Supporting patients by family education in psychotic illness: a longitudinal pre–post study protocol

Abstract

Introduction

A lack of education, resources and support for family carers of young adults with psychotic illnesses leaves them ill-equipped to support their loved one. By equipping families with skills and knowledge, public healthcare harnesses a powerful ally to support community stabilisation.

Aims

The primary goal is to study the effect of a psychoeducation intervention for family carers supporting young adults with psychosis on family burden and stabilisation of service users.

Methods and analysis

A longitudinal quantitative study with a pre–post design will be used to assess the long-term effectiveness of the psychoeducation intervention for family carers supporting a young adult with psychosis. 111 family carers will be recruited for the intervention and measures will be taken from family carer participants and their matched young adult service users. Nine evidence and family peer-informed and expert-reviewed psychoeducation modules are administered in 2-hour sessions over 9 weeks to family carers. Functional index measures are administered preintervention, and at 6-month, 12-month and 24-month follow-up. Service utilisation will be measured during a 12-month period preintervention, a 12-month period postintervention and during a 12–24-month period post-intervention.

Ethics and dissemination

The study has been reviewed and approved by the University of Alberta Research Ethics Board (Pro00110691). This novel methodological approach to studying family psychoeducation interventions addresses unique methodological challenges and limitations and will be disseminated through peer-reviewed publications and academic and medical conferences.

Trial registration number

NCT05500001; National Institutes of Health U.S. National Library of Medicine ClinicalTrials.gov

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Describes feasible and ecologically valid protocol for testing family outcomes in early intervention in psychosis.

• Directly recruits carers of young adults with psychosis to receive psychoeducational intervention without necessitating service user consent, bipassing barriers like paranoia and anosognosia.

• No randomisation and experimental control precludes drawing causative conclusions.

• Collects carer and service user data for up to 24 months postintervention.

• Small sample size of service users enrolled diminishes power to detect young adult outcome.

Background

A lack of education, resources and support for family carers of young adults with psychotic illnesses leaves them ill-equipped to support their loved one. Often family carers are excluded from appropriate involvement in safe hospital discharge planning or are blinded to information that is integral for effective and safe support of their loved one in the community despite best practice documents that champion family involvement.^1–3 The implementation of these guidelines seems to be challenged^{4 5} and many family carers continue to report experiences of exclusion and alienation.

Despite these hardships, family carers work with great tenacity towards improving their loved one’s stabilisation in the community and thus are predictably profoundly impacted by the disease themselves. The very nature of psychotic illness (especially symptoms of paranoia and anosognosia) can interfere with relational health and the natural maturation of family relationships over time.⁶ The profound emotional response experienced by families facing this challenge has been compared with a complex grief reaction.⁷ Specifically, first-degree family members experience post-traumatic stress and depression symptoms, poorer physical health, and social isolation.⁸

Through a positive lens, however, family carers remain a relatively untapped mental health resource for young adults with psychotic spectrum illnesses. By providing education and skill-based training to families, community stabilisation is fortified, transforming the mental health landscape by reducing the costly revolving door effect prevalent in mental health systems, and improving mental health trajectory.^{9 10} While family carer educational interventions do exist,^11–13 they are not widely integrated into early psychosis programmes⁴ nor necessarily endorsed at a public level.¹¹ Existing interventions focus on developing a supportive community and sharing experience-based knowledge but do not provide rigorous, evidence-based psychoeducation and skills delivered by expert facilitators. Furthermore, while it is accepted that family support is a critical factor affecting service user access to and benefit from treatment programmes,^{9 10} family is not explicitly integrated into care. In short, existing programmes lack synergy between the expertise of those with lived experience and that of mental health professionals. A synergistic and resilient foundation for recovery can be created by forming a triad of care between the young adult affected by psychosis, their family, and their care team.

While considerable literature exists showing efficacy of interventions delivered to those diagnosed with psychotic illness,^14–16 there is scant but emerging evidence regarding which programmes help family carers.¹⁷ Acceptance and Commitment Therapy (ACT¹⁸) is a rapidly growing approach that focuses on value-based living and fostering psychological flexibility, a perspective well suited to the unpredictable circumstances involved in supporting a young adult with psychotic illness. Currently two randomised control trials exist exploring the effects of ACT on supportive carers, one specifically directed to carers of individuals with psychosis.¹⁹ Both studies not only show promising preliminary results but also highlight methodological limitations. Another experimental study that delivered a combined psychoeducation and ACT programme to carers of loved ones with first episode psychosis found small and medium effects on reduction in perceived overload and burnout and improved active coping strategies.²⁰ One well-established technique for carers supporting individuals with psychosis is the Listen-Empathize-Agree-Partner (LEAP) Model of Communication,⁶ which focuses on adaptive ways of interacting with individuals experiencing psychosis.^{6 21} The LEAP approach improves communication and alliance, encouraging a strong relational bond to help the individual with psychotic illness accept and maintain adherence to medical treatment especially in the context of anosognosia.

