Abstract
Background:
Diabetes management requires complex and interdisciplinary cooperation of health care professionals (HCPs). To support this complex process, IT-support is recommended by clinical guidelines. The aim of this article is to report on results from a clinical feasibility study testing the prototype of a mobile, tablet-based client-server system for computerized decision and workflow support (GlucoTab®) and to discuss its impact on hypoglycemia prevention.
Methods:
The system was tested in a monocentric, open, noncontrolled intervention study in 30 patients with type 2 diabetes mellitus (T2DM). The system supports HCPs in performing a basal-bolus insulin therapy. Diabetes therapy, adverse events, software errors and user feedback were documented. Safety, efficacy and user acceptance of the system were investigated.
Results:
Only 1.3% of blood glucose (BG) measurements were <70 mg/dl and only 2.6% were >300 mg/dl. The availability of the system (97.3%) and the rate of treatment activities documented with the system (>93.5%) were high. Only few suggestions from the system were overruled by the users (>95.7% adherence). Evaluation of the 3 anonymous questionnaires showed that confidence in the system increased over time. The majority of users believed that treatment errors could be prevented by using this system.
Conclusions:
Data from our feasibility study show a significant reduction of hypoglycemia by implementing a computerized system for workflow and decision support for diabetes management, compared to a paper-based process. The system was well accepted by HCPs, which is shown in the user acceptance analysis and that users adhered to the insulin dose suggestions made by the system.
Keywords: basal-bolus insulin therapy, best practice, computerized decision support, diabetes mellitus type 2, hypoglycemia, medication management system
Current evidence supports the use of scheduled insulin regimens in the hospital for any patient with consistent hyperglycemia.1 However, the use of insulin for diabetes management is associated with risks for the patients. Recent evidence demonstrated that by using insulin in type 2 diabetes mellitus (T2DM) patients in the hospital, patients were exposed to significantly more medication errors, more hypoglycemic episodes and poorer glycemic control, compared to noninsulin treatment options.2 One has to add that comparing safety and efficacy of insulin and noninsulin treatment options is problematic because they target different patient populations, but the findings of the inpatient audit in Great Britain indicate room for improvement for inpatient diabetes management with insulin.
Although considerable efforts have been made to improve inpatient glycemic management, an adequate insulin therapy in clinical practice is still lacking in many hospitals despite its recommendation by diabetes experts and guidelines.3 Observational and randomized controlled studies indicate that improvement in diabetes management results in lower rates of hospital complications in general medicine and surgery wards.4,5 Therefore, international diabetes experts recommend a structured approach and an algorithm-driven basal-bolus insulin regimen in hospitalized T2DM patients.6
Diabetes management requires complex and interdisciplinary cooperation of health care professionals (HCPs) involving ordering doses and correction schemes, frequent blood glucose (BG) measurements and timely administration of resulting insulin doses. To support this complex process recently published guidelines and studies recommend the use of computerized decision support systems (CDSS)7 and physician order entry systems for diabetes therapy in hospitalized patients.8-10 The combination of physician order entry systems and CDSS has proven to reduce medication errors but clear evidence that this combination reduces clinical adverse drug events is still missing.11
Well-designed systems can improve glycemic control without increasing hypoglycemia.12 In contrast, systems which are ignoring certain design rules and necessary features are negatively impacting uptake by HCPs and health care providers. Consequently, ease of use (eg, quick access, minimal interactions), workflow integration, simplicity and visibility of messages are key drivers of user acceptance.13
We performed an iterative and interdisciplinary development process to develop a clinical workflow and decision support system, that is well accepted by users and which significantly improves and reduces risks in diabetes therapy with insulin: the GlucoTab® system.14-16
The aim of this article is to report on results from a clinical feasibility study testing the system and to discuss its impact on hypoglycemia prevention. Furthermore, lessons learned of the iterative and interdisciplinary development process and factors for a successful incorporation into clinical workflows are discussed.
