Problem: $124 Million in Lawsuit Payouts
From 1987 to 2000, the City of Detroit paid over $124 million in police misconduct lawsuits, averaging almost $10 million per year. In 2000, then-Mayor Dennis Archer requested the probe after police were involved in 47 fatal shootings between 1995 and 2000, including six of unarmed suspects. Between 1994 and 2000, 19 witnesses died while being detained by the Detroit Police Department (DPD). After federal review, on July 18, 2003 the DPD, the City of Detroit and the United States Department of Justice entered into two consent decrees. The first consent decree deals with the Use of Force, Arrest and Witness Detention.
The second decree concerns the Conditions of Confinement in our holding facilities. The Consent Decree was also very costly, where a monitored was paid upwards of $2.3 million per year. Within the Consent Decree, the DOJ mandated for Detroit to implement a Risk Management Database. This was a top priority for the City of Detroit, but very difficult for the city to implement on its own
The risk management database needed to be a new computerized relational database, integrating and retrieving data necessary for supervision and management of the police department as detailed in sections 79-99 of the Consent Decree. The purpose of the system was to acquire and evaluate the performance of DPD officers across all ranks, units and shifts, to manage risk and liability, and to promote civil rights and best police practices. The system was also required to gather and integrate with over 36 datasets such as canine bites, civil lawsuits, use of force reports, hold policy, and etc.
This was difficult to integrate into a system due to the nature of DPD’s paper-based records management. In 2009, DPD was no further along in their process of implementing the risk management system. By this time, the city paid over $13.8 million in monitoring fees. DPD’s internal IT staff had created a temporary system called the Interim Management Awareness System (IMAS) until the Management Awareness System (MAS) could be developed. The IMAS was primarily an Access database that required manual entry of data captured from over 2,500 police officers. This created inefficiencies, inaccuracies, and delays in identifying risk within the police department. The City of Detroit and DPD needed help. Without an effective solution, the city would continue to pay millions on monitoring fees.
The City of Detroit and DPD needed help. Without an effective solution, the city would continue to pay millions on monitoring fees.
Solution: Integrated Risk Management System
For MAS to be successful, we needed address DPD’s three primary areas of concern for an effective risk management system. The first is to create a system that met the compliance of the Consent Decree, the second is to create a system that is accurate in identifying risk, and the third is to have an organization of 2,500 officers adopt the technology efficiently and effectively. We embarked on this process in 2009.
In order to meet the compliance of the Consent Decree, the system needed to identify risk factors from 36 datasets in near real-time. The, then current, process with IMAS was arduous and inaccurate since it relied on human intervention to replicate data. Data replication was required because critical datasets such as use of force was captured in paper form. For our risk management software to function effectively, we needed to first digitize the police departments paper-based records into an electronic database.
Working the police department, we created a flexible digital forms solution that converted existing paper forms into digital computer entry. This saved officers time, increased accuracy in data capture, and allows us to develop the risk management system.
Once the paper-based records were transitioned to digital records, we were able to begin the data integration into the risk management system, including DPD’s existing digital-based systems such as the Crisnet police reports and human resource time keeping system. With digital data now available to DPD, we needed to ensure risk management was accurate in identifying risk. This was challenging because we needed to create a system that did not exist to meet the unique needs of DPD.
We benchmarked major cities across the nation and the best practice, at that time, of risk management was quarterly reviews and outdated methods of fixed rules or thresholds for analysis. The challenge with this is a potential for false positives or not enough risk positives. For example, a member of the gang squad in a police department may shoot a canine at a higher rate than a patrol officer on traffic duty. If we set a fixed threshold across the department for canine shootings, it will not be able to encompass the unique attributes that determine risk accuracy. To address this issue, we needed to create a flexible rule set that learned the uniqueness of the police department. We deployed an early version of an Artificial Intelligence (A.I.) rules engine that identified risk and adjusted based on the changing dynamics in the police department, reducing false positives.
