Predictors of criticism and emotional over-involvement in relatives of early psychosis patients.

Mechanisms underlying the manifestation of relatives' expressed emotion (EE) in the early stages of psychosis are still not properly understood. The present study aimed to examine whether relatives' psychological distress and subjective appraisals of the illness predicted EE dimensions over-and-above patients' poor clinical and functional status. Baseline patient-related variables and relatives attributes comprising criticism, emotional over-involvement (EOI), psychological distress, and illness attributions were assessed in 91 early psychosis patients and their respective relatives. Relatives were reassessed regarding EE dimensions at a 6-month follow-up. Relatives' psychological distress and illness attributions predicted criticism and EOI over-and-above patients' illness characteristics at both time points. Relatives' increased levels of anxiety, attributions of blame toward the patients, an emotional negative representation about the disorder, and decreased levels of self-blame attributions predicted EE-criticism at baseline. Relatives' anxiety and negative emotional representation of the disorder were the only significant predictors of EE-criticism at follow-up, whereas anxiety, attributions of control by the relative and an emotional negative representation about the disorder predicted EE-EOI both at baseline and follow-up assessments. Understanding the components that comprise and maintain EE attitudes should guide early psychosis caregivers in family interventions, enhancing proper management of psychological distress and reduction of negative appraisals about the illness. The prevention of high-EE attitudes over time in a sensitive period such as early psychosis might be critical in shaping the health of caregivers and the outcome of the affected relatives.

Introduction

Epidemiological research indicates that psychosocial factors impact the risk, progression and outcome of psychosis liability. In fact, some metanalyses have found odds ratios for macroenvironmental factors (e.g., urbanicity) that are similar to those attributed to genetic factors [1,2]. In the context of microenvironmental factors, one of the most significant factors in psychosocial research in psychosis has been expressed emotion (EE) [3], a measure of the family environment used to describe relatives’ attitudes toward an ill family member. High-EE attitudes, particularly criticism and emotional over-involvement (EOI), are considered the strongest psychosocial predictors of relapse in schizophrenia [4]. Given the heightened stress-sensitivity associated to psychosis liability [5], the impact of high negative emotionality and distress in the family environment appears particularly relevant. However, it is important to note that the EE construct does not aim to blame families for contributing unidirectionally to the patient’s clinical worsening. Instead, EE is best regarded as the product of a negative dynamic interaction between patients and their families that can play a central role in the course of the patient’s disorder [6,7].

Although over the last 50 years most studies on EE have involved patients with schizophrenia, recent research has focused on the study of EE in relation to schizotypy [8], and the early course of psychosis [9,10]. Preliminary studies indicate that high-EE is already present in over half of the relatives of persons with first-episode psychosis (FEP) [11,12], and is even present in relatives of at-risk mental state (ARMS) [13,14]. Importantly, early psychosis caregivers often report a heightened risk of psychological distress and more negative caregiving appraisals compared to family members of patients with chronic schizophrenia [15]. Besides, recent findings indicated that relatives’ high levels of depression and anxiety are associated with high levels of EE at these early stages [16-18]. Since the presence of EE is associated with multiple negative outcomes for relatives and patients, it is crucial to examine the mechanisms underlying the development of EE in the early stages of psychosis, when most of the changes are emerging and it is still possible to examine these factors without the bias created by chronic psychosis and relatives’ long-term burden.

Although patients’ poor clinical and functional status have been related with increased relatives’ EE in several early psychosis studies [10,19,20], other studies have suggested that patients’ symptoms/functioning have limited or no impact upon relatives’ EE [21-23]. The differences in the results among the above cited studies leaves unanswered the question as to the extent that EE is a reaction to the severity of the relative’s psychotic disorder [12].

The attributional model of EE suggests that relatives’ beliefs about the nature of the illness explains relatives’ emotional attitudes better than the unidirectional reactivity of relatives to patients’ illness characteristics [24]. Specifically, critical relatives are more prone to blame patients for their behaviors and perceive symptoms as controllable by patients rather than illness driven; as a result, they react with criticism to reduce undesired behaviors. In contrast, overinvolved relatives may believe that they have contributed in some way to the patients’ problems, so they often report high levels of self- control or self-blame attributions [25-27]. Some studies have empirically supported the attributional model in early psychosis. For example, critical relatives of FEP patients tend to believe that symptoms are within the patients’ control [28,29], and attributions of blame toward the patient have been shown to predict relatives’ criticism in the early stages of psychosis [16].

A different explanatory model of EE proposes that high-EE attitudes may represent a maladaptive attempt to cope with the stress of caring for an impaired ill relative; thus, EE behaviors could be conceived as a coping strategy used to reduce the perceived stress related to the caregiving role [21]. Accordingly, greater levels of psychological distress in caregivers seem to be related with increased levels of EE [16-18]. Converging evidence suggests that EOI is more related to distress than criticism [17,18,21], although criticism has also been linked to psychological distress in early psychosis caregivers [30]. Of note, some cross-sectional studies have pointed that relatives’ distress is an important predictor of criticism in early psychosis relatives [16,31].

Despite research aimed at disentangling the multiple correlates of EE, little is known about the developmental precursors of EE in early psychosis. To date, only a few longitudinal studies have explored potential predictors of criticism and EOI in relatives of FEP patients [21, 32]. Besides, the possible contribution of patients’ clinical and functional characteristics on EE in conjunction with relatives’ psychological factors has been scarcely investigated at these early stages [e.g., 21, 23]. Therefore, understanding the specific and common underlying factors of relatives’ criticism and EOI in the early stages of psychosis should improve the design of early family interventions, thereby enhancing the prognosis for both patients and their relatives.

The current study aims to examine the association of relatives’ EE with other family factors and patient’s clinical/functional status in incipient psychosis. It expands previous preliminary cross-sectional reports [16] by using an extended sample and a longitudinal design to test whether patient’s clinical/functional status and relatives’ psychological factors prospectively predict relatives’ EE at 6 months. The specific goals were to explore: (1) the association of patients’ clinical and functional status, as well as relatives’ psychological distress and illness attributions, with relatives’ EE dimensions (criticism and EOI) at baseline and at the 6-month follow-up; and (2) whether relatives’ psychological distress and illness attributions at the initial assessment predicted relatives’ EE dimensions both at baseline and follow-up assessments over-and-above patients’ baseline clinical and functional status.

Unlike previous preliminary cross-sectional studies, this study seeks to test a more comprehensive predictive model of EE in early psychosis by including both patients’ illness-related variables and relatives’ psychological factors. Moreover, the use of a longitudinal design allowed us to explore the assumption that relatives’ psychological factors predict EE levels (over-and-above patients’ clinical and functional features) across time at both subclinical and onset stages of psychosis.

Based on previous suggestions of the EE literature, we hypothesized that relatives’ baseline psychological distress and negative illness attributions would predict relatives’ EE dimensions over-and-above patients’ baseline clinical and functional variables at both time points. Based on the attributional model of EE and on previous findings, we also expected common as well as distinctive predictors of criticism and EOI at both baseline and follow-up assessments: (a) Both criticism and EOI would be predicted by relatives’ psychological distress (anxiety and depression); (b) Beliefs of self-blame, self-control, and emotional negative representation of the disorder would predict relatives’ EOI, whereas beliefs of control and blame toward the patient would predict relatives’ criticism.

Materials and methods

Participants and procedure

Ninety-one early psychosis patients (55 ARMS and 36 FEP) and their respective relatives were initially recruited in the present study. Of these, 46 family members (33 of ARMS and 13 of FEP) completed the 6-month follow-up assessment. Attrition results from the following reasons: 15 pilot participants completed the baseline assessment protocol but not the 6-months follow-up, 10 refused to participate, 10 did not complete the follow-up assessment, and 10 left the study because of patient withdrawal from the clinic or the study. Please note that patients’ data refers only to the baseline assessment.

Relatives were those who had most contact and/or the most significant relationship with the patient. Patients had to meet ARMS criteria as assessed by the Comprehensive Assessment of At-Risk Mental States (CAARMS) [33] and/or the Schizophrenia Proneness Instrument Adult-Version (SPI-A) [34]. FEP patients met DSM-IV-TR criteria [35] for any psychotic disorder or affective disorder with psychotic symptoms as established by the Structured Clinical Interview for DSM-IV (SCID-I) [36] and presented a first-episode of psychosis within the past two years. Mean duration of illness was 14 months (SD = 9.8), although 4 patients slightly exceed the 24-month period (range 1 to 29 months) and 2 reached a length of 33 and 34 months. Exclusion criteria for patients were (a) evidence of organically based psychosis, (b) any previous psychotic episode that involved pharmacotherapy, and (c) intellectual disability.