Methods
The Iterative and Interdisciplinary Development Process
An interdisciplinary team of engineers, physicians and nurses gathered requirements to improve glucose management in the hospital. Initially, we assessed the quality of glycemic management at 2 clinical wards at the Department of Internal Medicine at the Medical University of Graz.17 As a consequence a clinical data review was performed which identified an insulin dosing regimen for basal-bolus insulin therapy in hospitalized T2DM patients which demonstrated good glycemic control in noncritical care.18,19 This basal-bolus insulin regimen was customized to account for complex processes during inpatient care and was then integrated into the workflow of a general internal medicine ward. Before implementing the basal-bolus insulin regimen in medical software a proof-of-concept study was performed assessing the efficacy, workflow integration and usability of the paper-based protocol for basal-bolus insulin therapy in T2DM patients.14 The workflow-integrated insulin dosing regimen was effective in establishing glycemic control compared to standard care and was well accepted by medical staff. In a subsequent step the paper-based protocol was translated into a computerized system for workflow and decision support. In an iterative approach, involving end users in all design decisions, gradually more detailed prototypes were developed. This system aims to overcome shortcomings of manual procedures: specifically in preventing input, calculation, and double data entry errors, and providing automated therapy visualizations and traceable real-time documentation for time-critical tasks.
The GlucoTab System
The development process resulted in the design of a mobile, tablet-based client-server system for computerized decision and workflow support. It assists nurses and physicians during hospitalization of patients with T2DM requiring subcutaneous basal-bolus insulin therapy. We implemented a HL7 ADT interface to the hospital information system in the backend server to automatically retrieve patient master data. The backend was installed on a server at the hospital IT department. We used 2 Samsung Galaxy Tab 7” Plus with the systems frontend app at a general ward. The backend server and frontend devices communicate via the hospital Wi-Fi. We implemented a kiosk mode at the frontend avoiding the use of apps other than the decision and workflow application and any possibility for misconfiguration by the user.
The main function of the system is the provision of insulin dose recommendations for basal-bolus insulin treatment of T2DM patients. It comprises the following functionalities which aid physicians and nurses: (1) data entry at the bedside, (2) medication order entry with insulin dosing decision support for physicians, (3) drug administration support including insulin dose calculation for nurses, and (4) workflow management for physicians and nurses. The mobile system assists in organizing the treatment workflow, including display for open tasks, facilitating documentation and providing visualization of BG values, meals, and insulin doses.
The system provides 3 major functions for user input, comprising BG documentation, calculation of the meal insulin dose, and adjustment of the daily insulin dose. User login and a role-based service access provide functionality according to the users’ roles and support the tracking of already performed activities.
On the top left side of the main screen, medication, administrative and patient data are shown (Figure 1a). On the bottom left side of the main screen the 3 main functions, “Blood Glucose Measurement,” “Insulin Administration,” and “Daily Dose Adjustment,” can be selected. Glucose profile and administered insulin is displayed on the right side. Users can switch to the patient list, the list of open tasks, and the history of already performed tasks.
Figure 1.
(a) Main screen. (b) Screenshot of decision support for insulin dosing.
The following example shows the insulin dosing decision support functionality of system (Figure 1b). Based on the total daily dose of 28 units of insulin (U), a premeal BG value of 182 mg/dl, and the input that the patient will have lunch, the system proposes automatically 10 U of bolus and 14 U of basal insulin for lunch.
The current version of system is a CE marked medical device software (Class I according to the Medical Device Directive, risk class C according to IEC 62304).
Study Design
The feasibility study was designed as a monocentric, open, noncontrolled intervention study in patients with T2DM. Approval of the local ethical committee and legal authorities had been obtained prior to study start and the study was performed according to the Declaration of Helsinki (ICH-CCP). The investigational treatment was BG management with a basal-bolus insulin therapy by using computerized workflow and decision support. Adult patients (≥18 years) with T2DM who were treated with diet alone and/or with any oral or injectable antihyperglycemic therapy and who were admitted to the general ward (Division of Endocrinology and Diabetology at the Department of Internal Medicine, Medical University of Graz, Austria) were included in the clinical study. We excluded patients with significantly impaired renal function (serum creatinine ≥ 3.0 mg/dl), terminally ill patients, patients with known or suspected allergy to insulin, pregnant women, patients with type 1 diabetes mellitus, and patients already included in a clinical study within the past 3 months. All included patients gave their informed consent prior to any study related activity.
