A Systematic Review Exploring the Relationship Between Family Factors and Symptom Severity, Relapse and Social or Occupational Functioning in First‐Episode Psychosis

ABSTRACT

Aims

Research on family functioning in psychosis has typically focused on specific family‐related factors and their impact on symptomatology, finding strong associations between high expressed emotion and poor outcomes, especially in those with long‐term illness. The objective of this review is to examine the impact of a broad range of family‐related factors and their relationship with clinical, social, occupational and relational outcomes in first‐episode psychosis (FEP).

Method

A systematic search of databases PsycInfo, Pubmed, Embase and CINHAL between 1990 and August 2023 was completed. In total, 1408 articles were screened, and study quality was assessed using the JBI critical appraisal checklists for analytical cross‐sectional studies and cohort studies. A narrative synthesis approach was used to analyse the data.

Results

Of the 1408 identified articles, 80 full‐text articles were screened. Fifteen studies were included for data extraction. The objectives and scope of the selected studies varied considerably. Studies consisted of cross‐sectional and prospective cohort designs. Participants consisted of FEP patients and family members, with the patient age range varying from 15.9 to 30.8 (MD = 24.5). Significant associations were identified between family factors (high EE, family environment and carer burden) and symptom and social outcomes.

Discussion

While findings emphasised the significant impact of family factors on symptomatic and social outcomes, the literature had significant limitations due to the absence of underlying theoretical models and understanding of the dynamics of distress in families within FEP. Future research should seek to develop a model accounting for such processes.

1. Introduction

1.1. Psychosis

Psychosis is a major psychiatric disorder which has a significant impact on those affected (Heinze et al. 2018) and their support network (Koutra et al. 2014). Approximately 85% of patients experience a relapse following a FEP (Camacho‐Gomez and Castellvi 2020). Increased relapses lead to an increased risk of chronic illness, suicidal behaviours or compromised functional abilities (Fusar‐Poli, McGorry, and Kane 2017). There are several reasons why family factors are important in the context of improving outcomes in psychosis: (i) the onset of FEP often occurs while the patient is still residing at home with their parents (Jansen, Gleeson, and Cotton 2015); (ii) some people with psychosis struggle with transitioning to independence and long‐term romantic relationships; the family of origin can be their life‐long domestic milieu (Roy et al. 2013); (iii) relapse is strongly related to factors within the family environment (da Silva et al. 2021); (iv) family interventions are effective in reducing relapse and yet have a low level of service provision (Camacho‐Gomez and Castellvi 2020).

1.2. Family Functioning

Family functioning pertains to the overall social and structural characteristics of the family environment, which encompass the dynamics and connections within the family, focusing on elements such as conflict and cohesion, adaptability, organisation and the quality of communication (Lewandowski et al. 2010). Various models have been developed to explain the complex interactions necessary for effective functioning within families in general, such as The McMaster Model of Family Functioning (Epstein, Bishop, and Levin 1980) and The Circumplex Model of Family Functioning (Olson, Waldvogel, and Schlieff 2019).

Across models, optimal family functioning is often observed in an environment characterised by transparent communication, well‐defined roles, cohesion and effective emotional regulation. Conversely, poor family functioning is commonly marked by elevated conflict, lack of organisation and challenges in emotional and behavioural control (Lewandowski et al. 2010). The current review will focus on three areas of family functioning that are most examined in psychosis research: expressed emotion, family environment and family burden.

Expressed emotion refers to the emotional atmosphere or tone in a family or caregiving environment, particularly concerning how family members communicate and express their emotions towards a person with a mental health disorder (Zanetti et al. 2018). Components of expressed emotion include critical comments, emotional over‐involvement and hostility. High levels of expressed emotion among family members are a known risk factor for relapse in psychosis (Ng, Fung, and Gao 2020), as well as a significant family stressor (Amaresha and Venkatasubramanian 2012). A wide range of factors have been identified as predictors of high expressed emotion, including caregivers' personality factors, avoidant coping strategies, long duration of untreated illness and increased family burden (Koutra et al. 2014).

Research exploring the influence of family environment on outcomes in FEP has indicated that factors such as cohesion, positive remarks and warmth have been identified as protective factors (Lee, Barrowclough, and Lobban 2014).

Family burden refers to the challenges and obstacles experienced by families in response to a family member's illness (Ennis and Bunting 2013), which can be categorised into financial, physical, emotional, social and leisure burdens and has both objective and subjective aspects (Zanetti et al. 2018). Increased levels of family burden are associated with increased isolation, along with emotional and physical depletion in those caring for children with FEP (McCann, Lubman, and Clark 2011), along with reduced time to relapse in those experiencing FEP (Koutra et al. 2015).

1.3. Current Study

In psychosis, much of the research surrounding family function has focused on individual factors and their impact on symptomatology. It has also often focused on patients with long‐term illness trajectories. This perspective has often failed to explore the relationship between family factors and a range of social outcomes, especially early in the course of illness while familial roles may be preserved. In this current review, we aim to explore family functioning as a whole, including expressed emotion, positive relationships, family environment and family burden and their relationship with clinical and social outcomes.

2. Method

2.1. Design

The current study is a systematic review and narrative synthesis examining the association between family functioning and clinical and social outcomes, specifically symptom severity, general functioning and relapse. The current review used the PRISMA guidelines (Moher et al. 2009). The methods were predetermined and specified in a protocol submitted to the PROSPERO database of systematic reviews in May 2023 [CRD42023450229].

2.2. Search Strategy

Studies were identified through electronic searches conducted using four databases: PsycInfo, Pubmed, Embase and Cumulative Index to Nursing and Allied Health Literature (CINHAL). With support from the social sciences librarian, a PICO framework was used to identify relevant search terms. Initial search terms included ‘first episode psychosis’, ‘early psychosis’, ‘early schizophrenia’, ‘psychotic disorder’, ‘recent onset’, ‘family functioning’, ‘family conflict’, ‘family cohesion’, ‘family communication’, ‘family problem solving’, ‘family adaptability’, ‘expressed emotion’, ‘social’, ‘relational’, ‘occupational’, ‘psychosocial’, ‘functioning’, ‘quality of life’, ‘social adjustment’, which were used alone or as a combination of terms. Searches were completed up until July 21st, 2023. The complete list of search terms utilised in this study across databases is provided in Supporting Information.

Following the removal of duplicates, all papers were screened by title and abstract by the first author for relevance based on the inclusion criteria below. A second independent reviewer screened 25% of the papers, and a third independent review resolved conflicting decisions. Following completion of the initial screening, all papers were reviewed in full by both the author and the second reviewer. There was substantial agreement (89.47%: Cohen's K = 0.69) for full‐text review. Per PRISMA guidelines, the flow of papers during each phase of the search process is depicted in Figure 1.

FIGURE 1.

PRISMA flowchart of studies included in this review.

2.3. Inclusion and Exclusion Criteria

Requirements for inclusion were:

1. Patients diagnosed with an FEP and their parents or carers.

2. Family functioning, symptom severity and/or social/occupational/relational functioning are measured using clearly defined measures.

3. Exploring the relationship between family functioning, symptom severity and social or occupational functioning.

4. Observational peer‐reviewed. studies published in English after 1990.

Studies were excluded if: (1) the population of the study did not meet inclusion criteria, namely those described as ‘chronic presentations’ or ‘at risk of psychosis’; (2) the study design were intervention studies, systematic reviews, dissertations, case studies, editorials or articles not published in peer‐review journals.

2.4. Data Extraction

For each included study, the author extracted the following variables: the name and year of the study, the location of the study, demographic details of the participants, the study design, the main measures used and the principal findings (see Table 1). A second author extracted variables for 30% of the included studies, and any disagreements in the extracted data were resolved. A narrative synthesis approach was employed to analyse the findings (Popay et al. 2006).

TABLE 1.

Summary of main findings from relevant studies.

Author (date) country	Sample	Diagnosis	Sex	Mean age (SD)	Family measures	Symptom severity measures	Functioning measures	Main findings
Cotton et al. (2010) Australia	81 patients with 63 family members	SZ (33%) PD NOS (30%) Other (37%)	M (63.0%) F (37%)	20.1 (3.1)	FQ ECI	SANS MADRS BPRS	WHOQoL‐Bref PAS SOFAS	EOI and social relationships QoL (r = −0.25)* ECI difficult behaviours and psychological QoL (r = −0.27)* ECI difficult behaviours and environment QoL (r = −0.25)*
da Silva et al. (2021) Brazil	65 Patient‐family dyads	SZ (60%) Psy Dep (12%) BAD (17%) Sub ind (11%)	Overall breakdown not reported	Overall breakdown not reported	FQ‐BPV	BPRS		EOI associated with ↑ relapses (OR = 1.20, 95% confidence interval [CI]: 1.05–1.36, p = 0.007).
DeTore et al. (2018) USA	41 Patient‐family caregiver dyads	FEP (100%)	M (63%) F (27%)	24.6 (4.4)	QRS	BPRS	Mod‐SAS	At baseline, FB was not significantly associated with work status or hours worked at baseline. At 6 months, FB was associated with work status (r = −0.32)* and with ↓ hours worked (r = −0.39)* At 1 year, FB was associated with work status (r = −0.47)* with ↓ hours worked (r = −0.49)* LFO subscale was associated with work status at baseline (r = −0.42)*, 6 months (r = −0.49)* and 1 year (r = −0.51)* LFO subscale was associated with ↓ hours worked at baseline (r = −0.48) 8, 6 months (r = −0.52)* and 1 year (r = −0.50).
Domínguez‐Martínez et al. (2014) Spain	44 Patient‐family caregiver dyads	ARMS (45%) FEP (55%)	M (66%) F (34%)	23.7 (5.6)	FQ	PANNS CDSS	GFS SFS	Criticism was significantly associated with all clinical and functioning measures with the exception of GF‐S; ○PANNS—Positive symptoms (r = 0.42)*** ○PANNS—Negative symptoms (r = 0.35)* ○PANNS—General symptoms (r = 0.43)*** ○CDSS (r = 0.3)* ○SFS (r = −0.43)*** ○GF‐R (r = −0.43)*** EOI was significantly associated with; ○PANNS‐negative symptoms (r = 0.35)* ○PANNS‐general symptoms (r = 0.45)*** ○GF‐S (r = −0.49)*** ○GF‐R (r = −0.43)***
Halford, Schweitzer, and Varghese (1991) Australia	57 patient relative dyads	SZ (33%) MDD (39%) BAD (16%) Other (12%)	M (60%) F (40%)	23.7 (5.6)	FES	BPRS SANS	QLS	No association between family measures and outcome measures (QoL and BPRS) at 6 months. No association between patient and relative FES ratings with BPRS scores at 6 months. High relative‐rated emotional expressiveness was associated with ↓ symptom severity at 6 months (r = −0.42)* Relatives' rating of emotional expressiveness at 1 month after discharge predicted ↓ scores on; ○ SANNS (r = −0.46)** ○ QoL measure (r = −0.34)* Relative‐rated conflict at admission was significantly correlated with: ○ ↑ severity of psychopathology (r = 0.29)* and negative symptoms (r = 0.32)** at follow‐up. ○ However, when premorbid functioning was controlled for, conflict did not predict patient outcome.
Halford et al. (1999) Australia	52 patient‐relative dyads	SZ (31%) MDD (40%) BAD (19%) Other (10%)	M (60%) F (40%)	30.8 (12)	OFI	BPRS SANS	GAS QLS PAS	Living with an affect‐regulated family associated ○ ↓ relapse compared with living with an unregulated family (Χ 2 = 6.50)* ○ ↓ BPRS score compared with living with an unregulated family (F = 13.94)* ○ ↓ negative symptoms compared with living with an unregulated family (F = 6.35)* ○ ↑ GAS scores compared with living with unregulated (F = 14.71)** ○ ↑ QLS score compared with living with an unregulated family (F = 9.07)**
He et al. (2021) China	160 patients	FEP (25%) UHR (25%) FDR (25%) HC (25%)	M (60%) F (40%)	26.4 (6.5)	FACES II‐CV	MADRS PANSS SIPS	GAF	No association between family measures and symptom severity/functioning identified.
Lee, Barrowclough, and Lobban (2014) England	65 patient‐relative dyads	FEP (84%) AF PSY (6%) Sub in (5%) SZ (5%)	M (74%) F (26%)	24.5 (5.5)	CFI PBI PCM	PANSS		Overall EE status (r = 0.027)*, critical comments (r = 0.30)* and warmth (r[58] = −0.38)* associated with relapse at 6 months. Warmth was associated with ↓relapse (r = −0.29)* at 1 year. When covariates were controlled for, EE warmth significantly predicted relapse at 6 months (Χ 2 = 21.27)** Overall EE status and criticism no longer predicted relapse at 6 months.
Nuttall et al. (2019) USA	282 Patient relative dyad	SZ (53%) SZ AD (17%) SZPID (18%) Psy NOS (12%)	M (75%) F (25%)	22.5 (4.8)	BAS	CDSS PANSS	QLS	At baseline ↑ levels of initial FB were significantly associated with ↑ levels of positive symptoms (F = 7.32)* Over‐time, ↑ QoL associated with ↓ FB (F = −64.08)*
Otero et al. (2011) Spain	110 patient‐parent dyads	BPD (5%) SCZ (34%) PD NOS (25%) SZAD (6%) SZPID (9%) BAD (17%) DD (4%)	Not reported	15.5 (1.8)	PACI	K‐SADS‐PL CGI PANSS	CGAS	At baseline, open communication was significantly associated with ↑ PANNS (r = 0.209)* ↑ problems in communication were associated with ↑scores on the PANNS (r = −0.30)** at 6 months, and with ↓ scores on the CGI (r = −0.20)* at 12 months.
Ropi et al. 2021 Greece	40 patient‐parent dyad	SZ (68%) SZ AD (7%) BPD (5%) BPD NOS (2%) MDD (15%) BAD (3%).	M (68%) F (32%)	15.9 (1.2)	FAD FQ	K‐SADS‐PL PANSS	CGAS PAS‐GR KIDSCREEN 27	Scores on the SES (KIDSCREEN‐27) associated with ↓ roles (r = 0.64)*** and ↓ general family functioning (r = −0.39)* ↑ economic difficulties associated with ↓ psychological well‐being (F = 13.71)**
Stirling et al. (1993) England	30 patient‐relative dyads	SZ 80% BSAD (20%)	M (47%) F (53%)	25.2 (no SD)	CFI	SANS KGV scale, PSE change		No association between EE & relapse at baseline. EE measured at follow‐up associated with ↑relapse compared with low EE households (x 2 = 6.24)**
Tomlinson, Onwumere, and Kuipers (2014) England	33 patients 24 first‐degree relatives	SSD (100%)	M (67%) F (33%)	23.7 (3.6)	FAS ECI	HADS BPRS		Patients' perceived carer criticisms were strongly correlated with; ○↑ symptoms of psychosis on the BPRS (r = 0.53)** ○↑ anxiety (r = 0.57)** ○↑depression (r = 0.51)**
Verdolini et al. (2021) Spain	283 patients	FEP (100%)	M (66%) F (34%)	25.4 (5.3)	FES		FAST	↓ functioning in non‐affective patients was associated with; ○ ↓ rates of active‐recreational (F = −0.303)* ○ ↓achievement‐orientated family environment (F = −0.304)* ○ ↑ moral‐religious emphasis (F = 0.200)* ○ ↑ control (F = 0.268)* In the affective patients, ↓functioning was associated with: ○ Conflict (F = 0.718)*
Von Polier et al. (2014) Switzerland and Germany	31 patient parent dyads	SZ (58%) SZPID (22.3%) SZ AD (10%) DD (67) PD NOS (3%)	M (74) F (26)	15.8 (1)	FMSS	PANSS CGI‐SIS	PAS ILK	Patient level of perceived criticism was significantly associated with; ○Poorer quality of life (r = −0.43)* ○Poorer interactions with family (r = −0.47)* and peers (r = −0.48)*

Abbreviations: AF PSY = affective psychosis, ARMS = at risk mental state, BAD = bipolar affective disorder, BAS = burden assessment scale, BPD = brief psychotic disorder, BPD NOS = brief psychotic disorder not otherwise specified, BPRS = brief psychotic rating scale, BSAD = brief schizoaffective disorder, CDSS = Calgary depression scale for schizophrenia, CFI = Camberwell family inventory, CGAS = The children's global assessment scale, CGI‐S = clinical global impressions–severity of illness scale, DD = depressive disorder, ECI = experience of caregiving inventory, EOI = emotional over‐involvement, FACES II‐CV = family adaptability and cohesion scales second edition Chinese version, FAD = family assessment device, FAS = family attitudes scale, FAST = functioning assessment short test, FB = family burden, FDR = first‐degree relatives, FEP = first‐episode psychosis, FES = family environment scale, FMSS = 5 min speech sample, FQ = family questionnaire, FQ‐BPV = Brazilian Portuguese version of the family questionnaire, GAF = global assessment of functioning, GAS = Global Assessment Scale, GFS = global functioning scales, HADS = hospital anxiety and depression scale, HC = healthy controls, ILK = inventory of children's quality of life, KGV scale = modified Krawieka, Goldberg, and Vaughan psychiatric assessment scale, K‐SADS‐PL = the Kiddie schedule for affective disorders and schizophrenia, LFO = limits on family opportunity, MADRS = Montgomery–Asberg depression rating scale, MDD = major depressive disorder, Mod‐SAS = the modified social adjustment scale, OFI = observed family interaction, PACI = parent adolescent communication inventory, PANSS = The Positive And Negative Syndrome Scale, PAS = Premorbid Adjustment Scale, PAS‐GR = Premorbid Adjustment Scale Greek version, PBI = parental bonding instrument, PCI = perceived criticism measure, PD NOS = psychotic disorder not otherwise specified, PSE = the present state examination change rating scale, Psy Dep = psychotic depression, QLS = Quality Of Life Scales, QoL = quality of life, QRS = questionnaire of resources and stress for families with chronically ill or handicapped members, SANS = scale for assessment of negative symptoms, SES = school environment scale, SFS = social functioning scales, SIPS = structured interview for prodromal syndromes, SOFAS = social and occupational functioning assessment scale, SSD = schizophrenia spectrum disorder, Sub Ind = substance induced psychosis, SWN‐K = Subjective wellbeing under neuroleptics‐short form, SZ AD = schizoaffective disorder, SZ = schizophrenia, SZPID = schizophreniform disorder, UHR = ultra‐high risk for psychosis, WHOQoL‐Bref = World Health Organization quality of life scale‐brief.

^***< 0.001.

^**< 0.01.

^*< 0.05.

2.5. Study Quality Assessment

The author examined the quality of all studies with a second independent reviewer examining 30% of the studies using the JBI critical appraisal checklists for analytical cross‐sectional studies and cohort studies. A third independent review resolved conflicting ratings. Studies with a JBI score higher than 70% were classified as having high quality, those with a score between 50% and 70% as having medium quality and those with less than 50% as having low quality. Tables 2 and 3 describe the quality appraisal results.

TABLE 2.

JBI critical appraisal checklist for analytical cross‐sectional studies.

Study	Inclusion Criteria	Subjects and setting described	Valid and reliable measure of exposure	Measurement of condition	Confounding factors	Strategies to deal with confounding factors	Valid and reliable measure of outcome	Statistical analysis	Total (%)
Cotton et al. (2010)	x	x	x	x	x		x	x	87.5
Von Polier et al. (2014)	x	x		x			x	x	62.5
Tomlinson, Onwumere, and Kuipers (2014)	x	x	x	x			x		62.5
He et al. (2021)	x	x	x	x	x	x	x	x	100
Domínguez‐Martínez et al. (2014)	x		x	x			x	x	62.5
Ropi et al. (2021)		x	x	x			x	x	62.5

TABLE 3.

JBI critical appraisal checklist for cohort studies.

Study	Groups same population	Exposures measured similarly	Valid and reliable measure of exposure	Confounding factors	Strategies to deal with confounding factors	Groups free from outcome at the start	Outcomes measured in valid and reliable way	Follow‐up reported and sufficient time	Follow up complete	Incomplete follow‐up addressed	Statistical analysis	Total (%)
da Silva et al. (2021)	x		x				x	x	x		x	54.5
DeTore et al. (2018)	x	x	x	x	x		x	x	x		x	81.8
Halford, Schweitzer, and Varghese (1991)	x	x	x	x	x		x	x	x		x	81.8
Halford et al. (1999)	x	x	x				x	x	x		x	63.6
Lee, Barrowclough, and Lobban (2014)	x	x	x	x	x		x	x	x		x	81.8
Nuttall et al. (2019)	x	x	x	x	x		x	x	x		x	81.8
Otero et al. (2011)	x	x	x	x	x		x	x	x		x	81.8
Stirling et al. (1993)	x	x	x				x	x	x			54.5
Verdolini et al. (2021)	x	x	x				x	x	x		x	63.6

3. Results

3.1. Search and Screening Results

The initial search identified 1408 records. Following deduplication, 910 went through to the title/abstract screening. A total of 80 papers were screened at the full text, of which 64 papers were excluded. Studies were frequently excluded due to the population (n = 27), the focus of the research (n = 15), design (n = 15) and published language (n = 3). In total, 15 studies were included, with data extracted from all papers.

3.2. Included Studies

Overall, 15 studies were included in this review and a summary of the main findings can be found in Table 1. It was noticeable that included studies varied widely in their objectives and scope. Over 90% (n = 14) of studies took place in Western countries. Cross‐sectional (40%, n = 6) and prospective cohort designs (60%, n = 9) were used.

Eighty percent (n = 12) of studies included family members in the research, the remaining 20% (n = 3) of studies did not include family members, instead focused on the patient's experience of family functioning. There was a wide range of sample sizes ranging from 30 to 340 (M = 122.75, MD = 65), with 69% (n = 11) of the included studies having less than 70 subjects. Such sample sizes may indicate reduced power to detect small or subtle effects or relationships and may limit the generalizability of findings overall. With the exception of one study which did not provide a mean age of FEP subjects, the overall mean age of subjects was 23.09, with the mean age of participants ranging from 15.9 to 30.8 (MD = 24.5).

3.3. Measures

3.3.1. Family Factors

Over half of the studies explored expressed emotion (60%, n = 9), and the remaining studies explored family environment (27%, n = 4) and family burden (13%, n = 2). Two papers further explored the experience of caregiving, in addition to the above‐mentioned. A wide range of measures were utilised to explore these areas of family functioning, with the most frequently used including the family function questionnaire (FQ) (25%, n = 4), and The Camberwell Family Interview (CFI) (13%, n = 2).

3.3.2. Symptom Severity

Most studies (87%, n = 13) reported symptom severity, which was rated utilising a wide range of measures. The most used measures were the Positive and Negative Syndrome Scale (PANSS) (40%, n = 6) and the Brief Psychiatric Rating Scale (BPRS) (40%, n = 6).

3.3.3. Social/Occupational Functioning

The majority of studies (73%, n = 11) reported functional outcomes; several measures were utilised to do so including the Quality of Life Scales (25%, n = 3), the Global Assessment Scale—(GAS) (25%, n = 3), the Premorbid Adjustment Scale (PAS) (25%, n = 3).

3.4. Narrative Synthesis

High Expressed Emotion was found to be significantly associated with both psychopathology and functioning. Specifically, criticism was found to be significantly associated with symptoms severity (Domínguez‐Martínez et al. 2014; Tomlinson, Onwumere, and Kuipers 2014), social functioning (Domínguez‐Martínez et al. 2014), poorer quality of life (Von Polier et al. 2014) and relapse (Lee, Barrowclough, and Lobban 2014). Emotional over‐involvement was significantly associated with symptom severity (Domínguez‐Martínez et al. 2014), reduced quality of life, social relationships (Cotton et al. 2010), social functioning (Domínguez‐Martínez et al. 2014) and relapse (da Silva et al. 2021). In contrast, warmth was associated with reduced incidences of relapse, with a one‐point increase in the warmth scale associated with a 58% decrease in the likelihood of relapse (Lee, Barrowclough, and Lobban 2014). Communication style was found to have a significant relationship with patient outcome, with open communication found to be positively associated with general psychopathology scores (Halford et al. 1999). Communication problems were significantly associated with a higher psychopathology at follow‐up (Otero et al. 2011). Several family environment factors had an impact on functioning. Reduced recreational achievement, increased conflict, moral/religious and control within the family environment were related to poorer functioning (Verdolini et al. 2021). Positive family environment as reflected in affect regulation, and emotional expressiveness were related to improved functioning (Verdolini et al. 2021; Halford et al. 1999). Higher levels of family burden were found to be associated with increased levels of positive symptoms (Nuttall et al. 2019). Similarly, family burden, specifically economic difficulties, were found to be associated with worse outcomes in an adolescent sample, with increased economic difficulties significantly associated with lower psychological well‐being scores (Ropi et al. 2021). Positive experiences of caregiving questionnaire were positively associated with psychological, environmental and social relationships quality of life (Cotton et al. 2010).

4. Discussion

The present review identified 15 papers that investigated the relationship between family factors and outcomes in patients experiencing first‐episode psychosis (FEP). In line with previous reviews, the review emphasised the significant impact of family factors on outcomes such as symptom severity, relapse and social, occupational and relational functioning (Koutra et al. 2014). The experience of critical comments and overinvolvement, was extensive, and had detrimental implications on outcomes, such as increased symptom severity (Domínguez‐Martínez et al. 2014; Tomlinson, Onwumere, and Kuipers 2014), reduced functioning (Von Polier et al. 2014; Cotton et al. 2010), and increased risk of relapse (Lee, Barrowclough, and Lobban 2014; da Silva et al. 2021). Difficulties in communication style, similarly, were associated with increased symptom severity and reduced clinical improvement (Otero et al. 2011).

Perhaps most importantly, there was a ‘silence’ in the reviewed research papers on key contextual variables of family functioning (O'Cathain, Murphy, and Nicholl 2010). This silence was surprising, and points to a range of potential research targets for a better understanding of family dynamics and potential interventions.

4.1. Developmental Context

Across the papers, the mean age of those experiencing FEP was early adulthood (mean = 23.09 years). Yet, typically the research reviewed did not take a developmental perspective. Developmentally, this is a time when young people strive for autonomy, and due to their experience of FEP, their opportunities to gain independence are often hindered, placing parents and families in the role of enforcing boundaries, which has consequences for family dynamics (Kagitcibasi 2013). It is likely that the conflict between autonomy and need‐for‐support is the context for much of the high EE reported in the literature.

4.2. Families in Distress

The outcomes of this review paint a picture of families in significant distress in response to a family member experiencing FEP. As Van Os et al. memorably said ‘High expressed emotion: marker for a caring family?’ (van Os et al. 2001). It was noticeable that the literature on family factors in psychosis was not grounded in the broader field of systemic thinking, in particular how families react when under stress. This was an area of significant research in the early family and psychosis research (e.g., Birchwood and Cochrane 1990) but was not as evident in later papers. Understanding normal family dynamics, especially when distressed would allow researchers to contextualise, normalise and validate behaviours like criticism, or over‐involvement. There are models of family distress in other disorders for example, Waals et al. (2018) Family Distress Cascade Theory in self‐injury, or Family Stress Model in general psychopathology (Masarik and Conger 2017). Models such as these, identify broad economic and intergenerational stressors that families experience, as well as intra‐personal family dynamics, which may apply within psychosis. Yet, the current psychosis literature frequently failed to address the context or cause of distress for families.

4.3. Family as a Protective Factor

Just three studies reported results that indicated protective factors associated with outcomes in FEP, which included the positive association between warmth and relapse (Lee, Barrowclough, and Lobban 2014), open communication and outcomes (Otero et al. 2011) and supportive emotional expressiveness with symptom severity (Halford, Schweitzer, and Varghese 1991). While research frequently highlights family factors and negative outcomes, the protective nature of family factors was largely underexplored in the research.

4.4. Clinical Implications

The findings of the current review outline the necessity of considering the relationship between family factors and outcomes in terms of clinical practice. However, while research indicates the effectiveness of family interventions in FEP (Camacho‐Gomez and Castellvi 2020), barriers to the implementation of such interventions often exist within services Bucci et al. (2016) reported that family interventions were only offered in between 0% and 53% of EIP services. The barriers related to factors such as the service culture of involving family members, insufficient staff training, insufficient integration of family interventions into routine clinical practice, and inadequate access to supervision after family therapy training (Eassom et al. 2014). Equally, organisational barriers included limited finance and staff resources (Bucci et al. 2016). To effectively support those experiencing FEP and their families, a necessary shift in understanding and values surrounding systemic work is essential; failing to do such ultimately ignores evidence which suggests that mental health care should target populations, relationships, and other determinants, rather than only the individual (UN Human Rights Council 2020). Onwumere, Bebbington, and Kuipers (2011) recommended a tiered approach to family interventions with some early psychosis families only requiring low‐intensity interventions and while families where service users have persisting symptoms, where there were ongoing difficult relationships or where carers have complicated grief reactions to the psychotic illness that intensive interventions are probably most suitable for. Similarly, given the link between attachment difficulties and the development of psychosis (Grady et al. 2024), both individualised and family interventions emphasising repairing attachment ruptures could enhance individual recovery (Boden‐Stuart, Larkin, and Harrop 2021).

4.5. Research Implications

A wide number of family factors were identified by this review. However, these were primarily studied individually, prohibiting the examination of their potential interactions or cumulative effect. A cumulative effect has been documented in general psychopathology in the Family Stress Model (Masarik and Conger 2017). Future studies should examine potential interactions through network analysis or other methodologies. It was noticeable that family strengths were under‐researched but where studies supported positive outcomes. Additionally, family functioning measures often focused only on the present and rarely considered functioning before the onset of FEP. There is a need for the development of family functioning measures which include questions relating to differing life stages (Ramaswami et al. 2022).

Most of the reviewed papers in the current review were from Western industrialised countries with mainly white samples. It is important to consider the ethnicity of samples due to cultural differences in family systems. Family functioning measures are often based on nuclear families with high socioeconomic status, which may not reflect the values, ethnic identity and family structure of other families (Burkhard et al. 2021). For instance, Black individuals with schizophrenia experienced improved outcomes when exposed to higher levels of EE and lower levels of social support (Rosenfarb, Bellack, and Aziz 2006; Gurak and Weisman de Mamani 2017), a significant contrast to much of the research involving white samples. This highlights the need for the consideration of contextual and cultural factors in developing family interventions.

4.6. Limitations

Several limitations in the current literature exist. This review identified only 15 suitable research papers, and the selected papers varied considerably in their focus, follow‐up, samples and psychometric measures. This limits the generalisability of the findings. The majority of sample sizes of the selected papers were below 70 participants, posing an issue of generalizability and reduced statistical power to detect small or subtle effects. However, given the size and heterogeneity of the existing literature, the strength of the associations with clinical factors was noticeable, especially criticism and emotional over‐involvement.

The present review also had a number of limitations. Only English‐language research papers were included in the study. This is a significant limitations. The exclusion of non‐English language studies reduced the likelihood of samples with non‐Western family structures. The outcomes reported could therefore be vulnerable to cultural, language and publication bias. This is particularly important considering the diverse role of culture in forming family systems (Jones et al. 2021).

4.7. Conclusion

In conclusion, significant associations between family factors and outcomes in FEP were identified. The findings emphasised the significant impact of family factors on outcomes such as symptom severity, relapse and social and occupational functioning. The authors highlighted the need for a model reflecting family distress experienced as a result of FEP and the impact of family distress on family dynamics and processes.

Ethics Statement

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analysed in this study.