Attachment-based family therapy for sexual and gender minority young adults and their nonaccepting parents.

This pilot open trial examined the efficacy of attachment-based family therapy (ABFT) for Israeli sexual and gender minority (SGM) young adults and their persistently nonaccepting parents. Thirty families received up to 26 weeks of treatment, with parental rejection, parental acceptance, and young adults' attachment avoidance and attachment anxiety assessed at baseline, 8, 16, 24, and 36 weeks (three months post-treatment). Analyses using multilevel growth models revealed that both young adults and their mothers independently reported increases in mothers' acceptance of their young adult's same-sex orientation or noncisgender identity. In addition, young adults reported decreases in both parents' levels of rejection. Also, mothers, but not fathers, reported decreases in their own level of rejection. Finally, young adults reported a decrease in attachment avoidance in their relationships with both mothers and fathers, but not a decrease in attachment anxiety. Importantly, these treatment gains were maintained three months after the end of treatment. Together, these results suggest that ABFT-SGM, a manualized, affirmative, experiential, family-based treatment, may be effective in reducing long-standing parental rejection, promoting parental acceptance, and improving the quality of LGBTQ+ young adults' relationships with their parents. These findings are encouraging in light of the urgent need for efficacious interventions to reduce family generated minority stress and promote safer, more supportive environments for sexual and gender minority people.

INTRODUCTION

While in many countries societal attitudes toward same‐sex oriented and gender minority (SGM) people have become increasingly more accepting (Luhur et al., 2019; Pew Research Center, 2020), many LGBTQ+ 1 young adults, including in Israel, still report that their parents are rejecting or not accepting of their identity, even years after disclosure (Grossman et al., 2005; Samarova et al., 2014; VanBergen & Love, 2022). Such parental rejection can manifest in a variety of ways, including hurtful homo(trans)phobic comments, invalidation (e.g., “you are just confused”), explicit disapproval (e.g., “being gay is immoral and prohibited by the Bible”), guilt induction (e.g., “You are ruining our family”), and coercion/control (“Don't bring your girlfriend home for the holidays”). Parental acceptance, on the other hand, is reflected in affirmative behaviors, such as when parents tell their child that they are proud of them, show interest in their personal life, use their chosen pronouns, express a desire to get to know their friends and partners, encourage them to be their authentic selves, come out to extended family and friends, and advocate for LGBTQ+ rights. Importantly, although parental rejection and acceptance are related constructs, they are distinct (Kibrik et al., 2019; Perrin et al., 2004; Roe, 2016). For example, parents may explicitly state that they accept their child's same‐sex orientation or gender identity, while at the same time refuse to meet their child's romantic partner, or ask their child to continue to conceal their sexual orientation or gender identity from others (Johnson et al., 2020; Livingston & Fourie, 2016; Martin, 2016).

Needless to say, parental rejection of one's sexual orientation or gender identity can take a psychological toll. Such messages from parents can be internalized and negatively affect one's sense of self (Carastathis et al., 2017). They can also undermine the attachment bond by generating attachment anxiety, and lead to avoidance in the relationship. This leaves the young adult feeling disconnected. They feel that it is not safe or helpful to turn to their parents for support when experiencing minority stress (e.g., discrimination, rejection, victimization) outside of the family (e.g., at work, in the community). Indeed, research has shown that parental criticism, invalidation, and rejection of their child's sexual orientation or gender identity is associated with a number of negative psychological and health outcomes. These include internalized homophobia, expectations for future gay‐related rejection by others (Pachankis et al., 2008), increased risk for depression and suicidal ideation (Haas et al., 2010; Ryan et al., 2009; VanBergen & Love, 2022), and higher levels of drug and alcohol consumption (D’Amico & Julien, 2012; Padilla et al., 2010; Rothman et al., 2012). In contrast, parental support of their child's sexual orientation or gender identity is associated with greater self‐esteem and perceived social support, and has been found to buffer against psychopathology (D’augelli, 2002; Eisenberg & Resnick, 2006; Elizur & Ziv, 2001; Evans et al., 2004; Feinstein et al., 2014; Floyd et al., 1999; Grossman et al., 2005; Haas et al., 2010; Hershberger & D’Augelli, 1995; Needham & Austin, 2010; Ryan et al., 2010; Savin‐Williams, 1989; Shilo & Savaya, 2011). Not surprisingly, LGBTQ+ young adults report that their parents’ support is important to them, and that they want their parents’ genuine acceptance and active involvement in their lives (Denes & Afifi, 2014; Roe, 2016). Given the negative impact of parental rejection, and the positive impact of parental acceptance, on sexual and gender minority individuals, the development and testing of family‐based interventions is warranted (Pachankis & Goldfried, 2004; Parker et al., 2018). Such interventions should be designed to increase parental acceptance, decrease parental rejection, and promote more open, accepting, meaningful, supportive, loving relationships between SGM young adults and their parents.

