PROTOCOL: The Therapeutic Alliance and Psychotherapy Outcomes for Young Adults Aged 18 to 34: Protocol for a Systematic Review

In recent decades, important demographic, social, and cultural changes have affected the lives and needs of young adults in many countries around the world. In high-income countries, young adults are now taking longer to finish their education, assume full time employment, and undertake personal financial responsibility (Bound, Lovenheim, & Turner, 2010; Furstenberg, Rimbaut, & Settersten, 2005; Schoeni & Ross, 2004). For example, The Organization for Economic Co-Operation and Development (OECD) shows greatly increased rates of unemployment in the “youth“ (15–24) age group in countries such as Spain, Greece, and Ireland; and rising unemployment pushes many young people back home to live with their parents (Bell & Blanchflower, 2010). Based on OECD data, The Economist states that “46% of 18 to 34 year-olds in the European Union lived with at least one parent; in most countries the stay-at-homes were more likely to be unemployed than those who had moved out” (“The Jobless Young,” 2011). Newman and Newman (2011) indicate the developmental foci for 18 to 34 year-olds are the psychosocial tasks (from Erikson, 1959) of establishing facets of identity, and constructing the hallmarks of adulthood such as intimate relationships, childbearing, and mature work. For many, these tasks now take place after age 25, and into their early 30's. Hence, recent studies of young adulthood produced by groups such as The Transition to Adulthood (MacArthur Foundation), have focused on 18–34 year-olds. The Pew Foundation called the 25 to 34 year-old cohort, “The Boomerang Generation,“ for their repeated returns to their parent's home and financial dependence exacerbated by lower levels of economic opportunities (Parker, 2012). This preparatory period of life is putting additional strains on individuals, families of origin, and on institutions that support young adults (Berlin, Furstenberg, & Waters, 2010). At the same time, young adults are coping with marked levels of mental health issues. Psychiatric disorders in the U.S. are most prevalent in young adulthood for ages 18 to 25, with 29.9% of that group reporting serious mental illness (U.S. Substance Abuse and Mental Health Services Administration, 2010). The U.S. General Accounting Office (GAO) estimated that “at least 2.4 million young adults aged 18 through 26—or 6.5% of the 37 million non-institutionalized young adults in that age range—had a serious mental illness in 2006“ (U.S. GAO, 2008, p. 9). The U.S. 2007 National Survey on Drug Use and Health (NSDUH) indicates that younger adults (ages 18 to 25) demonstrate higher levels of psychological distress than other age groups in the United States (U.S. Substance Abuse and Mental Health Services Administration, 2010). Kessler et al. (2007) summarized the World Health Organization's (WHO) World Mental Health Survey data on the incidence and prevalence of major mental health disorders (anxiety, mood, impulse control, and substance use) across 17 countries. Examined by age group (18 to 34, 35 to 49, 50 to 64, 65+), the WHO data show high rates of multiple types of mental disorders for 18 to34 year olds, indicating significant levels of distress for this age group in many countries around the world (Kessler et al., 2007). These individuals with psychological distress are at heightened risk for mental health problems due to complex and changing societal factors, and yet they are less likely than other adults to receive mental health services (10.9 among those aged 18 to 25 vs. 14.8 percent among those aged 26 to 49 and 13.6 percent among those aged 50 or older) (U.S. Substance Abuse and Mental Health Services Administration, 2010). In the Worldwide Use of Mental Health Services for Anxiety, Mood, and Substance Disorders study based on the results from 17 countries in the WHO World Mental Health (WMH) Surveys (Kessler et al., 2007) the authors state, “age was a significant predictor of receiving mental health services in eight countries; in these, respondents in the middle years of life were generally more likely to receive services than either those younger or older“ (Wang et al., 2007, pg. 6). However, compared with the voluminous literature on adults, fewer studies and no systematic reviews examine the effects of the alliance on psychotherapeutic outcomes specifically for young adult clients. Among the elements that may lead to successful psychotherapy, the therapeutic alliance is considered both a central and a common factor linked to outcomes in psychotherapy by many leading psychotherapists (Bordin, 1979; Freud, 1912/1958; Rogers, 1951) as well as proponents of the “Common Factors“ orientation (Rosenzweig, 1936; Wampold, 2001). “Common Factors” refers to the perspective that effectiveness in psychotherapy is due to factors that are common to all forms of therapy rather than to specific techniques. Although considered integral to the process of psychotherapy, leading researchers have disagreed about exactly what the alliance is and how it works. Authors have used diverse theoretical conceptualizations to describe (Bordin, 1983; Freud 1912/1958; Greenson, 1965; Zetzel, 1956), operationalize, and measure the alliance (Gaston & Marmar, 1994; Horvath, 1981; Luborsky, Crits-Cristophe, Alexander, Margolis, & Cohen, 1983; Marmar, Horowitz, Weiss, & Marziali, 1986; Suh, Strupp, & O'Malley, 1986). The variety of therapeutic alliance constructs is reflected in the diversity of measurement instruments that have arisen out of different psychotherapeutic frameworks. For example, the Working Alliance Inventory (Horvath & Greenberg, 1986) is a 36-item