University of Utah student Lauren McCluskey was worried. She’d been scared by the man she broke up with days earlier, and she had just made a $1,000 payment after receiving a threat that compromising photos of them together would be posted online.
She contacted campus police on Oct. 13, for the second day in row, and the department ran its first background check on Melvin Rowland. But officers did not learn that Rowland was on parole until after he had killed McCluskey on campus on Oct. 22 and died by suicide hours later.
Independent reviewers blamed poor training and the department’s lack of direction to its officers about researching a suspect’s parole status. But they and a separate Utah Department of Public Safety team also examined three databases intended to help police and parole agents connect as they encounter and track offenders.
They found flaws in how the systems are designed, that police don’t know how to use them and that parole agents end up ignoring automated low-level alerts that pile up in the thousands within weeks.
While background checks on police procedurals can appear seamless — one-click access to every record — Utah parole records had been deliberately removed from one of the databases. Other databases couldn’t communicate: A search on Rowland’s driver’s license did not show his parole status because Utah Department of Corrections records didn’t include the license number.
John T. Nielsen, a former state public safety commissioner who led the independent team reviewing the university, said the systems could have made a difference in how Rowland’s case was investigated if they had worked as intended, and not as they currently operate.
Knowing a suspect is on probation or parole can be important information — Corrections agents typically don’t need a warrant to search a probationer or parolee’s home, phone or computer. Agents may know things about an offender that would be helpful to an investigating detective. Some of McCluskey’s information about Rowland — his possession of a gun, his use of social media — as well as her allegations of new crimes might have led to his arrest for violations of the terms of his release.
“What exactly would have been the difference is speculative,” Nielsen said, “and would depend on the actions of those receiving the information — police, [Adult Probation and Parole] agents and the Board of Pardons.”
Running Rowland’s criminal history
McCluskey, 21, had ended her monthlong dating relationship with Rowland, 37, after learning that he had lied to her about his age and was a sex offender who had hidden his criminal history. After a breakup, experts say, former partners can be at risk of dating or domestic violence. But the review found, among other shortcomings, university police were not viewing her case through that lens.
McCluskey met an officer at the public safety building on campus on Oct. 13 and showed him the $1,000 payment she had made on her phone. Afterward, the officer went to dispatchers and requested a criminal history check on Rowland.
A dispatcher searched for him in the Utah Criminal History Records database, which displays a list of past arrests or citations and how they were prosecuted. They learned of Rowland’s 2004 convictions for enticing a minor over the internet and attempted forcible sex abuse, and the officer noted them in his report.
Five months earlier, the same search would have alerted them that Rowland was still on parole.
But the Utah Bureau of Criminal Identification, which manages the criminal history database, had stripped parole and probation information out in a May 2018 update.
It was the cheapest way to resolve a complaint from the FBI — but a decision that should be reversed now, Nielsen said.
The university review team was initially surprised, he said, to find that the parole status was not in the criminal history report.
“It is my view, and I think shared by the other reviewers,” Nielsen said, “that this information needs to be included in criminal history reports, and the state needs to work with the FBI to correct this.”
The state bureau was getting parole information from the Department of Corrections and linking it to people in the state criminal history database. “It was an additional program that we provided,” said Jacob Dunn, field services manager for the bureau. “It would be packaged together.”
The FBI complaint, according to a 2015 audit, was that the state bureau wasn’t sharing that data with a federal, interstate index of criminal histories. And under federal reporting requirements, criminal history records are supposed to be uniformly shared with officers across the country.
The state bureau decided to strip the parole data out. It didn’t own the data; Corrections does. And it would take hours of programming and thousands of dollars, possibly hundreds of thousands, to reformat it so that it could be added to every search in the national index, Dunn said.
What’s next: The Department of Public Safety (DPS) report urges the state bureau to examine adding parole status back into the criminal history search.
Dunn said the bureau has scheduled meetings to discuss that. But, he added, he’s “not aware of any reports of [the lack of parole status] affecting” cases other than McCluskey’s. “That’s not to say there aren’t any, though,” he said.
Stumbling onto O-Track
The U. police officer still could have learned on Oct. 13 that Rowland was on parole, based on the dispatcher’s searches. Computer records show she tapped into another database — called O-Track — and clicked a link that would have displayed his status as a parolee.
O-Track is the Department of Corrections’ primary system for storing information about offenders, and it contains probation and parole data. One part of O-Track is accessible online to the public, and Rowland was listed there as being on parole.