One such treatment is attachment‐based family therapy for sexual and gender minority young adults and their nonaccepting parents (ABFT‐SGM; Diamond & Boruchovitz‐Zamir, in press; Diamond, Boruchovitz‐Zamir, et al., 2019). ABFT‐SGM is a manualized, time‐limited (26 weeks), empirically informed, family‐based treatment. It is specifically designed to reduce parental rejection and increase parental acceptance of their young adult's sexual orientation or gender identity, and improve the quality of the attachment relationship. The treatment is adapted from ABFT for depressed adolescents (Diamond et al., 2014), and from ABFT for LGB depressed and suicidal adolescents (Diamond et al., 2012). It is LGBTQ+ affirmative, experiential, and rooted in structural family therapy (Minuchin, 1974), multidimensional family therapy (Liddle, 2009), and emotion‐focused therapy and theory (Greenberg, 2011).

ABFT‐SGM is comprised of five tasks, conducted in sequence. The first task, conducted during the initial session of treatment, is to form an initial bond with each family member and establish building closer, more open, mutually respectful relationships as the shared goal of treatment. The second task is to build a strong therapeutic alliance with the young adult. This is done in the context of individual sessions. This task involves helping the young adult process their emotions (e.g., shame, fear, hurt, assertive anger) associated with their parents’ rejection, and with the rupture in their relationship with parents. It also involves helping the young adult identify and articulate their unmet attachment and identity needs (e.g., the need to feel safe, understood, validated, prized, and connected). Finally, the young adult is prepared to effectively communicate such emotions and unmet needs directly to their parents in subsequent conjoint sessions. The third task is to build a strong therapeutic alliance with each parent. This is done in the context of sessions alone with parents. This task involves helping parents process their emotions (e.g., shame, fear, anger, grief) associated with their young adult's sexual orientation or gender identity. Parents are also helped to reflect upon how their nonacceptance has affected their young adult, and their relationship with them. Finally, parents are prepared to reach out and listen openly and empathically to their young adult's pain and unmet needs in subsequent conjoint sessions. The fourth task, the attachment task, is the purported primary change mechanism of the treatment. This task is conducted in the context of conjoint sessions and consists of corrective attachment episodes. During such episodes, the young adult effectively communicates to their parents, often for the first time, their experience of feeling rejected. The young adult is helped to convey their vulnerable emotions (e.g., loss, sadness, loneliness, shame, sense of worthlessness, fear), legitimate assertive anger, and unmet attachment and identity needs (e.g., the need to feel understood, validated, accepted, prized, loved, connected, respected, and protected). At the same time, parents are helped to respond in an attuned, empathic, nondefensive, validating, and caring manner. Such positive parental responses, which are new and qualitatively different from past negative responses, lead the young adult to feel safer, less anxious, better understood, validated, cared about, accepted, more trusting, and connected. Parents, for their part, feel less distressed, more agency, and closer to their young adult. Together, these processes lead to deeper, more open, honest, loving, meaningful, and mutually respectful relationships. Finally, after tension in the relationship has diminished, the fifth task of the treatment involves helping family members work collaboratively as they consider future shared challenges (e.g., the young adult's continued transitioning, coming out to extended family, bringing a partner home for the holidays).

Three randomized clinical trials conducted in the United States have shown ABFT to be efficacious in treating depressed and suicidal adolescents (Diamond et al., 2002, 2010; Diamond, Kobak et al., 2019). ABFT has also been adapted for, and successfully tested with, LGBTQ+ suicidal and depressed adolescents in the United States (Diamond et al., 2012; Russon et al., 2021) and Israeli young adults suffering from unresolved anger toward a parent (Diamond et al., 2016). Moreover, findings from a number of process research studies lend support for many of the model's purported change mechanisms, including emotional processing (Diamond et al., 2016; Lifshitz et al., 2020), increases in parental psychological autonomy granting behaviors (Shpigel et al., 2012) and corrective attachment episodes (Tsvieli et al., 2021). Prior to the current study, however, ABFT had not yet been tested with SGM young adults reporting persistent parental rejection or nonacceptance of their sexual or gender identity.