We needed to ensure that the 2,500 officers would adopt MAS and the risk management worked to ensure accountability. With the many systems the police department was using, we established a philosophy to simplify, eliminate, and automate where possible, without creating additional work for the officers. A change from paper-based to digital records keeping created natural barriers for adoption.
Our approach to removing the barrier was to design the digital forms that are familiar to the officers, reducing training time and increased adoption. In addition, we automated many of the workflows that saved officers time from traditional paperwork. The Consent Decree also required accountability in the risk management process and a change in culture. We needed to also ensure command staff adoption of MAS.
Our approach to managing culture change was to leverage data and automated workflow for accountability. For example, when risk is identified, supervisors needed to intervene within 48 hours. Previously, this would have been difficult to enforce. We embedded compliance policy that would alert a sergeant if the supervisor did not intervene within the appropriate time frame. Each level of the organization was held accountable, including alerting the Mayor if the Chief of Police did not comply with certain policies. We also designed additional “public safety apps” bridge gaps in data collection and processes so the officers did not have to go outside of MAS to perform necessary functions. This allowed officers and command staff to stay in the MAS application as much as possible.
In 2011, the deployment and adoption of MAS at DPD met the compliance of the Consent Decree, reduced false-positives, and increased user adoption. This fundamentally transformed a police department and helped DPD leapfrog into data-drive risk management.
Results: Reduced Lawsuits, Reduced Citizen Complaints, Reduced Payouts
As a result of deploying MAS, DPD experienced a decline in citizens complaints, use of force, and lawsuit payouts. Since the implementation of MAS, DPD has seen a 36% average decline in citizen complaints. Use of force within the first year of deploying MAS, saw a 17.5% decline in use of force across all applicable categories. By the time DPD entered into a Consent Decree transition agreement, use of force was reduced by 22.5% across the spectrum. In a recent 2017 article published by the Detroit News, MAS reduced lawsuits filed against the city by 62%, compared to the previous year. Lawsuit payouts was $4.9 million compared to $10 million in years before the Consent Decree; this represents a $5.1 million save.
Lawsuit payouts was $4.9 million compared to $10 million in years before the Consent Decree; this represents a $5.1 million save.
In addition to the reduction in tangible risk categories, DPD also had access to their complete digital dataset, enabling additional benefits such as running time keeping reports from a single system as compared with multiple systems. MAS held supervisors accountable and ensured that proper intervention was performed and any remediation was completed. The risk management system worked.
Next Generation: Scalability for All Size Agencies with Arx Alert
Police departments of all sizes across the nation face risk. Although suburban police departments may not have the same levels of risk that Detroit had, the risks still exists. Categorically, the risks that departments face are use of force, citizen complaints, vehicular-incidents, overtime, intake, and lawsuits. In many cases, smaller police departments might have higher risks due to a lack of updated best-practice policies, outdated systems, and siloed systems purchased in different decades.
A system like MAS could benefit these smaller agencies, but the architecture and support required would not meet their financial budget. We needed to develop a new cloud-based platform that provided the benefits of a multi-million dollar risk management system only major cities can afford, at a fraction of the cost. In addition, we need to provide a new modern A.I. engine that was able to adapt to agencies of all sizes. Our solution is Arx Alert.
Arx Alert is a cloud-based risk management platform that is accessible and affordable to agencies of all sizes. We are able to provide the benefits of major city solution such as digitization of paper-based forms, A.I. risk identification, A.I. assisted remediation, and built-in business intelligence at a fraction of the cost. Whether you have 25 officers or 2,500 officers, Arx Alert is able to meet the unique needs of the agency within budget, saving money, reducing risks, and improving image.
References:
https://www.mlive.com/news/detroit/index.ssf/2014/08/us_attorney_babara_mcquade_to.html
http://josephsonsexemplarypolicing.org/2016/06/consent-decrees-use-of-force-detroit/
http://www.detroitmi.gov/How-Do-I/Find/Definitions-of-Consent-Decree
Get Started with Arx Community
Looking for help? Get in touch with us