We enrolled 30 patients into the system who were admitted at the Endocrinology ward and who matched the inclusion criteria of the study. Non-insulin-injected antidiabetic medicine and oral agents were stopped when treatment with the system was initiated. The clinical workflow and the algorithm-driven insulin dosing algorithm are described in full detail by Mader et al,14 and additional description is provided in the Supplemental Material.
The primary endpoint of the study was the percentage of actions the system supports, either to capture BG values or to provide insulin dose suggestions according to the insulin dosing algorithm.
Before the study start, we held workshops to train users in inpatient diabetes management with focus on insulin therapy of patients with T2DM and in the usage of the system. During the study, we supported users with hands-on training. We held regular study meetings to discuss medical and technical incidents and provided on-site and 24/7 phone support for users.
To evaluate user acceptance, 3 anonymous user questionnaires (UQ-1, UQ-2, UQ-3) were handed out to nurses and physicians using the system. Questionnaires were answered on a voluntary basis. UQ-1 had to be answered prior to study start, but after training on diabetes treatment and usage of decision and workflow support system; UQ-2 after half of the patients finished the study; and UQ-3 after all 30 patients had finished the study. The questionnaires were designed to inquire how users experienced the performance of the system and their perception of the effect on medication errors compared to routine care, using a combination of 5 point rating scales and yes or no questions.
Study Data Acquisition and Statistical Analysis
All manually entered data (eg, BG values, intended meals, insulin sensitivity) and all calculated data (eg, suggested/ordered insulin doses) were stored automatically on the backend server. In addition, we used paper source forms to verify the correct documentation of data in the electronic system. The paper source forms were then transcribed into electronic case report forms (eCRF) for data analysis.
The glucose profiles were analyzed based on recommendations for standardizing analysis and presentation of glucose monitoring data (ambulatory glucose profile).20 Glucose variability was calculated as standard deviation (SD). The level of glycemic control was calculated as patient-day-weighted mean, based on the daily premeal BG measurements. For the ambulatory glucose profile, “BG values in different ranges” was defined as “% of BG readings” within a well-defined range, such as 100-140 mg/dl.
Feedback regarding problems with the use of electronic system was collected with a 24/7 telephone support service and with a feedback form at the site (data entered by clinical and study personnel).
All metric outcome variables were checked for normality by means of Shapiro-Wilk’s test. In case of significant deviations from normality, results are represented as median and range instead of mean and SD.
For the analysis of the user acceptance yes and no answers were summarized to a “% of agreement” value and in case of the 5 point rating scale the average was calculated. Pearson’s χ2 test was used to analyze nominal data. The level of significance was set to 5% for all tests. Due to the descriptive character of this feasibility study no power calculation was performed. Statistical analysis was carried out using the statistical software R 3.0.1.21
Results
Safety and Efficacy
Thirty patients were treated on average 8.4 ± 4.5 days with the system (Table 1). The system was effective and safe in establishing glycemic control (Table 2). Only 1.3% of BG measurements were below 70 mg/dl and only 2.6% were above 300 mg/dl. Twelve hypoglycemic events (<70 mg/dl) were documented in 6 patients. No BG value below 40 mg/dl was recorded. During the study 24 adverse events occurred. Four of them were serious (2 patients died after study [respiratory failure and septic shock, cardiac failure with pulmonary edema], surgery was scheduled for 1 patient due to high grade mitral regulation, 1 patient was transferred to coronary care unit due to unstable angina). None of the adverse events was related to the electronic system.
Table 1.
