- Email: [email protected]
Using Standardized Insulin Orders to Improve Patient Safety in a Tertiary Care Centre
Can J Diabetes 38 (2014) 118e125
Contents lists available at ScienceDirect
Canadian Journal of Diabetes journal homepage: www.canadianjournalofdiabetes.com
Original Research
Using Standardized Insulin Orders to Improve Patient Safety in a Tertiary Care Centre Mary-Anne Doyle MD, Sharon Brez BScN, MA(Ed), Silvana Sicoli MHA, Filomena De Sousa MScN, Erin Keely MD, Janine C. Malcom MD * Division of Endocrinology and Metabolism, University of Ottawa, Ottawa, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 November 2013 Received in revised form 2 January 2014 Accepted 5 January 2014
Objective: To standardize insulin prescribing practices for inpatients, improve management of hypoglycemia, reduce reliance on sliding scales, increase use of basal-bolus insulin and improve patient safety. Methods: Patients with diabetes were admitted to 2 pilot inpatient units followed by corporate spread to all insulin-treated patients on noncritical care units in a Canadian tertiary care multicampus teaching hospital. Standardized preprinted insulin and hypoglycemia management orders, decision support tools and multidisciplinary education strategies were developed, tested and implemented by way of the Model for Improvement and The Ottawa Model for Research Process. Clinical and balance measures were evaluated through statistical process control. Results: Patient safety was improved through a reduction in hypoglycemia and decreased dependence on correctional scales. Utilization of the preprinted orders approached the target of 70% at the end of the test period and was sustained at 89% corporately 3 years post-implementation. Conclusions: The implementation of a standardized, preprinted insulin order set facilitates best practices for insulin therapy, improves patient safety and is highly supported by treating practitioners. The utilization of formal quality-improvement methodology promoted efficiency, enhanced sustainability, increased support among clinicians and senior administrators, and was effective in instituting sustained practice change in a complex care centre. Ó 2014 Canadian Diabetes Association
Keywords: diabetes decision support clinical patient safety quality improvement quality improvement methodology
r é s u m é Mots clés : diabète aide à la décision clinique sécurité des patients amélioration de la qualité méthodologie de l’amélioration de la qualité
Objectif : Standardiser les pratiques liées à la prescription d’insuline, améliorer la prise en charge de l’hypoglycémie, réduire la dépendance à l’échelle d’adaptation aux doses, augmenter l’utilisation d’insuline selon le schéma basal-bolus et améliorer la sécurité des patients. Méthodes : Des patients souffrant de diabète ont été admis dans 2 unités pilotes pour patients hospitalisés, et tous les patients traités par insuline ont été répartis sur les unités de soins non critiques d’un hôpital d’enseignement canadien de soins tertiaires canadien comptant plusieurs campus. Les ordonnances préimprimées standardisées d’insuline et de prise en charge de l’hypoglycémie, les outils d’aide à la décision et les stratégies d’enseignement multidisciplinaire ont été élaborés, testés et mis en application au moyen du Model for Improvement et du Modèle d’utilisation de la recherche d’Ottawa. Les mesures cliniques et de l’équilibre ont été évaluées par la maîtrise statistique des processus. Résultats : La sécurité des patients a été améliorée par une réduction de l’hypoglycémie et une diminution de la dépendance aux échelles de correction. L’utilisation des ordonnances préimprimées s’est approchée de l’objectif de 70 % à la fin de la période d’essai et a été maintenue à 89 % collectivement 3 ans après la mise en application. Conclusions : La mise en application d’un ensemble de modèles d’ordonnances préimprimées standardisées favorise de meilleures pratiques de traitement par insuline, améliore la sécurité des patients et bénéficie largement du soutien des praticiens traitants. L’utilisation de la méthodologie officielle de l’amélioration de la qualité a promu l’efficience, amélioré la durabilité, accru le soutien aux cliniciens et
* Address for correspondence: Janine C. Malcolm, MD, FRCPC, The Ottawa Hospital-Riverside Campus, Division of Endocrinology and Metabolism, 1967 Riverside Drive, Ottawa, Ontario K1H 7W9, Canada. E-mail address: [email protected]. 1499-2671/$ e see front matter Ó 2014 Canadian Diabetes Association http://dx.doi.org/10.1016/j.jcjd.2014.01.003
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
119
aux gestionnaires principaux, et s’est avérée efficace à la mise en place d’un changement durable de la pratique dans un centre de soins complexes. Ó 2014 Canadian Diabetes Association
Introduction
Project phases
Diabetes management of hospitalized patients is increasingly recognized as an important quality and safety priority (1). Although diabetes itself is usually not the primary reason for hospitalization, it is a common comorbidity affecting approximately 30% of hospitalized patients (2). Both hyperglycemia and hypoglycemia in hospital are associated with increased morbidity, mortality, length of stay and resource utilization (3e7). Consistent with findings in the literature, the daily point prevalence of diabetes at our institution, a multicampus, tertiary care, teaching hospital, is approximately 25%. The average length of stay for this population prior to our quality improvement initiative was almost twice as long as that of those without diabetes. Diabetes management and maintenance of glycemic control in the hospital setting are challenging (3,8). Several factors, including changes in diet, physical activity, medications and the stress of acute illness, all negatively impact blood glucose stability. Insulin, the drug most commonly recommended and prescribed to treat diabetes in hospitalized patients, is classified by the Institute for Safe Medication Practices as a high-risk medication secondary to the high incidence of prescription and administration errors and adverse patient outcomes (9). Identified barriers to achieving glycemic control in hospitalized patients who are prescribed insulin therapy include an over-reliance on sliding-scale insulin, lack of routine basal and mealtime insulin, lack of patient-specific supplemental scales, uneven knowledge and confidence among clinical staff, competing clinical and institutional priorities, and lack of standardized practices for appropriate treatment of hypoglycemia (10). At our institution, audits of diabetes care and needs assessments by nurses, physicians and pharmacists echoed many of the challenges associated with insulin therapy described in the literature. Although many barriers to glucose management in non-critical care settings have been identified, some solutions have been proposed. Use of a basal-bolus approach to insulin therapy with a supplemental correction sliding scale and implementation of standardized insulin order sets have been shown to increase patient safety by improving glycemic control, reducing dependence on insulin sliding scales and decreasing hypoglycemia in hospitalized patients (1,11e14). In response to these findings, we successfully implemented a systematic, theory-based quality improvement initiative across all non-critical care units in our hospital. The purpose of this initiative was to improve patient safety and the process of diabetes management for patients treated with insulin through the implementation of pre-printed insulin orders, decision support tools, education and standardization of hypoglycemia management.
The quality improvement project consisted of 2 pilot phases involving 2 inpatient units to test, monitor and refine interventions followed by a multisite corporate roll-out. This phased approach enabled rapid learning and evolution of the project through small tests of change.
Methods Two complimentary theoretical frameworks, the Ottawa Model of Research Utilization (OMRU) and the Model for Improvement, were chosen to guide this quality-improvement initiative (15,16). A series of Plan-Do-Study-Act (PDSA) cycles were undertaken in order to inform, refine and test changes related to insulin therapy in actual clinical practice.
