Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Oct 1.
Published in final edited form as: J Consult Clin Psychol. 2018 Oct;86(10):845–855. doi: 10.1037/ccp0000327

Three Perspectives on Mental Health Problems of Young Adults and their Parents at a 15-Year Follow-up of the Family Bereavement Program

Irwin Sandler 1, Heather Gunn 1, Gina Mazza 2, Jenn-Yun Tein 1, Sharlene Wolchik 1, Hanjoe Kim 3, Tim Ayers 1, Michele Porter 1
PMCID: PMC6166649  NIHMSID: NIHMS975067  PMID: 30265043

Abstract

Objective

Effects are reported of the Family Bereavement Program (FBP) on the mental health of bereaved youth and their surviving parent fifteen years following the program.

Method

One hundred and fifty six families (244 children ages 8-16; 54% male; 67% Non-Hispanic White) were randomly assigned to receive either the FBP (N=90) or a literature control condition (N=66). At the 15-year follow-up 80% of the youth and 76% of the bereaved parents were re-interviewed. Mental health problems and service use were self-reported by young adults and their parents. Key informants reported on mental health problems of young adults.

Results

Young adults in the FBP reported significantly less use of mental health services and of psychiatric medication than controls. Key informants reported significantly lower mental health problems for young adults who were in FBP as compared with controls and for those who were younger lower internalizing and externalizing problems for those in the FBP as compared with. controls. Bereaved parents reported a significantly lower rate of alcoholism and less use of support groups than controls.

Conclusions

The results provided evidence that FBP led to lower mental health problems and less service use by bereaved young adults and their parents as compared with controls.

Keywords: Randomized trial, bereaved children, bereaved parents

Multiple prevention trials have reported that family-focused preventive interventions delivered in childhood have significant effects that last into young adulthood (Hill et al., 2014; Spoth et al., 2013; Van Ryzin & Nowicka, 2013; Wolchik et al., 2013). Young adulthood is an important developmental period because many mental disorders have their median age of onset (Burke, Burke, Regier, & Rae, 1990) and/or increase in prevalence during this time (Kessler et al., 1994). Although multiple reporters are often used to study the effects of prevention programs on child mental health, studies of the effects of prevention programs on adult mental health have relied primarily on self-report measures. The current paper illustrates the use of three perspectives—self-report of problems, knowledgeable informant report of problems, and reports of service utilization—to evaluate the 15-year follow-up effects of the Family Bereavement Program (FBP), a preventive intervention with families who have experienced the death of a parent (Sandler, Wolchik, Ayers, Tein, & Luecken, 2013). The evaluation will report on the effects of the FBP on the mental health problems of the surviving offspring and their spousally-bereaved parent fifteen years following program participation.

Family Bereavement Program (FBP) Designed to Promote Resilience of Bereaved Children and Parents

The FBP is a 12-session program that seeks to prevent the development of mental health problems of bereaved children and their parent by teaching and supporting the use of skills that promote effective adaptation to the stressors that follow the death of a parent (Ayers et al., 2013-2014; Sandler et al., 2013). Theoretically, according to the dual-process model of bereavement (Stroebe & Schut, 1999) the bereaved are coping with loss-oriented stressors (e.g., feelings of grief) as well as restoration-oriented stressors (e.g., the everyday demands of finding safety, control and meaningful relationships in their changed world). More effective adaptation to these stressors is expected to lead to lower levels of mental health and other problems that sometimes follow the experience of parental bereavement (Brent, Melhem, Donohoe, & Walker, 2010; Melhem, Walker, Moritz, & Brent, 2008; Zisook & Schucter, 1991). The FBP consists of small groups that teach and support skills that promote the ability of parents and children/adolescents to meet the demands of these stressors or to reduce their occurrence (Ayers et al., 2013-2014; Sandler, et al., 2013). The program was designed to strengthen protective factors that theoretically promote effective adaptation to these stressors and that have been empirically associated with better outcomes for bereaved children (e.g., positive caregiver-child relationship quality, effective discipline, positive child coping) and to reduce risk factors that have been associated with worse outcomes for bereaved children (e.g., caregiver mental health problems, negative events, threat appraisals for negative events) (Ayers et al., 2013-2014; Sandler et al., 2013). The parent component of the program teaches parents skills to enable the family to restructure as a stable and supportive environment, in which positive experiences are shared, rules are fairly enforced, and in which their experiences and feelings can be understood. The program promotes parent resilience by providing a supportive group environment which helps them work on their own self-defined goals and by teaching them to challenge cognitive distortions that lead to depression. The child/adolescent component of the program teaches children to effectively cope with their feelings of grief as well as the stressors of everyday life, to challenge negative appraisals of stressors in their lives, and to use adaptive beliefs about their control over negative events (Haine, Ayers, Sandler, Wolchik & Weyer, 2003). Theoretically, strengthening the adaptive skills of bereaved parents and children will lead to a cascade of positive effects across time, resulting in lower levels of mental health problems fifteen years later for both parents and children (for prior examples of cascading effects of preventive intervention effects on children and parents over time see Sandler et al., 2016; Wolchik et al., 2016).

The approach of the FBP to teach and support parenting and coping skills distinguishes it from the primary approach of psychoeducation about grief and feeling expression that characterize most interventions for bereaved children represented in a recent meta-analysis (Currier, Holland, & Neimeyer, 2007). That meta-analysis reported a small and non-significant effect size of the child bereavement interventions. The focus on parenting skills distinguishes the FBP from programs for bereaved adults that were the subject of a meta-analytic review (Currier, Neimeyer, & Berman, 2008) and that found an overall small effect size at posttest, but no significant effect at follow-up.

Importance of Long-Term Follow-up and of Multiple Perspectives in the Evaluation of Program Effects on Mental Health Problems

This report contributes to the literature on programs for bereaved children and their parents in four ways. First, this is the longest term follow-up of a randomized trial of a program for the bereaved. Prior literature reviews have pointed to a dearth of evaluations of the long-term effects of such programs (Currier et al., 2008). Prior evaluations found lower levels of mental health problems in children and parents in the FBP as compared with controls at 11 months and six years after program participation (Sandler et al., 2003; Sandler et al., 2010). At the fifteen-year follow-up the children were young adults, a developmental period when the prevalence of mental disorders is elevated (Kessler et al., 1994). Second, this study uses both self-report and knowledgeable informant report on young adults’ mental health problems. Similar to most studies of program effects we assess the self-report of mental health problems by the youth and their parents. There is considerable evidence, however, that adult self-reports of mental health problems are subject to biases that may distort findings of program effects (Carlson, Vazire, & Oltmanns, 2013) as compared with the reports of a knowledgeable informant. Furthermore, self-report and knowledgeable informant reports of mental health problems have been found to uniquely predict other outcomes such as occupational success (Fiedler, Oltmanns, & Turkheimer, 2004) and physical health (Smith et al., 2008). In addition, reports from knowledgeable informants provide a source of information about adult mental health problems that is not affected by potential self-report bias. Third, the study assesses program effects on use of mental health service as compared to a control condition. Although use of services is determined by multiple factors, those with a mental disorder are four times more likely to receive services as those without a disorder (Wang et al., 2005). Some of the factors that influence use of mental health services such as severity of the problems and degree of disability from mental health problems (Bland, Newman, & Orn, 1997; Elhai & Ford, 2007; Leaf et al., 1988; Rayburn et al., 2005) theoretically would be reduced by a successful prevention program. Fourth, although the sample is seeking services, it is not selected by any screening measure to be high on mental health problems, grief, or other indicators of distress. Prior literature reviews have reported that programs for bereaved are more effective when delivered to those who have elevated problems and have questioned the benefit of programs for unselected bereaved populations (Currier et al., 2008). The current evaluation assesses long-term benefits of the program on this unscreened sample as well as whether the effects differ by level of problems when they entered the program.

