Minnesota’s child protective services system makes important decisions about whether (and how) government should intervene in families’ lives to protect maltreated children. For example, child protection agencies decide which allegations of child maltreatment to investigate, whether maltreatment occurred, and whether protective services should be offered. They also decide whether to initiate court actions that may lead to out-of-home placement or termination of parental rights. These are difficult decisions, and they are often made with minimal public scrutiny because the records of child protection agencies are private.
In May 1997, the Legislative Audit Commission asked us to examine child protective services in Minnesota. In our research, we asked:
How much variation is there among counties in the incidence of child maltreatment investigations, determinations, and services? To what extent do county policies and practices explain these variations?
Do people who work closely with Minnesota’s child protection system believe that it works effectively?
To what extent does maltreatment occur repeatedly within the same families? Are there additional steps that child protection agencies could take to reduce the incidence of repeated maltreatment?
How large are the caseloads of child protection workers? What types of education and experience do these workers have, and how much staff turnover is there?
An effective child protection system relies on the efforts of many people and agencies, including "mandated reporters" of child maltreatment, county child protection agencies, county attorneys, the courts, law enforcement agencies, and providers of services to families. In addition, relatives, neighbors, and the community at large bear a responsibility for reporting instances of suspected maltreatment and providing support to families in trouble. In response to legislative concerns, our study focused primarily on the role of county agencies in screening, investigating, and responding to reports of child maltreatment.
In 1996, Minnesota child protection agencies conducted 16,684 investigations and determined that maltreatment occurred in 6,725 cases (40 percent). The total number of investigations and maltreatment determinations in Minnesota has declined since 1993. Figure 1 shows trends in various types of maltreatment. Child neglect is the most common type of maltreatment, accounting for 54 percent of maltreatment determinations in 1996.
We wanted to examine trends in maltreatment-related deaths, but we found that statewide child mortality data in the Department of Human Service’ (DHS) maltreatment information system are unreliable. For example, the DHS information system indicated that 49 maltreatment-related child deaths occurred during 1994-96, but we found that half of these cases were erroneously reported as child deaths.
State agencies administer child protective services in most states, but in Minnesota these services are primarily administered by 84 county human services agencies. In fact, Minnesota is one of only 10 states with a county-administered child protection system. Furthermore, local property tax revenues pay for the majority of Minnesota’s $300 million in annual child welfare expenditures, while they pay for a much smaller percentage of child welfare costs nationwide. Minnesota laws and rules provide a framework for county services, but state definitions of maltreatment are broadly-stated and leave considerable room for county discretion.
Based on a survey of county human services directors, we estimated that Minnesota counties received about 50,000 allegations of child maltreatment in 1996. Figure 2 shows that counties investigated about one-third of these allegations statewide and "screened out" the remainder. According to our survey, the percentage of allegations investigated ranged from 20 percent or less in five county agencies to more than 90 percent in nine agencies.
Some counties have developed written screening criteria to help articulate local interpretations of state maltreatment laws, improve consistency in decision making, and inform the public and professionals about what types of cases will be investigated. But we found that:
Counties that used screening criteria reported to us that they investigated 28 percent of the allegations they received in 1996, while counties without criteria investigated 51 percent of the allegations.
During 1994-96, there were 14 reports of maltreatment investigated annually in Minnesota per 1,000 children under age 18. The rates of individual counties varied from 3 investigations per 1,000 children in Itasca County to 29 per 1,000 in neighboring Hubbard County. Variation in rates of investigation may partly reflect underlying differences in the incidence of maltreatment, but it was apparent from our interviews with county staff that variation also reflects differences in county philosophies and criteria about the types of reports that warrant investigations.
State rules require counties to begin all investigations within three days of receiving a report of maltreatment, and investigations must start sooner when children are alleged to be (1) in imminent danger or (2) victims of infant medical neglect. Information submitted by counties to DHS indicated that the state’s most populous county (Hennepin) started only 44 percent of its 1994-96 investigations within three days, while the remaining counties started 91 percent of their investigations within three days.
At the conclusion of an investigation, the law requires county agencies to determine whether maltreatment occurred. Table 1 shows that counties varied considerably in their number of determined maltreatment victims per 1,000 children in the population. This partly reflects the fact that:
For example, some county agencies require evidence of an injury "such as a bruise" before determining that maltreatment has occurred, while other agencies do not. Some county agencies think it is acceptable for children ages seven or older to be left unsupervised, while others do not. Some counties rarely if ever determine that caregivers have caused "mental injuries," while other counties frequently--and sometimes without psychiatric or psychological diagnoses--justify maltreatment determinations on the basis of mental injury.