Such research is warranted for a number of reasons. First, SGM young adults’ relationships with their parents are qualitatively different from those of SGM adolescents. Young adults typically do not live at home full‐time, have often gained some degree of financial independence, and have greater extrafamilial social support. Consequently, they approach their relationships with their parents from a position of greater psychological and behavioral autonomy. Second, young adults have typically identified as a sexual or gender minority person longer than adolescents (Calzo et al., 2011), making parents' questions regarding the validity and changeability of their young adult's identity less salient. Third, prior studies of ABFT for LGBTQ+ adolescents targeted depressed and suicidal adolescents rather than families with persistently rejecting or nonaccepting parents. Treating persistently rejecting or nonaccepting parents presents unique therapeutic challenges. For example, persistently rejecting or non‐accepting parents typically begin treatment with high levels of global distress, shame, and fear. Such maladaptive emotions are thought to fuel parents’ nonacceptance. Consequently, in ABFT‐SGM a substantial portion of sessions alone with parents (i.e., Task III) is devoted to helping parents process their shame and fear; access and work through their adaptive grief; and access their legitimate, adaptive assertive anger toward societal homo(trans)phobic messages. Another unique feature of ABFT‐SGM is the use of exposure techniques to combat parents’ maladaptive avoidant behaviors. This often involves preparing and supporting parents as they come out to friends or members of their extended family, or plan to meet their young adult's same‐sex partner. Fourth, in prior studies of ABFT for depressed and suicidal adolescents, the primary outcomes measured were adolescents’ depressive and suicidal symptoms. In contrast, the primary outcomes measured in ABFT‐SGM are relational: parental rejection and acceptance, and the quality of the young adult–parent attachment relationship. Finally, this particular study extends ABFT research to an Israeli sample. This provides an opportunity to pilot test the model in a different culture. On the one hand, Israel is a developed, industrialized, Western culture that emphasizes individualistic values. At the same time, Israelis also place a premium on family belonging, resulting in more frequent and intense contact between parents and their young adults (Lavee & Katz, 2003; Mayseless & Scharf, 2003). Therefore, Israeli young adults face the challenge of negotiating their increased autonomy within the context of ongoing relationships with their parents and other family members (Shulman et al., 2005).

This pilot open trial examined the efficacy of ABFT‐SGM for Israeli sexual and gender minority young adults and their persistently rejecting and/or nonaccepting parents. While a number of authors have published clinically rich, compelling guidelines for conducting family therapy with LGBTQ+ individuals and their parents (LaSala, 2010; Ryan, 2010; Stone Fish & Harvey, 2005), this is the first empirical study examining the efficacy of a manualized family‐based treatment for this population. We hypothesized that, over the course of 26 weeks of ABFT‐SGM, both young adults and their parents would independently report decreases in parental rejection, and increases in parental acceptance, of their young adult's sexual orientation or gender identity. At the same time, we hypothesized that young adults would report decreases in attachment anxiety and avoidance in their relationships with their parents. Based on previous research on ABFT and other family‐based treatments (Diamond et al., 2021; Friedlander et al., 2021; Liddle et al., 2001), we expected these gains to be maintained when measured three months post‐treatment.

METHOD

Participants

Clients

Thirty Israeli young adults self‐identifying as LGBTQ+, and their parents, received up to 26 weeks of ABFT‐SGM. In order to be included in the study, the young adult needed to: (a) have disclosed their identity to their parent(s) at least 1 year prior; (b) report feeling that their relationship with at least one of their parents was significantly negatively impacted due to their parent's persistent rejection/nonacceptance of their sexual orientation or gender identity; (c) endorse an above average score on perceived parental rejection and/or a below average score on perceived parental acceptance, as measured by the Parental Acceptance and Rejection of Sexual Orientation Scale (PARSOS; Kibrik et al., 2019), and/or above average levels of attachment avoidance or anxiety in their relationship with their nonaccepting parent, as measured by the parent version of the Experience in Close Relationships—Relationship Structures Questionnaire (ECR‐RS; Fraley et al., 2011); and (d) not be currently living with their parents.

The participating young adults ranged from 20 to 36 years of age (M = 25.98, SD = 4.09). Eighteen self‐identified as male and gay, six as female and lesbian, two as female and bisexual, one as a transgender man, one as a transgender woman, one as nonbinary, and one as genderqueer. In two of the families, parents identified as ultra‐orthodox Jews, in six other families, parents identified as orthodox Jews, and in the remaining 22 families, parents identified as secular. In 24 of the cases, both parents participated in the therapy and, in six cases, only the mother participated. On average, the parents in this study had known about their young adult's minority identity for 4.80 years (SD = 3.65).

Therapists

Treatment was delivered by eight therapists. Two of the therapists, including the first author, were licensed clinical psychologists. One was a licensed clinical social worker, and the remaining five were advanced graduate clinical psychology or social work interns. All of the therapists had previous training in family therapy. One of the therapists identified as a gay cisgender male, another as a cisgender lesbian, one as a cisgender heterosexual male, and the remaining five as cisgender heterosexual women.

Measures

Sexual orientation‐specific and gender identity‐specific parental rejection and acceptance were measured using the self‐report Parental Acceptance and Rejection of Sexual Orientation Scale (PARSOS; Kibrik et al., 2019). The PARSOS includes two subscales: parental rejection (13 items) and parental acceptance (15 items). The father and mother versions include identical items, simply worded differently to refer to the parent at hand. Examples of items from the parental rejection subscale include, “My father thinks that my sexual orientation is a wrong choice that I made” and “My mother makes me feel like my sexual orientation has caused trouble for her and for our family.” Examples from the parental acceptance subscale include: “My father shows an interest in really getting to know the people that I am dating/my partner” and “My mother proudly identifies herself as being the mother of a lesbian or bisexual daughter/gay or bisexual son.” For the purposes of this study, items were reworded for gender minority clients. More specifically the words “sexual orientation” were changed to “gender identity,” and the words “lesbian, gay, bisexual” were changed to “transgender” and “gender minority.”

