Business Intelligence for Better Insulin Therapy Management

Elevating Diabetes Care

Navigating Computerized Physician Order Entry (CPOE) Challenges

Main Line Health addressed the challenges of physicians circumventing CPOE guidelines for insulin therapy selections

Strategic Implementation of Endocrinology Dashboards

Leveraging Dimensional Insight Clinical Analytics Software, Main Line Health strategically implemented Endocrinology Dashboards for a profound analysis of insulin regimen ordering trends.

Efficient Patient Care

Patients on basal-bolus regimens experienced a shorter median length of stay, showcasing the positive impact on both clinical outcomes and care efficiency.

Optimized Glucose Management

Implementation of advanced basal-bolus regimens, guided by analytics, resulted in significantly improved glucose management for patients.

Physician Education and Training Identification

Analytics pinpointed physicians using sub-optimal insulin regimens, providing valuable insights for personalized education and training programs.

Personalized Coaching for Physicians

Drill-down capabilities empowered diabetes educators to provide personalized coaching and feedback directly to physicians based on detailed analytics.


CPOE guidelines for insulin therapy selections were being circumvented.


Endocrinology Dashboards that analyze insulin regimen ordering trends and demonstrate impact of insulin regimens on patient outcomes and care efficiency.


Better glucose management with basalbolus insulin regimens.

Shorter length of stay for patients on basalbolus regimens.

Identify physicians who could benefit from education and training


One out of every twelve Americans is afflicted with diabetes. Controlling inpatient diabetes is important to avoid complications and achieve efficient care. Insulin therapy for diabetes patients has changed over time, from the “old” sliding scale to newer basal-bolus regimens.

Main Line Health uses the Computerized Physician Order Entry (CPOE) system of their Siemens Soarian® EMR to guide physicians away from sliding scales insulin and towards basal-bolus regimens. While CPOE makes it harder to order sliding scales therapies, physicians nonetheless were using CPOE to “build” sliding scales and circumvent decision support interventions.

What Main Line Health needed was clinical intelligence to analyze the extent to which physicians were using sub-optimal insulin regimens. Physician leadership also wanted to demonstrate the impact insulin regimens had on patient outcomes and identify ways to coach physicians toward better diabetes management.

Leveraging the Diver Platform for Clinical Analytics

EvergreenHealth is committed to achieving the “Quadruple Aim” in healthcare: improving population health, improving patient care, reducing the per capita cost of care, and improving the work-life of its providers and staff. As a part of this, the organization has recently focused on what its leadership calls the “cost and revenue imperative.” In this initiative, departments are tasked with examining their processes and costs in order to figure out where they could
potentially save money in their operations.

One of the initiatives that has grown out of this directive is called care optimization. As a part of that, the health system is focusing on length of stay—examining where it is in relation to its peers, and where there are outliers that the organization could work to improve.


Complexities of Clinical Data

The EA team represented a collaboration of endocrinology physicians, patient educators, IT Analytics and the Hospitalist Service for all Main Line Health facilities. One challenge the team immediately faced was preparing and analyzing the clinical data to create actionable insights. Clinical data is complex and requires sophisticated algorithms and business rules to pre-process the data. Groupings and filters are needed extensively to evaluate multiple patients’ insulin orders and classify whether the physician ordered a sliding scale or any form of basal-bolus insulin therapy. Timing algorithms are used to determine the initial insulin treatment time-window and also pre- and post-glucose levels. Attribution algorithms are necessary to determine which physician is responsible for ordering and management of the patient.

Using Diver and its powerful data integration, business rules and modeling capabilities, the IT Analytics team was able to overcome this hurdle. Jamie Mitchell, Senior Applications Specialist noted, “Once we had access to our data sources, we were able to pull in and aggregate the data, create our business rules and calculations to transform the data, and build the data models that feed the EA dashboards.”

Communicating Insights

The interactive EA dashboards present physician leadership with:

  • Physician insulin regimen ordering patterns — identifies physicians who continue to order sliding scale insulin regimens vs. basal-bolus regimens. Views are presented for each of Main Line Health’s facilities with drill-down to the physician levels and trends over time.
  • Impact of insulin regimen on glucose trends — analyzes which regimens have the strongest positive impact on glucose levels, and demonstrates to physicians that the regimen choice has an immediate and measurable impact on patient care.
  • Length of stay — demonstrates the impact of the insulin regimen not only on quality but care efficiency.
  • Drill-down to ordering or attending physician — enables diabetes educators to personalize physician training and give direct, physician-specific feedback and coaching.
  • Ordering behaviors trends — shows the impact of physician education campaigns.

Demonstrable Benefits

The EA dashboard demonstrates the benefits of newer, basal-bolus over “sliding scales” insulin regimens. Physicians reviewing results can quickly see that for the majority of cases, a basal-bolus insulin regimen results in better glucose management.

The dashboard also shows that patients with a basal-bolus insulin regimen have an equal or shorter median length of stay, providing insights into improving both clinical and efficiency measures. The EA dashboard highlights large variances in physician ordering patterns to easily identify physicians who could benefit from in-service training.

Download case study


Main Line Health




Bryn Mawr, PA


4 acute care hospitals;

1,200 beds


Siemens Soarian



Diver Platform

Download case study