This paper reports on the effects of the FBP on the mental health problems of the young adults and their bereaved parents 15 years after the intervention. We hypothesized that young adults who participated in the FBP fifteen years earlier as compared with literature controls (LC) will have lower levels of mental health problems as assessed by self-report and knowledgeable informants report, and of mental health services utilization. Additionally, parents who participated in the FBP as compared to LC will have lower levels of mental health problems and mental health service utilization. The study will also examine whether program effects are moderated by baseline problems, gender, or age, which have moderated program effects in prior evaluations of family focused prevention programs (for a review see Sandler, Ingram, Wolchik, Tein, & Winslow, 2015).

Method

Participants

Participants were 244 young adults (135 FBP from 90 families, 109 LC from 66 families) and 131 spousally bereaved parents (72 FBP, 59 LC) who participated in the FBP trial fifteen years earlier. Consistent with our prior report of program effects (Sandler et al., 2016), only the parents who had been married to the deceased parent (131 out of the 156 caregivers in the FBP) were included in the evaluation of long-term program effects on the parents. In addition, 157 key informants (people the young adult reported knew them the best) provided information about the young adults; however, one of the key informants reported no face-to-face, telephone, or email contact with the young adult in the past year and thus was dropped from the analyses. The key informants were spouse or romantic partner (27.5%), relative (e.g., parent, brother, sister, 39.7%), or friend or other non-relative (32.7%).

Procedures for participant recruitment and randomization are more fully described in a previous publication (Sandler et al., 2003), but are briefly reviewed here1. Bereaved families were recruited from community agencies (e.g., schools, service agencies) in a large southwestern metropolitan area. Families were eligible for the program if they met the following criteria: family experienced parental death between four and thirty months prior to beginning the program; family had one or more children between the ages of 8 and 16; family was not currently receiving other mental health or bereavement services; family was willing to participate in either the group FBP or the self-study LC program; caregivers and youth were able to complete the assessment in English; youth were not in a special class for mentally challenged persons; and family planned to stay in the area for the next six months. After the pretest, families were excluded and referred for treatment if the youth or caregivers expressed suicidal ideation that included an intent or a plan or met diagnostic criteria for mental disorders that might interfere with their participation (i.e., major depressive disorder for caregiver; conduct disorder, oppositional defiant disorder or attention deficit/hyperactivity disorder not being treated with medication for youth). Following the pretest, families were randomly assigned (using computer generated randomization) to the FBP (90 families and 135 children) or to the LC condition (66 families and 109 children). Prior evaluations show that families in the FBP and LC were comparable on the demographic variables and baseline outcomes (Sandler et al., 2003), except that the percentage of non-Hispanic White participants were lower in the FBP than in the LC (64% vs. 72%). On average, parental death occurred 10.81 months (SD = 6.35, range = 4 – 30) prior to the program. Cause of death was 67% illness, 20% accident, and 13% homicide or suicide. Of the caregivers, 63% were mothers, 21% were fathers, and 16% were a relative or friend. Of the youths, 54% were males, mean age at program entry was 11.39 years (SD = 2.43, range = 8 – 16, 132 children [ages 8 – 11], 112 adolescents [ages 12 – 16]). Ethnicity of the families was 67% European American, 16% Hispanic American, 7% African American, 3% Native American, 1% Asian American or Pacific Islander, and 6% other. Median family income was between $30,000 and $35,000 and 15.9% were below the poverty line according to U.S. Health and Human Services poverty guidelines for 1996.

FBP and Literature Control Conditions

The FBP, which was developed by the authors and their colleagues, is fully described elsewhere and is only briefly described here (Ayers et al., 2013-2014; Sandler et al., 2013). The FBP is a 12-session program that included separate groups for parents, children, and adolescents, plus two individual family sessions. The parent component focused on supporting parents’ adaptive grief processes (e.g., identifying and working on their bereavement-related goals, reframing depressive cognitions about stress) and strengthening adaptive parenting for bereaved families (e.g., increasing positive interactions with their children, active listening, use of effective discipline). The child and adolescent components focused on strengthening effective coping skills (e.g., cognitive reframing, problem solving, adaptive control beliefs, emotional expression). The program was manualized and delivered by two group leaders with master’s or doctoral degrees in the helping professions. As reported previously, fidelity of implementation of the program was very high (Sandler et al., 2003); objective raters reported over 80% of the action items described in the manual were delivered by the group leaders. The self-study condition consisted of three self-help books dealing with grief and a syllabus for reading the books that were sent to the parents, children, and adolescents. Books were selected based on the evaluation by the FBP team that they contained helpful information for bereaved parents and age appropriate information for children and adolescents.2 The percentage of participants who reported reading 50% or more of the books was 42% of the parents, 38% of the adolescents, and 71% of the children.

Participants were interviewed (primarily in their homes) at pretest, posttest, 11 months, six years, and 15 years following completion of the program. Procedures used to retain the sample over time include periodic tracking to obtain updated contact information and periodic newsletters to maintain their interest in the study. Figure 1 provides the CONSORT diagram about participant recruitment and retention over the course of the 15 years of data collection.

Figure 1.

Figure 1

Open in a new tab

CONSORT diagram on recruitment, randomization, and assessment of Family Bereavement Program efficacy trial.

Assessment Procedures at the 15-Year Follow-Up

Interviews were conducted with 194 young adults between 2011 and 2014 (80% of the randomized sample of 244 randomized children) and 99 spousally-bereaved parents (76% of the 131 randomized caregivers who were spousally-bereaved parents). The young adult was asked to nominate “three people who know you the best” who could be contacted to complete an online survey about the young adult. One informant was selected to be interviewed based on the order in which the young adult gave us the name and contact information. Informants were told that the young adults were involved in a study, had given permission to ask them questions about the young adults’ adjustment, and that their responses were confidential. Data were obtained from an informant for 156 young adults (80% of the young adults who were interviewed). Young adults, parents, and informants signed informed consent prior to beginning the interviews. The study was approved by the IRB at Arizona State University.

