Virginia Mason Franciscan Health Average Cost per Case Reduction Project

Healthcare

For Virginia Mason Franciscan Health, one of the key components of scheduling surgery is the Preference Card, i.e., a list of all the supplies and instruments that a surgeon needs to complete a case. Our goal was to identify the cost-saving opportunities considering many outliers and variations in such supply costs between surgeons, locations, and procedures.
The team developed the analytical solution by analyzing the trends and delivering a prototype of the interactive dashboard to trace and identify such cost variations. The team also built a prediction calculator to predict the supplies cost for each surgical case to determine the optimal inventory levels by forecasting and planning the budget while meeting the demand.

#CostAnalysis, #HealthcareAnalytics

Virginia Mason Franciscan Health

Mission Control Artificial Intelligence and Analytics:

Healthcare

Virginia Mason Franciscan Health (VMFH) Mission Control Command Center has identified the ‘Ancillary Services Department’ is impacted by a delay in services provided to patients. Patients are waiting above the recommended time to receive their ancillary services. Command Center tasked Team A1 MSBA Consultants to perform an analysis on ‘Care Progression’ data, with regards to ancillary services to identify the overall turnaround time for each service, the total delay for each service, and/or recognize any patterns or trends to build an effective analytical solution. Measures included in the analysis are the hospital location, department, urgency for level of care, type of unit, volume on certain days/times, and the type of service requested.

#Operations/Optimization

Give InKind

Technology

Give InKind is a rising Tacoma-based startup with an empowering and comprehensive platform design that has transformed the way friends, families, and community members organize giving and receiving support during life’s challenging moments.   

With over 2 million users who have come to Give InKind to navigate actionable or financial support for loved ones, the UW MSBA team leveraged Give InKind’s data to identify insights to help answer overarching business opportunities: Where is help needed the most? What drives the virality of support?

The UW MSBA team’s multidimensional approach assessed user demographics, intangible prosocial support such as acts of service, and tangible prosocial support such as purchased products for recipients. These insights were delivered through multiple techniques including descriptive and diagnostic analytics configured by automated dashboards using Tableau and Google Analytics, predictive and prescriptive modeling of prosocial support using Regression and Classification methods, and Sentiment Analysis. In addition, industry and marketing research was conducted for project validation and community-based partnership development.

#UserAnalytics, #Operations/Optimization, #MarketingAnalytics

HeyKiddo

Sound Credit Union

Banking and Financial Services

Sound Credit Union (SCU) is one of the largest credit unions in Washington State. SCU serves 140,000 members at 29 full-service branches and online. As a credit union, any earnings that SCU makes are passed on to its members. As a result, credit unions face high rates of churn (the rate of attrition or members leaving). This is due to several factors including but not limited to the number of competitors, members’ ability to substitute products, membership restrictions, etc.

The purpose of this project is to carry out descriptive, diagnostic analysis of different variables and to create a predictive model to identify patterns in the member churn. From this, create a machine learning model to predict whether a member is going to churn or not.

#Finance/Accounting, #Customer Segmentation to define CLTV

CommonSpirit Health

Healthcare

CommonSpirit is the largest nonprofit health system in the U.S and is dedicated to advancing health for all people, advocating for those who are poor and vulnerable. They deliver clinical excellence across a system of 140 hospitals and more than 1,000 care sites in 21 states. The objective of the project is to conduct an evaluation of the differences in utilization patterns by patient demographics, geographic and practice setting, COVID-19 penetration, and community social vulnerability. The four phases of the project are business understanding, descriptive & diagnostic analysis, predictive analysis, Cognitive analysis, and deployment. The analysis provided by the MSBA project team will further allow them to identify potential opportunities to support and sustain equitable access to care across the health care system.

#TelehealthService

UW Tacoma

Education

University of Washington Tacoma (UWT) is a distinctive sized university with approximately 4600 undergraduates and 800 graduate students. Like many similar institutions of higher education, UWT is experiencing challenges with student retention. 

As such, UWT would like to better understand the precursors to undergraduate students leaving before completing their degree program and if there are any retention holes.

We generated valuable insights through descriptive and diagnostic analytics in Tableau. After some feature engineering, our data was fully ready, and we could identify factors affecting attrition and predict the at-risk students.

However, we realized there are many other factors that could affect attrition. But there was no data available on these aspects. We thought a little differently to come up with a cognitive analytical solution – a short cyclical survey for the students integrated with My UW or Canvas platforms. With the results of the survey UWT could proactively connect with these students to understand and address their area of concern. This would ultimately result in the increase of student retention rates.

#HR Analytics

Government

With an eye on recovering from the Pandemic and understanding what a good economic recovery model would look like, the City of Puyallup teamed up with the University of Washington Master’s of Business Analytics program to increase their economic resilience to recessions and dig deeper into sales tax and census data to understand the economic growth sectors within Puyallup. Our team accomplished descriptive, diagnostic, predictive, and prescriptive analytics to help the city gain deeper insights into their data during the four different phases of our project. Our descriptive, diagnostic, prescriptive and predictive analytics included social media analytics, demographic dashboards created within Tableau, sales tax analytics and modeling, and related city analysis.

#Tax Modeling and Predictive Analytics, #Social Media Analytics

Analytics Innovation Club 

Artificial Intelligence

Artificial Intelligence and Machine learning systems and algorithms are as perfect as the data they are built upon. Unfortunately, Data is often imperfect in the sense they incorporate the prejudices and bias of the decision maker or the system through how it’s collected or the prevalent traits in a society/business. When the data is biased the AI/ML algorithms become objectionable by unconsciously discriminating and putting certain privileged groups at advantage and certain unprivileged groups at disadvantage.

The project deals with testing and creating a use case presentation for the AI fairness 360 python package. The package consists of a comprehensive set of metrics and models to test for bias and a set of algorithms to mitigate the bias in the dataset and thus making the decision-making system fair.

#Bias Mitigation Algorithms, #AI Fairness