There is a paucity of high-quality evidence showing long-term efficacy for interventions directed to carers. For example, Day and colleagues provided a family intervention that was primarily knowledge based and support based.¹¹ They used a pre–post design to assess mental health knowledge but did not assess overall family function, nor long term effectiveness of the group intervention and did not look at functioning of the service user.¹¹ There are several methodological challenges that have made rigorous research in this area particularly difficult.^{4 11 19} Jolley and colleagues highlighted difficulties in recruitment and suggested programme dissemination be made through presentations to interested groups.¹⁹ They also identified logistic difficulties in randomising group assignment.¹⁹ Day and colleagues identified that stigma experienced by families and their need to juggle multiple responsibilities were barriers to their accessing support.^{4 11} Additionally, considerable resources are required to facilitate family psychoeducation groups while conducting high-quality, longitudinal outcome research.

In our completed phase 1 pilot project with 14 family carers, we identified challenges related to care team engagement, family readiness to participate and preservation of the family bond. First, the care team plays a vital role in disseminating the study information to family carers and their matched young adult. The care team, however, is often so over-tasked that they are challenged to take on additional duties that are outside of typical care responsibilities. Second, family carer readiness to participate and capacity for skill acquisition will vary according to family circumstances and where they find themselves in the parallel recovery process. Given that service user symptomatology is expected to fluctuate greatly during early stabilisation attempts around disease onset, we anticipate non-linear and variable readiness of family for intervention. Finally, the relationships among family members at the illness onset and throughout recovery are vulnerable, and considerable care is needed to preserve family bonds. It is important for researchers to have sensitivity to families’ specific situations and flexibility in allocation of resources to augment recruitment. Despite these challenges, there was consistent encouraging feedback from service providers as well as the pilot participants regarding the intervention (see online supplemental appendix 1).

The aim of this longitudinal pre–post quantitative study is to evaluate the effect of skills-based educational programming for family carers supporting young adults with psychosis on service user readmission rates and service user/family functional indexes. We hypothesise that system integration of a programme, which delivers early support to the family carers, may reduce family burden as well as improve short-term and long-term patient outcomes. To our knowledge, this study is the first to evaluate effectiveness of a skill-based intervention for family carers within a public healthcare system. This current paper presents a feasible methodology for costudying carers and service users in this environment with the aim to encourage further study of programme implementation and efficacy.

Methods

Study design

This longitudinal pretest–post-test design with a non-equivalent control group will evaluate the short-term and long-term effectiveness of a family psychoeducation intervention. Family functional index measures (ie, Positive and Negative Symptom Scale (PANSS), Manchester Short Assessment of Quality of Life (MANSA), Family Burden Interview Schedule (FBIS)) serve as primary outcomes and will be taken preintervention (ie, baseline), then again at 6-month, 12-month and 24-month postintervention (see figure 1A). Measurements of service utilisation are secondary outcomes and will be mined from the preintervention and postintervention periods for both the control groups and the experimental group (see figure 1B). Gift card incentives will be provided to family and young adult participants as compensation for their time to complete the measurement tools.

Figure 1.

(A) A pre–post study design with 6-month, 12-month and 24-month follow-ups will be used to evaluate quality of life and symptom severity of service users and family burden of family carers. (B) A pre–post design will be used to assess service utilisation of service users, comparing the period of 12-month pre-intervention to the period of 12-month post-intervention and that of 12–24 month post-intervention, and to a nonequivalent control group evaluated over the same time periods. O₁, observation pre-intervention; O₂, observation 6 months post-intervention; O₃, observation 12 months post-intervention; O₄, observation 24 months post-intervention; O_1T, treatment group 12-month period pre-intervention; O_2T, treatment group 12-month period post intervention; O_3T, treatment group 12–24 month period post intervention; O_1C, control group 12-month period pre-intervention; O_2C, control group 12-month period post-intervention; O_3C, control group 12–24 month period post intervention.