Following an investigation, county agencies are also required by law to determine whether the investigated family needs protective services. Families determined to need protective services must be monitored regularly by counties, and they may be offered services such as counseling, treatment, or placement of the children away from home. Statewide,
While most county human services directors told us that budget considerations did not play a role in their decisions to provide services, 71 percent of district court judges responding to our survey said that they perceived that budget considerations have at least "sometimes" affected county recommendations and actions in the past two years.
Counties may petition the court if they want children placed out-of-home involuntarily or to require families to comply with recommended services. The petitions, commonly called "CHIPS" petitions, allege that the children are in need of protection or services. We found that counties varied in the number of CHIPS petitions filed in 1994-96. For example, there were 2.7 maltreatment-related CHIPS petitions filed in the seven-county Twin Cities metropolitan area per 1,000 children, compared with 4.3 CHIPS petitions per 1,000 children in other counties. Some of the variation may reflect the willingness of individual county attorney offices and child protection agencies to bring maltreatment-related cases before the court.
Counties also vary in the child protection records they keep. For example, only 58 percent of county child protection agencies (accounting for 30 percent of 1996 investigations) keep logs of all of the allegations they receive. In addition, counties vary in the length of time they keep records of investigations that did not result in determinations of maltreatment or services needed. Most counties told us that the vast majority of such records from 1996 investigations were still on file in mid-1997, but 10 of the 84 county child protection agencies told us that at least 75 percent of these 1996 records were already destroyed.
According to state rules, "the purpose of child protective services is to protect children from maltreatment." Thus, counties not only determine whether allegations of prior maltreatment are valid, but they also aim to reduce the likelihood of future abuse or neglect.
We used data reported by counties to the Minnesota Department of Human Services (DHS) to determine the incidence of repeated investigations or maltreatment determinations within the same family. Unfortunately, it is not possible to use the DHS information system to determine whether a family with a maltreatment determination in one Minnesota county subsequently had a determination in a different county. This is a serious weakness of this system, and it means that our analysis likely understates the true incidence of repeated maltreatment statewide. In addition, we found that Hennepin County has not assigned case numbers to families in the manner prescribed by DHS, making it impossible to use the state maltreatment information system to track that county’s rates of repeated maltreatment.
As shown in Table 2, we found that:
Twenty-nine percent of families who were the subject of maltreatment investigations in 1993 were the subject of subsequent investigations in the same county within three years.
The rates of repeated maltreatment (and repeated investigation) were higher for cases that originally involved child neglect than those that originally involved physical or sexual abuse.
When counties conduct investigations, they assess families’ risks of subsequent maltreatment to help determine whether there is a need for protective services. All but one county agency use a DHS-recommended risk assessment instrument to classify families as "high," "intermediate", "low," or "no" risk. DHS has not validated its risk assessment instrument by examining whether rates of subsequent maltreatment correspond to the instrument’s classifications. We found that low and no risk families had lower rates of repeated maltreatment than families with higher risk classifications. However, intermediate risk families had slightly higher rates of repeated maltreatment than high risk families, even among families determined to need services. It is possible that the types of services provided to high risk families accounted for their lower rates of repeated maltreatment, but it is also possible that Minnesota’s risk assessment instrument is not sufficiently predictive. In addition, research in other states has indicated that other risk assessment instruments may be more reliable than the type Minnesota uses.
We reviewed county child protection records in detail for about 200 families in eight counties, including many families that were the subject of two or more maltreatment investigations or determinations. Our sample of cases was not statistically representative of cases statewide, but our reviews led us to conclude that some children might be more effectively protected from repeated maltreatment. For example, some chemically dependent parents repeatedly received "second chances," sometimes with little ongoing monitoring of their chemical use and spotty compliance with case plan requirements.
In general, we think it is possible that children could be more effectively protected if (1) counties had more predictive risk assessment approaches, (2) the behaviors of high-risk families were monitored by child protection agencies for longer periods, (3) child protection assessments were more comprehensive, rather than focusing solely on the incidents that initially prompted the investigations, and (4) counties petitioned the courts more quickly when families failed to comply with services. Recent changes in federal and state law are intended to expedite the process of finding permanent homes for children who have been removed from their families, and it is possible that these changes could reduce the opportunities for repeated maltreatment that some families have had.
There are limited statewide data that can be used to evaluate the performance of Minnesota’s child protection system. Lacking better measures, it is useful to consider whether the people who work closely with the system believe that it is operating effectively. We surveyed several groups of professionals required by law to report instances of suspected maltreatment--pediatricians, school social workers, and heads of local law enforcement agencies. We also surveyed district court judges, who hear CHIPS petitions, and county human services directors, who administer child protective services.
