Compliance Review of Child Protective Services

Update: Improvements Have Been Made in Timeliness of Child Abuse Investigations, But Further Progress Is Needed

This is a compliance review of Child Protective Services (CPS) as required by West Virginia Code §4-10-10a. Child Protective Services, within the Department of Health and Human Resources, was audited by the Legislative Auditor during 1996. The final results of the audit were released in February of 1997. A compliance review is intended to evaluate the extent to which the agency has complied with recommendations contained in the initial audit.


The 1996 audit showed relatively low response time

During the initial audit, the Legislative Auditor's Office sampled 663 CPS case records for fiscal year 1995 from twelve counties. The primary focus was to determine how effective CPS was in its investigation of child abuse allegations. As part of the agency's investigative response, West Virginia Code §49-6A-9 required that Child Protective Services have a face-to-face interview with the child or children within 14 days of the child abuse report. Figure 1 shows the CPS response time in meeting the 14 day standard reported in the 1996 audit.

As Figure 1 indicates, the Legislative Auditor found that in only one-third of the cases, CPS had face-to-face interviews with the alleged victims of child abuse. In 37% of the cases, CPS had no record of having a face-to-face interview. In the remaining 30% of the cases, face-to-face interviews were being conducted well in excess of the 14 day standard. In many cases the interviews were months after the initial referral.

The Compliance Review Involved Two Approaches

The Legislative Auditor's compliance review consisted of the following two approaches:

Approach 1: Examine CPS cases of the four counties that had the lowest response time among the twelve that were sampled in 1996. These four counties are: Kanawha, McDowell, Wood and Wyoming.
Approach 2: Examine CPS cases of four counties randomly selected that were not in the 1996 sample. These counties are: Barbour, Clay, Tyler, and Webster. The purpose for reviewing these new counties is to determine if the recommendations of the 1996 audit are being implemented statewide as opposed to only the counties that were reviewed in 1996.

Table 1 shows the counties with the lowest response time in the 1996 audit. As the table indicates, a large number of cases had no record of ever having a face-to-face interview. The 1996 audit revealed that the agency had a prioritization policy in which cases were given priority ranking. Cases with low priority were not immediately investigated. Many cases were going months without investigation. If a case was not investigated after six or more months and no additional referrals came in on the family, the case would be closed even though no investigation had been conducted.

For these counties, no more than one-fourth of the cases had face-to-face interviews in the required 14 days. One worker had indicated that during fiscal year 1995 it was not known that face-to-face interviews had to be conducted with alleged victims of child abuse within 14 days. The table also shows that a significant percentage of cases had interviews with the children several weeks or months after the initial referral was made.

Table 1
1996 Audit: Counties With Lowest Response Time

Region IRegion IIRegion IVRegion IV
WoodKanawhaWyomingMcDowell
Percentage of Cases without record of Face to Face interviews30.0%63.6%56.6%25.0%
Interviews within 14 days25.0%25.8%18.9%18.3%
Interviews in 15 to 90 days21.7%7.6%5.7%26.7%
Interviews above 90 days23.3%3.0%18.9%30.0%

In order to monitor the CPS response time since the 1996 audit, the Legislative Auditor reviewed 262 CPS cases from Kanawha, McDowell, Wood, and Wyoming counties. The CPS cases were randomly selected for the period of January 1, 1997 through September 30, 1997. Table 2 below shows the response time for these counties from the 1997 sample.

Table 2
1997 Review: Counties With Lowest Response Time in 1996

Region IRegion IIRegion IVRegion IV
WoodKanawhaWyomingMcDowell
Percentage of Cases without record of Face to Face interviews8.3%10.6%3.3%3.4%
Interviews within 14 days56.7%81.8%68.9%76.3%
Interviews in 15 to 90 days35.0%7.6%18.0%20.3%
Interviews above 90 days0.0%0.0%9.8%0.0%

As Table 2 shows, there has been improvement since the 1996 audit. No more than 10% of the cases had no record of a face-to-face interview. For some of those cases, there were indications that CPS workers made attempts to locate the family but they were unsuccessful. Other cases had no documentation to support an interview had occurred. This is a marked difference from the previous audit, in which most of the cases without record of face-to-face interviews had no evidence of there even being an attempt by CPS workers to interview children. Graph 1 illustrates the difference in this category for each county.

The percentage of cases that had interviews with children within 14 days of the referral was up substantially for all counties. Graph 2 shows the difference in compliance with the 14 day standard for each county between the 1996 audit results and the 1997 compliance review.