Baseline Patient Characteristics.
| Characteristic | |
|---|---|
| Number of patients (n) | 30 |
| Gender, F/M (n) | 11/19 |
| Age (years; mean ± SD) | 71.0 ± 10.5 |
| BMI (kg/m2; mean ± SD) | 29.6 ± 6.2 |
| Weight (kg; median (range)) | 80.0 (58.0-135.0) |
| Race: Caucasian (n) | 30 |
| Serum creatinine (mg/dl; median (range)) | 1.2 (0.7-2.9) |
| HbA1c (mmol/mol; median (range)) | 63.0 (44.0-157.0) |
| Diabetes duration (years; median (range)) | 14.5 (1.0-55.0) |
| Length of study (days; mean ± SD) | 8.4 ± 4.5 |
| Diabetes therapy (n) | |
| No previous insulin | 16 |
| Previous insulin | 14 |
| Admission diagnosis (%) | |
| Hematological disease | 3.3 |
| Endocrine disease | 33.3 |
| Cardiovascular disease | 33.3 |
| Infectious disease | 30.0 |
Table 2.
Glycemic Management, Adverse Events, Insulin Therapy.
| Glycemic management | |
|---|---|
| Mean premeal BG value (patient-day-weighted mg/dl) | Mean ± SD |
| Overall | 154.9 ± 31.6 |
| Morning | 150.8 ± 37.5 |
| Noon | 176.3 ± 50.7 |
| Evening | 141.0 ± 29.9 |
| Bedtime | 147.2 ± 32.6 |
| Premeal BG values in range (%) | % |
| 0 mg/dl to <40 mg/dl | 0.0 |
| 40 mg/dl to <70 mg/dl | 1.3 |
| 70 mg/dl to <100 mg/dl | 11.8 |
| 100 mg/dl to 140 mg/dl | 36.4 |
| >140 mg/dl to <180 mg/dl | 24.8 |
| 180 mg/dl to <300 mg/dl | 23.1 |
| ≥300 mg/dl | 2.6 |
| Hypoglycemic events (n) | n |
| Mild (≥40 mg/dl and <70 mg/dl) | 12 |
| Severe (<40 mg/dl) | 0 |
| Adverse events (n) | n |
| Adverse events not device related | 20 |
| Adverse device-related events | 0 |
| Serious adverse events | 4 |
| Serious adverse device-related events | 0 |
| Anticipated serious adverse device related events | 0 |
| Unexpected serious adverse device related events | 0 |
| Insulin therapy (patient-day-weighted) [U] | Mean ± SD |
| Daily ordered basal | 20.8 ± 12.2 |
| Daily ordered bolus | 26.3 ± 15.2 |
Performance of the Implemented Computerized Glycemic Management
The availability of the system and the rate of treatment activities documented with the system were high (Table 3). In addition, the mean number of performed and documented activities approached the recommended numbers as specified by the basal-bolus insulin dosing algorithm. Missing activities were most likely due to temporary absence of patients from the ward or due to days not fully spent at the ward (day of admission/discharge/transfer). The adherence to the insulin dose suggestions was very high. Only few suggestions from the insulin dosing algorithm were overruled by the users.
Table 3.
Treatment Process Support, Adherence.
| Treatment process support | |
|---|---|
| Documented treatment activities | Tablet/eCRF, n (%) |
| BG measurements | 866/883 (98.1%) |
| Basal injections | 229/235 (97.4%) |
| Bolus injections | 817/874 (93.5%) |
| New total daily dose | 198/207 (95.7%) |
| System uptime | Hours supported/total hours (%) |
| 5154/5295 (97.3%) | |
| Treatment activities (mean number of premeal activities/day) | Mean ± SD (max) |
| BG measurements | 3.4 ± 0.3 (4) |
| Basal injections | 0.9 ± 0.1 (1) |
| Bolus injections | 3.3 ± 0.3 (4) |
| Adherence to the insulin dose suggestions (%) | Section 1.01 % |
| Bolus | 96.2 |
| Basal | 95.7 |
| Adjustment of the total daily dose | 98.7 |
Users and study personnel collected 31 feedback entries during the clinical study. While 23 entries referred to user errors and functionality/usability improvements, 8 entries identified software errors and anomalies to be corrected. These errors were identified as (1) synchronization error of daytime ranges between backend and frontend, (2) unexpected frontend shutdown, if BG measurement details were selected in the frontend, (3) tasks in the task list were presented twice after restart of the user screen, (4) temporary Wi-Fi disconnection at ward during patient treatment, (5) multiple display of login screen at system start, (6) user login failure, (7) entering BG was not possible after a period of time due to a memory persistence issue, and (8) basal insulin was not stored during Wi-Fi loss (1). None of these software errors or anomalies led to a critical situation for the patients’ safety and all were corrected in the next release of the software.
