Description:
The KSFI assesses parents’ risk for child
maltreatment and/or caregiving difficulties. It is a 10-item
rating scale that is completed after a through psychosocial
interview with a trained professional. Service providers,
such as home visitors, may also rate families based on the
interactions they have had with them over a period of time
(Korfmacher, Younge, and Michalek 1996). Items on the KSFI
assess parents on a number of domains, such as psychiatric
and criminal history, childhood history of care, emotional
functioning, attitudes towards and perception of children,
discipline of children, and level of stress in the parent’s
life. Parents receive a raw score and are determined to be
at low, moderate, or high risk, depending on the cut-offs
established by the program administering the scale (see section
on interpretability, below).
Uses of Information: The KFSI is currently used with at-risk
families as an integral part of the screening and assessment
process for two home-visiting programs for families at-risk,
the Hawaii Healthy Start and the nation-wide Healthy Families
America program. It is primarily used as a second-level screening
tool for mothers who are considered at risk based on a 15-item
hospital chart review that assesses sociodemo-graphic risk
factors, such as maternal age and income.
Reliability: (1) Inter-rater reliability: a Healthy Families
America program site in Oregon collected data on 115 families,
and notes from each of their psychosocial interviews were
reviewed by independent raters. The reliability correlation
coefficient between raters on classifying parents into the
low, mild, or severe risk categories was .93. (Katzev et al.
1997).
Validity: Predictive validity:
three studies that compared KFSI scores to child maltreatment
based on hospital records, state child protection reports,
and scores on the Child Abuse Potential Inventory have shown
KFSI to have predictive validity (Murphy et al. 1985; Hawaii
Department of Health 1992; and Katzev et al. 1997). Three
studies of predictive validity were conducted using varying
cut-off scores and time periods. Positive predictive validity
reflected the percentage of mothers with high-risk scores
who later maltreated their children, and ranged from 25 to
52 in two of the studies. According to Korfmacher (2000),
the low percentage (3 percent) in a third study may be an
artifact of the study’s reliance on child protection
reports to measure the incidence of maltreatment. Negative
predictive validity measured the percentage of mothers with
low-risk scores who demonstrated no future evidence of child
maltreatment; it ranged from 85 to 100 percent. The KFSI sensitivity,
which is the percentage of maltreating mothers who were scored
at high risk, ranged from 80 percent to 97 percent across
the three studies. The specificity, which is the percentage
of non-maltreating mothers who were scored at low risk, ranged
from 21 to 89 percent.
Method of Scoring: A three-point scale, ranging from low
to high risk, is used to score both parents. Parents who receive
a high-risk score are considered at risk for child maltreatment.
Two methods have been employed in scoring the KFSI. Carroll
(1978) used weighted scoring, in which items that were determined
to be more immediate precedents to child abuse (such as violent
outbursts and harsh punishment of child) were given higher
scores, while lower weighting was given to items that were
perceived to be less important in assessing immediate risk
(such as parents’ history of child maltreatment, low
self-esteem or isolation, and having an unwanted child). If
a weighted scaling system is not used, items are assigned
values according to whether there is no risk (0), risk (5),
or high risk (10) of child maltreatment. The total score is
obtained by summing the values assigned to each item, although
the weighted system requires that examiners multiply raw scores
by 2.5. Total scores range from 0 to 100.
Interpretability: KFSI users have applied different cutoff
scores based on their clinical judgment to identify at-risk
parents. Carroll (1978) felt that scores above 30 were “concerning”
and that scores above 70 were “particularly differentiating”;
however, others have defined different cutoffs and have assigned
labels (such as, low, medium and severe risk) to scores in
a particular range. To aid in interpreting scores, Healthy
Families America and the Family Stress Center have created
and copyrighted supplemental rating criteria.
Training Support: Individuals must participate in a specialized
training on using the supplemental rating criteria, which
have been copyrighted by Healthy Families America and the
Family Stress Center.
Adaptations/Special Instructions for Individuals with Disabilities:
None described.
Report Preparation Support: None described.
References
Carroll, C.A. “The social worker’s evaluation.”
In The child protection team handbook, edited by B.D. Schmitt.
New York: Garland STM Press, 1978, pp. 83-108.
Center on Child Abuse Prevention Research. Intensive home
visitation: A randomized trial, follow-up, and risk assessment
study of Hawaii’s Healthy Start program. Final report
prepared for the National Center on Child Abuse and Neglect.
Chicago, IL: National Committee to Prevent Child Abuse, 1996.
Hawaii Department of Health. Report to the 16th Legislature,
State of Hawaii, on House Bill #139, c.d. 1: Requesting review
and recommendations from the Director of Health on the Healthy
Start Program. Honolulu, HI: Maternal and Child Health Branch,
Hawaii Department of Health, 1992.
Katzev, A., T. Henderson, and C. Pratt. Predicting child
maltreatment with the Kempe Family Stress Assessment. Document
presented at HFA: Rethinking the Assessment Process working
meeting. Chicago, IL: National Committee to Prevent Child
Abuse, 1997.
Korfmacher, J., P. Younge, and P. Michalek. Standardized
assessment of child abuse risk: A review. Paper presented
at 24th Annual Child Abuse and Neglect Symposium, Keystone,
CO, 1996.
Korfmacher, Jon. The Kempe Family Stress Inventory: A review.
Child Abuse and Neglect, vol. 24, 2000, pp. 129-140.
Murphy, S., B. Orkow, and R.M. Nicola. Prenatal prediction
of child abuse and neglect: A prospective study. Child Abuse
& Neglect, vol. 9, 1985, pp. 225-235.
|