Data from previous research support the reliability, factor structure, and validity of the measure (Kibrik et al., 2019). In this study, Cronbach's alpha reliability estimates were as follows: 0.94 for young adults’ report of paternal rejection, 0.90 for young adults’ report of paternal acceptance, 0.94 for young adults’ report of maternal rejection, 0.92 for young adults’ report of maternal acceptance, 0.87 for mother's report of their own rejection, 0.92 for mothers’ report of their own acceptance, 0.92 for fathers’ report of their own rejection, and 0.93 for father's report of their own acceptance.

Attachment anxiety and avoidance were measured using the Relationship Structures Questionnaire (ECR‐RS). The ECR‐RS is a revised version of the Experiences in Close Relationships Scale (Brennan et al., 1998; Fraley et al., 2000) and contains nine short statements rated on a 5‐point scale, ranging from 1 (strongly disagree) to 5 (strongly agree). Examples of items on the attachment avoidance scale include, “It helps to turn to this person in times of need (reverse scored)” and “I find it easy to depend on this person (reverse scored).” Examples of items on the attachment anxiety scale include “I’m afraid that this person may abandon me” and “I often worry that this person doesn't really care for me.” Young adults completed the questionnaire in relation to each participating parent. In previous studies, the measure has been shown to have good reliability and validity, across different types of relationships (Fraley et al., 2011). In this study, the Cronbach's alpha reliability estimate for the attachment anxiety scale was 0.90 and for the attachment avoidance scale was 0.96.

Procedure

Recruitment

Participants were recruited through advertisements on social media and relevant listservs. Thirty‐nine individuals inquired about the study. During the initial phone call, they were told about the treatment model and study conditions. After the initial phone call, five chose not to participate because they preferred individual therapy; two reported that their parents did not want to participate; and two did not contact us again.

Screening

The remaining 30 families were invited to our clinic for a more comprehensive screening, during which they underwent an abbreviated version of the Mini‐International Neuropsychiatric Interview (MINI; Lecrubier et al., 1997) to rule out psychotic disorders, bipolar disorder, and severe suicidal ideation. They also completed the PARSOS and ECR‐RS to ensure that they met inclusion criteria. None of the 30 families screened were excluded from the study. At the end of the screening procedure, family members completed the necessary consent forms and their first session of therapy was scheduled.

Treatment

Families were offered up to 26 weeks of therapy. Typically, sessions were conducted once a week, with the composition of the session depending on the task at hand. More specifically, Task I, IV, and V sessions included both the young adult and their parents, Task II sessions included the young adult only, and Task III sessions included parents only. In certain cases, for clinical reasons, more than one session was conducted in a given week. Only one family dropped out of therapy before reaching Task IV, attending only seven sessions. The remaining 29 cases completed a full dose of therapy, with their number of sessions attended ranging from 11 to 31 (M = 21.23, SD = 5.46).

Measurement procedure

Young adults completed the PARSOS and ECR‐RS at pretreatment, 8 weeks (typically just before the attachment task), 16 weeks (typically during or just after the attachment task), 26 weeks (end of treatment), and 38 weeks (typically 3 months post‐treatment). Parents also completed the PARSOS at those same time‐points. Measures were sent directly to family members’ email and were completed online, using Qualtrics. The data collection was overseen by an independent research assistant, naïve to the purpose and hypotheses of the study.

Therapist training and supervision

All of the therapists were trained in ABFT‐SGM by the first author—the primary developer of ABFT‐SGM. Training took place over the course of six months. During the training, therapists read the ABFT‐SGM manual (Diamond & Boruchovitz‐Zamir, in press; Diamond, Boruchovitz‐Zamir, et al., 2019), and clinical papers on working with SGM young adults and their persistently nonaccepting parents (Diamond & Shpigel, 2014). They also watched training tapes exemplifying how to complete each treatment task. They then joined the first author as a co‐therapist for one of the study cases, before treating cases independently. Throughout the entire course of the study, therapists participated in a weekly supervision group, led by the first author, and all sessions were videotaped, reviewed and discussed.

Therapist adherence

In order to assess therapists’ level of adherence to ABFT‐SGM, a group of eight independent raters, naïve to the purpose and hypotheses of the study, were trained to use the Therapist Behavior Rating Scale 3–Revised (TBRS‐3R; Diamond et al., 1998). The TBRS‐3R is an observational measure designed for rating the extent to which therapists employ a range of interventions over the course of a therapy session. For the purposes of this study, raters were trained to rate the five prescribed ABFT‐SGM interventions reflecting the core features of the treatment model. These interventions included focusing on vulnerable emotions, relational reframes, preparing for corrective attachment episodes, conducting corrective attachment episodes, and addressing the young adult's sexual orientation or gender identity.