"Mandated reporters" accounted for 62 percent of the reports investigated by child protection agencies in 1994-96. Consequently, it is especially important for child protection agencies to communicate effectively with these reporters and to have their confidence. We found that:
For example, 63 percent of pediatricians and 42 percent of school social workers statewide said that they were "sometimes, rarely, or never" adequately informed about county screening criteria for physical abuse. If the professionals who work regularly with the child protection system have limited knowledge about the criteria used by counties, we think it is safe to assume that the general public knows even less. In addition, state law requires counties to inform mandated reporters about the outcome of cases they report, but 69 percent of pediatricians and 54 percent of school social workers said they were "sometimes, rarely, or never" informed about case dispositions.
Our surveys also indicated that:
About 45 percent of school social workers and 18 percent of pediatricians statewide said they have considered not reporting an instance of suspected maltreatment during the past two years because they thought the child protection agency would not respond appropriately. Failure to report suspected maltreatment is a misdemeanor under Minnesota law, so the qualms indicated by reporters reflect serious concerns.
While our surveys revealed concerns about the effectiveness of child protection interventions in various types of cases, respondents expressed particular concerns about cases involving child neglect. For instance, 54 percent of school social workers and 38 percent of pediatricians said that child protection agencies have "sometimes, rarely, or never" conducted thorough investigations of child neglect. Likewise, 41 percent of county human services directors said that law enforcement agencies "sometimes, rarely or never" give sufficient attention to investigations of child neglect. Also, 55 percent of school social workers and 45 percent of pediatricians said that child protection agencies have "sometimes, rarely, or never" taken appropriate steps to protect victims of child neglect from further harm.
Many mandated reporters also expressed concerns about inconsistent child protection decisions. Only 38 percent of school social workers and 26 percent of pediatricians said that child protection staff "always" or "usually" use consistent criteria to make decisions.
The heads of law enforcement agencies expressed greater satisfaction than pediatricians and school social workers with child protection agency investigations and interventions. For example, 91 percent of the police chiefs and sheriffs we surveyed said that child protection agencies "always" or "usually" conducted thorough investigations. Also, we found that the heads of law enforcement agencies and child protection agencies generally believe they have established cooperative working relationships with each other.
For the most part, Minnesota judges told us that they do not believe that child protection staff have been too intrusive in the lives of families, and they usually think that staff have pursued reasonable options before recommending child placements or terminations of parental rights. But the majority of judges told us that child protection staff "sometimes" (or more frequently) give parents too many "second chances." In other words, judges were more likely to think that child protection agencies have been too timid in their family interventions than to think they have been too aggressive.
Nationally and in Minnesota, there has been debate about the goals of the child protection system. While state rules direct child protection agencies to protect children from maltreatment, federal and state laws have also directed agencies to make "reasonable efforts" to prevent out-of-home placements and reunite placed children with their families. Our surveys asked people who work closely with county child protection agencies to characterize what they perceive to be the goals of those agencies in practice. As shown in Figure 3, school social workers and pediatricians were more likely than judges or law enforcement officials to cite family preservation, rather than protection of children, as the goal that is more important to child protection staff. Large percentages of law enforcement staff and judges said that the goals of family preservation and protection of children were equally important.
Finally, we asked county human services directors about the adequacy of services for families they serve. Their most often cited "unmet need" was for truancy and educational support services, with 60 percent of directors indicating that existing services have not met their needs and one-third of directors identifying it as one of their top three needs. Of the various types of maltreatment, directors most often cited child neglect (including educational neglect and other types of neglect) as the type for which services are the least adequate.
The job of a child protection employee is a difficult one. Employees must make important judgments based on a wide variety of federal, state, and local laws and policies. They are also expected to work closely with the courts, law enforcement agencies, county attorneys, health professionals, school professionals, and others.
We collected information from counties in September 1997 to help us analyze child protection caseloads at that time. We examined the caseloads of staff who investigate allegations of child maltreatment, as well as the caseloads of staff who monitor families that have been determined to need protective services. We found that:
Statewide, there were 16 cases under investigation per full-time-equivalent (FTE) child protection investigator. Half of Minnesota counties had caseloads of 10 or more.