scale with three subscales reflecting Bordin's (1983) pantheoretical, tripartite conceptualization of the working alliance, which includes agreement on the goals of treatment, agreement on the tasks of treatment, and affective bonds. The Penn Helping Alliance, a shorter 10-item scale, evaluates two separate aspects of the alliance called Type I and Type II. Type I rates the client's feelings about the therapist from a psychoanalytically-influenced perspective and Type II measures therapist and client agreement on the tasks and goals of treatment, similar to the Working Alliance Inventory and Bordin's conceptualization (Fenton, Cecero, Nich, Frankforter, & Carroll, 2001). In their meta-analysis, Horvath et al. (2011) noted that there are at least 30 different alliance measures and multiple studies show the leading measures to be related to each other (Bachelor, 1991; Horvath & Luborsky, 1993; Tichenor & Hill, 1989). In their factor analysis of the patient version of three leading measures, Hatcher and Barends (1996, p. 1328) found “the total scores on the three measures correlated highly: CALPAS and WAI, r = .85; CALPAS, and HAQ, r = .74; WAI and HAQ, r = .74 (p<.0001, N= 231), indicating the presence of a strong general factor.“ However, the authors state “there has been little evidence to support the theoretical dimensions that underlie the measures.” Their analysis of the Working Alliance Inventory, California Psychotherapy Alliance Scales, and the Helping Alliance Questionnaire indicated that after removing the large general factor, only two of the six factors identified using principal component analysis, “Confident Collaboration and Idealized Relationship, correlated with patients' estimate of improvement (rs = .37 and -.23, respectively; p <.001)“ (Hatcher & Barends, 1996, p. 1326). Given the large body of empirical research on the therapeutic alliance and its relation to outcomes, meta-analyses have focused on specific populations such as adults (18+) (Horvath & Bedi, 2002; Horvath & Symonds, 1991; Horvath et al., 2011; Martin et al., 2000), youth (under 19) (McLeod, 2011; Shirk, Karver, & Brown, 2011; Shirk & Karver, 2003), youth and families (Karver, Handelsman, Fields, & Bickman, 2006), and couples and families (Friedlander, Escudero, Heatherington, & Diamond, 2011). We know of no meta-analyses on the alliance and outcomes specifically looking at young adults. Between 1991 and 2001 four major meta-analyses examined relationships between the therapeutic alliance and psychotherapy outcomes for adults in individual psychotherapy (Horvath & Bedi, 2002; Horvath & Symonds, 1991; Horvath et al., 2011; Martin et al., 2000). These meta-analyses produced similar, moderate correlational effect sizes (ES) ranging from r=.21 (Horvath & Bedi, 2002) to r = .28 (Horvath, Del Re, Fluckinger, & Symonds, 2011). Four additional meta-analyses examined correlations between the alliance and outcomes in psychotherapy for youth (children and adolescents) in individual and family treatment (Shirk & Karver, 2003; Karver et al., 2006; McLeod, 2011; Shirk, Karver, & Brown, 2011). These studies found correlational ESs of r=.24, r=.17, r=.14, and r=.22 respectively. Two authors noted methodological issues, such as study heterogeneity, might have impacted their effect sizes. We believe that all of the previous meta-analyses of research on the alliance and outcomes for adult clients in individual psychotherapy have limitations when compared to currently recommended methodological standards. The AMSTAR instrument, developed to assess the methodological quality of systematic reviews and meta-analyses (Shea et al., 2007), contains 11 items identified by exploratory factor analysis performed on over 150 studies to identify core components of review quality. Using the AMSTAR criteria, the four meta-analyses performed on studies of adults (Horvath & Bedi, 2002; Horvath & Symonds, 1991; Horvath, et al., 2011; Martin et al., 2000) appear to lack many of elements of rigorous and valid research syntheses. For example, none of the four meta-analyses reported that they had a public, a priori design, duplicate study selection and data extraction, or formal evaluation of study quality. The hazards of psychological setbacks to young adults entering employment, family commitments, and higher education can be severe: mental health issues increase risk for many long-term, negative consequences such as not finishing education, unplanned pregnancy, drug abuse, and unemployment (Gralinski-Bakker Hauser, Billings, & Allen, 2005). In addition, young adulthood is a period when some individuals are experiencing their first episodes of mental illness and are especially challenged (Pottick et al., 2008). Yet, two studies based on nationally representative US samples (Kessler et al., 2005; Pottick et al., 2008) indicate that individuals aged 18 – 24 were “significantly less likely to receive mental health services than adults in older age (Pottick et al., 2008, p. Pottick et al. indicate that multiple may for this age group of health to of and changes in the mental health needs of this For example, of is for individuals from mental health and services from which they are to fewer adult services (Pottick et al., 2008). evidence indicates that the prevalence and of mental health problems may be among young adults, as of with more complex and mental health problems 2010). This systematic review the between the therapeutic alliance and psychotherapy outcomes, using studies of young adults (ages that and valid measures of the therapeutic alliance and psychotherapy The of the alliance is to more studies identify the quality of the therapeutic alliance as one of the major factors that may