User Acceptance Analysis
Fourteen nurses and 10 physicians answered questionnaire 1 (UQ-1), 12 nurses and 3 physicians answered questionnaire 2 (UQ-2) and 12 nurses and 6 physicians answered questionnaire 3 (UQ-3). The majority of the users (92%) had practical experience in treating patients according to a basal-bolus therapy and two-thirds had already been working with the paper-based insulin titration protocol during the paper-based proof of concept study.14 Evaluation of the 3 questionnaires showed that confidence in the system increased over time (Figure 2). All users, except 1, felt safe with the electronic system and all users, except 1, believed that treatment errors could be prevented by using this system. During all phases of the study users believed that the system is suitable for daily routine. Users did not think that workload could be reduced with the system.
Figure 2.
Results of user questionnaires.
Discussion
Lessons Learned From the Clinical Feasibility Study
Implementing a computerized system for workflow and decision support into the workflow of a general ward in this feasibility study led to a decrease in hypoglycemia compared to a previous paper-based study (1.3% vs 3.0%, P = .01) while both studies achieved comparably good glycemic control.14 One reason for the reduction in hypoglycemia is probably due to an avoidance of manual insulin dose calculation errors by using computerized decision support.16
Our system showed similar BG control with fewer hypoglycemic episodes, when compared to computerized8,10 and paper-based19,22,23 best practice studies. Clinical users adhered to the insulin dose suggestions made by the system. More than 95% of all insulin dose suggestions were accepted without overruling. We did not observe any uncontrolled treatment processes due to nonadherence in this study.
Clinical users felt confident using the system. They thought that medical errors can be prevented and that the system is more effective in lowering the patients BG levels. At the end of the study all users felt confident that this system is able to prevent medical errors. In the preceding proof of concept study that tested the paper-based workflow-integrated insulin dosing algorithm only 8 of 12 HCPs answered that the used protocol was able prevent errors. 3 answered that the protocol would not help to prevent errors and 1 did not answer the question.14 These opinions were also confirmed in a post hoc analysis of a before and after study showing that manual dose calculations are prone to error and that they increase the risk of hypoglycemia in patients with diabetes.16
According to results of this feasibility study, more intensive diabetes management and a better quality of care can be established compared to routine care, without being more labor intensive. However, users did not feel that the use of the system leads to reduced workload. This is most likely due to the doubled documentation process which had to be performed during the clinical study to obtain study data. As a result of the feasibility study, improvements of the usability and the performance of the system were implemented: The view for the therapy-adjustment process for physicians was refined to fit on a single screen and can be completed in a single step. The login dialog was improved and an interface for using an existing directory service (Active Directory, LDAP) for user authentication was added. The reactivation of the decision support after missed mandatory workflow tasks was simplified. Measures were taken to keep bandwidth requirements as low as possible to reduce waiting times during data transmission (reduced data volume, reduced number of web service calls, data compression).
In a subsequent study which tested an improved version of the system, more than half of the users were confident that the system was able to reduce the workload compared to routine care.15 In this study double documentation was largely avoided.
Electronic treatment support for nursing staff and physicians was high with 97.3% of the treatment time supported with the electronic system. System downtime was caused by a software update of the system and software maintenance. Nonperformance of treatment support was mainly due to the inflexibility of the automatic workflow support. The system could not be set into pending mode when patients were not present at the ward. Consequently, the users were not able to perform the BG documentation on time. In that case, the decision support mode was disabled and no insulin dose suggestions were available.