It is possible that Minnesota child protection agencies are understaffed. A national child welfare organization has recommended that caseworkers not have more than 17 open cases and that investigators not have more than 12 cases. Many of the mandated reporters we surveyed suggested to us that child protection agencies need additional staff-to work with families before serious crises arise and to monitor troubled families for longer periods of time following maltreatment determinations or family reunifications. In addition, we saw evidence that some child protection agencies have not fulfilled important duties, such as communicating regularly with mandated reporters and keeping up-to-date records.
We also examined the education and training of child protection staff. We found that about 32 percent of Minnesota’s child protection staff have master’s degrees, typically in social work. Another 67 percent have bachelor’s degrees, and a majority of these employees had majored in social work. More than half (55 percent) of county child protection staff in the seven-county Twin Cities area have master’s degrees, compared with only 12 percent in other counties. Twin Cities child protection staff also tend to have more experience with their current agencies, averaging about 10.6 years of experience compared with 6.5 years for child protection employees elsewhere in the state. Most county human services directors told us in a survey that they "always" or "usually" have adequate training opportunities for their staff.
Partly because counties’ maltreatment-related records are private data, it has been difficult for the public, policy makers, and professionals who work with families to know whether the child protection system has been effective. We examined various options for improving the system’s accountability.
One option is external review of child protection agencies. State law requires DHS to "implement a method of monitoring and evaluating social services, including site visits that utilize quality control audits to assure county compliance with applicable standards, guidelines, and the county and state social services plans." Although DHS reviews county social services plans, we found that DHS has not systematically monitored county compliance with state child protection regulations since 1991. An alternative type of external review could focus on the appropriateness of child protection decisions, rather than compliance with regulations. The only such state-level case review has occurred through Minnesota’s child mortality review panel, which was created in 1989 but was inactive between 1995 and late 1997. External review of a county’s child protection agency could be done by (1) staff from DHS or the child protection agency of a similar county, (2) citizen review boards, such as those required (but not yet implemented in Minnesota) in states by a 1996 federal law, or (3) a special office created by the Legislature for this purpose--such as an ombudsman, case monitor, or inspector general. If such reviews are done, we think they should be conducted by people with a sufficient understanding of relevant laws, rules, and social work practices.
Another option for improving accountability is county agency self-monitoring and reporting. Since 1981, state law has required counties to prepare annual reports on "the effectiveness of the community social services programs in the county." Counties have prepared information on the number and type of social service recipients, but most have not regularly evaluated program effectiveness. Some counties have developed useful performance measures of child welfare services for their biennial social services plans, but most counties’ plans contain few measures and limited information on prior performance.
The 1997 Legislature considered but did not pass legislation to open CHIPS hearings to the public--another option for making the child protection system more accountable. Our study did not address the issue of open CHIPS hearings, but we did ask human services directors whether certain child protection agency records should be made public. Fifty-seven percent said they favor or might favor making records public in cases involving child deaths, and 39 percent said they favor or might favor opening records of cases involving serious injuries. Federal law requires states receiving federal grants to have methods of keeping child protection records confidential, but records may be released to persons "statutorily authorized by the State to receive such information pursuant to a legitimate State purpose" and states must publicly disclose "findings or information about" cases of maltreatment that result in child fatalities or near fatalities.
There may be other ways to make child protection agencies more accountable, such as improved staff supervision or stronger oversight by county boards. For example, only about one-third of county human services directors said that their child protection supervisors "always or almost always" review case evidence before maltreatment determinations are made. In addition, county policies for screening child protection cases have usually not been a subject of public discussion.
Child protection agencies throughout the nation make critically important decisions in the lives of families. In Minnesota, however, they do so with limited guidance in state laws and rules, considerable reliance on local property taxes, and little oversight by state government or others. The result is a system of widely varying practices and standards, sometimes operating without the full confidence of the public or the professionals who make many reports of maltreatment.
County variation can reflect differences in community norms and differences in local willingness or ability to pay for services. But variation sometimes reflects different interpretations of state laws and rules. In our view, these laws and rules provide insufficient direction to counties, and the definitions of maltreatment should be a topic of greater public discussion. We recommend:
Alternatively, the Legislature could require each county board to adopt its own maltreatment definitions to reflect local standards, without requiring definitions in state rules. But our survey of county human services directors indicated that 61 percent favored additional guidance in state rules about circumstances or evidence that justify a determination of maltreatment, and another 22 percent said they might favor such guidance. DHS should also consider developing training materials (and perhaps rules) that help child protection investigators evaluate the credibility of evidence and make decisions when evidence is conflicting.