lead to outcomes across many forms of psychotherapy & the process of not therapeutic the alliance has to be considered a or component of successful therapy (Rosenzweig, & Wampold, 2001). from Freud have to aspects of the therapeutic that may in outcomes for the Freud that the client's or and therapist with both and and lead to therapeutic This to that the into elements from the as the and the elements that are in time between the client and the of the therapeutic alliance when identified and and as for therapeutic recent a and orientation with three major agreement and of the goals of agreement on tasks to be and affective between therapist and client (Bordin, these factors are to be of the client's as or has We know of no systematic review or using that the between the therapeutic alliance and the outcomes of psychotherapy specifically for young adults. is for young adults to have psychotherapeutic needs based developmental with the to adulthood and recent and economic changes that have the and opportunities for young adults et al., 2010). In some countries, many of the developmental tasks that used to be have now been to the of the age additional on individuals who have high levels of psychological needs (Bound, Lovenheim, & Turner, 2010; Furstenberg, Rimbaut, & Settersten, 2005; Kessler et al., 2005; Schoeni & Ross, 2004). Using recent in this systematic review may be to important into the of the therapeutic alliance in psychological treatment of this important The of this review is to the literature and of between measures of the therapeutic alliance and outcomes of psychotherapy for 18 to 34 year This review studies of the therapeutic alliance and psychotherapy outcomes for young adults in psychotherapy with mental health studies at least one measure of the therapeutic alliance that at least one This is to time alliance one of that to be to support studies in which all alliance and measures were at the same in time be that to or the alliance in which the alliance is be the quality of the between therapist and client is not that can be at and time studies are studies be be young adults (ages that age groups be only they specifically for young adults as be no based on treatment or of illness as know of no evidence that indicates that of these factors the of This a of such as and substance We all to individuals in with mental health or be performed by mental health or those in psychotherapy by or mental health of formal mental health services as are by are mental health or of in leading to be to individuals in We studies in which therapy was performed using the of the therapist is in these We and group are more complex and to measure in these than in individual types of are the alliance may be central to of and of psychotherapy is some studies show that psychotherapy can be in as little as one & & The outcomes in this review are of disorders such as for example, by the et al., or the for for example, by the for example, by the of et al., or of specific outcomes or of drug and treatment or the client the of or et al., 2000). may multiple measures to outcomes in more than one of or We only measures for which there is evidence of and and this was in a to the one in the This in the of with no or as these additional or into This review has no Alliance measures are reported in the of Psychotherapy measures can be of of or We that most of the data in this review be in the of correlations between alliance and These are in of the and similar studies that for other in of between alliance and measures, for using or in may have been from indicated over 30 alliance measures have been used to measure the only studies that used alliance and measures that have some empirical evidence of and in the or a and for example, produced a of leading alliance measures and related theoretical frameworks. For each the authors evidence of or or with the of this and about the and of each such as the Mental similar for may studies that were in previous by to the of The process the of many the risk of & 2010). The multiple in previous and in are no be as The be and The Services of for as for other be as literature and and of studies to additional We authors of studies and in the to of studies that are to alliance studies be to studies not identified from We of alliance studies and reviews to identify additional are considered they be and to they the identify to be the studies in the most recent of the therapeutic alliance for adults (Horvath, and identified the most likely to on the therapeutic The results of the review indicate that of and Psychotherapy and Psychotherapy for the from In to the previous this review to We from The and for and in the US from National of Mental and and We literature for and and the for We with leading authors and in the of the therapeutic alliance and young adults through personal and for We alliance and young adult and for their additional that may have been in The alliance of Penn Horvath, Karver, of and Wampold, of of the young adult authors are Furstenberg, of and Settersten, These be used to a of about additional The studies in the therapeutic alliance literature are to research design, perspective of the the therapeutic alliance alliance and measures study that is and (2007) based on the age group types of by mental health alliance and measures with reported and of measures least one alliance measure and of reported data This study the between the therapeutic alliance and psychotherapy using patient outcomes from the of the of and Services in The authors data from clients by This only who had the Working Alliance to four and had at least four therapy were using a and In to a for alliance to this study to the of patient and therapist