Participants

Our population of study consists of family carers of young adults (aged 17–27) with psychosis who have been admitted or discharged within the past 12 months from AHS and Covenant Health governed inpatient psychiatric units within the Edmonton zone. When possible, the enrolled family carer’s matched young adult is then approached for recruitment into the study. Family carers who are unable to communicate fluently in English, and families whose young adult did not have an inpatient admission or discharge within the previous year will be excluded from the study. Multiple family carers from the same family are eligible to enroll as individual participants. Our study looks at short and long-term outcomes in both the carers enrolled in the curriculum as well as those of their corresponding young adult service users.

Study information will be disseminated by any member of the young adults’ clinical team and through community organisations, and family carers will self-identify for participation. Once recruited, and with family consent, service users matched to participating family carers will be contacted for recruitment into the treatment group of the study (figure 2).

Figure 2.

Recruitment schematic.

Sample size

Family recruitment

Literature shows sample sizes ranging from 55 to 135 for studies with 24-month follow-up.²² With approximately 154 eligible young adults (and corresponding family carers) discharged per year, we aim to recruit 111 family carers into the study, which is consistent with similar studies.¹⁹ While the researchers intend to offer participation to every eligible family, only a subset of those expressing initial interest are expected to consent and participate. During the pilot study, approximately 25% of families who were offered the intervention declined due to conflicting schedules or a lack of readiness to participate, which is consistent with previous studies.²³ A minimum sample size of n=111 was calculated for a population size of n=154, significance level of α=0.05 and accuracy of E=0.05 based on recommendations for small populations.²⁴

Attrition rates for 24-month follow-up have been reported as low as 4.4% and 11% in family intervention studies.^{23 25} However, other studies have indicated attrition rates of 25%–30% over 3-month follow-up.¹⁹ From these varied attrition rates, we expect to retain 78 family carers over the follow-up period. The current study will help to clarify how family carer attrition rates compare to other populations.

Young adult service user recruitment

While it is unclear how many families will provide consent to contact their young adult service user for data collection, a previous study indicated that 87% of psychosis patients approached will consent to participate.²⁶ This high rate may reflect service user relative wellness. The young adult service users for the current study are expected to be variably positioned in their recovery journeys. With the expectation that some carers will not consent to their young adult being contacted, and of those young adults contacted, some may be unwilling or too ill to consent, we estimated 50% recruitment rate for eligible service users, which yields approximately 56 participants. Attrition rates for young adults with psychosis at 24-month follow ups vary greatly, but generally fall between 11% and 45%.^25–27 A conservative estimate is to retain 39 young adult service users for the 24-month follow-up, which is a larger sample than seen elsewhere.²⁵

Non-equivalent control group

Including service user and carer control groups (ie, families who are not exposed to the intervention) is either (1) unethical, in the case that the carers desire the intervention but are relegated to the control group or (2) infeasible without considerable incentive, in the case that the carers are uninterested in the intervention at all and unlikely to engage in the study long term. Furthermore, because carer participants self-select for participation in the study, they may systematically differ from other carers in level of investment and engagement in the support of their young adult affected by psychosis.

One approach to mitigating this selection bias is to use secondary administrative health data to compare service utilisation of young people whose carers were exposed to the intervention versus those who were not. Because informed consent is not needed for secondary administrative data, the problems of denying the intervention to carers who would like it and coercion of those hesitant or uninterested in participating are removed. We will retrospectively mine archived data from health records of two groups of service users (one from Edmonton Zone and one from Calgary Zone) whose families were not enrolled in the study and did not receive the educational intervention for comparison to the data gathered from the families who were enrolled and received the intervention. We acknowledge that the Edmonton zone control group is particularly non-equivalent because families and service users who are unable to (or choose not to) engage in the study where it is offered may vary systematically from those who did. To deal with this selection bias and the fact that our Edmonton control group will be contaminated with data from service users whose family actually received the intervention but did not give consent to contact their young adult, we include a separate Calgary control group. We acknowledge that our study, which does compare service utilisation data with the non-equivalent control groups, does not look at control group comparison with respect to our measures of family burden, quality of life and service user symptomatology.

