Transcript: Government of Canada Data Conference 2023: Employment and Social Development Canada Integrity Services Branch's Use of Quantitative and Qualitative Data
Data Lakes, Machine Learning, AI (Artificial Intelligence) – these are just a few of the buzz words that not long ago seemed distant from most people working at ESDC. However, those terms are becoming increasingly more common and understood. Each year, there are well over 10 billion data points created in the Employment Insurance program alone. It is only natural for the department to gravitate to a more data-driven culture. Online client interactions, system-generated information, and improved internal processes all contribute to this growing, dynamically rich data ecosystem. Leveraging this data, along with people, becomes critical for innovation and innovative practices within the department; understanding the needs and requirements of Canadians as they interact with programs; Improving service standards to ensure a streamlined, low-friction client experience; and, ensuring that a robust integrity approach is at the forefront of service delivery.
The Integrity Services Branch (ISB) has well over a 1,000 employees across Canada, with functions ranging from investigations and administrative reviews to maintaining the Social Insurance Registry and conducting operations research. Managing these resources effectively and in coordination depends on the best available quantitative and qualitative information. As a result of the ever increasing importance of data, ISB has developed a first-class analytics team (The Integrity Business Analytics Directorate or IBAD for short!) this team is taking advantage of the ever-growing mountain of data as it relates to fraud detection, operational support, and reporting investigative outcomes. The team members come from a variety of backgrounds such as computer sciences, mathematics, programming, economics and even philosophy. The most important aspect of this team is they all have a passion for data!
So how do they make a difference for integrity? The team leverages technology to process large amounts of data coupled to the insights of our data analysts, intelligence officers, and regional investigators to create algorithms to detect patterns, discrepancies and anomalies that signal known or suspected fraudulent activities. Using data, technology, and people, ISB has been able to detect benefit fraud and threats in three main areas; Error or Misrepresentation – which means incorrect information provided on applications allowing someone to obtain benefits that they are not entitled for; Fraud - Individual in question makes statements that are not only misleading but outright false; and Organized Fraud - The most serious offence, characterized by large scale fraud on a benefit system. Involves multiple actors and sophisticated techniques to interact with the department's systems.
Specifically our branch has developed predictive models and algorithms that support a large portion of the integrity compliance activities in the EI program and pensions programs. These well-tuned algorithms ensure that investigators are able to focus their efforts on areas and files that are most likely fraud or have errors. In addition, data is extremely important for reporting which some say can be boring...or it can be exciting when it is visualized! This means arranging data in a meaningful and intuitive way for insights that are actionable. The analytics team has revamped the way in which they present information using leading-edge tools with the results being a more informed management group.
Another key data driven aspect of our work involves anomaly detection, which may signal fraud or errors. People tend to pick out anomalies naturally, however rarely can they do it on a larger scale. New techniques are constantly being tested and used by Integrity to keep pace with fraudsters For example, did you know that data-driven approaches such as anomaly detection and network analytics has uncovered fraud linked to identity theft and false Records of Employment? Indeed – new methods have shown to be good at detecting these schemes with a high degree of certainty. Utilizing data can have a disruptive effect on organized crime and large-scale fraud that saves the department millions of dollars and re-enforces confidence in the integrity function. These analytics activities played a significant role in detecting fraud in the EI-Emergency Response Benefit. Analytics were used to detect more than 30,000 cases of suspicious applications and stopped more than $42 million in payments from going to fraudsters. Incredible work!
This innovative work continues after the Emergency Response Benefits and a great example of this is the anomaly detection program that scans EI applications for known fraud indicators. Yes, we do! The Integrity Technology-Reinforced Anomaly Program – or as we call it iTRAP for short – is a suite of algorithms that try to triangulate identity theft. Since it was deployed in late 2021, it has shown to be a versatile tool to preventing money from going into a fraudster's bank account and ensure unknowing Canadians are not impacted. In terms of rough estimates, iTRAP has identified over 70,000 instances of potential identity theft and prevented about $1B in benefits from being paid fraudulently as well as prevent hundreds of thousands of work items from flooding Service Canada's operational delivery machinery. Just incredible work!
Analytics can also be used to make our work efficient, for example the analytics team created a tool for investigators that summarized weather data for their investigations. Exploratory analysis showed that regional staff spent a lot of time searching for information such as weather to corroborate what a client told them example; Did Mr. Smith really pick 400lbs of blueberries on Sunday when that area was experiencing heavy rainfall and thunderstorms? These new tools, and those that are inspired by data and innovation, will transform how integrity is performed so that employees get the most out of their jobs. This work and capacity to develop analytics was not developed overnight: the Integrity Services Branch (ISB) has been building the analytics function over a number of years and through a number of critical activities and fostering key relationships with stakeholders including working very closely with the Chief Data Officer and her team. For example, what better way to understand how fraud works than to ask investigators directly! Events such as 'Fun with Analytics' have been very successful where we bring together regional investigators, business experts and data scientists to explore ways to detect fraud using on-the-ground knowledge of investigations coupled with powerful data exploration.
Other key innovations that contribute to our knowledge base include hackathons for example with St.Mary's, University of British Columbia and Simon Fraser University where we bring together students to work collaboratively and explore solutions for real world integrity problems using data over a short period of time.
Moving forward, we are taking key lessons from the last several years and incorporating them into future counter fraud activities including; Embedding more analytics and intelligence functions upfront to reduce fraud without impacting client experiences. This includes the use of more data matching, supervised and unsupervised learning, and use of text mining in applications and network analysis. Increased collaboration and building multi-disciplinary teams made up of program and operational experts, data scientists, and investigative resources to assess gaps and react to evolving fraud. Increasing data to identify fraud through improved sharing among government departments and engaging private industry to complement existing data holdings and tools and investing in new technologies and platforms to continue to improve analytics capabilities, including through key initiatives such as the benefits delivery modernization which is building a new dynamic business environment to deliver the department's services of the future and incorporates an integrity by design approach to better react and prevent fraud. Preparation is Key – Establishing a good integrity framework from the beginning and knowing vulnerabilities and potential mitigation strategies which we're working on as well.
The ability to harvest the rich amount of data in our department, coupled with the collaboration of other key areas of the department such as intelligence, the Chief Data Officer and her team, business experts, IT, BISB, our colleagues in the transformation and integrated service management branch, and our regional investigators and regional leadership will ensure that this innovative work continues and that the department remains on the cutting edge of analytics to protect public funds. This type of collaborative effort demonstrates why Integrity is everyone's business!