to The produced for this study was r = indicating that were with outcomes for study In the most recent of Horvath et al. (2011) a of to r for studies that had reported data in other We these based on additional from authors in to r and for some We to effect sizes in most These are likely to repeated measures and multiple alliance measures and is the that multiple clients have the same is that the same researchers multiple to types of in this estimate & in all in In addition, may identify multiple from individual studies and may to authors are to results of multiple be into one of and the more than the of et al., and other to the two and and We the full for that may the We that from the same study to the of studies multiple Two review authors full and a of data from to a with the researchers to the data to a The review authors not be to on or the review authors about a they to the with from a who has methodological The be a of both and studies with a for each In addition, a the of studies and et al., Two review authors and the study using data forms in the forms be on two studies and be the Two review authors the study data and and be by the or changes in data forms be to all studies that have been be in the data the review authors may the study authors for more this to a review on all items in and risk of on forms and forms have been to data from the in this is used for to the study such as outcomes, and study is used to such as alliance and measures, and study data such as and We data on the of each and reported data and data study quality evaluation of multiple forms of and of The forms are in identify data in a the authors of the study to to on therapy and from research be for each We assess and the for reporting as of evaluation of study quality of have studies and high levels of in some to assess the of studies with high levels of for between risk of reporting and be for of to alliance measures, measures, of and reporting analysis all outcomes are reported as well as all individuals in the are for in the The be used to the quality of evidence related to each of the outcomes Working & 2008). The for each are or We effect sizes to be correlated studies each study likely multiple of between alliance and The of these effect sizes is not likely to be reported in of the studies & a developed to effect sizes through the of & 2010). The based on empirical of the the for about the of the effect The of the are studies demonstrate that and in samples as as studies estimate effect or between studies estimate the et al., 2010). This multiple effect from the same study to be in a The is improvement other such as one effect or a effect for each which about and their a to to all of the types of outcomes the same analysis treatment rather than using separate to the variety of alliance and measures, to a effects for all that for both study and between study the estimate of the between all the of effect sizes is reported in as recommended by et al. a to by of at different of ranging from to there is little in the of at different of a of = for the estimate to be used in all recommended by and a when the effects and assume = for these In this each effect a study is a effects to the of the of in that study by that the of the study the between studies This that study not receive additional it multiple effect sizes. The in this review and The that be used are and et al., 2008). there is a risk of a using which r to the of results across studies be by for of be based on the and the & The of the be due to its We some and a analysis with & to examine factors with the of The effect of are the age younger older alliance measure Helping Alliance Alliance Working Alliance alliance time of alliance and measure Inventory, of treatment of of and of therapist less than three more than three of the is that the of multiple effect sizes study of effects of that studies as well as between Four of the of may and between of alliance and of of these and between two for each the of the each study used to estimate the effect of the and a of that to estimate the effect of the The four of treatment, of of and of are to only between and be as in the We examine and the correlations between all to assess among We types of with different of in the to examine this had on the estimate of the and examine the of data on the We studies with high risk of of specific types high risk of to analysis of the studies with risk of a different the et al., and analysis, and et al., 2008) to for the of and lead the of the the the and some of the review for and criteria, and data, the analysis and of data, and lead the of the review and the the and the review as it and methodological and as and out the many of the the in the that to the and the of the the on data on the and review and the data of the review of of analysis issues. review and the as it review for and and and review the is estimate that the take six the and data take two to three and analysis take three the for the review is one year in be for this review three years in with Collaboration and as This study is by the to by the of and of Alliance and measures measures in the in which they are in the that a measure can be by multiple and at multiple in time from specific and be samples This or is know samples This or is know measures in the in which they are in the that a measure can be by multiple and at multiple in time from specific and be samples This or is know samples This or is know data in the outcomes refers to the measures identified the for a review on this the with and methodological from and the and the of the developed the forms based on forms by and The authors have no in the outcomes of this to in a for to from the of and