We think steps should be taken to help mandated maltreatment reporters regain confidence in the child protection system. In general, counties should place a higher priority on keeping mandated reporters informed about the cases they initially reported and the counties’ criteria for decisions. But we also recommend:
We think there may be times when mandated reporters could better serve children and families if they received information from the child protection agency in addition to case disposition information. For instance, school social workers might be better able to help children if they knew the status of a county investigation involving a family, the county’s assessment of a family’s strengths and problems, or whether a family has been complying with case plan requirements. We recommend:
To reduce the incidence of repeated maltreatment in Minnesota, it may be necessary to improve the way that child protection agencies assess families that are referred to them. Research has raised questions about whether the risk assessment instrument used by nearly all Minnesota counties is the most valid, reliable instrument available. We recommend that:
We think there is a need for county human services agencies to respond more effectively to cases involving child neglect. Several states are experimenting with alternative ways to respond to maltreatment reports. For instance, "dual track" child protection systems are based on the philosophy that some allegations require "investigations" that focus on whether maltreatment occurred while others (such as neglect cases) require less adversarial "assessments" of families’ needs and perhaps an offer of services. According to our survey, 85 percent of county human services directors favor or might favor such a system. The 1997 Legislature authorized county pilot projects to explore the feasibility of alternative methods of handling maltreatment allegations, and we think the Legislature should closely monitor their results. It is possible that these approaches could provide stronger assistance to families and perhaps allow counties to redirect some resources from investigations to services.
Earlier, we noted that there are probably steps that county agencies and courts could take to more effectively protect children from repeated maltreatment--such as longer home monitoring of parents with chemical problems who have neglected their children. In our view, these actions do not necessarily require changes in state law, although they would require continuing commitment and diligence by counties, the courts, and others. Improved case monitoring by counties and courts might also require additional resources.
Because the courts and counties sometimes terminate their involvement with families once the goals of case plans have been met, it might be helpful for state rules and laws to clarify the authority of counties to provide continued monitoring of certain families. For example, it may be reasonable to monitor for extended periods the behavior of caregivers with histories of repeated chemical abuse or maltreatment--as a way of better ensuring the children’s safety. We recommend:
There is no way to guarantee that counties and courts will always make decisions that protect the best interests of children, but there are several options for improving accountability for these decisions. At a minimum, we recommend:
The Department of Human Services should present to the Legislature by January 1999 a plan for periodic, external reviews of (1) county compliance with state requirements, and (2) the appropriateness of decisions made by county child protection agencies in selected individual cases.
The Legislature should direct DHS to establish a "performance measurement task force" of state and county officials to identify by January 1999 (1) statewide measures of the performance of child welfare services, and steps needed to collect reliable information on these measures, and (2) potentially useful practices that individual counties could use to monitor and evaluate child welfare services.
The Legislature should amend state law to require that the determinations made in all investigated cases be reviewed and approved by a county child protection supervisor.
Consistent with federal requirements, the Legislature should require state and local child mortality review panels to review "near fatalities" in addition to child deaths. Also, the Legislature should amend the statutory purpose of the panels to include examining, to the extent possible, whether public agencies took appropriate actions in individual cases. The Legislature should adopt policies (perhaps with input from the state child mortality review panel) for making public the child protection records in cases involving death or near death, including policies that indicate types of information that should not be made public.
In our view, some records of child protection investigations are destroyed too quickly. In many investigations, county staff are unable to assemble the preponderance of evidence required to determine that maltreatment occurred, yet there remains the possibility that it did. A record of these investigations can help county agencies if new evidence on these cases emerges, or if they investigate the same family for subsequent allegations. Such records can also help external reviewers evaluate an agency’s decisions. We think that records of cases that did not result in a determination of maltreatment should continue to be classified as private data, but we recommend that:
Finally, we think the Legislature should consider whether state financial support has been adequate for child protective services. Some Minnesota counties have difficulty adequately serving families for which they have documented abuse or neglect, and many also have difficulty finding resources to serve troubled families before children are harmed. Most state governments have played a more direct role in providing and paying for these services than has Minnesota’s. In light of Minnesota’s unusually high reliance on property taxes to pay for child welfare services, the Legislature should consider ways that state government could financially help counties if it concludes that there is a need to expand child welfare services or reduce child protection caseloads.More Information
The Program Evaluation Division was directed to conduct this study by the Legislative Audit Commission on May, 1998. For a copy of the full report, entitled "Child Protection Services", (98-01), 75 pp., published on January 13, 1997, please call 651/296-4708, e-mail Legislative.Auditor@state.mn.us, or write to Office of the Legislative Auditor, 658 Cedar St., St. Paul, MN 55155.Staff who worked on this project were Joel Alter (project manager), David Chein, Carrie Meyerhoff and Josh Halverson.