Intervention

The curriculum consists of nine modules delivered in 2-hour sessions over 9 weeks (table 1). The curriculum was vetted by family peers with lived experience as well as professionals from diverse backgrounds. Evidence-based content includes a complete overview and skills practice of LEAP communication technique,⁶ and an introduction to acceptance and commitment therapy.¹⁸ The essential skills of the LEAP communication tool are taught in two modules by facilitators trained specifically for our programme by the author of the technique, Dr Xavier Amador.^{6 21} The two ACT skill modules, codeveloped and facilitated by clinical psychologists, are modelled on ACT interventions for carers whose efficacy after four sessions (8 hours total) is demonstrated in ACT for Psychosis Recovery (ACTp) research.^{19 20} In line with the ACTp researchers’ suggestion that brief ACT interventions are beneficial in improving well-being for carers,²⁸ our ACT modules deliver 4 hours of skills with resources provided for further exploration if participants desire.

Table 1.

Module content for a 9-week psychoeducation programme for families supporting a young adult with psychosis

Session	Name	Content
1	Introduction	Programme overview, including facilitator and participant introductions.
2	Emotional stages on the family’s journey	Describes the emotional stages that are experienced by family carers learning to support a young adult with psychosis.
3	What is psychosis?	From a biopsychosocial perspective, describes symptomology, behaviours and disability that can be present in individuals with psychotic disease.
4	LEAP communication technique – part 1	Educates about anosognosia, the top predictor of noncompliance with treatment. The LEAP method (Listen, Empathize, Agree, Partner) is taught as a communication tool.
5	LEAP communication technique – part 2	Provides immersive roleplay activities to hone LEAP communication tool.
6	Psychosis and addictions	Teaches about the coexistence of psychosis and addictions.
7	ACT for carers—part 1	Introduces Acceptance and Commitment Therapy (ACT) approach to value-based living.
8	ACT for carers—part 2	Provides ACT activities and tools to maintain wellness while supporting a young adult with psychosis.
9	Programme closure and debriefing	Wraps up the education sessions and concludes with a story of family lived experience.

Module details are provided in table 1.

Patient and public involvement

The curriculum was vetted by Families Supporting Adults with Serious Mental Illness, an Edmonton-based grassroots community education and advocacy organisation formed by educated family carers supporting young adults with serious mental illness.

Measures

Family functional index

FBIS-24²⁹ assesses overall burden of the family related to supporting a loved one with addiction and mental health concerns, including the following domains: financial, routines and activities, leisure, relationships, physical health and mental health. Previous literature shows high inter-rater reliability between clinicians and family carers for all items (r ≥0.87³⁰).

Service user functional index

PANSS³¹ assesses overall symptom severity of service users, including seven positive symptoms (eg, hallucinations, grandiosity, etc), seven negative symptoms (eg, blunted affect, emotional withdrawal, etc) and general symptoms (eg, anxiety, lack of judgement/insight, poor impulse control, etc). The PANSS has demonstrated satisfactory inter-rater reliability (r ≥0.80) and is correlated with criterion measures.³²

MANSA³³ is a 16-item self-report scale assessing service user quality of life on a scale of 1 (could not be worse) to 7 (could not be better) and ‘yes/no’ for select items. MANSA has demonstrated satisfactory reliability with psychiatric populations (r=0.74).

Service utilisation

Service utilisation of the young adults whose family carers are enrolled in the educational intervention and matched controls will be measured using data from Alberta Health Services and Covenant Health administrative databases. Data will be mined during a period of up to 12-month preintervention, during the 12-month period postintervention, and during the 12–24 month period postintervention. See table 2 for a list of service utilisation variables mined.

Table 2.

Service utilisation variables to be mined from health records for young adult service users in the treatment and control groups

Setting	Variable
Hospital	Readmission rate for mental health diagnosis within 30 days of an inpatient discharge
Hospital	Frequency of hospital admissions for mental health diagnosis
Emergency department	Frequency of emergency department visits for acute mental health diagnosis within 30 days of last emergency presentation
Community service	Frequency of community service encounters for addition and mental health concerns (eg, CTO, ACT2, EEPIC3, etc.)
Community service	Frequency of crisis and urgent addiction and mental health services (eg, Access 24/7, PACT4, RPACT5)

¹Community Treatment Order, ²Assertive Community Treatment,³ Edmonton Early Psychosis Intervention Clinic, ⁴Police and Crisis Team, ⁵Regional Police and Crisis Team.