Ethics approval Ethics approval for this project was obtained from the Ottawa Hospital Research Ethics board. Patients and setting This project took place at a bilingual Canadian, multicampus tertiary care hospital with over 1100 beds and 47 000 patient admissions yearly. The cardiology and nephrology units were specifically selected as pilot locations to yield findings useful to the quality improvement process, including campus-specific variability, high volume of diabetes patients and a commitment to improving diabetes care. Inclusion and exclusion criteria Participants included all patients with either type 1 or type 2 diabetes admitted to a pilot unit during 1 of 2 intervention periods (intervention 1, a 6-week period; intervention 2, a 10-week period). Patients were included if they had been treated with subcutaneous insulin and had had a length of stay of 3 to 35 days. Patients were excluded if they had previously been admitted to the intensive care unit or had required IV insulin. Interventions Stakeholder engagement Engagement of partners, such as physicians, nurses, ward clerks, clinical program directors, inpatient unit managers, staff development trainers, forms developers and administrators was achieved through inclusion of representatives of these groups on our Corporate Inpatient Diabetes Quality Improvement committee. To enhance visibility within the institution, we were successful in having our initiative included in the Corporate Quality Improvement plan, and we actively searched for opportunities to communicate broadly the importance and progress of the project. A Corporate Inpatient Diabetes Quality Committee with broad, interdisciplinary and administrative stakeholder representation was established to guide and facilitate inpatient diabetes-practice improvements. A subset of this group, consisting of endocrinologists, internists, advanced practice nurse, diabetes nurse specialists, pharmacists, clinical managers, nursing educators and a qualityimprovement advisor was convened and mandated to focus specifically on improving insulin-ordering practices. Development of pre-printed insulin orders Informed by feedback, literature review, best practices and consensus expert opinion, a pre-printed subcutaneous insulin order form was developed. The goal of this tool was to standardize
120
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
Figure 1. Pre-printed subcutaneous insulin orders developed, implemented and evaluated using quality-improvement methodologies in a multicampus tertiary care centre.
insulin-prescribing practices, promote use of basal and mealtime insulin, reduce reliance on sliding-scale insulin as the sole form of diabetes treatment and standardize hypoglycemia management (Figure 1). To support the implementation of the pre-printed insulin orders, a diabetes medication administration record (MAR) and blood glucose monitoring form were revised and combined. This helped to facilitate assessment of patient response to treatment and pattern management, reduce transcription time, the potential for transcription errors and standardize hypoglycemia management.
Multidisciplinary education strategies To facilitate implementation of the orders, discipline- and recipient-appropriate educational sessions that addressed best practice for inpatient insulin therapy, management of hypoglycemia and instruction on use of the new standardized order form were developed. (Table 1) Education was facilitated by endocrinologists, medical residents, advanced practice nurses, diabetes nurse specialists, unit-based nurse educators and ward clerk trainers.
Implementation of pre-printed insulin orders on pilot units Two pilot units were identified to test and refine this process through participation in PDSA cycles over 2 pilot test periods. Pilot test period 1 Engagement of the pilot units in the first test cycle was achieved through a variety of strategies. Stakeholders on each unit were identified and discipline-specific “champions” for change were recruited. A baseline chart audit of 47 (21 cardiology, 26 nephrology) patients was conducted, and findings were shared with unit nurse managers, clinical educators, pharmacists and physicians. After 6 weeks, a post-intervention chart review and a user-satisfaction survey were undertaken. This process identified a broad range of opportunities to further improve the uptake of the orders and new forms. Pilot test period 2 In response to results and feedback in the first implementation period, the educational intervention and implementation process were modified (Table 2). Minor changes were made to the insulin order set to improve clarity and increase uptake. User education
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125 Table 1 Discipline and recipient appropriate learning sessions addressing inpatient insulin therapy, hypoglycemia management and instruction on the use of the new standardized order form Discipline Managers Ward clerks Nurses
Physicians
Specific education sessions provided General information provided concerning the goals, use and format of the orders 1:1 instruction on use and transcription of the form Facilitated discussions addressing challenges to hyperglycemia and hypoglycemia management Educational sessions on the format and use of the orders Case-based educational sessions on the use of the orders Sessions on approaches to glycemic management in the hospitalized patient
was enhanced with an emphasis on the goal of using the new forms for all insulin orders, and the new forms were relocated on each unit to improve access and promote use. Additional educational sessions focusing on the clinical impact of inpatient glucose control, case-based approaches to insulin dosing and selecting the appropriate correction scale were provided for physicians and nurses in the second test period. These sessions also reinforced the expectation that all orders be written on the new forms. Resident orientations to cardiology and nephrology were revised to include information related to insulin therapy. A supporting diabetes management pocket card with a stepwise approach for inpatient diabetes treatment, insulin action profiles, guidance for calculating doses and adjustments for special clinical situations was developed. This tool was distributed to medical residents and attending staff, and interested health professionals
121
on pilot units and was added as a quick reference tab on our electronic health record. During this implementation phase, the inpatient diabetes nurse specialists increased their availability on pilot units to provide additional educational support. They promoted the use of the forms and reinforced the goal of the project. This second test period lasted 10 weeks. Corporate spread Following the pilot phase, the pre-printed insulin orders, educational sessions and pocket card were implemented across all non-critical care inpatient units. Strategies successfully utilized in the pilot phases were implemented to support corporate spread. They included involvement and active support of leaders with reputation and position (i.e. the vice president of medical affairs, clinical and nursing professional practice directors, directors of laboratory medicine and pharmacy, physician champions and chief residents) and the provision of a range of educational opportunities for physicians, nurses and clerical staff tailored to their learning needs. Communication of the goals, targets and expectations of the project to clinical staff occurred via email from senior administration (the vice president of medical affairs) through news articles in the hospital journal and through department meetings. Endocrinologists, advanced practice nurses and nurse specialists continued to be available for clinical rounds and unit-based learning. However, to distribute the training load and encourage ownership by all programs, the trainer approach was used to include unitbased clinical nurse educators and residents in the training process. Timely, ongoing information related to utilization rates of the new insulin order forms and unit-specific progress was shared with all physician chiefs and clinical managers.