Adherence raters and training

Raters were undergraduate psychology students. They received four weeks of intensive training during which they were taught to identify each of the five interventions. They were then shown paradigmatic examples of each level of extensiveness for each intervention. Then, as a group, they practiced rating taped ABFT‐SGM sessions not included in the study sample. Once they reached adequate reliability (ICC[2,8] > 0.70), they were assigned actual study tapes to rate.

Tape selection and adherence coding procedure

Three taped sessions were drawn from each case (n = 90). Tapes were selected in a block randomized manner. More specifically, for each case, one session was drawn from the first task of therapy, another was randomly drawn from either the second or third task of treatment, and the third was randomly drawn from the fourth or fifth task of treatment. Tapes were then randomly assigned to raters. Raters were instructed to watch the entire tape of the session and then score the extent to which the therapist employed each of the five interventions over the course of the session. Scores were designated on a 7‐point Likert scale with the following anchors: 1 = not at all, 3 = somewhat, 5 = considerably, and 7 = extensively. Both thoroughness and frequency were considered in making each rating. The TBRS has been found to be reliable in a number of previous studies (Diamond et al., 2007; Hogue et al., 2008). In order to estimate reliability in this study, 30 of the sessions were coded by two raters. Pairs of coders for these 30 sessions were assigned on a rotating random basis and interrater reliability was calculated. The resulting ICC[2,1] = 0.956 (95% confident interval 0.939–0.968) indicated excellent inter‐rater reliability. Moreover, the mean adherence rating per session was 5.15 (SD = 2.60), indicating a high level of therapist adherence.

RESULTS

Data analytic strategy

The dataset had a hierarchical structure, with assessments nested within families. Therefore, we used a multilevel growth model (with assessments at level‐1 and families at level‐2). Using the maximum‐likelihood estimator, multilevel growth models can accommodate for nested unbalanced data resulting from missing values (Hoffman, 2015). When we attempted to estimate level‐3 models (i.e., taking into account therapist effects), the models did not converge. This is likely due to the low number of clients treated by each therapist in our sample, limiting the extent to which we could examine therapist effects. Thus, we retained the level‐2 model.

We first ran a set of multilevel growth models comparing the fit of a linear growth model versus a quadratic growth model for each outcome measure. We predicted that parental acceptance would increase, and that parental rejection, attachment avoidance, and attachment anxiety would decrease linearly from pretreatment to follow‐up based on findings from previous studies of family therapy showing continued improvement in symptoms and family functioning through follow‐up assessments (Butler et al., 2011; Liddle et al., 2001).

Deviance tests indicated that only for young adults' reports of their attachment anxiety with their fathers did the quadratic growth model show a better fit than the linear growth model (χ[2] = 6.66, p = 0.002). Therefore, for all other outcomes we estimated a linear growth model, with week number as the Time effect. The Time effect was recoded into a 0–1 scale (i.e., 0 = week #0 assessment; 0.211 = week #8 assessment; 0.421 = week #16 assessment; 0.684 = week #26 assessment; and 1 = week #38 assessment). This allowed us to interpret the Time effect as representing the average linear change over the entire assessment period, while retaining the ratio between the five assessment points.

The following generic linear growth model was estimated for each outcome: where the outcome reported at assessment t by participant i was predicted by the following parameters: the sample's average (i.e., fixed) intercept (i.e., γ00); the sample's average effect of Time (γ10); the deviations of the particular participant from the average intercepts and Time effects (i.e., the random effects for the intercept u0i and the slope u1 ), and the level‐1 residual terms (i.e., random effects at level 1; e ). A first‐order autoregressive structure was imposed on the level‐1 residual covariance matrix to account for the autoregression between adjacent assessments. Models were estimated using the R nlme package (Pinheiro et al., 2021). Effect sizes were estimated following the procedure outlined in Arntz et al. (2013); see also Shahar et al., (2017). Specifically, Cohen's d was calculated as the effect size of the average change (i.e., the fixed effect of Time) divided by the SD of the residual outcome variable (i.e., the squared root of the variance of the level‐2 random intercept + the variance of the level‐1 residual/number of measurements).

In addition, to reduce concerns regarding Type 1 error, we ran a secondary set of analyses: two additional multivariate models (Baldwin et al., 2014). In the first model, change in the eight young adult‐reported outcomes (i.e., parental acceptance, parental rejection, attachment anxiety, and attachment avoidance regarding their fathers and mothers) were modeled simultaneously. This allowed us to account for the covariation in the outcomes’ level‐1 and level‐2 residuals. In the second multivariate model, change in the four parent outcomes (i.e., fathers’ and mothers’ report of their acceptance and rejection) were modeled simultaneously. Note that given the complexity of these multivariate models, we were able to include only random intercepts. When random slopes were included in the models they did not converge.