There are still a relatively large percentage of cases that have face-to-face interviews beyond 14 days. This should be a concern for the obvious reason that children are at risk of further abuse the longer a case is not investigated, and it becomes more difficult to substantiate an allegation if abuse actually occurred. For example, a child's injury may heal before the CPS worker investigates. However, there is some improvement from the 1996 audit, in that the average length of time is shorter in cases where interviews were held between 15 and 90 days. Table 3 shows this difference.

Table 3
Average Number of Days to Conduct Interviews With Children
When Interviews Were Conducted Between 15 and 90 days of Referral
19961997Difference
Kanawha26 days18 days-8 days
McDowell55 days23 days-32 days
Wood41 days27 days-14 days
Wyoming35 days23 days-12 days

It is also a positive note that with the exception of Wyoming county, the other three counties did not conduct interviews with children beyond 90 days. Although Wyoming had nearly 10% of its cases fall into this category, it was still down from 19% in the 1996 audit. Overall, there has been improvement for the four counties with the poorest response times in the 1996 sample.


Review of Four New Counties Is Mixed

Table 4 shows the results of the 1997 sample of CPS cases for Barbour, Clay, Tyler, and Webster. For Tyler and Clay counties, over 90% of the cases were within 14 days of the referral. Every case in the sample for these counties had documentation of having a face-to-face interview with the children. A relatively small number of cases had interviews beyond the 14 day standard.

For Barbour and Webster counties, only about half the cases had interviews within 14 days of the referrals. A sizable percent of cases had interviews exceeding 14 days. The average length of time to conduct the interview when they exceeded the 14 day standard was 33 days for Barbour county and 27 days for Webster county. Workers at Barbour county informed Legislative Auditor staff that a staffing problem caused problems in responding within the appropriate timeframes.



Table 4
1997 Review: Response Time of New Counties

Region IRegion IIRegion IIIRegion IV
TylerClayBarbourWebster
Percentage of Cases without record of Face to Face interviews0.0%0.0%13.6%7.7%
Interviews within 14 days96.0%90.9%54.5%53.8%
Interviews in 15 to 90 days4.0%9.1%31.8%38.5%
Interviews above 90 days0.0%0.0%0.0%0.0%

The results from the 1997 review, with respect to response time for the eight counties surveyed, suggests that improvement has been made since the 1996 audit. The new counties surveyed show that there is some evidence of a Statewide emphasis on meeting the 14 day standard. Particularly, two counties were meeting the 14 day time period only about half the time. However, the results for all eight counties suggests while most cases are being responded to, CPS still has some difficulties in conducting interviews with children within 14 days. The Legislative Auditor acknowledges that inadequate staffing continues to present problems. The loss of one worker in a region can seriously hinder appropriate response time. The retention of CPS workers will require the agency's continued attention. The Legislative Auditor commends Child Protective Services for its improvement, and encourages the agency to further its progress.


The Investigative Process Also Has Improved

The 1996 performance audit identified the agency's inadequate use of a system that it purchased to improve its investigations of child abuse reports. The Child At Risk Field (CARF) System was developed by ACTION for Child Protection out of years of research in the field of child protection services. Two of the system's greatest assets are its ability to reduce the subjectivity in evaluating a child's environment, and its method of quantifying levels of risk of future child abuse from low to high risk. These parts of the CARF system comprise the process called the Initial Assessment. This process is important because if families have moderate to high risk of future child abuse, research shows that the family's situation will worsen without CPS intervention. When the initial assessment is used it provides a structured and objective approach to evaluating child abuse allegations.

In the 1996 audit, it was estimated that statewide, only 56% of CPS cases had initial assessment performed. Instead of using the initial assessment instrument, CPS workers would simply record their findings in narrative form and make a personal judgement on the risk of future child abuse. The narrative approach was used frequently because the initial assessment instrument requires more time to organize the information, rank the significance of the information, and determine the risk rating. The 1996 audit showed cases in which CPS workers indicated through the narrative approach that risk of future child abuse was present in a family. However, this approach cannot quantify the level of risk, and in many instances, cases were closed. The danger of not using the more rigorous technique of the initial assessment is that families in need of help will not be offered the services they need. Consequently, children could be abused again in the future.

In the 1997 compliance review, a similar case was identified which provides supportive evidence for using initial assessments instead of strictly using the narrative approach. A Wood county case was investigated by a CPS worker who did not perform an initial assessment. His original determination was to close the case because family members would not provide enough information to substantiate the allegations. Upon review of the case by a CPS supervisor, the CPS worker was instructed to use the initial assessment instrument to assess the risk of future abuse. When the initial assessment was used, it was found that there was high risk of future abuse and the family needed to receive child protection services. The case was opened for services.