Table 2 Barriers to uptake of order form and factors that potentially contributed to increased rates of hyperglycemia Barriers to uptake Lack of understanding of the purpose of the orders
High turnover of medical trainees Clinicians “not aware” of the orders
Access and availability of the order form Perception by clinicians that glycemic control in hospital was not achievable Fear of hypoglycemia and lack of confidence in how to safely adjust insulin to optimize glycemic control Lack of understanding that all insulin orders should be completed using the pre-printed orders Transcription times and potential for transcription errors would be increased Factors potentially contributing to hyperglycemia Omission of meal-time insulin Inappropriate and overly conservation selection of correctional scales
Lack of adjustment made to routine insulin doses despite inadequate glycemic control
Strategies implemented to improve uptake Unit-specific baseline findings and opportunities for improvements shared Discipline-specific educational sessions modified to emphasize goal of project and impact on patient safety Use of leaders of reputation and position (physician champions, chief residents, directors of nursing, clinical nurse educators, diabetes nurse specialists) to promote the orders Increased frequency of educational sessions Incorporation of sessions into orientation programs for new trainees Increased frequency of educational sessions Signage/posters created for workstations Communication via e-mail, meetings and rounds Reminders from pharmacists Units evaluated for optimal location to store order form Process developed to ensure adequate supply with unit clerks, nurse managers and physicians Modification to discipline-specific education sessions to address perception and enhance confidence Development of diabetes pocket card with a stepwise approach to achieving glycemic control Shared improvements of others with teams Development of diabetes pocket card Incorporation of case-based examples into teaching sessions Increased presence of diabetes nurse specialist on pilot units as educational resource to clinical staff Increased emphasis on purpose and goals of project during educational sessions Audit and feedback of order form usage to pilot units Development of supporting policies and procedures Revise diabetes MAR to include pre-printed sections for hypoglycemia orders, supplemental insulin scales and blood glucose monitoring orders Strategies to improve glycemic control Modification to educational sessions to include additional discussion on need for meal-time insulin and importance of proactive approach to diabetes management Modification to pre-printed insulin orders with removal of “very low” correctional scales Case-based examples during educational sessions Modification to diabetes pocket card to reflect changes in order set and inclusion of suggestions for appropriate correction scale choice Increased emphasis during educational sessions that insulin doses need to reassessed frequently and adjusted accordingly
122
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
Table 3 Measures used to evaluate improvement with implementation of standardized pre-printed insulin orders Measures of Improvement Process Percentage of insulin orders written using pre-printed orders measures Percentage of orders completed correctly Percentage of orders transcribed correctly to medication administration record (MAR) Percentage of exclusive use of sliding scale for >48 hours Percentage of episodes of hypoglycemia managed according to standards of practice Clinical Percentage of stay days classified as high days (1 or more measures BG >11 mmol/L). Percentage of stay days classified as low days (1 or more BG <4 mmol) Balance Consultation rates to endocrinology Measures Consultation rates to diabetes nurse specialists User satisfaction with new insulin orders, supporting resources
target was met and the process was stable. A point prevalence audit of pre-printed insulin order use across all 3 campuses was undertaken 3 years post-implementation to evaluate sustained utilization rates. Microsoft Excel and Chart Runner software (v 3.6) were used to analyze data. Surveys were undertaken to elicit user satisfaction with the new process, forms and decision-support tools. They were distributed to physicians (attending and resident physicians), nurses, pharmacists and ward clerks following each of the pilot test periods and 6 months after initiation of corporate spread. Various strategies were used to collect feedback, including semi-structured telephone and face-to-face interviews, anonymous electronic or written surveys, and facilitated unit-based user-group discussions. Results Pilot process measures
Measures of improvement Table 3 outlines clinical outcomes, process measures and balance measures evaluated during the pilot periods. A target of 70% of all subcutaneous insulin orders written on the new insulin order form was established a priori as an appropriate target for uptake of the pre-printed insulin orders. This goal recognized that a small proportion of patients would require an unusual or unique treatment regimen that the pre-printed form would not easily support and that in these situations a generic medication order form should continue to be used. Data collection and analysis The data collection strategies included standardized chart reviews, audits of order forms processed by the inpatient pharmacies and user surveys. The standardized chart reviews were completed by a clinical expert and audited by the advanced practice nurse to ensure consistent data collection and classification. Any discrepancies were reviewed together and consensus was achieved by team members. The data-collection plan was designed to support the use of statistical process control through development and analysis of Shewhart control charts and run charts. This enabled tracking of both the progress and the stability of the practice changes and identification of change related to special causes. When statistical process control was not feasible, a before-and-after study design was used. Data were collected for a 6-week period to establish a baseline and then every 2 weeks during the pilot phases. Corporate spread data were collected weekly for 17 weeks until the usage
During the first pilot test period, uptake of the new order forms was only 11% on the nephrology unit and 38% on the cardiology unit. With additional education, development and dissemination of decision-support tools and improved access to the form, the utilization rates increased to a sustained rate of 68% and 74%, respectively, after the second pilot test period. At the end of the second test period, the new pre-printed order forms were completed appropriately 80% to 88% of the time, and in all cases the orders were transcribed to the diabetes MAR correctly. Dependence on the insulin sliding scale as the only therapy was reduced from 30% to 4% on cardiology and from 32% to 0% on the nephrology unit, and the use of basal and meal-time insulin increased on both units (Figure 2). Pilot clinical measures The clinical impact of the pre-printed orders was assessed during the pilot phase of this improvement project. The rate of hypoglycemia (capillary blood glucose reading<4 mmol/L) was reduced after implementation of the intervention (Figure 3). No serious hypoglycemia was observed with use of the pre-printed form and supplemental insulin-adjustment scales. The number of high blood glucose days (2 or more documented blood glucose readings over 11 mmol/L in 24 hours) did not improve on either unit (Figure 3). Review of cases of hyperglycemia indicated that this was due to a variety of factors, including underprescription of meal-time insulin for patients who were eating regularly, the selection of an inappropriately conservative insulin-
Figure 2. Insulin therapy on pilot units at baseline and after 2 PDSA cycles of change.
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
123
Figure 3. Glycemic control on pilot units before and after implementation of standardized approach to insulin therapy after second test period.
correction scale or failure to increase basal insulin despite a persistent need for correction insulin. Pilot balance measures The intervention did not lead to increased consultation requests to the endocrine services. There was a reduction in the number of consults to the diabetes nurse specialists on both pilot units (cardiology 35% vs. 17%; nephrology, 12% vs. 6%) and no increase in the number of consults to the endocrine service on either unit (cardiology 35% vs. 17%; nephrology 16% vs. 16.7%). Satisfaction with the pre-printed orders was high. A total of 13 pilot unit physicians and 64 nurses were surveyed. Of them, 92% of physicians and 74% of nurses indicated that the orders were clear, and 85% of responders indicated that the pre-printed orders improved the diabetes treatment received by patients. The chart audit, surveys and feedback sessions following the first pilot test period identified a broad range of modifiable barriers to the uptake of the orders that were successfully addressed during the second improvement cycle on pilot units (Table 2). Corporate spread Following the pilot phase, the orders were spread corporately. Use of the pre-printed order form was monitored through weekly audits of all subcutaneous insulin orders received by the pharmacy over a 17-week period (Figure 4). Uptake of the new orders was progressive. After the first 4 months of the corporate launch, 60% to 70% of subcutaneous insulin orders were written on the new preprinted insulin forms. User satisfaction with the pre-printed orders for physicians, nurses and ward clerks was assessed 6 months after implementation through a short anonymous electronic survey. Of those surveyed (n¼54), 88.7% recommended continued use of the form, and 92.5% rated the form as useful to their practice. A point-prevalence audit of all subcutaneous insulin orders written at the hospital in a 24-hour period was performed 3 years after implementation to evaluate sustainability. The pre-printed insulin order form was used for 89% of orders written, surpassing the 70% target. Discussion Diabetes management using insulin therapy for hospitalized, non-critical-care patients continues to be a challenge for many practitioners and institutions. Despite the complexity of effecting multidisciplinary change in a large academic setting with rotating trainees, we were successful in designing and implementing preprinted insulin orders and accompanying decision support aids
that improved patient safety and led to sustained uptake of best practices. Improvements in patient safety may be attributed to a decrease in hypoglycemic events, fewer transcription errors, standardization of hypoglycemia treatment and reduced reliance on sliding-scale insulin as the sole insulin therapy. Reduced hypoglycemia was likely secondary to increased prescription of both long-acting basal insulin and rapid-acting meal-time insulin with the use of a more individualized correction dose when blood glucose was outside target. These findings resonate with those of others (3,11). It is also important to note that these improvements occurred through changes in the practices of non-diabetes specialists and without an increase in the demand for endocrinology or diabetes nurse specialist consults. Feedback from nurses, physicians and pharmacists indicated that confidence in treating diabetes using insulin was increased through this initiative. Although we observed a decrease in the frequency of hypoglycemia, an unanticipated small increase in the number of days with 2 blood glucose readings over 11 mmol/L was observed. Because we did not compare actual blood glucose readings, it is not possible to know how the overall degree of blood glucose control was impacted. A trend of increased hyperglycemia has previously been reported in other studies where standardized protocols for insulin dosing have been implemented (17,18). This increase has been attributed to poor protocol compliance and increased steroid use. In contrast to these reports, we obtained high levels of compliance with the use of the pre-printed insulin orders. It is possible that the increased rate of hyperglycemia we observed on the nephrology unit during the pilot was related to post-renal transplant steroid therapy. To better understand this finding, we undertook additional review of the pilot cases with frequent hyperglycemia. This review suggested that the increase in hyperglycemia may have been due to the omission of meal-time insulin orders or the selection of an inappropriately conservative correctional scale (for fear of hypoglycemia). Additionally, when a pattern of recurrent use of correction insulin dosing occurred, this additional insulin was not appropriately incorporated into the usual proactive insulin dosing. This analysis underlines the complexity and dynamic nature of insulin requirements during hospital admissions. It has been helpful in identifying further learning opportunities related to pattern management for insulin prescribers at our institution. Given the current evidence supporting the negative impact of hyperglycemia in acute illness, this remains an important finding for further investigation. It is also important to note that use of the standardized insulin orders was highly supported by treating practitioners. This positive response and ongoing support of the new insulin orders helped to facilitate uptake and increased awareness that inpatient
124
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
Figure 4. Corporate uptake of pre-printed insulin orders for the first 17 weeks after implementation.
