best results in terms of glucose control and safety (9)(10)(11) .For this reason, an accurate understanding of the insulin algorithms by nurses is essential.In particular, they should consider whether the algorithm is efficacious in terms of maintaining a controlled blood glucose level, safe and easy to use.We were able to demonstrate that CAIP was the protocol that best met these criteria.
We believe that our results can be generalized for most ICUs.The survey was performed in the setting of a trial involving five mixed ICUs.Eligibility criteria for the trial were very pragmatic and included critically ill hyperglycemic patients, who are ordinarily managed with strict blood glucose control.Finally, the algorithms evaluated are in current use (LP or CT) or may be easily implemented in the ICU routine (CAIP).
Several limitations exist in this study.First, the sample size of nurses is relatively small.Nevertheless, we observed statistically significant differences for most of the comparisons that were carried out.Second, the CAIP protocol was developed by one of the clinical trial authors.However, the study was carried out in five centers, and most nurses were not aware of this fact.Nevertheless, this may have influenced nurses' answers.Although we were not able to estimate the size of this possible bias, we do believe it would not have changed the main conclusions of the study.Third, the difference in the target blood glucose range between CAIP (100-130 mg/dL), LP (80-110 mg/dL) and CT (80-150 mg/dL) must have influenced the nurses' perceptions about the efficacy, complexity, and feasibility of the protocols.We cannot estimate to what extent the target blood glucose range independently influences the nurse's perception.

Table 1 .
How nurses evaluate the performance of Computer-Assisted Insulin Protocol, Leuven Protocol, and conventional treatment regarding their efficiency in maintaining glucose levels within the target range (% of time that blood glucose levels were under control) more than 90% of the time: p<0.001 for CAIP versus LP and p=0.16 for CAIP versus CT; ** more than 75% of the time: p<0.001 for CAIP versus LP and p=0.04 for CAIP versus CT.

Table 2 .
Nurses' evaluation of the complexity of the protocols