Comprehensive Community Mental Health Services for Children
and Their Families Program: Data Sources
1: Data source—the longitudinal outcome study conducted in
communities funded between 1997 and 2000. Findings based on 3,563 caregivers
with complete data at entry and 12 months on their child’s living situation in
the previous 6 months. Data were collected using the Restrictiveness of Living
Environments Scale (ROLES). Caregivers were asked at intake and after 12 months
to report the number of days their child received inpatient hospital care
during the previous 6 months. The number of days per child was calculated,
which was used to calculate the average number of days of inpatient
hospitalization per child at baseline and 12 months. These figures were
multiplied by the national daily cost estimate, resulting in the calculation of
a difference in costs from baseline to 12 months. According to the Agency for
Healthcare Research and Quality (AHRQ), the national estimate of the average
daily cost of inpatient hospital care in 2002 was $1,501 per day (AHRQ, 2004).
2: Data source—the longitudinal outcome study conducted in
communities funded between 1997 and 2000. Findings based on 2,602 caregivers
with complete data at entry and 18 months on their child’s living situation in
the previous 6 months. The Restrictiveness of Living Environments Scale (ROLES)
was used to collect information on types of living arrangements and amount of
time a child spent in each type during the previous 6 months. From these data,
children who spent any time in an inpatient hospital were identified and the
amount of time they spent there was calculated (in days). Because the data was
significantly skewed, the number of days was split so that one or more days
equaled 1. A generalized estimating equation (GEE) for assessing change in this
dichotomous longitudinal outcome was computed, and a statistically significant
change over the two time points (i.e., from baseline to 18 months) was obtained
(p<.0001).
3: Data source—the longitudinal outcome study conducted in
communities funded between 1997 and 2000. Findings based on 2,239 youth with
complete data at entry and 12 months on the Delinquency Survey (DS) for the
relevant question. The DS collected information from youth on engagement in
behaviors that could result in law enforcement involvement and on actual
contacts with law enforcement. One DS question asks youth how many times s/he
was arrested in the previous 6 months. The average number of arrests was
calculated per child at baseline and 12 months were calculated. These figures
were multiplied by the national cost estimate, and a difference in costs from
baseline to 12 months was calculated. According to data from the Bureau of
Justice Statistics, the average cost per juvenile arrest was $4,149 in 2000
(The National Center on Addiction and Substance Abuse, 2004).
4: Data source—the longitudinal outcome study conducted in
communities funded between 1997 and 2000. Findings based on 2,782 caregivers
with complete data at entry into services and at the 18-month follow up on the
Child Behavior Checklist (CBCL), which assesses a child’s behavioral and
emotional problems. Analysis involved calculation of the Reliable Change Index
(RCI) for the Total Problems scale of the CBCL, which compares a child’s scores
at two different points in time (in this case, at baseline and at the 18-month
follow up) while adjusting for the reliability of the measure. The RCI
indicates whether a change in scores shows clinically significant improvement,
stability, or deterioration. Improvement and deterioration are defined as a
difference in outcome scores, adjusted for measurement error of the outcome, so
that a difference of this size would not be due to the unreliability of the
measure.
5: Data source—the longitudinal outcome study conducted in
communities funded between 1997 and 2000. Findings based on 1,887 caregivers
with complete data at entry and 12 months on item 18 on the Child Behavior
Checklist, which asks respondents if their child had deliberately harmed
himself/herself or attempted suicide in the previous 6 months. Responses of
“very true or often true” and “somewhat or sometimes true” were combined.
Analysis involved the use of a generalized estimating equation (GEE) for
assessing change in this dichotomous longitudinal outcome; change over the two
time points was statistically significant (p < .0001).
6: Data source—the longitudinal outcome study conducted in
communities funded between 1997 and 2000. Findings based on 2,433 caregivers
whose children attended school in the previous 6 months and who provided
complete data at entry and 18 months on their child’s school-related outcomes.
Data were collected using the Educational Questionnaire developed specifically
for the national evaluation. The attendance item was split so that 75 percent
of the time or more equaled 1 and less than 75 percent of the time equaled 0.
Analysis involved the use of a generalized estimating equation (GEE) for
assessing change in these dichotomous longitudinal outcomes. Change over the
two time points (i.e., from baseline to 18 months) was statistically
significant (p < .0001).
7: Data source—the longitudinal outcome study conducted in
communities funded between 1997 and 2000. Findings based on 2,433 caregivers
whose children attended school in the previous 6 months and who provided
complete data at entry and 18 months on their child’s school-related outcomes.
Data were collected using the Educational Questionnaire developed specifically
for the national evaluation, including a series of items about the child’s
grades (if the school attended used grades). Grades were split so that A, B, or
C equaled 1 and D or F equaled 0. Analysis involved the use of a generalized
estimating equation (GEE) for assessing change in these dichotomous
longitudinal outcomes. Change over the two time points (i.e., from baseline to
18 months) was statistically significant (p < .0001).
8: Data source—the longitudinal outcome study conducted in
communities funded between 1997 and 2000. Findings based on 1,115 youth with
complete data at entry, 6 months, 12 months, and 18 months on the Delinquency
Survey (DS) for the relevant question. The DS collects information from youth
on engagement in behaviors that could result in law enforcement involvement and
on actual contacts with law enforcement. One DS question asks the youth whether
s/he had been in juvenile detention or other secure facilities in the previous
6 months. Positive and negative responses were coded as 1 and 0, respectively.
Analysis involved the use of a generalized estimating equation (GEE) for
assessing change in this dichotomous longitudinal outcome. Change over the four
time points (i.e., baseline, 6, 12, and 18 months) was statistically
significant (p < .0001).
References
Agency for Healthcare Research and Quality, Healthcare Cost and Utilization
Project. (2004). The HCUP Nationwide Inpatient Sample (NIS), 2002.
Rockville, MD.
The National Center on Addiction and Substance Abuse. (2004). Criminal neglect:
Substance abuse, juvenile justice and the children left behind. New
York: The National Center on Addiction and Substance Abuse (CASA) at Columbia
University.
Zeger, S.L. & Liang, K.Y. (1986). Longitudinal data analysis for discrete
and continuous outcomes. Biometrics, 42, 121-30.
|