diabetes management is an important, shared patient quality and safety issue. These tools made diabetes management easier for practitioners. Our improvement experience has additionally provided some learning related to the implementation of complex change that may help others who are contemplating such a project. The OMRU was helpful in enabling an understanding of the context of the planned change and in supporting the marketing of the proposal throughout the project. While use of OMRU assisted us to understand and manage our larger environment, the Model for Improvement provided more detailed guidance for the PDSA cycles required to inform, refine and test changes related to insulin therapy in actual clinical practice and provided us with the tools to determine whether our change was resulting in an improvement. Attributes of the planned innovation, the characteristics of potential adopters (i.e. current practices, knowledge, skill, attitudes and beliefs) and the practice environment (culture, relationships, priorities, patients and resources) were all considered early in the process. Early engagement of potential adopters, key stakeholders and opinion leaders fostered trust that enabled a transparent establishment of shared goals, awareness of competing demands and the needs of each partner to find benefit associated with the proposed change. Use of project teams and charters enabled all partners to understand their roles in the process and enhanced ongoing commitment and accountability. Maintaining timely, respectful and clear communication was key to understanding the issues and helped to avoid conflict. Finally, in working in an academic setting, team members had to learn and value the differences between research and quality-improvement methods. Having access to a trained quality-improvement advisor was a valuable asset in this area. We identified 2 recurrent and serious challenges to our success that may be common to other teams considering this type of quality-improvement project. Ensuring easy access to forms was critical to their uptake by busy practitioners; however, the best place to locate the forms was not always intuitively evident and differed among disciplines and practice settings. This challenge was overcome largely through detailed mapping of the process of writing and processing the orders on target units and then determining the most suitable location for the forms.
The second and more sensitive challenge was identification and acknowledgement of attitudes, beliefs and vulnerabilities related to the diabetes management of the hospitalized patients. These included fear of severe hypoglycemia, lack of confidence related to insulin adjustments and beliefs that improvement in glycemic control in hospital was not possible, based on past experience (Table 2). These challenges are not unique and have been noted in other studies (10). Identifying these barriers early through solicitation of anonymous feedback, developing alliances and non-judgemental relationships, and tailoring education and information to the needs of the recipients were important in reducing these barriers successfully. Through our improvement efforts, we have demonstrated that implementation of a standardized, pre-printed insulin order set facilitates best practices for insulin therapy, improves patient safety and is highly is supported by treating practitioners. The utilization of formal quality-improvement methodology promoted efficiency and was effective in instituting sustainable change in a complex environment. Acknowledgements This work was supported by The Ottawa Hospital Quality Improvement Plan. References 1. Umpierrez GE, Hellman R, Korytkowski MR, et al. Management of hyperglycemia in hospitalized patients in non-critical care setting: an endocrine society clinical practice guideline. J Clin Endocrinol Metabol 2012;97:16e38. 2. Cook CB, Kongale GL, Potter DJ, et al. Inpatient glucose control: a glycemic survey of 126 US hospitals. J Hosp Med 2009;4:E7e14. 3. Clement S, Braithwaite S, Magee MF, et al. Management of diabetes and hyperglycemia in hospitals. Diabetes Care 2004;27:553e91. 4. Campbell KB, Braithwaite SS. Hospital management of hyperglycemia. Clin Diabetes 2004;22:81e8. 5. Umpierrez GE, Isaacs SD, Bazargan N, et al. Hyperglycemia: an independent marker of in-hospital mortality in patients with undiagnosed diabetes. J Clin Endocrinol Metab 2002;87:978e82. 6. McAlister FA, Majumdar SR, Blitz S, et al. The relation between hyperglycemia and outcomes in 2471 patients admitted to hospital with community-acquired pneumonia. Diabetes Care 2005;28:810e5. 7. Krinsley JS. Association between hyperglycemia and increased hospital mortality in a heterogeneous population of critically ill patients. Mayo Clin Proc 2003;78:1471e8.
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125 8. Inzucchi SE. Clinical practice: management of hyperglycemia in the hospital setting. N Engl J Med 2006;355:1903e11. 9. Patient Safety Alert. “High-alert” medications and patient safety. Int J Qual Health Care 2001;13:339e40. 10. Korytkowski M, Dinardo M, Donihi AC, Bigi L, Devita M. Evolution of a diabetes inpatient safety committee. Endocr Prac 2006;12(Suppl 3):91e9. 11. 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;30:2181e6. 12. Trujillo J, Barsky E, Greenwood B, et al. Improving glycemic control in medical inpatients: a pilot study. J Hosp Med 2008;3:55e63. 13. Maynard G, Lee J. Improved inpatient use of basal insulin, reduced hypoglycemia, and improved glycemic control: effect of structured subcutaneous insulin orders and an insulin management algorithm. J Hosp Med 2009;4:3e15.
125
14. Schnipper J, Ndumele CD, Lian CL, et al. Effects of a subcutaneous insulin protocol, clinical education and computerized order set on the quality of inpatient management of hyperglycemia: results of a clinical trial. J Hosp Med 2009;4:16e27. 15. Logan J, Graham I. Towards a comprehensive interdisciplinary model of health care research use. Sci Commun 1998;20:227e46. 16. Langley GJ, Moen RD, Nolan KM, et al. A practical approach to enhancing organizational performance, 2nd ed. Jossey-Bass Books, 2009, pp 23e5. 17. Noschese M, Donihi AC, Koerbel G, et al. Effect of a diabetes order set on glycemic management and control in the hospital. Qual Saf Health Care 2008;17: 464e8. 18. Chen HJ, Steinke D, Karounos DG, et al. Intensive insulin protocol implementation and outcomes in the medical and surgical wards at a Veterans Affairs medical centre. Ann Pharmacother 2010;44:249e56.