Measures

Young adult outcomes

The Adult Self-Report (ASR; Achenbach & Rescorla, 2003) was used to obtain young adult self-report of mental health problems during the past month. Items are answered using a three-point scale (Not true, Somewhat or sometimes true, and Very true or often true) during the past month. T-scores were obtained on Internalizing, Externalizing, and Total Problems. Reliability and validity of these measures are good in the normative population (Achenbach & Rescorla, 2003) and in the current sample (Internalizing α = .93; Externalizing α = .90; Total problems α = .97). We also created binary variables to indicate whether internalizing problems, externalizing problems, or total problems exceeded the clinical cut point score of 64 as Achenbach and Rescorla (2003) suggested.

We used the Composite International Diagnostic Interview (CIDI; Robins et al., 1988) to assess diagnosis of mental disorders as defined by the DSM-IV and ICD-10 in the 15 years since participating in the program. The CIDI is a comprehensive, fully structured interview used to assess mental disorders. Participants are asked questions about symptoms of psychiatric disorders that are not due to medication, drugs, alcohol, physical illness, or injury, and additional questions to establish the onset and recency of the symptoms. In order to have power to detect effects on a diagnosed disorder, modules were collapsed into four categories: substance abuse, internalizing, externalizing, and total disorders.

Informants completed the Young Adult Behavior Checklist (YABCL; Achenbach & Rescorla, 2003) to report young adults’ problem behaviors over the past month. T-scores were derived for Internalizing, Externalizing, and Total Problems. Reliability and validity of these measures are good in the normative population (Achenbach and Rescola, 2003) and in the current sample (Internalizing α = .92; Externalizing α = .94; Total problems α = .97). We used the clinical cut point to create binary variables of internalizing, externalizing, and total problems. The correlations between young adults and informants ranged from .43-.45 across the three T-scores.

Parent outcomes

The 21-item Beck Depression Inventory (BDI; Beck & Steer, 1993) was used to assess the severity of depression in the past week. The scale has good reliability and validity (Beck, Steer, & Garbin, 1988) in the scale development samples and good reliability in the current sample (α = .87). Twenty-five of the original 27 items from the Psychiatric Epidemiology Research Interview (PERI; Dohrenwend, Shrout, Ergi & Mendelsohn, 1980) were used to measure non-specific psychiatric distress in the past month. The scale consists of items assessing dread, anxiety, sadness, helplessness, hopelessness, psychophysiologic symptoms, perceived physical health, poor self-esteem, and confused thinking, which are summed to yield an overall score indicating general psychiatric distress. The score has good reliability and validity in the scale development sample (Shrout et al., 1988) and good reliability in the current sample (α = .94).

The Short Michigan Alcoholism Screening Measure (SMAST) was used to assess parental symptoms of alcoholism in the past nine years after the previous interview. Based on recommendations from Crews and Sher (1992) an additional item (“Do you think you were/are an alcoholic?”) was added to the original 13-item scale. The scale had good reliability in the current sample (α = .82). Agreeing with three or more symptoms has been proposed as the cut point for alcoholism (Selzer, Vinokur, & van Rooijen, 1975). Because four of the 14 items reflect self-perceptions of being an alcoholic (e.g., “Do you think you were/are an alcoholic?”) and not actual problem drinking behavior (e.g., “Have you gotten into trouble at work because of drinking during the past nine years?”), we also analyzed the SMAST using only the ten items which reported on problem drinking behaviors. This was done to assess program effects on reports of problem drinking behaviors during the past nine years, separately from self-perception of being an alcoholic which might have been captured at the 6-year follow-up (Sandler et al., 2016). This 10-item scale was treated as a count variable, assessing the number of areas of problem drinking.

Parent and young adult service utilization

An adapted version of the Services Assessment for Children and Adolescents (SACA; Stiffman et al., 2000) was used to assess young adult and parent reports of their own use of 13 services (e.g., mental and behavioral health services, psychiatric prescription drug use) in the past year. The SACA has good test-retest reliability for previous year service use (Horowitz et al. 2001). Because of the low number of endorsements in some areas, we collapsed and re-categorized the areas to total number (i.e., count) of doctor visits for mental health problems, counselor/advisor (counselor, spiritual advisor, and other healer) visits for mental health problems, psychiatric related medications, and attendance at support group meetings. We omitted areas that had very low base rates (e.g., psychic visits).

Data Analytic Strategy

Analyses were done separately for young adults and spousally-bereaved parent outcomes. To determine if either internal or external validity were affected by loss to follow-up, we ran attrition analyses based on the method proposed by Jurs and Glass (1971). Young adults were considered to be present if they or their key informant provided data. We used a chi-square test to examine differential attrition rates across intervention groups, and analysis of variance or logistic regression to examine group x attrition status interactions effects on continuous and categorical pretest variables, respectively. Additionally, we flagged potential outliers based on having a Cook’s distance over 1.0 (Cook, 1977).

To test the intervention main or moderation effects on the 15-year follow-up outcomes, we used regression analysis for continuous outcomes, logistic regression for binary outcomes, and zero-inflated Poisson models for count outcomes (i.e., accounting for a large number of zeros), controlling for baseline risk. For young adult outcomes (as reported by young adults and key informants), baseline risk was defined as the average of the caregiver and child reports of internalizing [caregiver: CBCL (Achenbach, 1991a); child: Children’s Depression Inventory (Kovacs, 1981) and Revised Children’s Manifest Anxiety Scale (Reynolds & Richmond, 1978)] and externalizing problems [caregiver: CBCL (Achenbach, 1991a); child: Youth Self Report (Achenbach, 1991b)]. For parent outcomes, baseline risk was defined as the average of PERI, BDI, and the Texas Revised Inventory of Grief (Faschingbauer, DeVaul, & Zisook, 1977). There are two components of the zero-inflated Poisson model: a Poisson model, which analyzes the count outcome including those who produce counts of zero with some probability (e.g., who might not have a zero count at different assessment points), and a logit model, which analyzes the odds ratio of the individuals with structural zeros (e.g., those who would always have a zero count at any assessment points) versus the others (Coxe, West, & Aiken, 2009).

We used an intent-to-treat approach and handled missing data by using full information maximum likelihood with Mplus 7.3 (Muthén & Muthén, 1998-2014). Because children were nested within families and families were nested within intervention group, nonindependence of the data points could impact the standard errors (i.e., biased downward). The intraclass correlations (ICCs) and design effects were low by intervention group (average youth group size = 6.75 – MICC = .005 for youth self-report and MICC = 0 for key informant report of young adult variables; average parent group size = 5.54 – MICC = .002 for parent variables) but not by family (average children per family = 1.56 – MICC = .13 for youth self-report and MICC = .007 for key informant report of young adult variables). To account for the dependency of young adult scores clustered by family, we used a sandwich estimator to calculate the robust standard errors for the young adult outcomes (Yuan & Bentler, 2000).