But the dispatcher told investigators she didn’t notice Rowland’s parole status; neither did her supervisor, who was still training her after her first month on the job.
The likely explanation: The department did not routinely access O-Track — its officers were not trained to use it and the dispatcher said she had never seen it before. The university reviewers concluded she had accessed it inadvertently.
There’s still another way O-Track could have revealed Rowland’s status that day.
The dispatcher had ended up accessing the O-Track database directly, by accident. But she also had run at least one other search using his driver’s license number. Had Rowland’s driver’s license number been entered in O-Track, the search would have been able to pull his parole status from O-Track.
Instead, the Department of Corrections had only entered Rowland’s ID card number. And the systems only connect to each other if they have same number identifier — including a Social Security number or an FBI number. Using a different number, or a name alone, will not make the connection.
The same officer ran the search again the next day, but since it was again based on Rowland’s driver’s license, it again failed to uncover his parole status.
Former DPS Commissioner Keith Squires, who was a part of the independent review team, said that was the most glaring shortcoming he discovered.
“They just need to further develop that system, so that whether it's a driver's license or an ID number, if it's been assigned to that individual, either should flag, whether one's run or the other,” he said.
What’s next: The DPS report and the report from the university reviewers urge training for officers — at the U. and elsewhere — on how to use O-Track and other databases. The university reviewers’ report urged the department to train its officers to check criminal history and Corrections records when investigating anyone other than a driver in a routine traffic stop.
One alert ignored
University police and parole agents still could have been connected by a third system — Public Safety Alert Notification, or PSAN.
Yet the way the alert system is designed meant the one low-level notification sent to parole agents was overlooked in a flood of such notices. And the bar for a priority alert — an arrest or citation — was set too high to trigger another.
The system sends an electronic notice to a log in O-Track every time someone — whether a police officer or a potential employer — conducts a background check or driver’s license query on someone on probation or parole.
But the log where those searches are stored has become so massive that it’s effectively useless, and agents rarely consult it, according to the DPS report.
The U. dispatcher’s Oct. 13 check, and the second search the next day, did not generate a log entry, since Rowland’s license wasn’t in O-Track. Her direct access of his records in O-Track on Oct. 13 did create an entry — but it was one of 18,988 transactions entered between Oct. 1 and Oct. 22.
Corrections has prioritized some alerts: If a parolee is arrested or cited, the agent supervising them gets an email and desktop notice. But university police had not arrested or cited Rowland — and had not discovered that they could have reached out to his parole agent.
So when an agent spoke to Rowland Oct. 16, he did not know Rowland had been accused of new crimes and was potentially violating the conditions of his release.
Squires said the volume of inquiries flooding the alert system is a shortcoming that needs to be addressed.
“It's a case of them needing to refine … what is most important for the agents with their caseload to be able to receive so that it's realistic for them to be able to be aware of it and look into it,” he said.
What’s next: The DPS report says the alert system should be evaluated for “enhancements or modifications” that would make it more useful. One specific suggestion: Set up a messaging system, so a person making a query could contact Corrections or parole officers to ask questions.
‘Not widely used’
Aside from changes to computer systems, the DPS report noted, law enforcement officers and parole agents could simply talk to each other more often.
Nielsen said there appears to be a barrier between police agencies and Adult Probation and Parole (AP&P), which isn't generally treated as a resource and partner as investigations are underway.
“We were told that there are times that officers are reluctant to notify AP&P that they might be looking at a person under supervision,” he said. “… These issues are not unique to the university police but appear to be more widespread.”
And police don’t necessarily check whether someone is on probation or parole. At Salt Lake City police, for example, detectives might only stumble on that information while doing more general database searches about someone suspected of a serious crime, said Salt Lake City Detective Greg Wilking, or a crime that carries enhanced penalties for each new conviction.
“It’s a tool that’s in our capability, but it’s not widely used by us,” Wilking said.
Besides, Wilking said, under the PSAN alert system, Utah’s probation and parole agents are supposed to be getting notified when offender names are searched in law enforcement databases.
So agents can reach out when they get an alert that an officer has checked on someone under supervision, Wilking reasoned.
But as the DPS report showed, officers counting on those PSAN notifications will not hear from agents — who are ignoring the voluminous low-level alerts, with no way to tell if an investigating officer or a potential employer is triggering them.
What’s next: The DPS report suggested adding AP&P outreach officers to train and build relationships with law enforcement, hoping they will work together more often when probationers and parolees are under investigation.