None of the adverse events identified during this clinical study was related to the use of electronic system. Software errors did occur, but none of them led to a critical safety situation for any patient. Nevertheless, these errors limited the usability of the system and were corrected in the next release of the system. Sufficient workflow support is 1 key issue in terms of acceptance of CDSS13 and therefore a more flexible workflow integration was addressed in subsequent releases of the system.15,16 However, most of the users were convinced that the version of decision and workflow support system tested in the feasibility study was already suitable for daily routine.
Factors for a Successful Development and Implementation of a Clinical Decision Support System
A poor user interface is the most common cause for technology related errors.24 Therefore, the iterative development process of the system included usability tests in a clinical environment with diabetes specialists.25 We think paper mock-ups in particular proved to be very valuable in gaining insight into the clinical workflows and processes. User interfaces were designed exactly to the specific needs of HCPs in their daily routine. Simple menus and few loops in the medication process on the mobile tablet device were preferred over ample customization options.
Some problems with system performance were due to low Wi-Fi signal strength. The system stayed connected to a Wi-Fi access point with a weak signal even though access points with better signal strength were available. Therefore, we developed a routine which continuously measures signal strength and automatically switches to the strongest signal. This was implemented in the current version of the system and significantly improved system performance.
Paper-based documentation allows deviations from workflow and is thus very flexible. The implementation of computerized systems supporting clinical workflows is challenging and a large number of special cases have to be considered without compromising the usability of the systems. The system allows manual correction of BG levels and insulin doses and belated entry of values with a time stamp. Impact on the algorithm is automatically handled and appropriate user interaction is initiated if required.
For complex therapies such as a basal-bolus insulin regimen all relevant HCPs have to be included in the medication management.12,16 Access to therapy relevant information has to be available to all HCPs on duty at all times and at multiple locations.26 Therefore, a future task is the development of a web-frontend to allow access to the GlucoTab system from a web browser on a desktop PC, such as in a nurses station, to facilitate an improved integration into hospital workflows.
Limitations
Our analysis of the user acceptance was influenced by parallel documentation of the study data on paper source data forms to verify the correct system operation. This parallel documentation is not required in a real-life clinical scenario. The double documentation increased the workload and consequently negatively influenced the user acceptance of the system. Unfortunately the last user acceptance questionnaire was erroneous and the statement “Doctor to be called less often by the nurse” was missing. The analysis of safety and efficacy may be influenced by the fact that the distribution of bolus insulin for meals was modified during the study. The first 15 patients were treated with an equally distributed amount of bolus insulin per meal. In the next 15 patients was the daily bolus insulin distribution modified to 45% of daily bolus insulin in the morning, 25% at noon and 30% in the evening. However, the difference in distribution of bolus insulin did not influence the investigation of the primary endpoint.
Conclusion
Data from our feasibility study show a significant reduction of hypoglycemia by implementing a computerized system for workflow and decision support for diabetes management of T2DM patients, compared to a paper-based process.14 The system was well accepted by HCPs, which is documented by the high confidence in the system in the user acceptance analysis and that users adhered to the insulin dose suggestions made by the system.
Supplementary Material
Acknowledgments
We thank Beate Boulgaropoulos (Health–Institute for Biomedicine and Health Sciences, Joanneum Research GmbH, Graz, Austria) for critical review and editorial assistance with the manuscript. Martina Buttinger and her nursing team (Division of Endocrinology and Diabetology, Department of Internal Medicine at Medical University of Graz) supported the clinical trial with enthusiasm and gave valuable feedback for the further development of the system. We thank Bernd Tschapeller and Christian Krainer (both Health–Institute for Biomedicine and Health Sciences, Joanneum Research GmbH, Graz, Austria) for clinical data management and Andrea Berghofer (Endocrinology and Metabolism, Department of Internal Medicine at Medical University of Graz) for clinical data monitoring. We thank Sarah Raudner and Karin Pickl (both Health–Institute for Biomedicine and Health Sciences, Joanneum Research GmbH, Graz, Austria), who supported us with quality management for medical software. We also thank Matthias Enzmann and Frederik Franke (both Institute for Secure Information Technology, Fraunhofer Society, Darmstadt, Germany) for contributing to the security and user management of the system. Finally, we specially thank Thomas Truskaller, Reinhard Moser (Health–Institute for Biomedicine and Health Sciences, Joanneum Research GmbH, Graz, Austria) and Georg Petritsch (Technical University of Graz, Austria) for ensuring a high software quality as part of the development team. The study (ClinDiab03) was conducted with the approval of the Ethics Committee of the Medical University of Graz. Clinical trial reg no NCT01766752, http://www.clinicaltrials.gov.
