Does the Transtheoretical Model Predict Attrition in Domestic Violence Treatment Programs?
Brodeur, Normand, Rondeau, Gilles, Brochu, Serge, Lindsay, Jocelyn, Phelps, Jason, Violence and Victims
Attrition in intervention programs for domestically violent men is considered to be a serious and enduring problem. Researchers have found a number of sociodemographic variables that partially explain this phenomenon; however, models based on these variables have a limited predictive power. Scott (2004) argues that a firm theoretical base is needed in future investigations of the problem and suggests the use of the transtheoretical model of behavior change (TTM), which was found to predict dropout with accuracy in other areas of behavioral change. This study investigated the relationship between four TTM constructs (Stages of Change, Decisional Balance, Self-Efficacy, and Processes of Change) and premature termination with a sample of Canadian French-speaking men (N = 302) in five domestic violence treatment programs. Contrary to the initial hypotheses, the TTM constructs did not predict dropout. Discussion investigates how social desirability bias affects results being obtained by current TTM measures and whether more motivation to change at intake necessarily relates to involvement in treatment for longer periods of time.
Keywords: transtheoretical model; intimate partner violence; treatment dropouts; human males
Over the past 20 years, conjugal violence prevention in North America has evolved multilayered responses that include treatment programs mandated to help men cease their violent behavior toward their female partners. One issue of concern regarding the efficacy of these programs has been their particularly high dropout rates and the impact this has on ensuring the safety of the women and children in these men's lives. Men who end treatment prematurely have been shown to have higher rearrest rates for reassaulting their partners and are reported to be more controlling and psychologically abusive than men who complete (Gondolf, 1997).
THE PROBLEM OF ATTRITION
Attrition in intervention programs for domestically violent men ranges from 22% to 99% (for review, see Daly & Pelowski, 2000) and is considered to be a serious and enduring problem. The fact that most men who enter treatment programs terminate prematurely highlights the pressing need to identify program and client factors common to unsuccessful completion. Over the years, there have been a variety of studies intended to address this question, but results have been of limited utility for intervention planning. In their comprehensive review of 16 studies dealing specifically with dropout, Daly and Pelowski (2000) were able to identify the reoccurrence of several sociodemographic factors that predicted dropout but with only moderate accuracy. Specifically, early terminators were found more likely to be unmarried and/or childless and unemployed and to have lower incomes and less education than men who remained in treatment. Other variables that correlated with premature termination in the various studies included substance abuse, criminal history, marital status, psychopathology, and court or correctional services involvement. However, the latter variables presented less consistently and may be influenced by a combination of factors not yet identified. In a later study, Daly, Power, and Gondolf (2001) went beyond the dropout/completion dichotomy to identify how long clients with different sociodemographic characteristics actually stayed in treatment. In addition to confirming results from their earlier study, Daly and her colleagues were also able to show that clients with low socioeconomic status (SES) and alcohol problems were not only dropping out but also terminating earlier than clients in other categories.
These results appear to mirror findings within the field of general psychotherapy that indicate that SES and lifestyle instability have a significant effect on treatment outcome (e.g., Garfield, 1994), especially when it comes to other serious public health issues like drinking and driving, substance abuse, and sexual assault (Gondolf, 1997). …