Descriptive results

Table 1 presents the descriptive statistics of the study's variables. Table 2 presents the correlation matrix of the study's variables. As Table 2 shows, within reporter (i.e., young adults, fathers or mothers), levels of acceptance and rejection were strongly associated with each other. In addition, young adults tended to perceive their mothers and fathers similarly in terms of their acceptance and rejection. Finally, parents’ and their young adults’ reports of parents’ acceptance and rejection converged (high positive correlation).

TABLE 1

Descriptive statistics for all dependent variables

	Young Adults’ reports						Parents’ reports
	Fathers			Mothers			Fathers			Mothers
	N	M	SD	N	M	SD	N	M	SD	N	M	SD
Assessment #1
Acceptance	25	1.99	0.94	29	2.28	0.97	23	2.42	1.07	28	2.81	1.06
Rejection	25	2.95	1.24	29	2.98	1.12	23	2.70	1.07	28	2.32	0.88
Anxiety	24	2.00	1.05	28	1.95	1.36
Avoidance	24	4.72	1.58	28	4.47	1.69
Assessment #2
Acceptance	21	2.21	0.87	26	2.6	0.91	20	2.61	1.14	26	3.12	1.03
Rejection	21	2.88	1.10	26	2.73	1.04	20	2.39	0.96	26	2.07	0.68
Anxiety	21	2.55	1.73	26	2.29	1.48
Avoidance	21	4.75	1.59	26	4.37	1.73
Assessment #3
Acceptance	22	2.37	1.05	27	2.78	1.1	20	2.68	1.11	27	3.09	1.01
Rejection	22	2.81	1.21	27	2.48	1.05	20	2.54	0.99	27	2.02	0.66
Anxiety	23	2.91	2.09	27	2.25	1.71
Avoidance	23	4.70	1.66	27	4.18	1.64
Assessment #4
Acceptance	21	2.61	1.22	27	2.99	1.04	22	2.90	1.14	27	3.41	1.04
Rejection	21	2.69	1.26	27	2.44	1	22	2.57	1.03	27	1.90	0.69
Anxiety	21	2.27	1.64	26	1.85	1.4
Avoidance	21	4.09	1.40	26	3.9	1.64
Assessment #5
Acceptance	20	2.82	1.14	24	3.22	0.97	20	2.93	1.01	24	3.53	0.99
Rejection	19	2.33	1.12	24	2.23	1.07	20	2.34	0.92	24	1.75	0.57
Anxiety	19	1.77	1.19	25	1.59	0.96
Avoidance	19	3.72	1.46	25	3.6	1.49

TABLE 2

Correlation matrix of all dependent variables

	Young Adults’ reports								Parents’ reports
	Fathers				Mothers				Fathers		Mothers
	1 Acc.	2 Rej.	3 Anx.	4 Avd.	5 Acc.	6 Rej.	7 Anx.	8 Avd.	9 Acc.	10 Rej.	11 Acc.	12 Rej.
1.		−0.699	−0.359	−0.449	0.655	−0.517	−0.243	−0.345	0.747	−0.512	0.473	−0.417
2.	<0.001		0.485	0.246	−0.494	0.629	0.144	0.212	−0.722	0.750	−0.385	0.404
3.	<0.001	<0.001		0.340	−0.112	0.202	0.448	0.188	−0.410	0.380	−0.144	0.379
4.	<0.001	0.011	<0.001		−0.163	0.041	0.254	0.481	−0.174	0.167	−0.078	0.308
5.	<0.001	<0.001	0.251	0.092		−0.690	−0.372	−0.333	0.596	−0.297	0.785	−0.546
6.	<0.001	<0.001	0.036	0.674	<0.001		0.164	0.139	−0.546	0.470	−0.586	0.538
7.	0.012	0.142	<0.001	0.008	<0.001	0.062		0.444	−0.154	−0.071	−0.317	0.292
8.	<0.001	0.029	0.051	<0.001	<0.001	0.115	<0.001		−0.169	−0.007	−0.298	0.197
9.	<0.001	<0.001	<0.001	0.094	<0.001	<0.001	0.137	0.103		−0.656	0.662	−0.366
10.	<0.001	<0.001	<0.001	0.108	0.003	<0.001	0.496	0.946	<0.001		−0.309	0.412
11.	<0.001	<0.001	0.155	0.444	<0.001	<0.001	<0.001	0.001	<0.001	0.002		−0.598
12.	<0.001	<0.001	<0.001	0.002	<0.001	<0.001	0.001	0.031	<0.001	<0.001	<0.001

Correlation coefficients are shown above the diagonal; p values are shown below the diagonal.

Abbreviations: Acc., Acceptance; Anx., Attachment anxiety; Avd., Attachment avoidance; Rej., Rejection.

Primary results

Table 3 presents the results of the multilevel growth models.