Footnotes
Abbreviations: BG, blood glucose; CDSS, computerized decision support system; eCRF, electronic case report form; HCP, health care professional; T2DM, type 2 diabetes mellitus; U, units of insulin.
Declaration of Conflicting Interests: The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: PB is CTO and SS, KD, TRP, and JKM are cofounders of decide Clinical Software GmbH. The feasibility study was performed and completed before foundation of the company. TRP is a member in the advisory board of NovoNordisk A/S and received speaker honoraria from NovoNordisk A/S. JKM and JP received speaker honoraria from NovoNordisk A/S. The remaining authors declare no duality of interest associated with this article.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was partly funded by the European Commission under the 7th Framework Programme in the area of Personal Health Systems under Grant Agreement no. 248590 (REACTION project). Preparation of the manuscript was funded by the Austrian Federal Ministry of Science, Research and Economy–Austrian Research Promotion Agency project number 844737 (Research Studio Austria “GlucoTab”).
Supplemental Material: The supplemental material is available at http://journals.sagepub.com/doi/suppl/10.1177/1932296816676501
References
- 1. McDonnell ME, Umpierrez GE. Insulin therapy for the management of hyperglycemia in hospitalized patients. Endocrinol Metab Clin North Am. 2012;41:175-201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Rayman G, National Health Service. National diabetes inpatient audit 2013. 2014. Available at: http://www.hscic.gov.uk/catalogue/PUB13662/nati-diab-inp-audi-13-nat-rep.pdf.
- 3. Draznin B, Gilden J, Golden SH, Inzucchi SE. Pathways to quality inpatient management of hyperglycemia and diabetes: a call to action. Diabetes Care. 2013;36:1807-1814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Umpierrez GE, Hellman R, Korytkowski MT, et al. Management of hyperglycemia in hospitalized patients in non-critical care setting: an endocrine society clinical practice guideline. J Clin Endocrinol Metab. 2012;97:16-38. [DOI] [PubMed] [Google Scholar]
- 5. Murad MH, Coburn JA, Coto-Yglesias F, et al. Glycemic control in non-critically ill hospitalized patients: a systematic review and meta-analysis. J Clin Endocrinol Metab. 2012;97:49-58. [DOI] [PubMed] [Google Scholar]
- 6. American Diabetes Association. Standards of medical care in diabetes—2015. Diabetes Care. 2015;38:S1-S93. Available at: http://care.diabetesjournals.org/cgi/doi/10.2337/dc15-S001. [DOI] [PubMed] [Google Scholar]
- 7. Moghissi ES, Korytkowski MT, DiNardo M, et al. American Association of Clinical Endocrinologists and American Diabetes Association consensus statement on inpatient glycemic control. Diabetes Care. 2009;32:1119-1131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Schnipper JL, Liang CL, Ndumele CD, Pendergrass ML. Effects of a computerized order set on the inpatient management of hyperglycemia: a cluster-randomized controlled trial. Endocr Pract. 2010;16:209-218. [DOI] [PubMed] [Google Scholar]
- 9. Guerra YS, Das K, Antonopoulos P, et al. Computerized physician order entry- based hyperglycemia inpatient protocol and glycemic outcomes: the CPOE-HIP study. Endocr Pract. 2010;16:389-397. [DOI] [PubMed] [Google Scholar]
- 10. Wexler DJ, Shrader P, Burns SM, Cagliero E. Effectiveness of a computerized insulin order template in general medical inpatients with type 2 diabetes: a cluster randomized trial. Diabetes Care. 2010;33:2181-2183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Ranji SR, Rennke S, Wachter RM. Computerised provider order entry combined with clinical decision support systems to improve medication safety: a narrative review. BMJ Qual Saf. 2014;23:773-780. [DOI] [PubMed] [Google Scholar]