Contents lists available at ScienceDirect
Canadian Journal of Diabetes journal homepage: www.canadianjournalofdiabetes.com
Original Research
Using Standardized Insulin Orders to Improve Patient Safety in a Tertiary Care Centre Mary-Anne Doyle MD, Sharon Brez BScN, MA(Ed), Silvana Sicoli MHA, Filomena De Sousa MScN, Erin Keely MD, Janine C. Malcom MD * Division of Endocrinology and Metabolism, University of Ottawa, Ottawa, Canada
a r t i c l e i n f o
a b s t r a c t
Article history: Received 25 November 2013 Received in revised form 2 January 2014 Accepted 5 January 2014
Objective: To standardize insulin prescribing practices for inpatients, improve management of hypoglycemia, reduce reliance on sliding scales, increase use of basal-bolus insulin and improve patient safety. Methods: Patients with diabetes were admitted to 2 pilot inpatient units followed by corporate spread to all insulin-treated patients on noncritical care units in a Canadian tertiary care multicampus teaching hospital. Standardized preprinted insulin and hypoglycemia management orders, decision support tools and multidisciplinary education strategies were developed, tested and implemented by way of the Model for Improvement and The Ottawa Model for Research Process. Clinical and balance measures were evaluated through statistical process control. Results: Patient safety was improved through a reduction in hypoglycemia and decreased dependence on correctional scales. Utilization of the preprinted orders approached the target of 70% at the end of the test period and was sustained at 89% corporately 3 years post-implementation. Conclusions: The implementation of a standardized, preprinted insulin order set facilitates best practices for insulin therapy, improves patient safety and is highly supported by treating practitioners. The utilization of formal quality-improvement methodology promoted efficiency, enhanced sustainability, increased support among clinicians and senior administrators, and was effective in instituting sustained practice change in a complex care centre. Ó 2014 Canadian Diabetes Association
Keywords: diabetes decision support clinical patient safety quality improvement quality improvement methodology
r é s u m é Mots clés : diabète aide à la décision clinique sécurité des patients amélioration de la qualité méthodologie de l’amélioration de la qualité
Objectif : Standardiser les pratiques liées à la prescription d’insuline, améliorer la prise en charge de l’hypoglycémie, réduire la dépendance à l’échelle d’adaptation aux doses, augmenter l’utilisation d’insuline selon le schéma basal-bolus et améliorer la sécurité des patients. Méthodes : Des patients souffrant de diabète ont été admis dans 2 unités pilotes pour patients hospitalisés, et tous les patients traités par insuline ont été répartis sur les unités de soins non critiques d’un hôpital d’enseignement canadien de soins tertiaires canadien comptant plusieurs campus. Les ordonnances préimprimées standardisées d’insuline et de prise en charge de l’hypoglycémie, les outils d’aide à la décision et les stratégies d’enseignement multidisciplinaire ont été élaborés, testés et mis en application au moyen du Model for Improvement et du Modèle d’utilisation de la recherche d’Ottawa. Les mesures cliniques et de l’équilibre ont été évaluées par la maîtrise statistique des processus. Résultats : La sécurité des patients a été améliorée par une réduction de l’hypoglycémie et une diminution de la dépendance aux échelles de correction. L’utilisation des ordonnances préimprimées s’est approchée de l’objectif de 70 % à la fin de la période d’essai et a été maintenue à 89 % collectivement 3 ans après la mise en application. Conclusions : La mise en application d’un ensemble de modèles d’ordonnances préimprimées standardisées favorise de meilleures pratiques de traitement par insuline, améliore la sécurité des patients et bénéficie largement du soutien des praticiens traitants. L’utilisation de la méthodologie officielle de l’amélioration de la qualité a promu l’efficience, amélioré la durabilité, accru le soutien aux cliniciens et
* Address for correspondence: Janine C. Malcolm, MD, FRCPC, The Ottawa Hospital-Riverside Campus, Division of Endocrinology and Metabolism, 1967 Riverside Drive, Ottawa, Ontario K1H 7W9, Canada. E-mail address: [email protected]. 1499-2671/$ e see front matter Ó 2014 Canadian Diabetes Association http://dx.doi.org/10.1016/j.jcjd.2014.01.003
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
119
aux gestionnaires principaux, et s’est avérée efficace à la mise en place d’un changement durable de la pratique dans un centre de soins complexes. Ó 2014 Canadian Diabetes Association
Introduction
Project phases
Diabetes management of hospitalized patients is increasingly recognized as an important quality and safety priority (1). Although diabetes itself is usually not the primary reason for hospitalization, it is a common comorbidity affecting approximately 30% of hospitalized patients (2). Both hyperglycemia and hypoglycemia in hospital are associated with increased morbidity, mortality, length of stay and resource utilization (3e7). Consistent with findings in the literature, the daily point prevalence of diabetes at our institution, a multicampus, tertiary care, teaching hospital, is approximately 25%. The average length of stay for this population prior to our quality improvement initiative was almost twice as long as that of those without diabetes. Diabetes management and maintenance of glycemic control in the hospital setting are challenging (3,8). Several factors, including changes in diet, physical activity, medications and the stress of acute illness, all negatively impact blood glucose stability. Insulin, the drug most commonly recommended and prescribed to treat diabetes in hospitalized patients, is classified by the Institute for Safe Medication Practices as a high-risk medication secondary to the high incidence of prescription and administration errors and adverse patient outcomes (9). Identified barriers to achieving glycemic control in hospitalized patients who are prescribed insulin therapy include an over-reliance on sliding-scale insulin, lack of routine basal and mealtime insulin, lack of patient-specific supplemental scales, uneven knowledge and confidence among clinical staff, competing clinical and institutional priorities, and lack of standardized practices for appropriate treatment of hypoglycemia (10). At our institution, audits of diabetes care and needs assessments by nurses, physicians and pharmacists echoed many of the challenges associated with insulin therapy described in the literature. Although many barriers to glucose management in non-critical care settings have been identified, some solutions have been proposed. Use of a basal-bolus approach to insulin therapy with a supplemental correction sliding scale and implementation of standardized insulin order sets have been shown to increase patient safety by improving glycemic control, reducing dependence on insulin sliding scales and decreasing hypoglycemia in hospitalized patients (1,11e14). In response to these findings, we successfully implemented a systematic, theory-based quality improvement initiative across all non-critical care units in our hospital. The purpose of this initiative was to improve patient safety and the process of diabetes management for patients treated with insulin through the implementation of pre-printed insulin orders, decision support tools, education and standardization of hypoglycemia management.
The quality improvement project consisted of 2 pilot phases involving 2 inpatient units to test, monitor and refine interventions followed by a multisite corporate roll-out. This phased approach enabled rapid learning and evolution of the project through small tests of change.
Methods Two complimentary theoretical frameworks, the Ottawa Model of Research Utilization (OMRU) and the Model for Improvement, were chosen to guide this quality-improvement initiative (15,16). A series of Plan-Do-Study-Act (PDSA) cycles were undertaken in order to inform, refine and test changes related to insulin therapy in actual clinical practice.
