At the Root

Published in the Spring 2007 issue of SSA Magazine

"Root-cause analysis" can help family-services programs work better

Adapted from a November-December 2006 University of Chicago Magazine article

By Lydialyle Gibson

Tina Rzepnicki first encountered root-cause analysis in an article on failed military operations missions doomed by bad decisions, bad communication, unforeseen or unforeseeable obstacles, errors of action and inaction. "Things started looking familiar," says Rzepnicki, who is the principal investigator for the Program Practices Investigation Project, which examines and develops better child-welfare practice for the inspector general's office of the Illinois Department of Children and Family Services (DCFS).

Root-cause analysis traces an outcome to its most basic triggers: an unmade phone call, an unheeded order, a missing part, a miscalculated step. It's typically used to assess disasters like airplane crashes, industrial accidents, and hospital mistakes, but Rzepnicki saw the possibility to use the technique to help prevent smaller but no less terrible events.

"There are a lot of parallels," she says. "Any human-service agency that occasionally experiences tragic client outcomes needs a systematic way to take incremental steps back and analyze what went wrong. While it is easy to blame individual caseworkers for poor decision making, it is increasingly recognized that errors are as likely to result from problems at multiple levels of the organizational process. Remedies should be directed at improving defenses and removing error traps."

Rzepnicki, SSA's inaugural David and Mary Winton Green Professor and director of the University's Center for Social Work Practice, decided to test the method in the inspector general's office. Headed by Denise Kane, A.M. '78, Ph.D. '01, the office examines child deaths in families that have warranted recent DCFS attention. Three years ago Rzepnicki established a root-cause analysis pilot program there, adapting the approach to increase its applicability to the specific needs and large number of cases investigated by Kane's office.

The computer-assisted approach incorporates systematic questions about why and how specific events occurred. Getting the answers is a labor-intensive process that yields six- or eight footlong, color-coded diagrams. "You look at each discrete point, rather than lumping it all together. At each point in the chain of events, questions are asked, like, 'What led to this event?' 'What contextual factors were evident?' and 'What things should have happened but didn't?' Pretty quickly you end up with a lot of these red squares, which highlight what didn't happen that should have at a particular point in time," Rzepnicki says.

To see how it all works, take the case of an eight-month-old baby—let's call her Sarah—who died four years ago in Chicago from abuse by her father. Sarah had already been the subject of a DCFS investigation when her grandmother brought her to the hospital with bruises. At that time, the baby's mother first said her husband had hit Sarah, but later she claimed responsibility. Investigators took her at her word.

The analysis by Rzepnicki uncovered several opportunities for DCFS to have altered events. For example, the caseworker, who had provided therapy to Sarah's mother in the past and argued the infant should be kept with her parents, felt torn between being Sarah's advocate and her mother's former therapist. The supervisor, dealing with too many cases, failed to request reports on the mother's mental history and the father's prior violence. Deferring to the therapist-caseworker's judgment, others didn't verify the mother's claims with neighbors or family members.

"Hindsight bias" presents a constant temptation, and sometimes the conclusions are for changes in workplace culture, not policy. "We find that with a lot of these cases, the mid-level is where a lot of shortcuts are being taken, or where there are disincentives for good practice," Rzepnicki says.

Rzepnicki, with SSA's Dean of Students Penny Johnson, published a study of root-cause analysis in the January 2005 Children and Youth Services Review. Based in part on their work at DCFS, state legislation was recently proposed that establishes error reduction teams. These teams will work with agencies serving child welfare clients to plan, apply, and monitor corrective actions when patterns of serious error are detected. Rzepnicki is currently working with Kane's office to develop implementation details. In the future, she'd like to look at successful cases as well—children returned happily to reformed parents or rescued from deteriorating situations. "What actions, conditions, decisions led to a good outcome?" she asks.