We tested whether the intervention effects were moderated by young adult gender or age for young adult outcomes, parent gender or age for parent outcomes, and baseline risk for all outcomes one moderator at a time. Significant interaction effects were probed to examine simple main effects. For continuous moderators, we tested the simple main effect at one standard deviation above and below the mean of the moderator. For categorical moderators, we tested the intervention effect at each level of the moderator. We used the false discovery rate (FDR) procedure (Benjamini & Hochberg, 1995) to correct for multiple tests for each set of analyses (e.g., attrition, mental health outcomes). We interpreted effects and reported effect sizes for effects that had FDR corrected p-values < .10 (Benjamini, Drai, Elmer, Kafkafi, & Golani, 2001). Effect size indicators are Cohen’s d for continuous outcomes, odds ratios (ORs) for binary outcomes, and incidence rate ratios (IRRs) for count outcomes (Cohen, 1988; Pearce, 2004)

Results

Preliminary Analyses

There were no differential attrition rates across groups, no significant group by attrition status interaction effects, and no group or attrition status main effects. Additionally, all Cook’s D values were below 1.0, indicating that none of the data points were flagged as outliers.

Outcome Analyses

Young adult self-report of mental health problems

As shown in Table 1, there were no significant main or moderated effects of the FBP group on young adult self-reported mental health problems after the FDR correction.

Table 1.

Effects of the FBP on Young Adult and Parent Mental Health Problems at 15 Years

Outcome Raw Mean/Proportion (FBP) Raw Mean/Proportion (LC) Unstandardized Program Main Effect [95% CI] Main Effect p-value/P-FDR Significant Moderators (p-value/p-FDR) Effect Size [95% CI]
Young Adult Report
Internalizing
 Continuous ASR 48.65 49.06 −0.23 [−4.17, 3.71] .91/95
 Clinical ASR .11 .18 −0.55 [−1.42, 0.32] .21/.78
 Clinical CIDI .56 .56 0.02 [−0.65, 0.69] .95/.95
Externalizing
 Continuous ASR 48.62 49.77 −0.94 [−4.03, 2.16] .55/.95 Gender (.03/.28)
 Clinical ASR .10 .09 0.14 [−0.87, 1.17] .78/.95
 Clinical CIDI .20 .33 −0.68 [−1.45, 0.08] .08/.75
Total
 Continuous ASR 47.73 48.46 −0.52 [−4.01, 2.98] .77/.95
 Clinical ASR .08 .10 −0.25 [−1.32, 0.81] .64/.95
 Clinical CIDI .61 .66 −0.17 [−0.86, 0.51] .62/.95
Any Clinical Diagnosis ASR .14 .23 −0.59 [−1.37, 0.19] .14/.75
Substance Use CIDI .39 .46 −0.24 [−0.86, 0.38] .44/.95
Service Utilization
 Number of Visits to Doctor for Mental Health 0.26 0.48 −1.00 [−1.67, −.32] .004/.01 IRR = 0.37 [0.19, 0.73]
 Number of Counselor/Advisor Visits 0.40 0.88 −0.48 [−1.65, .70] .43/.49
 Number of psychiatric Medications 0.07 0.16 −2.95 [−4.50, −1.33] <.001/.008 IRR = 0.05 [0.01, 0.26]
 Number of Support Group Meetings 0.94 4.62 −1.86 [−3.09, −0.62] .003/.01 IRR = 0.16 [0.05, 0.54]
Key Informant Report
Internalizing YABCL
 Continuous 52.78 56.01 −3.10 [−6.54, 0.34] .08/.14 Child Age (.014/.03) d−1SDage = 0.46 [0.14, 0.78]
 Clinical .12 .26 −0.97 [−1.83, −0.10] .03/.09 Child Age (.009/.03) OR−1SDage = 0.16 [0.05, 0.54]
Externalizing YABCL
 Continuous 50.62 52.81 −1.96 [−4.85, 0.93] .18/21 Child Age (.05/.09) d−1SDage = 0.35 [0.02, 0.66]
 Clinical .09 .10 −0.09 [−1.12, 0.95] .87/.87
Total YABCL
 Continuous 51.00 53.36 −2.69 [−5.09, −0.30] .03/.09 d = 0.36 [0.04, 0.67]
 Clinical .08 .16 −0.76 [−1.73, 0.21] .13/18
Any Clinical Diagnosis YABCL .15 .29 −0.83 [−1.62, −0.05] .04/.09 Child Age (.01/.03) OR−1SDage = 0.20 [0.07, 0.58]
Parent Report
BDI 7.78 8.33 −0.34 [−2.99, 2.31] .80/.88
PERI 2.01 2.03 −0.02 [−0.24, 0.21] .88/.88
SMAST (14 items)
 Continuous 0.52 1.28 −0.74 [−1.45, −0.03] .04/.08 Age (.02/11) d = 0.36 [0.01, 0.71]
 Alcoholism .04 .20 −1.88 [−3.47, −0.29] .02/.05 OR = 0.15 [0.03, 0.75]
SMAST (10 items) 0.24 0.65
 Count model −1.24 [−2.01, −0.47] .002/.02 IRR = 0.29 [0.13, 0.63]
Service Utilization
 Number of Visits to Doctor for Mental Health 1.57 3.00 −0.74 [−1.59, 0.10] .08/.32
 Number of Counselor/Advisor Visits 1.24 2.05 −0.68 [−1.96, 0.60] .30/.64
 Number of Psychiatric Medications 0.37 0.14 −0.10 [−1.25, 1.05] .86/.86
 Number of Support Group Meetings 0.98 1.84 −2.35 [−4.06, −0.63] .007/.06 IRR= 0.10 [0.02, 0.53]
Open in a new tab

Notes. p-FDR refers to the p-value after running the FDR correction, IRR refers to the Incidence Rate Ratio, d refers to Cohen’s d, OR refers to the Odds Ratio,

ASR = Adult Self-Report, CIDI = Composite International Diagnostic Interview, YABCL = Young Adult Behavior Checklist, BDI = Beck Depression

Inventory, PERI = Psychiatric Epidemiology Research Interview, SMAST = Short Michigan Alcoholism Screening Measure

Key informant report of young adult mental health problems

A significant program by age effect was found for internalizing T-scores on the YABCL [(unstandardized) B = 1.83, SEB = 0.74, z = 2.46, p = .01]. For young adults who were between 8 and 11 years old at pretest, those in the FBP were rated by key informants to have significantly lower internalizing T-scores than those in the LC (at −1SD below mean age: MFBP = 51.50, MLC = 57.97, z = −2.86, p = .004, Cohen’s d = 0.46, 95% CI[0.14, 0.78]). A significant program by age interaction was also found for the clinical cutoff of internalizing problems on the YABCL (B = 0.52, SEB = 0.20, z = 2.60, p = .009). For young adults who were between 8 and 11 years old at pretest, those in the FBP were over six times less likely of having a clinical diagnosis on internalizing problems compared to those in the LC group (at −1SD below mean age: OR = 0.16, 95% CI[0.05,0.54]; NNT = 7.37). A significant program by age effect was also found for externalizing T-scores on the YABCL (B = 1.28, SEB = 0.66, z = 1.93, p = .05). For young adults who were between ages 8 and 10 at pretest, those who were in the FBP were reported to have fewer externalizing problems than those in the LC (at −1SD below mean age: MFBP = 50.23, MLC = 54.49, z = −2.13, p = .03, Cohen’s d = 0.35, 95% CI[0.02,0.66]). There was a significant main effect for total problems T-scores on the YABCL (B = −2.69, SEB = 1.22, z = −2.20, p = .03) that was not moderated by child’s age, such that young adults in the FBP were rated to have fewer total problems than young adults in the LC (MFBP = 54.96, MLC = 57.65, Cohen’s d = 0.36, 95% CI[0.04,0.67]). There was a significant program by age interaction for the clinical diagnosis of internalizing, externalizing, or total problems (B = 0.43, SEB = 0.17, z = 2.49, p = .01). For young adults who were between ages 8 and 11 at pretest, those in FBP had lower odds of exceeding the cut point for a clinical level of problems than those in the LC (at −1SD below mean age: z = −2.97, p = .003, OR = 0.20, 95% CI[0.07,0.58]; NNT = 5.58).

Parent self-report of mental health problems

There was a significant main effect on the 14-item SMAST (B = −0.74, SEB =0.36, z = −2.05, p = .04) such that parents in the FBP indicated fewer symptoms of alcoholism (MFBP = 0.54, MLC = 1.28, Cohen’s d = 0.36, 95% CI[0.01,0.71]). There was a significant main effect on the SMAST cutoff score of three such that bereaved parents in the LC were over six times more likely to exceed the cut point for a diagnosis of alcoholism than those in the FBP (B = −1.88, SEB = 0.81, z = −2.32, p = .02; OR = 0.15, 95% CI[0.03,0.75]; NNT = 6.25). For the assessment of the 10-item SMAST count scores of problem drinking behavior during the past nine years, there was a significant main effect in the count model (B = −1.24, SEB = 0.39, z = −3.17, p = .002; IRR = .29) such that the number of problem drinking behaviors for the parents in the FBP was .29 times the number of problem drinking behaviors for the parents in the LC. No significant effects were found for depression or psychiatric symptoms.

Service utilization

As shown in Table 1, young adults in the FBP as compared with controls had fewer visits to the doctor for mental health problems (B = −1.00, SEB = 0.35, z = −2.89, p = .004), attended fewer support group meetings (B = −1.86, SEB = 0.63, z = −2.95, p = .003), and took less psychiatric medications (B = −2.92, SEB = 0.81, z = −3.60, p < .001). Parents in the FBP were less likely to have attended support group meetings than those in the LC (B = −2.35, SEB = 0.88, z = −2.68, p = .007). The IRR showed that, compared with young aduts in the FBP in the past year, young adults in the LC had 2.7 times the number of visits to the doctor for mental health, attended 6.25 times the number of support group meetings, and took 20 times the number of psychiatric medications. Parents in the LC as compared with those in the FBP attended 10 times the number of support group meetings. There was also one main effect in the logit model. The likelihood of young adults in the FBP to take psychiatric medications was significantly smaller than for those in the LC [B = −3.59, SEB = 1.69, z = −2.12, p = .03 (FDR corrected-p = .06); OR = .03].

Discussion

The findings provide support for the hypotheses that young adults who were in FBP as compared to those who were in the LC would have lower mental health problems fifteen years following the program as reported by knowledgeable informants and lower self-reported use of mental health services, and that the bereaved parents in FBP as compared with LC would have lower self-report of alcohol problems and lower attendance at support groups over 15 years. The results did not support the hypothesized benefit for young adults in the FBP as compared to LC to have lower self-reported mental health problems. The clinically significant effects on several measures (i.e. parental alcohol abuse, key informant report of internalizing problems) as well as the effects on young adult and parent service use support the public health significance of the findings. The results will be discussed in terms of the utility of assessing program effects from the perspective of multiple measures and multiple reporters, alternative explanations for divergent findings across measures, implications for services provided to parentally-bereaved families, and limitations and future directions for research.

Although there were no program main or moderated effects on young adult self-report of their mental health problems, key informants reported lower levels of total mental health problems, and lower internalizing and externalizing problems for young adults in the FBP as compared with controls. These findings extend findings at the six-year follow-up of lower key-informant reports of mental health problems for youth in FBP as compared with controls [i.e., parent reports of youth externalizing problems and teacher reports of youth internalizing and externalizing problems (Sandler et al., 2010)]. Similarly, the findings that bereaved parents as compared with controls had lower levels of alcohol problems and alcohol abuse at the 15-year follow-up extends similar findings at the six-year follow-up. One limitation of the findings for youth and parents is that there was no pretest measure on these variables, precluding any interpretation as to program effects to reduce the levels of these problems over time, and somewhat weakening inferences as to the causal effect of the program. However, the limitation on inferences that the differences between the FBP and control groups on these variables is due to the FBP is mitigated by minimizing confounding effects through the randomization of the groups to FBP and control conditions, the use of an ITT approach to data analysis, the lack of attrition differences between groups over time, and the use of other measures of mental health at baseline as covariates in the analyses (Sackett, 1979; Sedgwick, 2013).

Although there was a main effect for the FBP to have lower levels of key informant reports of young adult total mental health problems at 15 years, the benefits of the FBP on key informant reports of both internalizing and externalizing problems were primarily found for those who were age 10 or 11 or younger at the time they participated in the program. One potential interpretation of the effects for younger participants in the FBP is that bereaved children were more responsive than bereaved adolescents to the coping and parenting skills taught in the program. However, this interpretation does not explain why there was very little evidence of an age by program interaction at the posttest, 11-month or six-year follow-ups (Sandler et al., 2003; Sandler et al., 2010). An alternative interpretation of the age-related program effects is that the program impacted mental health problems that are a response to the stressors that occur earlier in the young adult transition process (e.g., age 23 – 26; adjustment to leaving home and school) rather than later (age 27 – 31). A third potential explanation for the age-related effects is that different informants reported for younger and older participants. We explored this explanation, but found no difference across age in informants from whom reports were obtained.

One possible explanation why significant effects were not found for self-report of mental health problems by those in the FBP is that the LC participants’ greater use of mental health services and psychiatric medication reduced their reports of symptoms of mental health problems. Similarly, the lack of differences on parent report of their depression and psychiatric symptoms may be due to those in the control condition having benefitted from their greater use of support groups. It also may be that there were significant moderated effects on self-report of young adult or parent report of mental health problems, but that the study was underpowered to detect such effects. Post hoc power analyses suggest that power to detect a significant interaction effect was approximately .45 for a small effect size. In support of the possibility of moderated effects, however, we did find a significant gender by program interaction effect on young adult report of externalizing problems prior to the FDR correction which showed a positive effect for women in the FBP condition as compared with controls. Another possibility is that biases (either over-reporting or under-reporting) contributed error variance to the self-report of mental health problems, reducing the likelihood for detecting true program effects.

The finding that FBP participants reported using fewer mental health services in the past year than controls is consistent with reports that other preventive interventions have long-term effects to reduce service use (e.g., Herman, Mahrer, Wolchik, Jones, & Sandler, 2015; Ialongo, Poduska, Werthamer, & Kellam, 2001). It is interesting to note that although participation in the FBP might be expected to reduce stigma around seeking services and thus increase service use, it was instead associated with lower rates of use for a wide range of mental health and support group services.

The findings have several significant implications for services for parentally bereaved families. First, prior meta-analyses of interventions for bereaved children and adults (Currier, Holland, et al., 2007; Currier, Neimeyer, et al., 2008) reported larger effects for the sub-sample with higher levels of distress or symptoms when they entered treatment then for those who had lower levels of distress when they began treatment. The treatment effects were not significant for interventions delivered to those who were not selected as high in symptoms. We found, however, that the long-term benefits of the program were not moderated by baseline levels of problems, indicating that the FBP is equally appropriate for both high and low risk bereaved families. The lack of greater benefits at the 15-year follow-up for those with higher baseline problems is consistent with the lack of program x baseline effects on mental health problems and grief at the six-year follow-up (Sandler et al., 2010; Sandler Ma, et al., 2010). Second, the findings are supportive of a skills based approach to services with bereaved families. The theory of the intervention was that teaching coping and parenting skills would promote resilience of children and parents, leading to lower levels of mental health problems over time. This is consistent with prior evidence of long-term benefits of skill focused parenting and child coping programs with children and parents experiencing other stressful life situations (for a review see Sandler, Schoenfelder, Wolchik, & MacKinnon, 2011). It is also consistent with prior evidence that previously reported program effects on youth and parent mental health were mediated by program effects on quality of parenting and children’s coping (Tein, Sandler, Ayers, & Wolchik, 2006; Sandler et al., 2016). A third implication is to support the value of involving the surviving bereaved parents in services for bereaved children. The current findings extend findings of benefits for bereaved parents and youth at the 11-month (Sandler et al. 2003), and six-year follow-ups (Sandler et al., 2010; Sandler et al., 2016) to support a “double prevention” effect in which both the children and their bereaved parents benefit from participating in the FBP.

There are several limitations of this evaluation that point to directions for future research. First, longitudinal mediation analysis are needed to better understand which of the coping and parenting processes targeted by the intervention account for the lower levels of long-term mental health problems experienced by these young adults and their parents. As in previous long-term follow-ups of preventive interventions these analyses should test cascading effects in which short-term program effects to increase protective factors (e.g. positive parenting) lead to lower levels of problems several years later, which in turn lead to lower levels of other problems or higher levels of functioning at still later developmental periods (for a review of such findings see Sandler, et al., 2015). In addition, it may be that changes in family contexts following the program (e.g, economic losses, remarriage, etc.) may also contribute to program effects on child and parent mental health problems over time. Understanding the pathways through which family focused prevention and mental health promotion programs benefit children is a critical next step for research (Sandler et al., 2011). Second, a qualitative study of parent and child experiences in the program and their use of program skills following their participation is needed to provide a better understanding of how the program is helpful. Third, it is possible that key informants were knowledgeable about program conditions, and that this knowledge affected their ratings. Fourth, the lack of a more active control condition does not allow inferences as to how effects of the FBP might compare to those obtained from alternative programs, such as a supportive group that does not involve teaching coping or parenting skills. The fact that many parents and children did not read the books assigned to the control group indicates that the potency of our control condition was likely limited. Fifth, the effects obtained on service use need to be extended to a full cost-benefit analysis, which should include assessment of service use as documented by service provider records (e.g. Herman et al., 2015). Sixth, the 15-year effects on additional outcomes need to be assessed including effects on biological processes (e.g., indicators of emotion regulation), physical health, grief, and success in developmentally appropriate life tasks such as developing satisfying family relationships and success at work.

Public Health Significance.

The public health significance of the study is supported by the findings of lower mental health problems by bereaved young adults, lower alcohol abuse by their parents, and less mental health service use by young adults and bereaved parents fifteen years following program participation.

Acknowledgments

This research was supported by grant R01 MH 049155-11A1 from the National Institute of Mental Health which is gratefully acknowledged

Footnotes

1

A more detailed description of the procedures used to retain families over time will be provided by the authors on request.

2

A list of the three books provided to parents, adolescents and children will be provided on request.

References

  1. Achenbach TM. Manual for the Child Behavior Checklist/4-18 and 1991 Profile. Burlington, VT: University of Vermont, Department of Psychiatry; 1991a. [Google Scholar]
  2. Achenbach TM. Manual for the Youth Self-Report and 1991 Profile. Burlington, VT: University of Vermont, Department of Psychiatry; 1991b. [Google Scholar]
  3. Achenbach TM, Rescorla LA. Manual for the ASEBA Adult Forms & Profiles. Burlington, VT: University of Vermont, Research Center for Children, Youth, & Families; 2003. [Google Scholar]
  4. Ayers TS, Wolchik SA, Sandler IN, Towhey JL, Lutzke Weyer JR, Jones S, Weiss L, Cole E, Kriege G. The Family Bereavement Program: Description of a theory-based prevention program for parentally-bereaved children and adolescent. Omega: Journal of Death and Dying. 2013–2014;68:293–314. doi: 10.2190/om.68.4.a. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Beck AT, Steer RA. Manual for the Beck Depression Inventory. San Antonio, TX: Psychological Corporation; 1993. [Google Scholar]
  6. Beck AT, Steer RA, Garbin MG. Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review. 1988;8:77–100. doi: 10.1016/0272-7358(88)90050-5. [DOI] [Google Scholar]
  7. Benjamini Y, Hochberg Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. Journal of the Royal Statistical Society, Series B. 1995;57:289–300. doi: 10.2307/2346101. [DOI] [Google Scholar]
  8. Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I. Controlling the false discovery rate in behavior genetics research. Behavioural Brain Research. 2001;125:279–284. doi: 10.1016/S0166-4328(01)00297-2. [DOI] [PubMed] [Google Scholar]
  9. Bland RC, Newman SC, Orn H. Help-seeking for psychiatric disorders. Canadian Journal of Psychiatry. 1997;42:935–942. doi: 10.1177/070674379704200904. [DOI] [PubMed] [Google Scholar]
  10. Brent D, Melhem N, Donohoe MB, Walker M. The incidence and course of depression in bereaved youth 21 months after the loss of a parent to suicide, accident, or sudden natural death. American Journal of Psychiatry. 2009;166:786–794. doi: 10.1176/appi.ajp.2009.08081244. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Burke KC, Burke JD, Regier DA, Rae DS. Age at onset of selected mental disorders in five community populations. Archives of General Psychiatry. 1990;47:511–518. doi: 10.1001/archpsyc.1990.01810180011002. [DOI] [PubMed] [Google Scholar]
  12. Carlson EN, Vazire S, Oltmanns TF. Self-other knowledge asymmetries in personality pathology. Journal of Personality. 2013;81:155–170. doi: 10.1111/j.1467-6494.2012.00794. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cohen J. Statistical Power Analysis for the Behavioral Sciences. Hillsdate, NJ: Lawrence Erlbaun; 1988. [Google Scholar]
  14. Cook RD. Detection of influential observations in linear regression. Technometrics. 1977;19:15–18. doi: 10.2307/1268249. [DOI] [Google Scholar]
  15. Coxe S, West SG, Aiken LS. The analysis of count data: A gentle introduction to Poisson regression and its alternatives. Journal of Personality Assessment. 2009;91:121–136. doi: 10.1080/00223890802634175. [DOI] [PubMed] [Google Scholar]
  16. Crews TM, Sher KJ. Using adapted short MASTs for assessing parental alcoholism: Reliability and validity. Alcoholism: Clinical and Experimental Research. 1992;16:576–584. doi: 10.1111/j.1530-0277.1992.tb01420.x. [DOI] [PubMed] [Google Scholar]
  17. Currier JM, Holland JM, Neimeyer RA. The effectiveness of bereavement interventions with children: A meta-analytic review of controlled outcome research. Journal of Clinical Child and Adolescent Psychology. 2007;36:253–259. doi: 10.1080/15374410701279669. [DOI] [PubMed] [Google Scholar]
  18. Currier JM, Neimeyer RA, Berman JS. The effectiveness of psychotherapeutic interventions for bereaved persons: A comprehensive quantitative review. Psycholgoical Bulletin. 2008;134:648–661. doi: 10.1037/0033-2909.134.5.648. [DOI] [PubMed] [Google Scholar]
  19. Dohrenwend BP, Shrout PE, Ergi G, Mendelsohn FS. Non-specific psychological distress and other dimensions of psychopathology. Archives of General Psychiatry. 1980;39:1229–1236. doi: 10.1001/archpsyc.1980.01780240027003. [DOI] [PubMed] [Google Scholar]
  20. Elhai JD, Ford JD. Correlates of mental health service use intensity in the National Comorbidity Survey and National Comorbidity Survey Replication. Psychiatric Services. 2007;58:1108–1115. doi: 10.1176/ps.2007.58.8.1108. [DOI] [PubMed] [Google Scholar]
  21. Faschingbauer TR, DeVaul RA, Zisook S. The Development of the Texas Inventory of Grief. American Journal of Psychiatry. 1977;134:696–698. doi: 10.1176/ajp.134.6.696. [DOI] [PubMed] [Google Scholar]
  22. Fiedler ER, Oltmanns TF, Turkheimer E. Traits associated with personality disorders and adjustment in military life: Predictive validity of self and peer reports. Military Medicine. 2004;169:207–211. doi: 10.7205/MILMED.169.3.207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Haine RA, Ayers TS, Sandler IN, Wolchik SA, Weyer JL. Locus of control and self-esteem as stress-moderators or stress-mediators in parentally bereaved children. Death Studies. 2003;27:619–640. doi: 10.1080/07481180302894. [DOI] [PubMed] [Google Scholar]
  24. Herman PM, Mahrer NE, Wolchik SA, Jones S, Sandler IN. Cost-benefit analysis of a preventive intervention for divorced families: Reduction in mental health and justice system service use costs 15 years later. Prevention Science. 2015;16:586–608. doi: 10.1007/s11121-014-0527-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hill KG, Bailey JA, Hawkins JD, Catalano RF, Kosterman R, Oesterle S, Abbott RD. The onset of STI diagnosis through age 30: Results from the Seattle Social Development Project intervention. Prevention Science. 2014;15:19–32. doi: 10.1007/s11121-013-0382-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Horwitz SM, Hoagwood K, Stiffman AR, Summerfeld T, Weisz JR, Costello EJ, Norquist G. Reliability of the services assessment of children and adolescents. Psychiatric Services. 2001;52:1088–1094. doi: 10.1176/appi.ps.52.8.1088. [DOI] [PubMed] [Google Scholar]
  27. Ialongo N, Poduska J, Werthamer L, Kellam S. The distal impact of two first-grade preventive interventions on the conduct problems and disorder in early adolescence. Journal of Emotional and Behavioral Disorders. 2001;9:146–160. doi: 10.1177/106342660100900301. [DOI] [Google Scholar]
  28. Jurs DG, Glass GV. The effect of experimental mortality on the internal and external validity of the randomized comparative experiment. Journal of Experimental Education. 1971;40:62–66. doi: 10.1080/00220973.1971.11011304. [DOI] [Google Scholar]
  29. Kessler RC, McGonagle KA, Zhao S, Nelson C, Hughes M, Eshleman S, Kendler KS. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Archives of General Psychiatry. 1994;51:8–19. doi: 10.1001/archpsyc.1994.03950010008002. [DOI] [PubMed] [Google Scholar]
  30. Kovacs M. Rating scales to assess depression in school-aged children. Acta Paedopsychiatrica: International Journal of Child & Adolescent Psychiatry. 1981;46:305–315. [PubMed] [Google Scholar]
  31. Leaf PJ, Bruce ML, Tischler GL, Freeman DH, Weissman MM, Myers JK. Factors affecting the utilization of specialty and general medical mental health services. Medical Care. 1988;26:9–26. doi: 10.1097/00005650-198801000-00002. [DOI] [PubMed] [Google Scholar]
  32. Lin K, Sandler I, Ayers T, Wolchik S, Leucken L. Resilience in parentally bereaved children and adolescents seeking preventive services. Journal of Clinical Child and Adolescent Psychology. 2004;33:673–683. doi: 10.1207/s15374424jccp3304_3. [DOI] [PubMed] [Google Scholar]
  33. Melhem NM, Walker M, Moritz G, Brent DA. Antecedents and sequalae of sudden parental death in offspring and surviving parents. Archives of Pediatric and Adolescent Medicine. 2008;162:403–410. doi: 10.1001/archpedi.162.5.403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Muthén LK, Muthén BO. Mplus user’s guide. 7th. Los Angeles, CA: Muthén & Muthén; 1998–2014. [Google Scholar]
  35. Pearce N. Effect measures in prevalence studies. Environmental Health Perspectives. 2004;112:1047–1050. doi: 10.1289/ehp.6927. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Rayburn NR, Wenzel SL, Elliott MN, Hambarsoomians K, Marshall GN, Tucker JS. Trauma, depression, coping, and mental health service seeking among impoverished women. Journal of Consulting and Clinical Psychology. 2005;73:667–677. doi: 10.1037/0022-006X.73.4.667. [DOI] [PubMed] [Google Scholar]
  37. Reynolds CR, Richmond BO. What I think and feel: A revised measure of children’s manifest anxiety. Journal of Abnormal Child Psychology. 1978;6:271–280. doi: 10.1007/BF00919131. [DOI] [PubMed] [Google Scholar]
  38. RobinsL N, Wing J, Wittchen HU, Helzer JE, Babor TF, Burke J, Towle LH. The Composite International Diagnostic Interview: An epidemiological instrument suitable for use in conjunction with different diagnostic systems and in different cultures. Archives of General Psychiatry. 1988;45:1069–1077. doi: 10.1001/archpsyc.1988.01800360017003. [DOI] [PubMed] [Google Scholar]
  39. Sackett DL. Bias in analytic research. Journal of Cronical Disease. 1979;32:51–63. doi: 10.1016/0021-9681(79)90012-2. [DOI] [PubMed] [Google Scholar]
  40. Sandler IN, Ayers TS, Tein JY, Wolchik S, Millsap R, Khoo ST, Kaplan D, Ma Y, Luecken L, Schoenfelder E, Coxe S. Six-year follow-up of a preventive intervention for parentally-bereaved youth: A randomized controlled trial. Archives of Pediatrics and Adolescent Medicine. 2010;164:907–914. doi: 10.1001/archpediatrics.2010.173. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Sandler IN, Ingram A, Wolchik S, Tein J-Y, Winslow E. Long-term effects of parenting-focused preventive interventions to promote resilience of children and adolescents. Child Development Perspectives. 2015;9:164–171. doi: 10.1111/cdep.12126. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Sandler IN, Ma Y, Tein JY, Ayers TS, Wolchik S, Kennedy C, Millsap R. Long-term effects of the Family Bereavement Program on multiple indicators of grief in parentally bereaved children and adolescents. Journal of Consulting and Clinical Psychology. 2010;78:131–144. doi: 10.1037/a0018393. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sandler IN, Schoenfelder EN, Wolchik SA, MacKinnon DP. Long-term impact of prevention programs to promote effective parenting: Lasting effects but uncertain processes. Annual Review of Psychology. 2011;62:299–329. doi: 10.1146/annurev.psych.121208.131619. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Sandler IN, Tein JY, Cham H, Wolchik S, Ayers T. Long-term effects of the Family Bereavement Program on spousally-bereaved parents: Grief, mental health, alcohol problems and coping efficacy. Development and Psychopathology. 2016;28:801–818. doi: 10.1017/S0954579416000328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sandler IN, Wolchik SA, Ayers TS, Tein JY, Luecken L. Family Bereavement Program (FBP) approach to promoting resilience following the death of a parent. Family Science. 2013;4:87–95. doi: 10.1080/19424620.2013.821763. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Selzer ML, Vinokur A, van Rooijen L. A self-administered Short Michigan Alcoholism Screening Test (SMAST) Journal of Studies on Alcohol. 1975;36:117–126. doi: 10.15288/jsa.1975.36.117. [DOI] [PubMed] [Google Scholar]
  47. Sedgwick P. What is intention to treat analysis? BMJ. 2013;346:f3662. doi: 10.1136/bmj.f3662. [DOI] [Google Scholar]
  48. Shrout PE, Lyons M, Dohrenwend BP, Skodol AE, Soloman M, Klass F. Changing time frameworks on symptom inventories: Effects on the Psychiatric Epidemiology Research Interview. Journal of Consulting and Clinical Psychology. 1988;56:267–272. doi: 10.1037/0022-006X.56.2.267. [DOI] [PubMed] [Google Scholar]
  49. Smith TW, Uchino BN, Berg CA, Florsheim P, Pearce G, Hawkins M, Hopkins PN. Associations of self-reports versus spouse ratings of negative affectivity, dominance, and affiliation with coronary artery disease: Where should we look and who should we ask when studying personality and health? Health Psychology. 2008;27:676–684. doi: 10.1037/0278-6133.27.6.676. [DOI] [PubMed] [Google Scholar]
  50. Spoth R, Trudeau L, Shin C, Ralston E, Redmond C, Greenberg M, Feinberg M. Longitudinal effects of universal preventive intervention on prescription drug misuse: Three randomized controlled trials with late adolescents and young adults. American Journal of Public Health. 2013;103:665–672. doi: 10.2105/AJPH.2012.301209. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Stiffman AR, Horwitz SM, Hoagwood K, Compton W, III, Cottler L, Bean DL, Weisz JR. The service assessment for children and adolescents (SACA): Adult and child reports. Journal of the Academy of Child and Adolescent Psychiatry. 2000;39:1032–1039. doi: 10.1097/00004583-200008000-00019. [DOI] [PubMed] [Google Scholar]
  52. Stroebe M, Schut H. The Dual Process Model of coping with bereavement: Rationale and description. Death Studies. 1999;23:197–224. doi: 10.1080/074811899201046. [DOI] [PubMed] [Google Scholar]
  53. Tein JY, Sandler IN, Ayers TS, Wolchik SA. Mediation of the effects of the Family Bereavement Program on mental health problems of bereaved children and adolescents. Prevention Science. 2006;7:179–197. doi: 10.1007/s11121-006-0037-2. [DOI] [PubMed] [Google Scholar]
  54. Van Ryzin MJ, Nowicka P. Direct and indirect effects of a family-based intervention in early adolescence on parent-youth relationship quality, late adolescent health, and early adult obesity. Journal of Family Psychology. 2013;27:106–116. doi: 10.1037/a0031428. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Wang PS, Lane M, Olfson M, Pincus HA, Wells KB, Kessler RC. Twelve-month use of mental health services in the United States: Results from the National Comorbidity Survey Replication. Archives of General Psychiatry. 2005;62:629–640. doi: 10.1001/archpsyc.62.6.629. [DOI] [PubMed] [Google Scholar]
  56. Wolchik SA, Sandler IN, Tein JY, Mahrer NE, Millsap RE, Winslow E, Reed A. Fifteen-year follow-up of a randomized trial of a preventive intervention for divorced families: Effects on mental health and substance use outcomes in young adulthood. Journal of Consulting and Clinical Psychology. 2013;81:660–673. doi: 10.1037/a0033235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wolchik SA, Tein J-Y, Sandler IN, Kim H-J. Developmental cascade models of a parenting-focused program for divorced families on mental health problems and substance use in emerging adulthood. Development and Psychopathology. 2016;28:869–888. doi: 10.1017/S0954579416000365. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Yuan K-H, Bentler PM. Three likelihood-based methods for mean and covariance structure analysis with nonnormal missing data. Sociological Methodology. 2000;30(1):165–200. doi: 10.1111/0081-1750.00078. [DOI] [Google Scholar]
  59. Zisook S, Schucter SR. Depression through the first year after the death of a spouse. American Journal of Psychiatry. 1991;148:1346–1352. doi: 10.1176/ajp.148.10.1346. [DOI] [PubMed] [Google Scholar]

RESOURCES