A highly touted computer system designed by the
The report by the Police Commission's inspector general, Alex Bustamante, scrutinized an early warning computer program that the LAPD has used since 2007 to track patterns of excessive force and other misconduct by its roughly 10,000 officers. The analysis casts doubt on the usefulness of the computer system, which federal officials forced the LAPD to build after years of corruption and abuse.
In light of similar concerns raised internally, department officials have asked an outside research group to conduct a comprehensive review of an entire network of databases that contain information on officers' performance and are used to trigger the early warnings. That review, officials said, began last month and will determine what changes, if any, should be made to more effectively focus on potentially trouble-prone officers. It will also examine whether the system has had any effect on improving officers' conduct.
The Police Commission, which oversees the LAPD, will discuss the inspector general's report at a meeting Tuesday. Commissioner Robert Saltzman said the department's current tracking system appears to be "providing limited predictive capabilities," adding that Bustamante's report raises "significant questions."
"I look forward to understanding how the department is responding to correct the issues," he said.
In his report, Bustamante examined nearly 750 warnings about officers generated over a recent four-month period. In 70% of the cases, supervisors took no action after determining that the conduct flagged by the computer system did not point to any problems, the report found.
The lack of action after so many red flag notifications raises questions about the criteria being used to trigger warnings — called "action items" in LAPD jargon. Currently, the system attempts to compare several aspects of an officer's conduct to that of other officers in similar assignments. A warning is triggered when an officer exceeds acceptable limits for each benchmark. The various benchmarks include the number of times an officer uses force on a suspect, as well as complaints and lawsuits filed against the officer.
Maggie Goodrich, the LAPD's chief information officer, said it could be that the system currently is too quick to issue a warning. The risk, she said, is that the department might narrow its assessment of officers too much and, in doing so, miss some misconduct.
"The challenge is finding a balance," she said.
However, the inspector general's report also found that the early warning system has been relatively useless in making early identifications of officers who went on to commit serious misconduct.
Of 40 officers who were fired from the department over a recent two-year period, 30 of them had never been flagged by the computer system or had only one warning over the five years the system had been in place, Bustamante found.
In general, he concluded, the warning system flagged behavior that was unrelated to conduct that resulted in the officer being fired.
The findings call into question what was a cornerstone of the department's effort to reform itself in the wake of the Rampart scandal and other episodes that pointed to widespread abuses. As part of a federal monitoring program, the department began in the early 2000s to implement a sweeping set of safeguards, including the computer monitoring system.
Major technological problems delayed the completion of the so-called TEAMS II — Training, Evaluation and Management System. For years, city officials were critical of the department's slow pace, as well as the many millions of dollars that contractors were charging to build the computer system, which went online in 2007. At one hearing, a high-ranking LAPD official rebuffed criticisms from City Council members, telling them that something so important could not be rushed.
"The key for us is it needs to be accurate and done right," he said.
The eventual completion of TEAMS II was a major factor in the decision by a judge to free the LAPD from the federal monitoring.
Since the program was implemented, the department has not conducted any meaningful evaluation of the warning system's effectiveness.
Under the current system, supervisors are notified whenever a warning is issued about an officer under their command. After reviewing the issue, a supervisor can choose from a range of options, including dismissing the warning as unfounded, ordering retraining or restricting the officer's work duties. The supervisor's decision must then be reviewed and approved by several higher-ranking officials.
Along with the cases the supervisors dismissed, Bustamante found that about a quarter of the warnings were resolved with an "informal meeting" between the officer and supervisor. Such meetings are typically called when something inappropriate about the officer's behavior is identified, but the supervisor decides it is not serious enough to warrant a more formal response, Goodrich said.
In only about 25 of the nearly 750 cases, officers were ordered to undergo retraining, were reprimanded or had some other action as a result of the warning system.