Ethics approval Ethics approval for this project was obtained from the Ottawa Hospital Research Ethics board. Patients and setting This project took place at a bilingual Canadian, multicampus tertiary care hospital with over 1100 beds and 47 000 patient admissions yearly. The cardiology and nephrology units were specifically selected as pilot locations to yield findings useful to the quality improvement process, including campus-specific variability, high volume of diabetes patients and a commitment to improving diabetes care. Inclusion and exclusion criteria Participants included all patients with either type 1 or type 2 diabetes admitted to a pilot unit during 1 of 2 intervention periods (intervention 1, a 6-week period; intervention 2, a 10-week period). Patients were included if they had been treated with subcutaneous insulin and had had a length of stay of 3 to 35 days. Patients were excluded if they had previously been admitted to the intensive care unit or had required IV insulin. Interventions Stakeholder engagement Engagement of partners, such as physicians, nurses, ward clerks, clinical program directors, inpatient unit managers, staff development trainers, forms developers and administrators was achieved through inclusion of representatives of these groups on our Corporate Inpatient Diabetes Quality Improvement committee. To enhance visibility within the institution, we were successful in having our initiative included in the Corporate Quality Improvement plan, and we actively searched for opportunities to communicate broadly the importance and progress of the project. A Corporate Inpatient Diabetes Quality Committee with broad, interdisciplinary and administrative stakeholder representation was established to guide and facilitate inpatient diabetes-practice improvements. A subset of this group, consisting of endocrinologists, internists, advanced practice nurse, diabetes nurse specialists, pharmacists, clinical managers, nursing educators and a qualityimprovement advisor was convened and mandated to focus specifically on improving insulin-ordering practices. Development of pre-printed insulin orders Informed by feedback, literature review, best practices and consensus expert opinion, a pre-printed subcutaneous insulin order form was developed. The goal of this tool was to standardize
120
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
Figure 1. Pre-printed subcutaneous insulin orders developed, implemented and evaluated using quality-improvement methodologies in a multicampus tertiary care centre.
insulin-prescribing practices, promote use of basal and mealtime insulin, reduce reliance on sliding-scale insulin as the sole form of diabetes treatment and standardize hypoglycemia management (Figure 1). To support the implementation of the pre-printed insulin orders, a diabetes medication administration record (MAR) and blood glucose monitoring form were revised and combined. This helped to facilitate assessment of patient response to treatment and pattern management, reduce transcription time, the potential for transcription errors and standardize hypoglycemia management.
Multidisciplinary education strategies To facilitate implementation of the orders, discipline- and recipient-appropriate educational sessions that addressed best practice for inpatient insulin therapy, management of hypoglycemia and instruction on use of the new standardized order form were developed. (Table 1) Education was facilitated by endocrinologists, medical residents, advanced practice nurses, diabetes nurse specialists, unit-based nurse educators and ward clerk trainers.
Implementation of pre-printed insulin orders on pilot units Two pilot units were identified to test and refine this process through participation in PDSA cycles over 2 pilot test periods. Pilot test period 1 Engagement of the pilot units in the first test cycle was achieved through a variety of strategies. Stakeholders on each unit were identified and discipline-specific “champions” for change were recruited. A baseline chart audit of 47 (21 cardiology, 26 nephrology) patients was conducted, and findings were shared with unit nurse managers, clinical educators, pharmacists and physicians. After 6 weeks, a post-intervention chart review and a user-satisfaction survey were undertaken. This process identified a broad range of opportunities to further improve the uptake of the orders and new forms. Pilot test period 2 In response to results and feedback in the first implementation period, the educational intervention and implementation process were modified (Table 2). Minor changes were made to the insulin order set to improve clarity and increase uptake. User education
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125 Table 1 Discipline and recipient appropriate learning sessions addressing inpatient insulin therapy, hypoglycemia management and instruction on the use of the new standardized order form Discipline Managers Ward clerks Nurses
Physicians
Specific education sessions provided General information provided concerning the goals, use and format of the orders 1:1 instruction on use and transcription of the form Facilitated discussions addressing challenges to hyperglycemia and hypoglycemia management Educational sessions on the format and use of the orders Case-based educational sessions on the use of the orders Sessions on approaches to glycemic management in the hospitalized patient
was enhanced with an emphasis on the goal of using the new forms for all insulin orders, and the new forms were relocated on each unit to improve access and promote use. Additional educational sessions focusing on the clinical impact of inpatient glucose control, case-based approaches to insulin dosing and selecting the appropriate correction scale were provided for physicians and nurses in the second test period. These sessions also reinforced the expectation that all orders be written on the new forms. Resident orientations to cardiology and nephrology were revised to include information related to insulin therapy. A supporting diabetes management pocket card with a stepwise approach for inpatient diabetes treatment, insulin action profiles, guidance for calculating doses and adjustments for special clinical situations was developed. This tool was distributed to medical residents and attending staff, and interested health professionals
121
on pilot units and was added as a quick reference tab on our electronic health record. During this implementation phase, the inpatient diabetes nurse specialists increased their availability on pilot units to provide additional educational support. They promoted the use of the forms and reinforced the goal of the project. This second test period lasted 10 weeks. Corporate spread Following the pilot phase, the pre-printed insulin orders, educational sessions and pocket card were implemented across all non-critical care inpatient units. Strategies successfully utilized in the pilot phases were implemented to support corporate spread. They included involvement and active support of leaders with reputation and position (i.e. the vice president of medical affairs, clinical and nursing professional practice directors, directors of laboratory medicine and pharmacy, physician champions and chief residents) and the provision of a range of educational opportunities for physicians, nurses and clerical staff tailored to their learning needs. Communication of the goals, targets and expectations of the project to clinical staff occurred via email from senior administration (the vice president of medical affairs) through news articles in the hospital journal and through department meetings. Endocrinologists, advanced practice nurses and nurse specialists continued to be available for clinical rounds and unit-based learning. However, to distribute the training load and encourage ownership by all programs, the trainer approach was used to include unitbased clinical nurse educators and residents in the training process. Timely, ongoing information related to utilization rates of the new insulin order forms and unit-specific progress was shared with all physician chiefs and clinical managers.
