Clinical Decision Support for Diabetes Care in the Hospital: A Time for Change Toward Improvement of Management and Outcomes

Abstract

The increasing prevalence of diabetes permeates hospitals and dysglycemia is associated with poor clinical and economic outcomes. Despite endorsed guidelines, barriers to optimal management and gaps in care prevail. Providers’ limitations on knowledge, attitudes, and decision-making about hospital diabetes management are common. This adds to the complexity of dispersed glucose and insulin dosing data within medical records. This creates a dichotomy as safe and effective care are key objectives of healthcare organizations. This perspective highlights evidence of the benefits of clinical decision support (CDS) in hospital glycemic management. It elaborates on barriers CDS can help resolve, and factors driving its success. CDS represents a resource to individualize care and improve outcomes. It can help overcome a multifactorial problem impacting patients’ lives on a daily basis.

In alignment with the increasing prevalence of diabetes mellitus, the number of hospital discharges with diabetes listed as a diagnosis increased from 2.8 to 5.5 million from 1988 to 2009 in the United States. In 2014, 7.2 million hospital admissions were incurred by adults 18 years or older with diabetes. ¹ Hyperglycemia and hypoglycemia in the inpatient setting is common ² and associated with poor clinical and economic outcomes including a greater risk of morbidity, mortality, hospital length of stay, and use of hospital resources which has been comprehensively summarized in consensus statements, practice guidelines, and standards of medical practice.^3-5 There are evidence-based endorsed strategies for the management of dysglycemia in the hospital that are considered best clinical practice. Benefits of strategic action for glycemic control are also supported by scientific evidence on clinical and economic outcomes.^3-5 Van den Berghe and colleagues ⁶ demonstrated that controlling hyperglycemia in critically ill surgical patients to a target glucose range between 80 and 110 mg/dL (4.4-6.1 mmol/L) resulted in a reduction in mortality by approximately 40% in comparison to conventional glucose targets between 180 and 215 mg/dL (10-12 mmol/L). This evidence guided glycemic goals and practice guidelines for the critically ill patient for several years, which was reconsidered following the findings of the Normoglycemia in Intensive Care Evaluation and Survival Using Glucose Algorithm Regulation (NICE-SUGAR) trial in 2009. ⁷ This multinational, multicenter, randomized trial showed that the intensive glycemic control arm treated to a glucose target between 80 and 110 mg/dL (4.4-6.1 mmol/L) provided no significant benefit in comparison to less stringent glucose control between 140 and 180 mg/dL (7.8-10.0 mmol/L). The intensively treated group experienced a small but significant increase in mortality as well as an increase in hypoglycemic rates.^7,8 Additional substantial evidence from meta-analyses including the NICE-SUGAR trial support that stringent glycemic control results in higher rates of hypoglycemia and has either no benefit, or an increase in the risk of mortality in comparison to moderate glucose control.^9-11 Therefore, recommended glucose targets are now 140-180 mg/dL (7.8-10.0 mmol/L) for most critically ill patients, and insulin therapy is recommended to manage persistent hyperglycemia ≥180 mg/dL (10.0 mmol/L).^3,5

More stringent glucose target between 110 and 140 mg/dL (6.1-7.8 mmol/L) is considered appropriate for certain populations such as cardiac surgery patients and critically ill postoperative patients. However, in non-intensive care settings, glucose targets or the effect of intensive glycemic control on outcomes has not been examined through randomized control trials. Observational studies do suggest a strong association between hyperglycemia and poor clinical outcomes among medical and surgical patients including infectious complications, longer hospital stay, and mortality.^3,12 The recommendation for non-critically ill patients is fasting glucose ≤140 mg/dL and random or postprandial ≤ 180 mg/dL.^{3, 4}

Diabetes leading societies and The Joint Commission on the Accreditation of Healthcare Organizations strongly recommend that hospitals implement programs to optimize the management of inpatient hyperglycemia, hypoglycemia, and glycemic variability and to facilitate a framework of quality in the hospital and the continuum of care.^3-5,13 Despite recommendations and the state of evidence, glycemic control and diabetes management in the hospital are often regarded as suboptimal. Regardless of providers’ and organizations’ efforts, barriers to managing dysglycemia and diabetes in the hospital abound. These barriers added to the many tasks of demanding systems of practice limit clinicians’ focus on addressing glycemic issues. There is often much complexity in the evaluation of diabetes, in the assessment of glucose control, and in the appraisal of response to insulin treatment in acutely ill patients. ¹⁴ Existing gaps in the management of diabetes relate to various domains, including knowledge, attitudes, and decision-making. Barriers are often related to inadequate initiation and adjustment of insulin regimens; failure to recognize glucose abnormalities; inadequate communication, education, or coordination of care; poor perception of the relevance of glucose problems in the hospital; missing documentation of dysglycemia; and poor sustainability of practice behaviors. These barriers can potentially jeopardize acute management and continuity of care following hospital stay.^15-19