TABLE 3

Results of multi‐level growth models

	Mothers			Fathers
	Est. (SE)	p	d	Est. (SE)	p	d
Young adults’ reports
Acceptance
Intercept	2.333 (0.169)	<0.001		1.992 (0.157)	<0.001
Time	0.864 (0.159)	<0.001	1.01	0.762 (0.151)	<0.001	1.02
Rejection
Intercept	2.937 (0.186)	<0.001		2.992 (0.223)	<0.001
Time	−0.652 (0.148)	<0.001	−0.72	−0.568 (0.169)	0.001	−0.53
Anxiety a
Intercept	2.213 (0.235)	<0.001		2.137 (0.303)	<0.001
Time	−0.452 (0.246)	0.069	−0.41	2.429 (0.918)	0.010
Time2				−2.705 (0.875)	0.003	−0.22
Avoidance
Intercept	4.574 (0.298)	<0.001		4.889 (0.295)	<0.001
Time	−0.811 (0.206)	<0.001	−0.53	−0.786 (0.259)	0.003	−0.61
Parents’ reports
Acceptance
Intercept	2.859 (0.190)	<0.001		2.470 (0.218)	<0.001
Time	0.618 (0.129)	<0.001	0.63	0.322 (0.197)	0.106	0.69
Rejection
Intercept	2.282 (0.146)	<0.001		2.645 (0.215)	<0.001
Time	−0.518 (0.106)	<0.001	−0.68	−0.116 (0.119)	0.330	−0.12

For mothers, the linear model including random time effects did not converge. Therefore, a simplified version including only random effects for the intercept terms was estimated. For fathers, we estimated a quadratic growth model, which fitted the data better than a linear growth model.

Young adults’ reports

As Table 3 shows (see also Figure 1), young adults reported increases in both maternal and paternal acceptance. In addition, they reported decreases in both maternal and paternal rejection and attachment avoidance. Note that the effect size for Time was large for acceptance, and medium for rejection and avoidance. In contrast to our prediction, no significant change was found in attachment anxiety regarding mothers. For fathers, we found a significant quadratic effect. As Figure 1 shows, attachment anxiety levels showed an increase toward the middle assessment point (week #16), and then declined toward the final assessment point (week #38).

FIGURE 1

Young adults’ reports regarding their fathers’ (gray bars) and mothers’ (black bars) acceptance and rejection, and regarding their own attachment anxiety and attachment avoidance. Dots represent means; bars represent 95% confidence interval

The results of the multivariate growth model replicated the pattern of results obtained with the univariate models (see Appendix A1). Notably, the multivariate model allowed us to run a set of post hoc analyses in which we contrasted the estimates pertaining to mothers’ and fathers’ variables. These analyses revealed no significant differences in mothers’ and fathers’ baseline levels (i.e., intercept) or change (Time slope; p’s > 0.387).

Parents’ reports

As evident in Table 3 (see also Figure 2), mothers reported an increase in their acceptance and a decrease in their rejection over time. The effect sizes were medium–large. In contrast to our prediction, no such changes were observed for fathers. The results of the multivariate growth model mostly replicated the pattern of results obtained with the univariate models (see Appendix A1). However, although the results from the univariate growth model showed that fathers’ self‐reported acceptance did not increase, the results from the multivariate model did show a significant increase in fathers’ self‐reported acceptance. Notably, this model allowed us to run a set of post hoc analyses in which we contrasted the estimates pertaining to mothers’ and fathers’ variables. The only significant difference was in regard to the Time effect for mothers’ and fathers’ rejection (Est. = −0.398, SE = 0.133, p = .011).

FIGURE 2

Fathers’ (gray bars) and mothers’ (black bars) reports regarding their acceptance and rejection of their young adults. Dots represent means; bars represent 95% confidence interval

DISCUSSION

High levels of parental rejection, and low levels of parental acceptance, have been shown to negatively impact the welfare of sexual and gender minority individuals, and undermine their relationships with their parents. This pilot open trial was the first to examine the efficacy of a manualized family‐based treatment specifically designed to reduce parental rejection, increase parental acceptance, and improve the quality of the young adult–parent attachment bond in a sample of Israeli sexual and gender minority young adults and their persistently rejecting or nonaccepting parents. Results showed that both young adults and their mothers independently reported increases in mothers’ acceptance and decreases in mothers’ rejection of the young adult's same‐sex orientation or noncisgender identity. Young adults, but not their fathers, also reported increases in fathers’ acceptance and decreases in fathers’ rejection. Finally, young adults reported a decrease in attachment avoidance in their relationships with both mothers and fathers, but not a decrease in attachment anxiety. Importantly, all of these treatment gains continued to increase when measured three months after the end of treatment. Together, these results suggest that ABFT‐SGM, a manualized, affirmative, experiential, family‐based treatment, may be effective in reducing long‐standing parental rejection, promoting parental acceptance, and improving the quality of LGBTQ+ young adults’ relationships with their parents. These findings are encouraging in light of the urgent need for efficacious interventions to reduce family‐generated minority stress and promote safer, more supportive environments for LGBTQ+ people (Parker et al., 2018).

