The Logic Behind Data-Driven Early Intervention Systems

Accountability and transparency are the “name of the game” in police reform. As much as that’s a contemporary topic of the reform dialog, it has its roots in the evolution of modern policing. As Sir Robert Peel put it, “The ability of the police to perform their duties is dependent upon public approval of police existence, actions, behavior and the ability of the police to secure and maintain public respect.”

Providing support to officers so that they may live up to those high ideals described by Peel is a crucial component of maintaining the public trust in policing. In an era of smartphone cameras and social media, agency leaders must utilize the most promising, evidence-based tools available to prevent misconduct and excessive use of force before it occurs and becomes an issue of public concern. Early intervention systems (EIS) emerged some 50 years ago and are an essential and effective means of demonstrating accountability by addressing performance issues. This article looks at how these systems have risen to prominence and how they’ve evolved to meet the needs of 21st century policing.

Early Intervention as Accountability

Thought leaders in law enforcement began to recognize the importance of early invention as far back as the early 1980s. The United States Commission on Civil Rights (USCCR) published a report in 1981 (republished in 2000) entitled Who is Guarding the Guardians? which took an in-depth look at official misconduct, community trust, and accountability in American law enforcement. Notably, the report detailed nascent efforts at systematically analyzing officer performance as a part of a performance-improving early intervention philosophy.

Building on the first early intervention recommendation from the USCCR, The International Association of Chiefs of Police (IACP) issued guidance for early intervention in 1989, citing them to control corruption and build integrity in policing. Later in 2001, both the Department of Justice (DOJ) and the Commission on Accreditation for Law Enforcement Agencies (CALEA) recommended them as a best practice, with CALEA stating, “[a] comprehensive Personnel Early Warning System is an essential component of good discipline in a well-managed law enforcement agency. The early identification of potential problem employees and a menu of remedial actions can increase agency accountability and offer employees a better opportunity to meet the agency’s values and mission statement.”

In addition to the recommendations from scholars, government entities, and accreditation organizations, early intervention systems are present in consent decrees. Since 1997, most consent decrees have required adopting an early intervention system to improve accountability and reduce citizen complaints.

Diving Deeper

Broadly speaking, an EIS seeks to identify potentially problematic patterns of officer behavior and offer them support before an adverse event occurs. Typically, these systems track disciplinary action rising to the level of suspension, excessive or repeated use of force, citizen complaints, and incidents that may be associated with drug or alcohol abuse. The underlying philosophy is to envision these factors as opportunities for support – additional training, coaching, mentorship, or other forms of professional and personal support. Furthermore, early intervention systems conclude that many if not most of these data points suggest knowable risk factors can be mitigated outside of the disciplinary process, leading to better outcomes for the officer, the agency, and ultimately the community.

There are different methodologies early intervention systems use to achieve these goals. At its most basic, early intervention can be rooted in the social interactions within a department, with supervisors recognizing changes in an officer’s pattern of behavior that could indicate an underlying issue. This methodology is evident in agencies with fewer sworn officers, where a smaller and consistent headcount can lead to close relationships. However, this approach is by its nature anecdotal and does readily lend itself towards data analysis or easy reproduction in other agencies.

More formal systems rely on analyzing a set of data points to determine when prescribed action is necessary. The systems analyze a limited set of indicators, usually expressed in the form of a threshold. Once that established threshold is reached, an officer is “flagged” and typically referred to a supervisor or internal affairs for additional support. Systems vary widely, but common thresholds include the number of use of force incidents in a given period, a pre-determined number of citizen complaints, or other data points.

A drawback of these systems is that they are limited by the number of data points they utilize in the analytical phase. Limited datasets can offer a somewhat narrow assessment of an officer’s activities while on duty, often not fully considering the nuance and context involved in an officer’s patrol. Once a threshold is reached, prescribed actions can lead to instances in which a supervisor’s best judgment or knowledge of circumstances is not incorporated into the analysis – leading to a false positive (or false negative) created by the system. This can lead to the perception among officers that it is “only a matter of time” before they are flagged and prescribed corrective actions that may not be useful or appropriate.

Reform-Minded Early Intervention

Benchmark Analytics’ First Sign® Early Intervention takes a different approach to early intervention compared to informal or threshold-based systems. Building on decades of knowledge and experience derived from generational iterations of early intervention systems and a research consortium dedicated to studying policing transparency and accountability, First Sign is the only EIS that uses a research-based statistical model that is preventative by design.

This means that First Sign does not simply analyze an officer’s performance in a vacuum but instead contextualizes it by using peer analysis, which leverages multiple data sources to present a more holistic picture of the officer and pinpoint where they can best be supported. These data points can be what time of day their shifts are, what areas they patrol, and their assigned unit. An example of this could be the difference between an officer assigned to a drug task force in a higher crime neighborhood at night is likely to have different performance data points than one patrolling a relatively quiet business district on a daytime shift. A threshold-based system doesn’t take this into account. It would instead assign the same thresholds to both officers which, lacking context, could present an inaccurate portrait of an officer’s behavior and potential for risky behavior.

First Sign may be predictive, but it is not prescriptive by nature. Instead, using the Case Action Response Engine (C.A.R.E.), it gives departmental leaders and supervisors a suite of tools and data analysis dashboards to understand an officer’s performance to facilitate thoughtful and effective intervention. It is non-disciplinary and non-punitive by design, leading to better alignment among all stakeholders – departmental leaders, supervisors, civic leaders and advocates, and officers – in building a more meaningful and lasting sense of buy-in.

Advancing reform

Though there are distinct differences in early intervention systems, they all seek to advance the reform dialog by providing law enforcement with a valuable tool to address off-track officer behavior before it manifests as misconduct, excessive use-of-force, or other actions that have a detrimental affect on community trust of police. First Sign distinguishes itself in the field by offering a reform-minded EIS that supports forward-thinking policy decisions designed to build and strengthen that community trust that is fundamental to 21st-century policing.