Table 2 Barriers to uptake of order form and factors that potentially contributed to increased rates of hyperglycemia Barriers to uptake Lack of understanding of the purpose of the orders
High turnover of medical trainees Clinicians “not aware” of the orders
Access and availability of the order form Perception by clinicians that glycemic control in hospital was not achievable Fear of hypoglycemia and lack of confidence in how to safely adjust insulin to optimize glycemic control Lack of understanding that all insulin orders should be completed using the pre-printed orders Transcription times and potential for transcription errors would be increased Factors potentially contributing to hyperglycemia Omission of meal-time insulin Inappropriate and overly conservation selection of correctional scales
Lack of adjustment made to routine insulin doses despite inadequate glycemic control
Strategies implemented to improve uptake Unit-specific baseline findings and opportunities for improvements shared Discipline-specific educational sessions modified to emphasize goal of project and impact on patient safety Use of leaders of reputation and position (physician champions, chief residents, directors of nursing, clinical nurse educators, diabetes nurse specialists) to promote the orders Increased frequency of educational sessions Incorporation of sessions into orientation programs for new trainees Increased frequency of educational sessions Signage/posters created for workstations Communication via e-mail, meetings and rounds Reminders from pharmacists Units evaluated for optimal location to store order form Process developed to ensure adequate supply with unit clerks, nurse managers and physicians Modification to discipline-specific education sessions to address perception and enhance confidence Development of diabetes pocket card with a stepwise approach to achieving glycemic control Shared improvements of others with teams Development of diabetes pocket card Incorporation of case-based examples into teaching sessions Increased presence of diabetes nurse specialist on pilot units as educational resource to clinical staff Increased emphasis on purpose and goals of project during educational sessions Audit and feedback of order form usage to pilot units Development of supporting policies and procedures Revise diabetes MAR to include pre-printed sections for hypoglycemia orders, supplemental insulin scales and blood glucose monitoring orders Strategies to improve glycemic control Modification to educational sessions to include additional discussion on need for meal-time insulin and importance of proactive approach to diabetes management Modification to pre-printed insulin orders with removal of “very low” correctional scales Case-based examples during educational sessions Modification to diabetes pocket card to reflect changes in order set and inclusion of suggestions for appropriate correction scale choice Increased emphasis during educational sessions that insulin doses need to reassessed frequently and adjusted accordingly
122
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
Table 3 Measures used to evaluate improvement with implementation of standardized pre-printed insulin orders Measures of Improvement Process Percentage of insulin orders written using pre-printed orders measures Percentage of orders completed correctly Percentage of orders transcribed correctly to medication administration record (MAR) Percentage of exclusive use of sliding scale for >48 hours Percentage of episodes of hypoglycemia managed according to standards of practice Clinical Percentage of stay days classified as high days (1 or more measures BG >11 mmol/L). Percentage of stay days classified as low days (1 or more BG <4 mmol) Balance Consultation rates to endocrinology Measures Consultation rates to diabetes nurse specialists User satisfaction with new insulin orders, supporting resources
target was met and the process was stable. A point prevalence audit of pre-printed insulin order use across all 3 campuses was undertaken 3 years post-implementation to evaluate sustained utilization rates. Microsoft Excel and Chart Runner software (v 3.6) were used to analyze data. Surveys were undertaken to elicit user satisfaction with the new process, forms and decision-support tools. They were distributed to physicians (attending and resident physicians), nurses, pharmacists and ward clerks following each of the pilot test periods and 6 months after initiation of corporate spread. Various strategies were used to collect feedback, including semi-structured telephone and face-to-face interviews, anonymous electronic or written surveys, and facilitated unit-based user-group discussions. Results Pilot process measures
Measures of improvement Table 3 outlines clinical outcomes, process measures and balance measures evaluated during the pilot periods. A target of 70% of all subcutaneous insulin orders written on the new insulin order form was established a priori as an appropriate target for uptake of the pre-printed insulin orders. This goal recognized that a small proportion of patients would require an unusual or unique treatment regimen that the pre-printed form would not easily support and that in these situations a generic medication order form should continue to be used. Data collection and analysis The data collection strategies included standardized chart reviews, audits of order forms processed by the inpatient pharmacies and user surveys. The standardized chart reviews were completed by a clinical expert and audited by the advanced practice nurse to ensure consistent data collection and classification. Any discrepancies were reviewed together and consensus was achieved by team members. The data-collection plan was designed to support the use of statistical process control through development and analysis of Shewhart control charts and run charts. This enabled tracking of both the progress and the stability of the practice changes and identification of change related to special causes. When statistical process control was not feasible, a before-and-after study design was used. Data were collected for a 6-week period to establish a baseline and then every 2 weeks during the pilot phases. Corporate spread data were collected weekly for 17 weeks until the usage
During the first pilot test period, uptake of the new order forms was only 11% on the nephrology unit and 38% on the cardiology unit. With additional education, development and dissemination of decision-support tools and improved access to the form, the utilization rates increased to a sustained rate of 68% and 74%, respectively, after the second pilot test period. At the end of the second test period, the new pre-printed order forms were completed appropriately 80% to 88% of the time, and in all cases the orders were transcribed to the diabetes MAR correctly. Dependence on the insulin sliding scale as the only therapy was reduced from 30% to 4% on cardiology and from 32% to 0% on the nephrology unit, and the use of basal and meal-time insulin increased on both units (Figure 2). Pilot clinical measures The clinical impact of the pre-printed orders was assessed during the pilot phase of this improvement project. The rate of hypoglycemia (capillary blood glucose reading<4 mmol/L) was reduced after implementation of the intervention (Figure 3). No serious hypoglycemia was observed with use of the pre-printed form and supplemental insulin-adjustment scales. The number of high blood glucose days (2 or more documented blood glucose readings over 11 mmol/L in 24 hours) did not improve on either unit (Figure 3). Review of cases of hyperglycemia indicated that this was due to a variety of factors, including underprescription of meal-time insulin for patients who were eating regularly, the selection of an inappropriately conservative insulin-
Figure 2. Insulin therapy on pilot units at baseline and after 2 PDSA cycles of change.
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
123
Figure 3. Glycemic control on pilot units before and after implementation of standardized approach to insulin therapy after second test period.
correction scale or failure to increase basal insulin despite a persistent need for correction insulin. Pilot balance measures The intervention did not lead to increased consultation requests to the endocrine services. There was a reduction in the number of consults to the diabetes nurse specialists on both pilot units (cardiology 35% vs. 17%; nephrology, 12% vs. 6%) and no increase in the number of consults to the endocrine service on either unit (cardiology 35% vs. 17%; nephrology 16% vs. 16.7%). Satisfaction with the pre-printed orders was high. A total of 13 pilot unit physicians and 64 nurses were surveyed. Of them, 92% of physicians and 74% of nurses indicated that the orders were clear, and 85% of responders indicated that the pre-printed orders improved the diabetes treatment received by patients. The chart audit, surveys and feedback sessions following the first pilot test period identified a broad range of modifiable barriers to the uptake of the orders that were successfully addressed during the second improvement cycle on pilot units (Table 2). Corporate spread Following the pilot phase, the orders were spread corporately. Use of the pre-printed order form was monitored through weekly audits of all subcutaneous insulin orders received by the pharmacy over a 17-week period (Figure 4). Uptake of the new orders was progressive. After the first 4 months of the corporate launch, 60% to 70% of subcutaneous insulin orders were written on the new preprinted insulin forms. User satisfaction with the pre-printed orders for physicians, nurses and ward clerks was assessed 6 months after implementation through a short anonymous electronic survey. Of those surveyed (n¼54), 88.7% recommended continued use of the form, and 92.5% rated the form as useful to their practice. A point-prevalence audit of all subcutaneous insulin orders written at the hospital in a 24-hour period was performed 3 years after implementation to evaluate sustainability. The pre-printed insulin order form was used for 89% of orders written, surpassing the 70% target. Discussion Diabetes management using insulin therapy for hospitalized, non-critical-care patients continues to be a challenge for many practitioners and institutions. Despite the complexity of effecting multidisciplinary change in a large academic setting with rotating trainees, we were successful in designing and implementing preprinted insulin orders and accompanying decision support aids
that improved patient safety and led to sustained uptake of best practices. Improvements in patient safety may be attributed to a decrease in hypoglycemic events, fewer transcription errors, standardization of hypoglycemia treatment and reduced reliance on sliding-scale insulin as the sole insulin therapy. Reduced hypoglycemia was likely secondary to increased prescription of both long-acting basal insulin and rapid-acting meal-time insulin with the use of a more individualized correction dose when blood glucose was outside target. These findings resonate with those of others (3,11). It is also important to note that these improvements occurred through changes in the practices of non-diabetes specialists and without an increase in the demand for endocrinology or diabetes nurse specialist consults. Feedback from nurses, physicians and pharmacists indicated that confidence in treating diabetes using insulin was increased through this initiative. Although we observed a decrease in the frequency of hypoglycemia, an unanticipated small increase in the number of days with 2 blood glucose readings over 11 mmol/L was observed. Because we did not compare actual blood glucose readings, it is not possible to know how the overall degree of blood glucose control was impacted. A trend of increased hyperglycemia has previously been reported in other studies where standardized protocols for insulin dosing have been implemented (17,18). This increase has been attributed to poor protocol compliance and increased steroid use. In contrast to these reports, we obtained high levels of compliance with the use of the pre-printed insulin orders. It is possible that the increased rate of hyperglycemia we observed on the nephrology unit during the pilot was related to post-renal transplant steroid therapy. To better understand this finding, we undertook additional review of the pilot cases with frequent hyperglycemia. This review suggested that the increase in hyperglycemia may have been due to the omission of meal-time insulin orders or the selection of an inappropriately conservative correctional scale (for fear of hypoglycemia). Additionally, when a pattern of recurrent use of correction insulin dosing occurred, this additional insulin was not appropriately incorporated into the usual proactive insulin dosing. This analysis underlines the complexity and dynamic nature of insulin requirements during hospital admissions. It has been helpful in identifying further learning opportunities related to pattern management for insulin prescribers at our institution. Given the current evidence supporting the negative impact of hyperglycemia in acute illness, this remains an important finding for further investigation. It is also important to note that use of the standardized insulin orders was highly supported by treating practitioners. This positive response and ongoing support of the new insulin orders helped to facilitate uptake and increased awareness that inpatient
124
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125
Figure 4. Corporate uptake of pre-printed insulin orders for the first 17 weeks after implementation.