It is worth noting that the parents in our sample were rejecting or nonaccepting, on average, approximately five years after their child's disclosure. They differed from most parents, who tend to become more accepting over time even without therapy (Beals & Peplau, 2006; Grossman et al., 2005; Samarova et al., 2014; Savin‐Williams & Ream, 2003). The fact that ABFT‐SGM was efficacious with parents’ who were rejecting or nonaccepting even after a long period of time underscores the potency of the treatment model and provides hope for families who feel stuck, are suffering, and who are searching for a way to reconnect in a mutually respectful, loving manner.

The finding that young adults reported significant decreases in attachment avoidance is of particular import. The ultimate goal of ABFT‐SGM is to create safer, more open, honest relationships, in which the young adult can communicate directly with their parents and express both vulnerability (e.g., fear, hurt, the need to be seen, loved, supported, etc.) and adaptive assertive anger when their boundaries have been violated. The decrease in attachment avoidance reported by the young adults in this study suggests that they may have felt increasingly able to open up to their parents, show them how they feel deep down, share their problems and concerns, and depend on their parents. When young adults are able to utilize their parents for support, comfort, and guidance, they feel heard, validated, more connected and less alone, buffering them against the insidious effects of minority stress.

In contrast to our expectations, there were no reductions in young adults’ attachment anxiety over the course of the treatment. This may have been, at least in part, due to young adults’ relatively low levels of attachment anxiety at the start of treatment. Indeed, young adults’ baseline scores on attachment anxiety were less than half the size in magnitude than were their baseline scores on attachment avoidance. This finding is consistent with our clinical impression that most of the young adults who presented for treatment in this trial generally avoided speaking openly about their sexual orientation or gender identity, and emotions such as fear and hurt.

Interestingly, while there was no linear decrease in young adults’ attachment anxiety, there was a quadratic pattern. More specifically, young adults reported an increase in attachment anxiety in their relationships with both parents from the start of therapy through week 16 of treatment (typically during or just after the attachment task), and then a decrease from week 16 through the end of treatment and follow‐up. This quadratic effect was significant for young adults’ report of their attachment anxiety viz‐a‐viz their fathers. Although we did not predict this quadratic effect, in retrospect such a pattern is consistent with ABFT theory and practice. In ABFT, young adults’ often become anxious in preparation for, and during, the attachment task. The attachment task requires the young adult to be more vulnerable and open in the presence of their parents and, thus, exposes them to the risk of further parental rejection. Such anxiety then typically decreases after ruptures have been resolved, acceptance has increased, and trust and safety in the relationship have been established or restored.

The validity of our findings is bolstered by a number of methodological strengths of the study. First, ABFT‐SGM is a manualized therapy, therapists were intensively trained, and all sessions were live supervised or supervised based on videtape recordings, ensuring treatment adherence and a high level of internal validity. Indeed, the results from our analysis of our observer‐rated adherence data indicated that the treatment was delivered in the manner prescribed. Second, participants in this study reflected a range of sexual orientations and gender identities, socioeconomic backgrounds, and degrees of religiosity, increasing the external validity of the findings. Third, the use of five measurement points allowed us to examine the trajectories of parental rejection, acceptance, and attachment avoidance and anxiety over time, including up to three months post‐treatment. Finally, the use of multiple reporters (i.e., young adults and parents) provided convergent validity for our findings regarding parental rejection and acceptance. It is also worth noting that this is the first study to use the PARSOS to measure parental acceptance and rejection from both the young adults’ and their parents’ perspectives. The high intercorrelations between the young adults’ and parents’ ratings provides additional evidence for the validity of the measure.

With that said, our results should be interpreted within the context of a number of methodological limitations of the study. First and foremost, there was no control group. Future studies should include a comparison condition to control for common factors. Second, because of the sample size, there were not enough participants to examine whether the treatment was differentially efficacious for same‐sex oriented versus transgender/nonbinary/genderqueer clients. This is worth noting since the literature suggests that parents have more difficulty accepting gender minority children than same‐sex oriented children (Abreu et al., 2019; Johnson et al., 2020). Also, there are no data yet regarding the reliability or predictive ability of the PARSOS specifically for gender minority individuals. The original PARSOS was developed for use with an LGB sample.

While a number of recent studies have successfully tested affirmative individual cognitive behavioral therapy to target sexual minorities’ mental and physical health (Pachankis et al., 2015, 2020), this is the first study to empirically examine the efficacy of a family‐based treatment specifically designed to reduce parental rejection, increase parental acceptance, and improve relationships between sexual and gender minority young adults and their nonaccepting parents. In light of the documented link between parental rejection, parental acceptance, and the welfare of LGBTQ+ individuals, and in light of the lack of family focused intervention research for this population, our findings represent an encouraging first step in filling this critical research–practice gap. More research examining the effectiveness of ABFT‐SGM with larger and more diverse samples, including other nationalities and ethnic and cultural groups, is warranted.

Table S1

Click here for additional data file.