Understanding what drives these multiple barriers in practice, and identifying strategies to solve them is a work in progress. This understanding has been informed by surveys, quantitative and qualitative studies trying to address this conundrum. In the realm of education and competencies, suboptimal management of dysglycemia may seem related to providers’ deficits. A needs assessment study of diabetes care in the hospital showed shortfalls in the domains of knowledge, attitudes, clinical decision-making, confidence, and familiarity with hospital resources among clinicians. ¹⁵ While knowledge deficit in inpatient glycemic management is widely recognized among various disciplines, it remains questionable whether the knowledge gap indeed correlates with poor glucose control of hospitalized patients. It is also known that knowledge gained during diabetes education programs may wane over time, which may potentially contribute to practice inconsistencies. A review of learning programs highlights that knowledge and confidence in diabetes management increase with the implementation of educational programs, but when faced with complex tasks, providers are less confident in their management decisions. Educational programs facilitate clinical actions leading to greater patient safety, glucose monitoring, and improvements in insulin selection and use. However, suboptimal practices such as persisting on the use of insulin sliding scales as monotherapy prevail despite education. ¹⁶ These findings suggest that education alone is unlikely to address the multifaceted responsibility of diabetes care for hospitalized patients. Another important factor that adds complexity to the matter and is superimposed on the demanding work of hospital clinicians is the task of gathering and interpreting glycemic data from different sources in the electronic medical records. Searching and interpreting glucose patterns and their relation to insulin treatment need to take place daily in complex systems of practice. ¹⁴ Such a situation may be perceived as an overpowering pursuit for glucose management that often results in unaddressed scenarios in clinical practice.

Systems of practice are interjected with obstacles which are in need of innovative solutions. Irrespective of limitations related to complex workflows, educational gaps, or clinical inertia, meeting safe and effective patients’ care goals and maintaining organizational accountability is the mission of most healthcare institutions. Therefore, planning and executing interventions to manage dysglycemia in the hospital needs to encompass innovative strategies that include but do not entirely depend on providers’ knowledge and performance for their execution. These strategies need to support hospitals by lessening the limitations of complex and taxing healthcare systems while promoting learning, optimal practice, and good clinical judgment. This perspective aims to provide substantiating evidence of the benefits of clinical decision support (CDS) in the management of diabetes and dysglycemia in the hospital setting. This report examines barriers that hinder practice and factors that may promote CDS systems’ effectiveness.

In the evolving landscape of health systems, CDS is becoming a critical tool for care. CDS entails the use of person-specific information that is categorized and presented in the appropriate context of practice to aid in decision making, and to enhance the care provided to patients and improve their outcomes. ²⁰ Health Informatics as a discipline, and electronic medical records as resources are infrastructures for the integration of decision support tools in practice which can facilitate innovative approaches to enable processes and sustain programs to improve care. There is a growing body of evidence demonstrating the positive impact of clinical decision support systems on healthcare processes, adoption of guidelines, clinical outcomes, and economic outcomes. ²¹ There are defined blueprints for the development, optimization, and use of CDS. ²¹ A scoping review of systematic reviews examined the effects of CDS systems on diabetes care in ambulatory settings. A significant impact on the process of care was found in 82% of the studies that examined this effect. Additionally, evidence of impact on patient outcomes was recognized in 31% of the studies that examined this effect. A concern exists related to the lack of consistency in the reporting of studies and methodological quality of CDS systems which could be concealing the accuracy of effectiveness in improving diabetes care. ²² Adoption of CDS in the ambulatory setting relying on providers and patient's participation are proliferating. CDS helps personalize diabetes management by adapting therapy in response to glucose data. It can also help address variability in responses to glucose management and counteract limited expertise in management. ²³ Acceptance and adoption of CDS likely play an important role in its effectiveness. Proof of concept for the investigation of efficacy, safety, and user’s acceptance of decision support in ambulatory participants receiving home health care is undergoing. ²⁴ More evidence is needed to determine whether CDS for diabetes in the ambulatory setting influences rates of hospitalization or glycemic control achieved during hospital stays.

Hospital-based CDS programs for diabetes management have been provided through computerized physician order entry systems, case finding of in-need patients, integrated algorithms for insulin management and clinical recommendations, and technology-based intervention of point of care blood glucose transfer for analysis.^25-30 Improvement in glycemic control reflecting mean glucose, time in target, and frequency of glycemic abnormality among hospitalized patients has been effectively demonstrated through the use of insulin dose computerized calculators, electronic order templates, insulin management algorithms, automated insulin dosing protocols, and glycemic control reminders in contrast to conventional management. ²⁶ Achieving target glycemia safely and effectively using CDS with high acceptability and adherence to recommendations by providers has been demonstrated. ³¹ Applying continuous glucose sensor data for decision-making and analysis integrated with CDS has been shown. ³¹ Innovative approaches capitalizing on electronic medical records data to implement CDS can help reduce barriers to care and promote safer and more effective practices. Real-time CDS tool synchronized to clinicians’ workflow that recognized gaps in glycemic control and management and provided evidence-based recommendations for management considerations improved practice performance and reduced hyperglycemic events in hospitalized patients. ¹² The future appears promising for the use of CDS. How it will continue to enhance patient-centered clinical decision-making, optimize care, and improve clinical and economic outcomes is of great interest and promise

Electronic Health Records adoption and evidence of benefits are progressively increasing since the inception of the Health Information Technology for Economic and Clinical Health Act in 2010. The Office of the National Coordinator for Health IT supports the nationwide implementation of health IT resources including CDS. Among its various goals are to attain better clinical outcomes, to strengthen population health results, and to generate robust research data on health systems. ²⁰ Factors that can promote successful implementation of CDS include adequate validation of programs, evidence and knowledge-based assimilation, users’ feedback, widespread implementation in collaboration with stakeholders, and consistent evaluation of programs' impact. The time is just right and calls for significant and sustainable efforts toward optimization of diabetes care in the hospital and transitions of care. Firm steps toward designing comprehensive programs evaluating the impact of CDS on hospital diabetes care and corresponding clinical and economic outcomes will help overcome a multifaceted problem impacting the lives of millions of people on a daily basis.