Caregivers’ readiness for change: predictive validity in a child welfare sample

Abstract

Objective:

To assess the predictive validity of continuous measures of problem recognition (PR), intentions to change (ITC), and overall readiness for change (RFC) among primary caregivers who received in-home services following substantiated reports of child abuse or neglect.

Method:

A modified version of the University of Rhode Island Change Assessment scale was included in interviews with a sample of 353 primary caregivers at 4 weeks, 16 weeks, and 1 year after referral for in-home services. Additional data were obtained from administrative records and caseworker surveys. Hierarchical linear and nonlinear models were used to assess relationships between PR, ITC, RFC and changes over time in measures of individual and family functioning (e.g., parenting behaviors, children's behaviors, housing and economic problems, social support, and life events). Bivariate probit regression analysis was used to examine relationships between PR, ITC, RFC and the likelihood of subsequent reports of child maltreatment and out-of-home placements within 1 year after referral.

Results:

Initial problem recognition and intentions to change predict a few improvements in individual and family functioning, along with significant reductions in the likelihood of additional reports of child maltreatment within 1 year. Initial intentions to change also predict reductions in the substantiation of subsequent reports of maltreatment. An overall measure of readiness for change predicts reductions in the likelihood of out-of-home placement.

Conclusions:

Problem recognition and intentions to change predict somewhat different outcomes; hence, there are few advantages of a combined measure of readiness for change. Further inquiry is needed to determine whether and how these associations are mediated by intervention processes or other factors in child welfare services populations.

Résumé

Objectif:

Evaluer la validité prédictive de mesures continues de la reconnaissance du problème (RP), des intentions de changer (IC), et de l’aptitude générale au changement (AGC) chez ceux qui élèvent un enfant et qui ont reçu des aides à domicile suite à des signalements de maltraitance ou de négligence infantile.

Méthode:

Une version modifiée de «l’Echelle d’Evaluation du Changement» de l’Université de Rhode Island a été incluse dans les entretiens d’un échantillon de 353 personnes en charge d’enfants, 4 semaines, 16 semaines, et un an après un signalement aux services sociaux. Des données supplémentaires étaient fournies par les rapports administratifs et les enquêtes des travailleurs sociaux. Des modèles hiérarchiques linéaires et non-linéaires ont été utilisés pour évaluer les relations entre la RP, l’IC, l’AGC et les changements dans le temps des mesures du fonctionnement individuel et familial (c’est-à-dire les comportements parentaux, les comportements des enfants, les problèmes économiques et de logement, l’aide sociale, et les événements de la vie). Une analyse de régression bivariée a été utilisée pour étudier les relations entre la PR, les IC, l’AGC et la probabilité de nouveaux signalements de maltraitance infantile et de placements extérieurs au cours de l’année qui suit le premier signalement.

Résultats:

La reconnaissance initiale du problème et les intentions de changer augurent de quelques améliorations dans le fonctionnement individuel et familial, ainsi que de réductions significatives de la probabilité d’autres signalements de maltraitance infantile au cours de l’année. Les intentions initiales de changement augurent également de réductions de signalements ultérieurs. Une mesure globale de bonne volonté de changement présage de réductions dans la probabilité de placement externe.

Conclusions:

La reconnaissance du problème et les intentions de changer présagent de résultats quelque peu différents ; de là peu d’avantages à associer une évaluation de la bonne volonté de changement. On a besoin d’une investigation supplémentaire pour affirmer si et comment ces résultats sont modifiés par des processus interventionnels ou d’autres facteurs provenant des services qui s’occupent de bien-être infantile.

Resumen

Objetivo:

Evaluar la validez predictiva de medidas repetidas de reconocimiento de problemas (PR), intenciones de cambio (IC) y disponibilidad general para el cambio (DGR) entre cuidadores primarios que reciben tratamiento en domicilio después de notificaciones confirmadas de maltrato o abandono infantil.

Método:

En las entrevistas realizadas con una muestra de 353 cuidadores a las 4 semanas, 16 semanas y un año después de la derivación a recursos de atención domiciliaria se aplicó una versión modificada de la Escala de evaluación del cambio de la Universidad de Rhode Island. Se obtuvieron datos adicionales de los expedientes administrativos y las entrevistas de los coordinadores de caso. Para evaluar la relación entre PR, IC y DGR y los cambios a lo largo del tiempo en medidas de funcionamiento individual y familiar (p.e., conductas parentales, conductas de los niños, problemas en el hogar y económicos, apoyo social y acontecimientos vitales) se utilizaron modelos jerárquicos lineales y no lineales. Los análisis de regresión bivariada probit fueron utilizados para examinar la relación entre PR, IC y DGR y la posibilidad de notificaciones posteriores de maltrato y medidas de salida del domicilio en el año posterior a la derivación.

Resultados:

El reconocimiento inicial de los problemas y las intenciones de cambio predicen una ligera mejora en el funcionamiento individual y familiar y una reducción significativa en la posibilidad de notificaciones posteriores de maltrato infantil en el año siguiente. Las intenciones iniciales de cambio también predicen reducciones en la confirmación de posteriores notificaciones de maltrato. Una medida general de disponibilidad para el cambio predice reducciones en la posibilidad de que se tome una medida de separación del menor de su domicilio.

Conclusiones:

El reconocimiento del problema y las intenciones de cambio predicen consecuencias algo diferentes por lo que se observan pocas ventajas de una medida combinada de disponibilidad para el cambio. Se necesita más investigación para determinar si y cómo estas asociaciones están mediadas por procesos de intervención o otros factores en los servicios de protección infantil.