Analyses

Functional index analysis (primary outcomes)

Linear mixed models are recommended when data points are dependent on one another (ie, a within group factor such as time) and when there is expected to be missing data, such as longitudinal designs.³⁴ A linear mixed model will be used to determine statistically significant changes in outcome variables (ie, family burden, young adult quality of life and young adult symptomatology), with time as a fixed effect and subjects as a random effect, at four time points: preintervention, 6-month, 12-month and 24-month follow-up.

Service utilisation data (secondary outcomes)

Service utilisation data will be captured for three periods: 12 months preintervention, 12 months postintervention and 12–24 months postintervention. Service utilisation data will be mined to provide a control group comparison on service utilisation metrics. First, an analysis will be conducted to confirm there are no discernable differences between the treatment group and control group on variables of interest in the preintervention period. Next, a linear mixed model will be used to predict statistically significant changes in public health services use (table 2 contains a list of outcome variables), with time and intervention group as fixed effects and subjects as random effects, in a 2 (Intervention: family psychoeducation vs treatment as usual) × 3 (time: 12 months preintervention vs 12 months postintervention vs 12–24 months post-intervention) design.

Strengths and limitations

The described protocol is one of the first attempts to evaluate a psychoeducation group that draws on material from evidence-based interventions for carers of young adults with psychotic illness. The current study builds on others^{11 17} by incorporating psychoeducational elements, the LEAP technique⁶ and ACT components.¹⁸ Furthermore, the design of the study intends to gather longitudinal 24-month postintervention data from carers and service users, including health service utilisation data.

It has long been recognised that there are considerable challenges to studying psychotic illness as it affects families in an ecologically valid, ethical and feasible way.^{17 19 27} Historically, intervention and outcome measurements have focused on the young adults affected by psychosis.¹⁷ One strength of this protocol is that while it acknowledges that there are sources of error, such as shared information among the caregiver community and that multiple family members supporting a single young adult may participate in the intervention, the study considers a broader socioenvironmental approach to treatment that allows for greater ecological validity. Researchers must continue to develop methodologies that allow for study of the support network despite complexities.^{17 19} Providing longitudinal data for indexes of family function is vital to understand the enduring changes of this intervention on the family system.

A second strength of the study is that family member participation does not depend on the consent of their young adult service user. Well-recognised barriers to implementation of family psychoeducational interventions include the complexity in the ecosystem of existing family relationships (eg, different family members being at different levels of acceptance of the disease), symptom-informed interaction between family and loved ones (eg, paranoia of which the family carer is the subject) and concern around the need for exclusive patient–professional relationship.⁵ Our study manages these challenges by directly recruiting family participants, creating both breadth in the study sample and more opportunity for families to benefit from the education. On the other hand, the ethical necessity of gaining family carers’ consent to recruit their young adult restricts the number of young adults we are able to enrol, which in turn limits the power to detect young adult outcome.

Several limitations must be considered in the current protocol. First, to ethically provide an equal opportunity to receive the intervention, we were not able to randomly assign participants to treatment and control groups. It follows that a lack of random assignment and experimental control precludes causal conclusions regarding the intervention, which makes it difficult to ascertain true effectiveness. However, by using health utilisation data from two non-equivalent control groups, we gain greater confidence in long-term effects of the programme on public health system burden. Second, while this protocol focuses on family carer and service user functional indexes and young adult service utilisation, it is possible that this intervention affects other outcome variables (eg, clinician attitude towards family involvement). Nonetheless, our focused interest is in assessing programme efficacy in an ecologically valid system as reflected by family functional outcomes in the context of real-life experience. Third, we acknowledge that attrition can be a considerable source of bias in longitudinal data sets. To account for this, systematic patterns in those who drop out of the programme will be explored.

Despite the challenge of conducting high-quality family-centred research in this population, our protocol describes a feasible and ecologically valid methodology to evaluate a skills and knowledge-based psychoeducational intervention for family carers of young adults with psychosis using both carer and service user measurement outcomes over 2 years.

Ethics and dissemination

All procedures involved in the study are consistent with the generally accepted standards of ethical practice. All study participants provided written informed consent to be eligible for study inclusion. The study was approved by the University of Alberta Research Ethics Board (Pro00110691). Study findings will be disseminated through peer-reviewed publications, academic and medical conferences and direct reporting to senior administration of health systems serving population of interest. A programme manual is under development for dissemination of programme content to be administered by other care teams.

Ethics statements

Patient consent for publication

Not applicable.