Graph 3 illustrates the use of initial assessments by Kanawha, McDowell, Wood, and Wyoming counties. The graph reveals that there has been a substantial increase in the use of the instrument. The 1996 audit showed that of the four counties with the lowest response times, only McDowell county used the initial assessment procedure frequently. The other three counties used it infrequently. The greater use of the initial assessment should provide for more accurate assessment of a child's environment and the family's need.


Although Wood county's use of the initial assessment improved, it is still relatively low. A similar finding for Wood county was made by the agency in an internal audit initiated by the regional director of region one. Twenty-six cases were sampled in Wood county in April of 1997. Eighteen of the twenty-six cases did not have initial assessments. Therefore, only 30% of the sample had initial assessments performed. The report indicated that "They are not addressing risk since they are not completing Initial assessments on all accepted referrals. Instead, workers were using the narrative approach on service documentation logs. This method is clearly inferior to initial assessments for reasons stated previously.

The four new counties had a relatively high percentage of cases with initial assessments completed. This is illustrated in Graph 4. Barbour county was the one exception. However, Barbour county is part of a pilot project called the Family Option Initiative that is examining certain modifications to the CARF system. Part of this modification involves substituting the initial assessment instrument with another form for certain cases. In these cases, only safety is evaluated, not risk. This explains the relatively low use of initial assessments in Barbour county. It was not in the scope of this compliance review to evaluate the effectiveness of the Family Option Initiative. The relatively high use of initial assessments in other counties suggests that there is a Statewide emphasis in this area. However, Wood county's results suggests that there may be other counties that still need to improve in the use of initial assessments.

Variation Still Exists By County

Although improvements are evident, the 1997 compliance review reveals that CPS performance still varies significantly by county. This is true with respect to response times and in the use of initial assessments. This reinforces the need for a statutory requirement which mandates the Department of Health and Human Resources to conduct detailed performance evaluations. This was recommended in the 1996 audit. The following is a restatement of that recommendation.

The Legislature should consider amending state law to require the Department of Health and Human Resources to conduct detailed performance evaluations, to include, but not to be limited to, the child protective services program of every local county office, once every two years. The Legislature should consider requiring such evaluations to be conducted with a sample size that is statistically significant. The Legislature should also consider requiring DHHR to prepare a full and detailed report of its findings and include any proposals to rectify any deficiencies noted, upon completion of each county audit. [Recommendation 6 of the 1996 audit.]

Without statutory authority, the improvements may not continue after changes in administrations. Also, agency priorities could de-emphasize the need to monitor the performance of CPS. It is important to note that the improvements made were accomplished without a significant increase in staffing. According to the agency, total staff Statewide as of November 30, 1997 was 252. This is only two more than the total at the end of December of 1996. Although in the 1996 audit the Legislative Auditor acknowledged the need for additional staff, however, it was indicated that the development of management information was more important.

The agency has provided the Legislative Auditor's Office with certain aggregated Statewide performance statistics. Some of the statistics showed that for the period between July 1996 and April 1997, the Statewide average of 10% of the cases did not have a face-to-face interview within 14 days.
This was close to the estimates for three of the eight counties sampled for this review, but the other five counties sampled show higher percentages of cases without face-to-face interviews within 14 days. This reflects the variation that exist by county. The Legislative Auditor commends DHHR for compiling such statistics. Nevertheless, these statistics cannot possibly provide the same type of valuable information as can a detailed performance evaluation that involves sampling of cases. Case sampling could identify program needs and deficiency across the entire process of child protection services.

A good example of this is the Family and Children Service Delivery System Review conducted in DHHR's region one. This self review process evaluated counties along several aspects of child protection and foster care. The process also required the county to report to the regional manager a corrective action plan. It was indicated that this was only being done in Region One. Despite the efforts initiated in region one within the DHHR, this effort could lose priority if other agency needs take greater priority. Furthermore, changes in administration may not necessarily continue a comprehensive review process without statutory authority. Therefore, the Legislative Auditor makes the following recommendations.





Recommendation 1:

The Legislative Auditor recommends that the Legislature consider implementing recommendation 6 of the 1996 audit on Child Protective Services, which reads as follows.

The Legislature should consider amending state law to require the Department of Health and Human Resources to conduct detailed performance evaluations, to include, but not to be limited to, the child protective services program of every local county office, once every two years. The Legislature should consider requiring such evaluations to be conducted with a sample size that is statistically significant. The Legislature should also consider requiring DHHR to prepare a full and detailed report of its findings and include any proposals to rectify any deficiencies noted, upon completion of each county audit.

Recommendation 2:

The Legislative Auditor also recommends an additional compliance review of CPS within the next year.