diabetes management is an important, shared patient quality and safety issue. These tools made diabetes management easier for practitioners. Our improvement experience has additionally provided some learning related to the implementation of complex change that may help others who are contemplating such a project. The OMRU was helpful in enabling an understanding of the context of the planned change and in supporting the marketing of the proposal throughout the project. While use of OMRU assisted us to understand and manage our larger environment, the Model for Improvement provided more detailed guidance for the PDSA cycles required to inform, refine and test changes related to insulin therapy in actual clinical practice and provided us with the tools to determine whether our change was resulting in an improvement. Attributes of the planned innovation, the characteristics of potential adopters (i.e. current practices, knowledge, skill, attitudes and beliefs) and the practice environment (culture, relationships, priorities, patients and resources) were all considered early in the process. Early engagement of potential adopters, key stakeholders and opinion leaders fostered trust that enabled a transparent establishment of shared goals, awareness of competing demands and the needs of each partner to find benefit associated with the proposed change. Use of project teams and charters enabled all partners to understand their roles in the process and enhanced ongoing commitment and accountability. Maintaining timely, respectful and clear communication was key to understanding the issues and helped to avoid conflict. Finally, in working in an academic setting, team members had to learn and value the differences between research and quality-improvement methods. Having access to a trained quality-improvement advisor was a valuable asset in this area. We identified 2 recurrent and serious challenges to our success that may be common to other teams considering this type of quality-improvement project. Ensuring easy access to forms was critical to their uptake by busy practitioners; however, the best place to locate the forms was not always intuitively evident and differed among disciplines and practice settings. This challenge was overcome largely through detailed mapping of the process of writing and processing the orders on target units and then determining the most suitable location for the forms.
The second and more sensitive challenge was identification and acknowledgement of attitudes, beliefs and vulnerabilities related to the diabetes management of the hospitalized patients. These included fear of severe hypoglycemia, lack of confidence related to insulin adjustments and beliefs that improvement in glycemic control in hospital was not possible, based on past experience (Table 2). These challenges are not unique and have been noted in other studies (10). Identifying these barriers early through solicitation of anonymous feedback, developing alliances and non-judgemental relationships, and tailoring education and information to the needs of the recipients were important in reducing these barriers successfully. Through our improvement efforts, we have demonstrated that implementation of a standardized, pre-printed insulin order set facilitates best practices for insulin therapy, improves patient safety and is highly is supported by treating practitioners. The utilization of formal quality-improvement methodology promoted efficiency and was effective in instituting sustainable change in a complex environment. Acknowledgements This work was supported by The Ottawa Hospital Quality Improvement Plan. References 1. Umpierrez GE, Hellman R, Korytkowski MR, et al. Management of hyperglycemia in hospitalized patients in non-critical care setting: an endocrine society clinical practice guideline. J Clin Endocrinol Metabol 2012;97:16e38. 2. Cook CB, Kongale GL, Potter DJ, et al. Inpatient glucose control: a glycemic survey of 126 US hospitals. J Hosp Med 2009;4:E7e14. 3. Clement S, Braithwaite S, Magee MF, et al. Management of diabetes and hyperglycemia in hospitals. Diabetes Care 2004;27:553e91. 4. Campbell KB, Braithwaite SS. Hospital management of hyperglycemia. Clin Diabetes 2004;22:81e8. 5. Umpierrez GE, Isaacs SD, Bazargan N, et al. Hyperglycemia: an independent marker of in-hospital mortality in patients with undiagnosed diabetes. J Clin Endocrinol Metab 2002;87:978e82. 6. McAlister FA, Majumdar SR, Blitz S, et al. The relation between hyperglycemia and outcomes in 2471 patients admitted to hospital with community-acquired pneumonia. Diabetes Care 2005;28:810e5. 7. Krinsley JS. Association between hyperglycemia and increased hospital mortality in a heterogeneous population of critically ill patients. Mayo Clin Proc 2003;78:1471e8.
M.-A. Doyle et al. / Can J Diabetes 38 (2014) 118e125 8. Inzucchi SE. Clinical practice: management of hyperglycemia in the hospital setting. N Engl J Med 2006;355:1903e11. 9. Patient Safety Alert. “High-alert” medications and patient safety. Int J Qual Health Care 2001;13:339e40. 10. Korytkowski M, Dinardo M, Donihi AC, Bigi L, Devita M. Evolution of a diabetes inpatient safety committee. Endocr Prac 2006;12(Suppl 3):91e9. 11. 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;30:2181e6. 12. Trujillo J, Barsky E, Greenwood B, et al. Improving glycemic control in medical inpatients: a pilot study. J Hosp Med 2008;3:55e63. 13. Maynard G, Lee J. Improved inpatient use of basal insulin, reduced hypoglycemia, and improved glycemic control: effect of structured subcutaneous insulin orders and an insulin management algorithm. J Hosp Med 2009;4:3e15.
125
14. Schnipper J, Ndumele CD, Lian CL, et al. Effects of a subcutaneous insulin protocol, clinical education and computerized order set on the quality of inpatient management of hyperglycemia: results of a clinical trial. J Hosp Med 2009;4:16e27. 15. Logan J, Graham I. Towards a comprehensive interdisciplinary model of health care research use. Sci Commun 1998;20:227e46. 16. Langley GJ, Moen RD, Nolan KM, et al. A practical approach to enhancing organizational performance, 2nd ed. Jossey-Bass Books, 2009, pp 23e5. 17. Noschese M, Donihi AC, Koerbel G, et al. Effect of a diabetes order set on glycemic management and control in the hospital. Qual Saf Health Care 2008;17: 464e8. 18. Chen HJ, Steinke D, Karounos DG, et al. Intensive insulin protocol implementation and outcomes in the medical and surgical wards at a Veterans Affairs medical centre. Ann Pharmacother 2010;44:249e56.