Introduction

For assessment and case planning purposes, child welfare workers and other helping professionals usually consider parents’ perceptions of the events that led to their involvement in the child welfare system (Dubowitz & DePanfilis, 2000). Parents and other caregivers who acknowledge abusive or neglectful practices and express a desire to change those practices are generally thought to pose less risk of future harm to children than caregivers in similar cases who deny caregiving problems or express little interest in addressing these issues (Gelles, 1995, Gelles, 1996, Gelles, 2000). Caregivers who seem ready for change at the outset are also expected to respond better to treatment than those who appear to be unwilling to change. Hence, readiness for change is considered an important component of risk assessment, case planning, decision-making, and allocation of treatment and out-of-home placement resources in child welfare (Gelles, 1995, Gelles, 1996, Gelles, 2000).

Although clients’ readiness for change may affect intervention processes and outcomes, there is some evidence that problem recognition and stated intentions to change are not good predictors of behavior (Miller, 1985). As with outward compliance (Littell, 2001), it is possible that clients’ apparent readiness for change affects casework decisions, which in turn influence outcomes. Other factors that are associated with apparent readiness could also account for differences in outcomes.

In this article, we view caregivers’ readiness to change as a potential predictor of the outcomes of in-home child welfare services. Following substantiated reports of child abuse or neglect, caregivers who reported more readiness for change were expected to demonstrate greater improvement in targeted outcomes within several months and be more likely to maintain those gains at a 1-year follow-up.

Client motivation or readiness for change has been conceptualized in several ways: as part of a stage model of cognitive and behavioral change, as a single continuum, and as a multi-dimensional construct. Most of the recent empirical evidence on the predictive validity of readiness for change comes from research on the “stages of change,” which is the central organizing construct in Prochaska and DiClemente, 1984, Prochaska and DiClemente, 1986, Prochaska and DiClemente, 1992, Prochaska and DiClemente, 1998 Transtheoretical Model of behavior change.

According to the Transtheoretical Model, readiness for change and behavioral change develop in an ordered sequence of discrete stages. These stages are termed precontemplation, contemplation, preparation, action, and maintenance (DiClemente & Prochaska, 1998). Although stage status changes over time, at any given moment a person is assumed to be in a single stage; hence, the stages are considered mutually exclusive (Martin, Velicer, & Fava, 1996). People are thought to “pass through each stage” in an orderly fashion (Prochaska, DiClemente, Velicer, & Rossi, 1992, p. 825, emphasis in the original), but this progression is not always linear. People may relapse, returning to a previous stage, and may cycle through the stage sequence more than once (Prochaska et al., 1992).

The early stages of change are thought to be practical predictors of treatment participation and outcomes (Prochaska & Velicer, 1997). For example, people in the precontemplation stage are expected to be less likely than those in the contemplation and preparation stages to enter treatment, remain in treatment, and succeed in changing their behavior. However, empirical evidence does not provide consistent support for the predictive validity of these stages (for a review, see Littell & Girvin, 2002). Some studies found no significant relationship between stages of change and measures of treatment attendance, duration, or program completion. In other studies, measures of precontemplation (indicating that the participant was not considering change) were predictive of drop-out (Smith, Subich, & Kalodner, 1995), but sometimes in the “wrong” direction (i.e., longer retention in treatment; Belding, Iguchi, Lamb, Lakin, & Terry, 1995; Jefferson, 1991). Similarly, baseline stage has been predictive of outcomes in some studies and not others, and there are mixed results within some studies. In a review of prospective studies of stages of change, Belding, Iguchi, and Lamb (1997) found that “none clearly and consistently supports the predictive validity” of the stages of change model (p. 65).

Some reviewers have noted that the empirical evidence is more consistent with a continuous model of readiness for change (Carey, Purnine, Maisto, & Carey, 1999; Davidson, 1998, Sutton, 1996; Weinstein, Rothman, & Sutton, 1998). A few studies provide evidence that a single continuous measure fits the data at least as well as a stage model (Budd & Rollnick, 1996; Satterfield, Buelow, Lyddon, & Johnson, 1995; Tsoh, 1995). Using a single continuous motivation score, the Project MATCH Research Group (1997) found that initial motivation for change was predictive of better outcomes among outpatients but not those in aftercare treatment for alcoholism.

Multidimensional models of readiness for change have been proposed by Bandura, 1997, Bandura, 1998, Miller and Tonigan (1996), and others. Bandura suggested that the stages of change reflected fluctuations along two separate dimensions: intentions to change and gradations of the problem behavior. Others have identified continuous factors that appear to represent various cognitive, affective, and behavioral dimensions of readiness for change (Hemphill & Howell, 2000; Miller & Tonnigan, 1996). Isenhart (1997) reported that three continuous pretreatment measures of readiness predicted some outcomes for alcohol dependent men, but not the quantity and frequency of their alcohol consumption.

There is little empirical research on readiness for change among caregivers of abused and neglected children. In a previous analysis of data on readiness for change among caregivers in a child welfare sample, we found that a two-dimensional model fit the data better than either a stage model or a single continuum (Littell & Girvin, submitted for publication). These dimensions, termed problem recognition (PR, the inverse of precontemplation) and intentions to change (ITC, self-reported contemplation and action), were not correlated in our sample. Here, we examine the predictive validity of PR and ITC in comparison with an overall measure of readiness for change.