- 12. Wei NJ, Wexler DJ. Basal-bolus insulin protocols enter the computer age. Curr Diab Rep. 2012;12:119-126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Moxey A, Robertson J, Newby D, Hains I, Williamson M, Pearson SA. Computerized clinical decision support for prescribing: provision does not guarantee uptake. J Am Med Informatics Assoc. 2010;17:25-33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Mader JK, Neubauer KM, Schaupp L, et al. Efficacy, usability and sequence of operations of a workflow-integrated algorithm for basal-bolus insulin therapy in hospitalized type 2 diabetes patients. Diabetes Obes Metab. 2014;16:137-146. [DOI] [PubMed] [Google Scholar]
- 15. Neubauer KM, Mader JK, Höll B, et al. Standardized glycemic management with a computerized workflow and decision support system for hospitalized patients with type 2 diabetes on different wards. Diabetes Technol Ther. 2015;17:685-692. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Donsa K, Beck P, Höll B, et al. Impact of errors in paper-based and computerized diabetes management with decision support for hospitalized patients with type 2 diabetes. A post-hoc analysis of a before and after study. Int J Med Inform. 2016;90:58-67. [DOI] [PubMed] [Google Scholar]
- 17. Neubauer KM, Schaupp L, Plank J, et al. Failure to control hyperglycemia in noncritically ill diabetes patients despite standard glycemic management in a hospital setting. J Diabetes Sci Technol. 2013;7:402-409. Available at: http://www.ncbi.nlm.nih.gov/pubmed/23566999. Accessed July 15, 2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Umpierrez GE, Smiley D, Zisman A, et al. Randomized study of basal-bolus insulin therapy in the inpatient management of patients with type 2 diabetes (RABBIT 2 trial). Diabetes Care. 2007;30:2181-2186. [DOI] [PubMed] [Google Scholar]
- 19. Umpierrez GE, Smiley D, Jacobs S, et al. Randomized study of basal-bolus insulin therapy in the inpatient management of patients with type 2 diabetes undergoing general surgery (RABBIT 2 surgery). Diabetes Care. 2011;34:256-261. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Bergenstal RM, Ahmann AJ, Bailey T, et al. Recommendations for standardizing glucose reporting and analysis to optimize clinical decision making in diabetes: the ambulatory glucose profile. J Diabetes Sci Technol. 2013;7:562-578. Available at: http://www.ncbi.nlm.nih.gov/pubmed/23567014. Accessed February 10, 2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. R Development Core Team. R: A Language and Environment for Statistical Computing. Vienna, Austria; 2008. [Google Scholar]
- 22. Umpierrez GE, Smiley D, Hermayer K, et al. Randomized study comparing a basal-bolus with a basal plus correction insulin regimen for the hospital management of medical and surgical patients with type 2 diabetes: basal plus trial. Diabetes Care. 2013;36:2169-2174. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Umpierrez GE, Gianchandani R, Smiley D, et al. Safety and efficacy of sitagliptin therapy for the inpatient management of general medicine and surgery patients with type 2 diabetes: a pilot, randomized, controlled study. Diabetes Care. 2013;36:3430-3435. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Samaranayake NR, Cheung STD, Chui WCM, Cheung BMY. Technology-related medication errors in a tertiary hospital: a 5-year analysis of reported medication incidents. Int J Med Inform. 2012;81:828-833. [DOI] [PubMed] [Google Scholar]
- 25. Höll B, Spat S, Plank J, et al. Design of a mobile, safety-critical in-patient glucose management system. Stud Health Technol Inform. 2011;169:950-954. Available at: http://www.ncbi.nlm.nih.gov/pubmed/21893886. Accessed January 22, 2015. [PubMed] [Google Scholar]
- 26. Moss J, Berner ES. Evaluating clinical decision support tools for medication administration safety in a simulated environment. Int J Med Inform. 2015;84:308-318. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials