Collaborative Alliance of Parent and Child Welfare Caseworker

Abstract

This secondary analysis of data describing 3,035 parents, drawn from the National Survey of Child and Adolescent Well-Being II, identified factors fostering the collaborative alliance of parents and caseworkers within the child welfare system. We used generalized least squares random effects modeling for panel data. We sought associations between caseworker engagement as perceived by parent and parent’s interpersonal capacities, intrapersonal dynamics, problem severity, and racial/ethnic background, and between that perception and caseworker turnover. Parents in our sample had been substantiated for maltreatment of their children. Results showed that parent’s perceived caseworker engagement was associated positively with seven factors: parent’s social support, parent’s mental health, kinship care, out-of-home placement, parent’s African American ethnicity, parent’s Hispanic ethnicity, parent/caseworker shared ethnicity, and family income. Perceived engagement was associated negatively with caseworker turnover (i.e., number of caseworkers assigned, by turns, to parent’s case). Implications for practicing social work within the child welfare system are discussed.

Keywords

child welfare services child welfare workers parents/adults longitudinal research

In 2016, almost 20% of this nation’s 3.4 million reports of child maltreatment were substantiated, and nearly 23% of maltreated children were removed from home (Children’s Bureau, 2018). Over 395,000 parents or legal guardians were substantiated for child maltreatment (Children’s Bureau, 2018). Most families who were substantiated for maltreatment (95%) received child welfare services (Cheng & Lo, 2012); parents were required to complete recommended or mandatory services to either keep a family intact or become reunified with children. Achieving such permanency goals relies heavily on child welfare caseworkers assisting parents (Zlotnik, Strand, & Anderson, 2009). Many child welfare caseworkers invite parents to participate with them in this type of goal-oriented collaborative relationship (Dore & Alexander, 1996; Drake, 1994; Littell, Alexander, & Reynolds, 2001), leading to permanency (Yatchmenoff, 2005). Building parent–caseworker collaborative alliances promotes receipt of needed services by parents (Cheng & Lo, 2012), fosters making progress on the case plan (Cheng & Lo, 2016), fosters permanency outcomes (Cheng, 2010; Cheng & Lo, 2012; Grella, Needell, Shi, & Hser, 2009; Yatchmenoff, 2005), and reduces the likelihood of a substantiated re-report (Cheng & Lo, 2015). It is crucial to focus on client perception of such collaborative alliance because that perception mediates any change in client behavior (Gurman, 1977).

A working or collaborative alliance is a relationship between client and therapist that is founded on respect, empathy, shared goals, and shared participation (Bordin, 1979; Castonguay, Constantino, & Holtforth, 2006; Crits-Christoph, Gibbons, & Hearon, 2006; Duncan, Miller, & Sparks, 2004; Graybeal, 2007; Horvath, 2006). Parallel with client–therapist collaborative alliances, parent–caseworker collaborative alliances in child welfare services are founded on these, and on connection, trust, and self-worth (Buck & Alexander, 2006; Gladstone et al., 2014; Redko, Rapp, Elms, Snyder, & Carlson, 2007). Within the child welfare system, collaborative alliance is characterized by a parent’s active participation in decision-making and implementation of the case plan, the parent acting as the caseworker’s partner in the intervention process (Alexander & Dore, 1999).

The present study represented an application of several prior researchers’ conceptualization of client involvement in working or collaborative alliance (Gurman, 1977; Horvath, 1994, 2001, 2006; Horvath & Luborsky, 1993). As they conceptualized it—and they did so within a domain of individual therapy—client involvement reflected three factors describing clients: interpersonal capacities, intrapersonal dynamics, and severity of problem(s). These prior researchers confirmed a significant association between working alliance as perceived by client and the outcome of individual therapy (Gurman, 1977; Horvath & Luborsky, 1993). Furthermore, that development of client–therapist collaborative working alliance is influenced by the interpersonal capacities (e.g., social support), intrapersonal dynamics (e.g., motivation to change), and problems (e.g., substance dependence) of the client (Horvath, 1994, 2001, 2006). In child welfare, too, then, parent’s perception of the parent–caseworker collaborative alliance can be related to the parent’s interpersonal capacities, intrapersonal dynamics, and problem severity. Thus identifying factors linked to parent’s perception of strong collaborative alliance is important.

Literature Review

Interpersonal Capacities

A parent’s interactions with a child welfare caseworker as the parent participates in intervention are influenced by any difficulty the parent has with interpersonal relationships. A collaborative alliance can be considered a manifestation of an object relation; an object relation is defined as the individual’s internal tendency to develop relationships with others (Piper et al., 1991). Those who build supportive interpersonal relationships with people demonstrate the capacity to engage in object relation (Piper et al., 1991). Establishing collaborative alliance indicates a parent’s capacity to form object relations, which is influenced by early experiences the parent had (Bordin, 1979, 1994; Horvath, 1994; Horvath & Luborsky, 1993; Luborsky, 1976, 1994). The presence of social support in one’s experiences indicates one’s success at forming relationships with others. Demonstrated ability to form supportive social relationships should, in turn, facilitate the development of working alliance with a therapist (Horvath, 1994). In interacting with others, however, some parents cannot adequately regulate and express emotion (Gross, 1998). When they cannot, their social support networks are limited, affecting the child welfare helping process. A parent’s social support network indicates his/her capacity to form supportive relationships or object relations (Piper et al., 1991). Among parents involved in child welfare, this network may comprise members of the immediate and extended family, friends, neighbors, child welfare workers, a faith community, and/or support groups (Child & McIntyre, 2015; Lalayants, Baier, Benedict, & Mera, 2014). Frequently when permanency outcomes like nonremoval or reunification are achieved, the parent has enjoyed social support from these sources (Child & McIntyre, 2015; McWey, Holtrop, Wojciak, & Claridge, 2015; Rajendran, Smith, & Videka, 2015). We speculated that, among parents involved in child welfare, likelihood of strong parent–caseworker collaborative alliance rises when parents have access to social support and declines when they do not.

Intrapersonal Dynamics

Augmenting the parent’s interpersonal capacities are individual intrapersonal dynamics, which embody each individual’s internal contextualization of relation events; examples of such contextualization include motivation to change and traumatic experience (Horvath, 2006). Motivation to change is the parent’s internal desire to change whatever behavior the helping process is addressing (Drieschner, Lammers, & van der Staak, 2004). Consistently, the literature reports a positive association between client motivation to change and client perception of working or collaborative alliance (Urbanoski, Kelly, Hoeppner, & Slaymaker, 2012; Wolfe, Kay-Lambkin, Bowman, & Childs, 2013). Furthermore, within the child welfare system, parent compliance with mandated services can be conceptualized as a function of parent willingness or motivation to change (Littell, 2001; Smith, 2008). Throughout an intervention, parent motivation to change influences the character and depth of the parent–caseworker relationship—especially the degree of the parent’s participation in the helping process (de Greef et al., 2018; Kemp, Marcenko, Lyons, & Kruzich, 2014). Weak motivation to change often characterizes clients whose services are court-mandated or otherwise coerced (Johnson et al., 2017).

Another potentially negative influence on parent–caseworker collaborative alliance is any traumatic experience during childhood and/or experience of intimate partner violence in adulthood the parent has had (Doran, Doukas, D’Andrea, & Nnamdi, 2011; Paivio & Patterson, 1999). Such experience of complex trauma fosters mistrust of others and negative expectations concerning social relationships, among which would be included collaborative relationships with caseworkers (Doukas, D’Andrea, Doran, & Pole, 2014).

Problem Severity

If severe enough, clients’ problems (such as substance abuse, mental illness, intimate partner violence, being stigmatized, and child maltreatment substantiation) can distort client perceptions, possibly hindering a relationship with a caseworker.

Prior studies (Fluckiger et al., 2013; Horvath, 2001; O’Brien, Fahmy, & Singh, 2009) have reported that clients with relatively severe problems, such as substance abuse, exhibit relatively weak collaborative alliances with therapists. Substance-abusing clients show some characteristics—mistrust, poor emotional regulation, difficulty in interpersonal relationships—that jeopardize collaborative alliance (Wolfe et al., 2013). Not surprisingly, clients with dual diagnosis (co-occurring mental disorder and substance abuse) experience serious emotional disturbance that distorts perception of the client–therapist relationship (Gurman, 1977).

Individuals who have experienced intimate partner violence or mental health symptoms tend to have weak collaborative alliance (Goodman, Fauci, Sullivan, DiGiovanni, & Wilson, 2016). Poor collaborative alliance is also associated with clients’ internalization of the stigma of, for example, psychological problems (Kendra, Mohr, & Pollard, 2014); one study, however, found such stigma affected only therapists’, not clients’, perceptions (Nakash, Nagar, & Levav, 2015), and another found no link between such stigma and collaborative alliance (Kondrat & Early, 2010). Nevertheless, stigma attaching to the mandating of child welfare services may hinder the development of client–caseworker collaborative alliance. Not surprisingly, within the child welfare system, parents who have multiple problems are likely to lack strong collaborative alliances with their caseworkers (Kemp, Marcenko, Hoagwood, & Vesneski, 2009).

Substantiated child maltreatment itself is a severe problem for a parent; moreover, to remove a child from home suggests the maltreatment’s severity. At the same time, parents with children who were placed out of home are relatively unlikely to view caseworkers as collaborative partners in the helping process (Kemp et al., 2014). Losing custody of a child is stressful and has been found to hamper parents’ verbal expression of emotion, which exacerbates their interpersonal difficulties and further hinders building of collaborative alliance (de Tychey, Garnier, Lighezzolo-Alnot, Claudon, & Rebourg-Roesler, 2010; Horvath & Luborsky, 1993).

Demographic Characteristics of Parents

Parent–caseworker collaborative alliance can be affected by the demographic characteristics of the parents. African American clients often mistrust their mental health providers, reflecting prior experiences with culturally incompetent therapists and discriminatory agencies (Alvidrez, Snowden, & Kaiser, 2010; Ward, Clark, & Heidrich, 2009). Hispanic clients often come from cultures emphasizing engagement with extended family and peers, a focus discording with individual engagement on which collaborative alliance depends (Paris, Anez, Bedregal, Andres-Hyman, & Davidson, 2005). Additionally, according to one study, African American and Latina survivors of domestic violence tended to lack collaborative alliance with therapists (Goodman et al., 2016), although another examining specifically mental health case management reported no such link involving race/ethnicity (Kondrat & Early, 2010). A study focusing on the child welfare system reported frequent mistrust of the system by African American parents, leading to their weaker alliances with caseworkers versus White parents’ alliances (Kemp et al., 2009). Similarly, versus White parents, Hispanic parents have been found less satisfied by the depth of collaboration with and empathy from caseworkers (Cheng & Lo, 2012). At least one study (McBeath, Chuang, Bunger, & Blakeslee, 2014) found successful child welfare outcomes to be relatively likely when a parent and caseworker shared a race/ethnicity.

In addition, a study of therapeutic intervention showed female clients to have stronger working alliances with therapists than male clients (Urbanoski et al., 2012), although that finding is contradicted by another study reporting no association between gender and collaborative alliance (Kondrat & Early, 2010). Moreover, one study focusing on the child welfare system found no association between collaborative alliance and parents’ education (Cheng & Lo, 2018), while several suggest a positive association between such alliance and parents’ income (suggesting parents in poverty often lack strong alliances; Cheng & Lo, 2018; Goodman et al., 2016; Kemp et al., 2009).

Caseworker Turnover

A fifth factor that describes child welfare experiences, caseworker turnover, also may affect parent–caseworker collaborative alliances. Repeated turnover—indicated by the number of caseworkers assigned, by turns, to the parent’s case—may impede alliance building. Studies have consistently shown such turnover to diminish parent–caseworker relationships and hinder desirable outcomes (Chambers, Brocato, Fatemi, & Rodriguez, 2016; Chambers, Crutchfield, Harper, Fatemi, & Rodriguez, 2018). Frequent caseworker turnover introduces instability into the helping process because the time needed to build the trust distinguishing alliance is lacking.

Most studies in the literature have investigated relevant concepts or constructs separately. Moreover, just two studies (Cheng & Lo, 2018; McBeath et al., 2014) from the literature reviewed used national data (or even large statewide samples) to examine parent–caseworker relationships and alliances. Our study, then, with its national data set and attempted exploration of multiple constructs jointly via a single model, addresses a gap in the literature. Our study hypothesized that, in the child welfare setting, parent–caseworker collaborative alliance would be associated with parents’ interpersonal capacities, intrapersonal dynamics, problem severity, and demographic characteristics, as well as with caseworker turnover.

Method

Sample

We extracted a sample of parents from the National Survey of Child and Adolescent Well-Being II (NSCAW-II) public use data set. The data set describes a nationally representative sample of 5,872 children and their caregivers who were involved in child protective services (CPS; National Data Archive on Child Abuse and Neglect, 2013). Through three waves of interviews between 2008 and 2012, NSCAW-II researchers collected information from the children as well as from their caregivers, caseworkers, and teachers regarding physical and mental health, social functioning, academic achievement, behavioral adjustment, and receipt of services. Caregivers were either the biological parents, adoptive parents, foster parents, or other relatives. The original data set contains three waves of interviews. Each longitudinal record in the original data set contains information collected during the Wave-1 interview, collected 18 months after Wave 1, and collected 36 months after Wave 1.

The present study’s sample included only those records indicating substantiation of child maltreatment during an initial CPS investigation. Such records constituted 61.5% of the original NSCAW-II sample. We excluded all other records (i.e., those not indicating substantiation). We did so because we surmised that those cases either provided evidence insufficient to substantiate maltreatment, or ruled maltreatment out altogether. In records lacking substantiation, any need for services or interaction with caseworkers was unrelated to maltreatment. In those records in which child maltreatment was substantiated, children and their families were experiencing dire problems that required multiple services (Maikovich-Fong & Jaffee, 2010). Ultimately, our study evaluated data from 3,035 biological or adoptive parents or other permanent caregivers interviewed for NSCAW-II.

Since our study analyzed data from three interview waves, we used the discrete-time method of longitudinal analysis (Singer & Willett, 2003). We divided each longitudinal record into person-waves. Each person-wave served as one case in our data file, and each contained our outcome and explanatory variables. Our outcome variable and most of our explanatory variables, including placement type, substance use, and refusal of services, could vary across interview waves, while others, including gender and ethnicity, remained constant. Our final analytical sample consisted of 6,142 person-waves, derived from the 3,035 parents; these person-waves were our units of analysis. We allowed each interview wave to serve as a time indicator, allowing simultaneous comparison of the results across waves.

Measures

The outcome variable parent’s perceived caseworker engagement served as a proxy measure of each parent’s perception of the degree of collaborative engagement demonstrated by the caseworker during the helping process. This variable reflected the total score from 6 items asking parents whether the caseworker explained problems well, listened to parent concerns, treated parent respectfully, maintained contact with parent, invited parent to meetings, and involved parent in decision-making. Explaining problems well was indicated with a 3-point response scale comprising 1 (not well), 2 (somewhat well), and 3 (very well). A 4-point response scale comprising 1 (never), 2 (not very often), 3 (sometimes), and 4 (all the time) measured the remaining engagement behaviors. To standardize our outcome variable, we created a z-score for each item and summed them to obtain the total score. The Cronbach’s α of the 6 items was .90. Higher total scores indicated more sufficient caseworker collaborative engagement perceived by parents.

There were six groups of explanatory variables: interpersonal capacities, intrapersonal dynamics, problem severity, parent demographic characteristics, caseworker turnover rate, and interview wave. The first group, interpersonal capacities, contained only one variable. Social support reflected the total score from 11 items querying parents about forms of support they received from people they trusted. Each item was responded to via a 5-point scale. Support took any of 11 resulting forms: parent knowing people who care about what happens to him/her, parent being invited to spend time and do things with others, parent talking to others about personal and family problems, parent talking to others about problems with job or housework, parent talking to others about finances, parent being offered useful advice about life matters, parent receiving help with transportation from others, parent receiving others’ help when confined to bed due to illness, parent receiving love and affection from others, parent receiving others help with household duties, and parent receiving others’ help to care for children. The data set provided no Wave-2 data describing parents’ social support; to overcome this deficit, we averaged the Wave-1 and Wave-3 measures and allowed these averages to provide Wave-2 measures of social support. Higher scores suggested greater support received.

The second group of explanatory variables consisted of but one construct that measured an intrapersonal dynamics factor. Refusal of services drew on caseworker reports to state the proportion of case plan–indicated services not received by a parent because the parent refused to participate in the service. Nine services were considered: housing services, income assistance, job placement, domestic violence services, legal services, alcohol treatment, drug treatment, mental health services, and health services. The set of services refused by a parent might or might not be unique. Refusing higher proportions of the services implied parents’ lower motivation or compliance.

The third group of explanatory variables included five variables measuring the severity of problems parents faced. Number of alcoholic drinks consumed gave the total number of drinks each parent consumed on a typical day; offered responses were 0 (1 or 2 drinks), 1 (3 or 4 drinks), 2 (5 or 6 drinks), 3 (7 to 9 drinks), and 4 (10 drinks or more). Nonmedical use of medication, the next variable in the group, was dichotomous and stated whether a parent had used prescription or over-the-counter medicine for nonmedical reasons in the 12 months preceding interview. The third variable, mental health, gave for each parent the standardized mental score from the Short-Form Health Survey (SF-12; Ware, Kosinski, & Keller, 1996), calculated and provided by NSCAW-II researchers; higher scores implied better mental health. The data set provided no specification of items from the scale but did report test–retest reliability of .86 (Ware et al., 1996). The fourth variable was domestic violence experience, giving the total number of domestic violence episodes experienced by a parent in the year preceding interview (whether such assaults were minor or severe). Finally, the variable number of needed services stated the total number of services a caseworker noted down in a case plan, which implied how many problems a parent had. Caseworkers had specified which of nine services a parent “needed”: housing services, income assistance, job placement, domestic violence services, legal services, alcohol treatment, drug treatment, mental health services, and health services. Each parent might have a unique combination of services she/he needed. Higher numbers of services implied that relatively more problems existed.

The variable current placement setting (developed by NSCAW-II researchers) included three dummy variables: (a) home of biological/adoptive parent (reference group), (b) kinship care, and (c) out-of-home care (foster care with nonrelative caretaker, group home, residential facility, and the like). In the present study, our focus was the current placement, not change in placement setting from one interview wave to another. Additionally, we used some dummy variables describing maltreatment type as controls during data analysis. (Maltreatment type data came from caseworker reports obtained at NSCAW-II’s Wave-1 interviews; we ourselves did not attempt to rank the relative severity of the various types.) For each parent, the dummy variables indicated the most serious maltreatment substantiated at initial CPS investigation, whether (a) physical maltreatment, (b) neglect (physical neglect, inadequate supervision, abandonment), (c) sexual maltreatment, or (d) other maltreatment (emotional, moral, educational, exploitation, other), provided the reference during data analysis. We used these types as control variables in light of a prior finding that parents substantiated for neglect perceived greater caseworker engagement versus those substantiated for other maltreatment of a nonphysical, nonsexual nature (Cheng & Lo, 2018).

Two variables measured parent’s individual characteristics: gender, either male or female; and race/ethnicity, either White (reference group), African American, Hispanic, or other racial/ethnic minority. Shared race/ethnicity (yes/no) indicated if a parent and her/his caseworker self-reported being of the same ethnicity. Additional variables described certain socioeconomic characteristics of parents. The continuous variable education level comprised five levels: 1 (no schooling), 2 (general equivalency diploma [GED] or high school diploma), 3 (vocational/technical training), 4 (associate’s/bachelor’s degree), and 5 (graduate/professional degree). Another continuous variable, family income, had four levels: 1 (below 50% of federal poverty level), 2 (50–99% of federal poverty level), 3 (100–200% of federal poverty level), and 4 (above 200% of federal poverty level). The variable employed parent (yes/no) indicated if a parent had been employed at some point in 6 months preceding interview. Measures of socioeconomic characteristics were determined by the NSCAW-II researchers.

The caseworker turnover variables group featured a single variable, number of caseworkers parent worked with. It stated how many caseworkers a parent reported meeting or talking with during the helping process to date. The final group of explanatory variables comprised two dummy variables, second and third interview wave. These specified the interview wave at which a given measure had been obtained; first interview wave was the reference. While we did not emphasize, in our study, the potential association between time and the outcome, we did seek the greatest possible flexibility for the time function within longitudinal analysis, by treating discrete periods as time indicators (Singer & Willett, 2003).

Data Analysis

For each interview wave, any missing values were replaced by their midpoints. Our study emphasized descriptive statistics and multivariate analysis, employing Stata generalized least squares (GLS) random effects modeling for panel data (with robust standard errors) to examine outcome–explanatory variable associations. GLS random effects modeling is more efficient than GLS fixed effects modeling in estimating coefficients (Rabe-Hesketh & Skrondal, 2008). Furthermore, GLS random effects modeling takes autocorrelations into account (i.e., correlations of a respondent’s measures across multiple interviews; Rabe-Hesketh & Skrondal, 2008; Singer & Willett, 2003). Tolerance statistics (≥ .63) and correlation coefficients (−.49 ≤ r ≤ .43) found in preliminary analysis suggested no multicollinearity problems.

Results

Of 3,035 parents in our sample, 92.2% were female. Additionally, 31.0% of these parents were White, 29.3% were Hispanic, 33.1% were African American, and the rest (6.6%) were other ethnic minority. Initial CPS investigations had substantiated 3.7% of the sample for physical maltreatment, 1.5% for sexual maltreatment, 10.8 for neglect, and the rest (84.0%) for other maltreatment.

On average across the person-waves, parents’ perceived caseworker engagement score was 0.08, social support score was 42.5 (of 55 possible), proportion of services refused was 2.86%, number of alcoholic drinks was 0.52 (1–4 drinks), mental health score was 50.23 (of 70.48 possible), number of domestic violence episodes was 2.78, number of services needed was 1.11 (of 9 possible), education level was 2.16 (high school diploma or GED), family income was 2.39 (50–99% of federal poverty level), and number of assigned caseworkers was 2.30 (see Table 1). In a large majority—86.6%—of the 6,142 person-waves analyzed, children had not been removed from a parent’s home. Just 11.3% of person-waves indicated kinship care, and just 2.1% indicated out-of-home placement. In addition, in 7.8% of person-waves, the parent reported nonmedical use of a prescription; in 14.1%, the parent was employed; and in 26.0%, the parent and caseworker shared an ethnicity. As for the time indicator, in 34.2% of the person-waves, data were from the Wave-1 interview, while in 34.8%, it was from Wave 2, and in 31.0%, it was from Wave 3.

Table 1.

Descriptive Statistics of Time-Varying Variables.

Variables	Percentage	Mean	Range	SD
Parent’s perceived caseworker engagement		.08	−27.9 to –10.07	6.52
Interpersonal capacities
Social support		42.51	11 to 55	9.81
Intrapersonal dynamics
Refusal of services		2.86%	0 to 100%	14.40%
Problem severity
Number of alcoholic drinks consumed		.52	0 to 4	.51
Nonmedical use of medication (Yes)	7.8
(No)	92.2
Mental health		50.23	7.64 to 70.48	10.30
Domestic violence experience		2.78	0 to 300	15.64
Number of needed services		1.11	0 to 9	1.78
Home of biological/adoptive parent	86.6
Kinship care	11.3
Out-of-home care	2.1
Parent demographic characteristics
Shared race/ethnicity (yes)	26.0
(No)	74.0
Education level		2.16	1 to 5	1.00
Family income		2.39	1 to 4	1.02
Employed parent (yes)	14.1
(No)	85.9
Caseworker turnover
Number of caseworkers parent worked with		2.30	0 to 97	2.09
Interview wave
First interview wave	34.2
Second interview wave	34.8
Third interview wave	31.0

Note. n = 6,142 person-waves.

Our multivariate analysis results confirmed that the hypothesized GLS random effects model differed significantly from the null model (Wald’s χ² = 228.23, p < .01; see Table 2). The intraclass correlation (ρ) from the random effects model was .34; ρ denotes the proportion of between-subjects variance explained by the overall variance (Rabe-Hesketh & Skrondal, 2008). Our multivariate analysis showed parent’s perceived caseworker engagement was associated positively with social support (b = .08, p < .01), mental health (b = .05, p < .01), kinship care (b = .64, p < .05), out-of-home placement (b = 1.04, p < .05), parent’s African American ethnicity (b = .83, p < .01), parent’s Hispanic ethnicity (b = .66, p < .05), parent/caseworker shared ethnicity (b = .55, p < .01), and family income (b = .18, p < .05). In contrast, the outcome was negatively associated with number of caseworkers parent worked with (b = −.39, p < .01). The remaining explanatory variables were not significantly associated with our study’s outcome variable.

Table 2.

Multivariate Analysis Results of Parent’s Perceived Caseworker Engagement.

Variables	b	Robust Standard Error
Interpersonal capacities
Social support	.08**	.01
Intrapersonal dynamics
Refusal of services	−.00	.01
Problem severity
Number of alcoholic drinks consumed	.06	.17
Nonmedical use of medication (no)	.15	.33
Mental health	.05**	.01
Domestic violence experience	−.00	.01
Number of needed services	.01	.05
Kinship care (home of biological/adoptive parent)	.64*	.28
Out-of-home care (home of biological/adoptive parent)	1.04*	.51
Neglect (other maltreatment)	−.34	.34
Sexual maltreatment (other maltreatment)	−1.25	.80
Physical maltreatment (other maltreatment)	−.92	.60
Parent demographic characteristics
Female (male)	−.30	.31
African American (White)	.83**	.24
Hispanic (White)	.66*	.26
Other racial/ethnic minority (White)	.45	.44
Shared race/ethnicity (no)	.55**	.20
Education level	.05	.10
Family income	.18*	.10
Employed parent (no)	−.46	.23
Caseworker turnover
Number of caseworkers parent worked with	−.39**	.06
Interview wave
Second interview wave (first interview wave)	.09	.18
Third interview wave (first interview wave)	−.05	.22
Constant	−5.71*	.72
Wald’s χ2	228.23**
Overall R²	.06
√ψ	3.74
√θ	5.22
ρ	.34

Note. n = 6,142 person-waves; √ψ = between-subjects standard deviation; √θ = within-subject standard deviation; ρ = intraclass correlation; reference groups in parentheses.

*p < .05. **p <.01.

Discussion

The z-scores for parent’s perceived caseworker engagement ranged from −.27.99 to 10.07. Further examination of the raw scores for the 6 items constituting the outcome measure yielded an average score of 18.33; or an average score of 3.05 (or sometimes) was obtained for an item with a 4-point response scale. In our study, the average raw score for the outcome was lower than the average score obtained by a prior study of collaborative alliance employing a standardized scale (de Greef et al., 2018). Furthermore, on average in our study, parents had sufficient social support, and the proportion of services they refused was low, implying relatively strong motivation or compliance. Overall, most parents needed very few services, very few drank alcohol, mental health was good for most, and most did not misuse prescribed medicine. A large majority of parents in our study cared for their children in their homes. Such findings suggest the parent sample had few problems, and problems of minor or moderate degree, not the severe problems tending to undermine collaborative alliance. Still, parents in the sample reported an average of two-plus episodes of domestic violence within the year. Parents in our sample did seem able to function fairly well; for example, the majority had retained custody of children at home. A plausible reason for such functioning is that this group of parents was characterized by few problems or at least by problems that were not especially severe and thus could be managed.

The findings offer some support for our hypothesis that in the child welfare setting, parent–caseworker collaborative alliance would be linked to a parent’s interpersonal capacities, intrapersonal dynamics, problem severity, and demographic characteristics, as well as caseworker turnover. Our results confirmed an association between parents’ high scores for social support and their high scores for perceived caseworker engagement. This finding supports the notion that parents with interpersonal difficulties have trouble building collaborative alliance with caseworkers (Bordin, 1979, 1994; Horvath, 1994; Horvath & Luborsky, 1993; Luborsky, 1976, 1994). The finding implies too that caseworkers should help parents build informal networks of support that remain in place once child welfare services have ended.

Contrary to prior findings (de Tychey et al., 2010; Horvath & Luborsky, 1993; Kemp et al., 2014), our findings suggest that parents with children in out-of-home placement and kinship care perceived greater caseworker engagement than did parents with children not removed from home. A plausible explanation is that earlier research focused on parents who had lost custody of their children. It is also probable that caseworkers worked especially hard to win such parents’ trust and cooperation in order to reunite families, while paying less attention to collaborative alliance with parents retaining custody of their children. In general, per our findings, caseworkers should stay vigorously involved in collaborating with biological and adoptive parents to promote lasting successful in-home placement.

Results of our study showed that parent refusal of services was not associated with perceived caseworker engagement, the outcome. This implies that parents’ motivation (i.e., their compliance) is unrelated to perceived caseworker engagement—and to collaborative alliance. The finding contradicts prior results (de Greef et al., 2018; Kemp et al., 2014; Littell, 2001; Smith, 2008), perhaps for the plausible reason that the number of services a parent refused provided only a proxy measure for motivation or compliance; it did not come close to providing a robust conceptualization of motivation. Moreover, it seems likely that the child welfare system’s mandatory and thus coercive nature commonly makes internal motivation to change a moot point. The finding nevertheless implies caseworkers need competence in building collaborative alliance with parents who resist, or are otherwise uncooperative in, the child welfare intervention process. The skills required within solution-focused therapy may be of considerable use here (O’Connell, 2012).

Our present findings show that parent’s perceived caseworker engagement was not significantly associated with level of alcohol consumption or nonmedical use of medication (indicators of parent’s problem severity). This contradicts some of the literature on client–therapist collaborative alliance (Fluckiger et al., 2013; Horvath, 2001; O’Brien et al., 2009; Wolfe et al., 2013). The discrepancy may be explained by the fact that we measured substance abuse by proxy, completely leaving out use of illicit drugs. As well, we found that the number of services needed was not associated with perceived caseworker engagement, contradicting prior results (Kemp et al., 2009) and implying that having multiple problems does not impede collaborative alliance; our measures might not, however, have captured problems’ intensity, acuteness, or chronicity. On the other hand, we observed an association between parents’ relatively good mental health and higher scores for perceived caseworker engagement. In this, our findings confirm prior research (Gurman, 1977). They imply, moreover, that some parents may need caseworkers’ help to access needed mental health services. Such care can be crucial to the development of robust caseworker–caregiver relationships.

Where parent demographics are concerned, in our study, only race/ethnicity was significantly associated with the outcome variable. Contrary to prior results (Kemp et al., 2009), our study found parent–caseworker collaborative alliances involving African American parents to be stronger than those involving White parents. It furthermore found Hispanic ethnicity to be associated positively with parent’s perceived caseworker engagement, contradicting prior results (Cheng & Lo, 2012). In light of some earlier related evidence (Cheng & Lo, 2018), it is reasonable to argue that cultural competence training helps caseworkers overcome cultural barriers and gain the trust of parents of minority ethnicity; matching parents’ and caseworkers’ ethnic backgrounds fosters, our study shows, successful working relationships that may presage desirable outcomes (McBeath et al., 2014). Implications of this result include a mandate that child welfare caseworkers complete cultural competency training and that agencies recruit caseworkers of minority ethnicity. Successful child welfare outcomes are fostered by caseworker–caregiver shared ethnicity (McBeath et al., 2014).

Like prior studies (Cheng & Lo, 2018; Kemp et al., 2009), our study observed a positive association between family income and parent’s perceived alliance. The implication is that low-income parents experience less sense of collaborative alliance with a caseworker. In turn, the present study observed no link between its outcome variable and parent’s education, consistent with prior results (Cheng & Lo, 2018). This may suggest that how diligently a caseworker pursues collaborative alliance with parents has little or nothing to do with their educational backgrounds. Such findings imply that securing parents’ collaboration may sometimes require caseworkers to recognize the presence of financial struggles and know how to help struggling families obtain beneficial resources even outside the child welfare domain (e.g., housing and childcare subsidies, job training, and employment at a living wage).

Like some earlier studies (Chambers et al., 2016; Chambers et al., 2018; Ryan, Garnier, Zyphur, & Zhai, 2006), our study indicates that caseworker turnover has an attenuating effect on collaborative alliance, throughout the helping process. Frequent changing of a parent’s caseworker should be avoided because it quashes desirable outcomes. Two more implications of our results for agency administrators are the need to ensure caseloads are reasonable and manageable and the need for supervision of caseworkers that is both supportive and sufficient.

Conclusion

The present study successfully identified factors in collaborative alliance between parents in the child welfare system and their caseworkers. Among other things, it showed it can be vital for caseworkers to help parents strengthen and broaden their networks of social support, improving parents’ interpersonal relationships, including the parent–caseworker collaborative alliance. In particular, caseworkers need to continue striving for alliance with parents whose children were removed from home. Family reunification needs to be diligently pursued, but perhaps even greater effort is needed on behalf of intact families so that they remain intact.

Throughout the helping process, caseworkers need to monitor parents’ mental health, tailoring relationship-building efforts to parents’ mental state; mentally ill clients, of course, must be helped to obtain requisite services. Domestic violence was not a significant factor in our study. Nevertheless, caseworkers need to explore all clients’ experiences of such (especially their recent experiences) and make prompt referrals as appropriate. Additionally, agencies and caseworkers alike need to show special commitment to the needs of clients of minority ethnicity. Pairing a parent with a caseworker who shares his/her ethnicity can be beneficial. Caseworkers also need to be diligent about building collaborative alliances with low-income clients; sensitivity to their needs is called for, as they may face numerous social structural barriers as well as stereotyping. Strong collaborative alliances can empower low-income parents to tackle some difficulties on their own.

Our findings also implied that caseworkers need to balance various demands entailed in practices and training. To gain such balance requires child welfare agencies to set manageable caseloads and provide ample supportive supervision to caseworkers pursuing collaborative alliance with parents.

Our study was limited by certain characteristics of its sample. That is, over 84% of respondents were substantiated for other maltreatment types, and over 92% were female. In addition, we were constrained to address missing values with a tactic increasingly outdated: the replacement of any such value with the midpoint of the given variable. Generalizing from our results, therefore, must be approached very cautiously. Caution in generalizing is also necessary in light of our proxy measuring of collaborative alliance, motivation to change, and substance use. Furthermore, the data set did not provide information regarding the intensity, acuteness, and chronicity of clients’ problems that might have affected measuring of mental health, substance use, and number of documented problems. Future research might develop comprehensive measures/diagnoses of these variables to overcome limitations like these. As well, while we chose to study caseworker engagement as perceived by parent, comparing parents’ perceptions to those of caseworkers could also be fruitful, perhaps yielding fresh understanding of collaborative alliance within the child welfare system. Our study observed no significant associations between collaborative alliance and maltreatment types; we treated maltreatment types strictly as controls. Future research, though, could profitably explore how past child welfare involvement impacts current collaborative alliance. It could also assess how caregiver roles—that of biological parent, or adoptive parent, or other—affect collaborative alliance.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Tyrone C. Cheng

References

Alexander

L. B.

Dore

M. M.

(1999). Making the parents as partners principle a reality: The role of the alliance. Journal of Child and Family Studies, 8, 255–270.

Alvidrez

Snowden

L. R.

Kaiser

D. M.

(2010). Involving consumers in the development of a psychoeducational booklet about stigma for Black mental health clients. Health Promotion Practice, 11, 249–258.

Bordin

E. S.

(1979). The generalizability of the psychoanalytic concept of the working alliance. Psychotherapy: Theory, Research & Practice, 16, 252–260. doi:10.1037/h0085885

Bordin

E. S.

(1994). Theory and research on the therapeutic working alliance: New directions. In Horvath

A. O.

Greenberg

L. S.

(Eds.), The working alliance: Theory, research, and practice. New York, NY: Wiley.

Buck

P. W.

Alexander

L. B.

(2006). Neglected voices: Consumers with serious mental illness speak about intensive case management. Administration and Policy in Mental Health, 33, 470–481.

Castonguay

L. G.

Constantino

M. J.

Holtforth

M. G.

(2006). The working alliance: Where are we and where should we go? Psychotherapy, 43, 271–279.

Chambers

R. M.

Brocato

Fatemi

Rodriguez

A. Y.

(2016). An innovative child welfare pilot initiative: Results and outcomes. Children and Youth Services Review, 70, 143–151.

Chambers

R. M.

Crutchfield

R. M.

Harper

S. G. G.

Fatemi

Rodriguez

A. Y.

(2018). Family reunification in child welfare practice: A pilot study of parent and staff experiences. Children and Youth Services Review, 91, 221–231.

Cheng

T. C.

(2010). Factors associated with reunification: A longitudinal analysis of long-term foster care. Children and Youth Services Review, 32, 1311–1316.

10.

Cheng

T. C.

C. C.

(2012). Racial disparities in access to needed child welfare services and worker–client engagement. Children and Youth Services Review, 34, 1624–1632.

11.

Cheng

T. C.

C. C.

(2015). A longitudinal causal analysis of impact made by collaborative engagement and service receipt on likelihood of substantiated re-report. Child Maltreatment, 20, 258–267.

12.

Cheng

T. C.

C. C.

(2016). Linking worker-parent working alliance to parent progress in child welfare: A longitudinal analysis. Children and Youth Services Review, 71, 10–16.

13.

Cheng

T. C.

C. C.

(2018). Do social work education, job description, and cultural competence foster child-welfare caseworkers’ therapeutic alliances? Child & Family Social Work, 23, 435–442.

14.

Child

McIntyre

(2015). Examining the relationships between family drug court program compliance and child welfare outcomes. Child Welfare, 94, 67–87.

15.

Children’s Bureau. (2018). Child maltreatment 2016. Washington, DC: U.S. Department of Health & Human Services.

16.

Crits-Christoph

Gibbons

M. B. C.

Hearon

(2006). Does the alliance cause good outcome? Recommendations for future research on the alliance. Psychotherapy, 43, 280–285.

17.

de Greef

van Hattum

M. J. C.

Granger

K. L.

McLeod

B. D.

Pijnenburg

H. M.

Scholte

R. H. J.

(2018). Predictors of parent-professional alliance in home-based parenting support. Children and Youth Services Review, 89, 171–178.

18.

de Tychey

Garnier

Lighezzolo-Alnot

Claudon

Rebourg-Roesler

(2010). An accumulation of negative life events and the construction of alexithymia: A longitudinal and clinical approach. Journal of Personality Assessment, 92, 189–206.

19.

Doran

Doukas

D’Andrea

Nnamdi

(2011). Psychophysiological correlates of complex trauma and their relationship to psychotherapy process. Psychophysiology, 48, S58–S59.

20.

Dore

M. M.

Alexander

L. B.

(1996). Preserving families at risk of child abuse and neglect: The role of the helping alliance. Child Abuse & Neglect, 20, 349–361.

21.

Doukas

D’Andrea

Doran

Pole

(2014). Psychophysiological predictors of working alliance among treatment-seeking women with complex trauma exposure. Journal of Traumatic Stress, 27, 672–679.

22.

Drake

(1994). Relationship competences in child-welfare services. Social Work, 39, 595–602.

23.

Drieschner

K. H.

Lammers

S. M. M.

van der Staak

C. P. F.

(2004). Treatment motivation: An attempt for clarification of an ambiguous concept. Clinical Psychology Review, 23, 1115–1137.

24.

Duncan

B. L.

Miller

S. D.

Sparks

J. A.

(2004). The heroic client: A revolutionary way to improve effectiveness through client-directed, outcome-informed therapy. San Francisco, CA: Jossey-Bass.

25.

Fluckiger

Del Re

A. C.

Horvath

A. O.

Symonds

Ackert

Wampold

B. E.

(2013). Substance use disorders and racial/ethnic minorities matter: A meta-analytic examination of the relation between alliance and outcome. Journal of Counseling Psychology, 60, 610–616. doi:10.1037/a0033161

26.

Gladstone

Dumbrill

Leslie

Koster

Young

Ismaila

(2014). Understanding worker-parent engagement in child protection casework. Children and Youth Services Review, 44, 56–64. doi:10.1016/j.childyouth.2014.06.002

27.

Goodman

L. A.

Fauci

J. E.

Sullivan

C. M.

DiGiovanni

C. D.

Wilson

J. M.

(2016). Domestic Violence survivors’ empowerment and mental health: Exploring the role of the alliance with advocates. American Journal of Orthopsychiatry, 86, 286–296. doi:10.1037/ort0000137

28.

Graybeal

C. T.

(2007). Evidence for the art of social work. Families in Society: The Journal of Contemporary Social Services, 88, 513–523. doi:10.1606/1044-3894.3673

29.

Grella

C. E.

Needell

Shi

Y. F.

Hser

Y. I.

(2009). Do drug treatment services predict reunification outcomes of mothers and their children in child welfare? Journal of Substance Abuse Treatment, 36, 278–293. doi:10.1016/j.jsat.2008.06.010

30.

Gross

J. J.

(1998). The emerging field of emotion regulation: An integrative review. Review of General Psychology, 2, 271–299.

31.

Gurman

A. S.

(1977). The patient’s perception of the therapeutic relationship. In Gurman

A. S.

Razin

A. M.

(Eds.), Effective psychotherapy (pp. 503–543). New York, NY: Pergamon Press.

32.

Horvath

A. O.

(1994). Empirical validation of Bordin’s pantheoretical model of the alliance: The working alliance inventory perspective. In Horvath

A. O.

Greenberg

L. S.

(Eds.), The working alliance: Theory, research, and practice (pp. 109–128). New York, NY: Wiley.

33.

Horvath

A. O.

(2001). The alliance. Psychotherapy, 38, 365–372.

34.

Horvath

A. O.

(2006). The alliance in context: Accomplishments, challenges, and future directions. Psychotherapy, 43, 258–263.

35.

Horvath

A. O.

Luborsky

(1993). The role of the therapeutic alliance in psychotherapy. Journal of Consulting and Clinical Psychology, 61, 561–573.

36.

Johnson

R. S.

Fowler

J. C.

Jani

S. N.

Eichelberger

H. L.

Oldham

J. M.

Poa

Graham

D. P.

(2017). A systematic analysis of treatment effects on depressive symptom severity by level of coercion. Psychiatric Quarterly, 88, 39–46.

37.

Kemp

S. P.

Marcenko

M. O.

Hoagwood

Vesneski

(2009). Engaging parents in child welfare services: Bridging family needs and child welfare mandates. Child Welfare, 88, 101–126.

38.

Kemp

S. P.

Marcenko

M. O.

Lyons

S. J.

Kruzich

J. M.

(2014). Strength-based practice and parental engagement in child welfare services: An empirical examination. Children and Youth Services Review, 47, 27–35.

39.

Kendra

M. S.

Mohr

J. J.

Pollard

J. W.

(2014). The stigma of having psychological problems: Relations with engagement, working alliance, and depression in psychotherapy. Psychotherapy, 51, 563–573.

40.

Kondrat

D. C.

Early

T. J.

(2010). An exploration of the working alliance in mental health case management. Social Work Research, 34, 201–211. doi:10.1093/swr/34.4.201

41.

Lalayants

Baier

Benedict

Mera

(2014). Social supports in the lives of child welfare-involved families. Child Welfare, 93, 93–118.

42.

Littell

J. H.

(2001). Client participation and outcomes of intensive family preservation services. Social Work Research, 25, 103–113.

43.

Littell

J. H.

Alexander

L. B.

Reynolds

W. W.

(2001). Client participation: Central and underinvestigated elements of intervention. Social Service Review, 75, 1–28.

44.

Luborsky

(1976). Helping alliances in psychotherapy. In Cleghhorn

J. L.

(Ed.), Successful psychotherapy (pp. 92–116). New York, NY: Brunner/Mazel.

45.

Luborsky

(1994). Therapeutic alliances as predictors of psychotherapy outcomes: Factors explaining the predictive success. In Horvath

A. O.

Greenberg

L. S.

(Eds.), The working alliance: Theory, research, and practice (pp. 38–50). New York, NY: Wiley-Interscience.

46.

Maikovich-Fong

A. K.

Jaffee

S. R.

(2010). Sex differences in childhood sexual abuse characteristics and victims’ emotional and behavioral problems: Findings from a national sample of youth. Child Abuse & Neglect, 34, 429–437.

47.

McBeath

Chuang

Bunger

Blakeslee

(2014). Under what conditions does caseworker-caregiver racial/ethnic similarity matter for housing service provision? An application of representative bureaucracy theory. Social Service Review, 88, 134–165.

48.

McWey

L. M.

Holtrop

Wojciak

A. S.

Claridge

A. M.

(2015). Retention in a parenting intervention among parents involved with the child welfare system. Journal of Child and Family Studies, 24, 1073–1087.

49.

Nakash

Nagar

Levav

(2015). Predictors of mental health care stigma and its association with the therapeutic alliance during the initial intake session. Psychotherapy Research, 25, 214–221.

50.

National Data Archive on Child Abuse and Neglect. (2013). National survey of child and adolescent well-being (Combined waves 1–3 data file user’s manual, general release version). Ithaca, NY: Administration on Children and Families.

51.

O’Brien

Fahmy

Singh

S. P.

(2009). Disengagement from mental health services. Social Psychiatry and Psychiatric Epidemiology, 44, 558–568.

52.

O’Connell

(2012). Solution-focused therapy. Los Angeles, CA: Sage.

53.

Paivio

S. C.

Patterson

L. A.

(1999). Alliance development in therapy for resolving child abuse issues. Psychotherapy, 36, 343–354.

54.

Paris

Anez

L. M.

Bedregal

L. E.

Andres-Hyman

R. C.

Davidson

(2005). Help seeking and satisfaction among Latinas: The roles of setting, ethnic identity, and therapeutic alliance. Journal of Community Psychology, 33, 299–312.

55.

Piper

W. E.

Azim

H. F. A.

Joyce

A. S.

McCallum

Nixon

G. W. H.

Segal

P. S.

(1991). Quality of object relations versus interpersonal functioning as predictors of therapeutic alliance and psychotherapy outcome. Journal of Nervous and Mental Disease, 179, 432–438.

56.

Rabe-Hesketh

Skrondal

(2008). Multilevel and Longitudinal modeling using Stata. College Station, TX: Stata Press.

57.

Rajendran

Smith

B. D.

Videka

(2015). Association of caregiver social support with the safety, permanency, and well-being of children in child welfare. Children and Youth Services Review, 48, 150–158.

58.

Redko

Rapp

R. C.

Elms

Snyder

Carlson

R. G.

(2007). Understanding the working alliance between persons with substance abuse problems and strengths-based case managers. Journal of Psychoactive Drugs, 39, 241–250.

59.

Ryan

J. P.

Garnier

Zyphur

Zhai

(2006). Investigating the effects of caseworker characteristics in child welfare. Children and Youth Services Review, 28, 993–1006.

60.

Singer

J. D.

Willett

J. B.

(2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York, NY: Oxford University Press.

61.

Smith

B. D.

(2008). Child welfare service plan compliance: Perceptions of parents and caseworkers. Families in Society: The Journal of Contemporary Social Services, 89, 521–532.

62.

Urbanoski

K. A.

Kelly

J. F.

Hoeppner

B. B.

Slaymaker

(2012). The role of therapeutic alliance in substance use disorder treatment for young adults. Journal of Substance Abuse Treatment, 43, 344–351.

63.

Ward

E. C.

Clark

L. O.

Heidrich

(2009). African American women’s beliefs, coping behaviors, and barriers to seeking mental health services. Qualitative Health Research, 19, 1589–1601.

64.

Ware

J. E.

Kosinski

Keller

S. D.

(1996). A 12-item short-form health survey: Construction of scales and preliminary tests of reliability and validity. Medical Care, 34, 220–233.

65.

Wolfe

Kay-Lambkin

Bowman

Childs

(2013). To enforce or engage: The relationship between coercion, treatment motivation and therapeutic alliance within community-based drug and alcohol clients. Addictive Behaviors, 38, 2187–2195.

66.

Yatchmenoff

D. K.

(2005). Measuring client engagement from the client’s perspective in nonvoluntary child protective services. Research on Social Work Practice, 15, 84–96.

67.

Zlotnik

J. L.

Strand

V. C.

Anderson

G. R.

(2009). Achieving positive outcomes for children and families: Recruiting and retaining a competent child welfare workforce introduction. Child Welfare, 88, 7–21.