Seventy-eight relatives and forty-four patient-relative dyads were included in previously published preliminary studies exploring cross-sectional associations of the EE dimensions with patients' clinical/functional status [19] and differences on the relationship of EE with distress and illness attributions in early psychosis stages [16].

All participants were recruited within the Sant Pere Claver-Early Psychosis Program conducted in Barcelona (Spain) [37] and provided written consent to participate. The project was developed in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki). Ethical approval was granted by the Ethics Committee of the Unió Catalana d’Hospitals (Comitè d’Ètica d’Investigació Clínica (CEIC); number 09–40) and by the Ethics Committee of the Universitat Autònoma de Barcelona (Comissió d'Ètica en l'Experimentació Animal i Humana (CEEAH); number 2679). All the interviews were conducted by experienced clinical psychologists. The time gap between patients and relatives’ assessments ranged from 3 to 15 days.

Measures

Relatives’ measure at baseline and 6-month follow-up

EE was measured with the Family Questionnaire (FQ) [38] which consists of 20 items equally distributed in two subscales (criticism and EOI) scored on a 4-point Likert scale ranging from ‘never/very rarely’ to ‘very often’.

Patients and relatives’ measures at baseline

ARMS and FEP patients’ clinical status at baseline was rated with the Positive and Negative Syndrome Scale (PANSS) [39]. Patients’ current functional status was measured with the short version of the Social Functioning Scale (SFS) [40], a self-reported measure that assesses multiple facets of social adjustment.

Relatives’ distress was measured with the Depression and Anxiety subscales of the Symptom Checklist (SCL-90-R) [41], a self-report inventory intended to measure symptom intensity on a 5-point Likert scale. The Illness Perceptions Questionnaire for Schizophrenia-Relatives version (IPQS-R) [42] was used to measure relatives' beliefs about the disorder. Each item is rated from 1 ‘strongly disagree’ to 5 ‘strongly agree’. For the purposes of this study, we used the following subscales: personal control-patient and personal control-relative (control over the disorder), personal blame-patient and personal blame relative (blame toward the patient or self-blame about the disorder), and emotional representation of the illness (negative emotions about the disorder including sense of fear, frustration, anger, worry).

Statistical analysis

Pearson correlations were used to analyze the associations between each of the baseline predictors and relatives’ outcome variables (EE-criticism and EE-EOI) at baseline and at the 6-month follow-up. Hierarchical regression analyses were computed to predict relatives’ EE-criticism and EE-EOI at baseline and 6-month assessments using patients’ and relatives’ baseline predictors. The goal of the regression analyses was to test the extent to which relatives’ baseline predictors accounted for variance in relatives’ EE-criticism and EE-EOI (at baseline and at the 6-month follow-up) over-and-above patients’ baseline symptom severity and social functioning. The following steps were entered in all regression analyses. The PANNS total score was entered at step 1 to examine the variance accounted for by patients’ symptom severity. Patients’ SFS score was entered at step 2 to examine the variance accounted for by patients’ social functioning. Relatives’ Depression and Anxiety SCL-90-R subscales were entered at step 3 to examine the variance accounted for by relatives’ psychological distress. The five IPQ-S subscales were entered as a block at step 4 to examine the variance accounted for by relatives’ illness attributions.

Results

Sample characteristics

At baseline, relatives were mainly female (74.7%), particularly patients' mothers (71.4%), with the remaining caregivers being fathers (17.6%), siblings (5.5%), partners (4.4%) or adoptive parents (1.1%). Mean age of the relatives was 51.7 years old (S.D = 10.0). Patients were predominantly male (69.2%) and lived with their families (91.2%). The mean age of the patients was 22.4 years old (SD = 4.9). Over half of them (58.3%) were studying or had a job, 36.3% were unemployed/unoccupied, and 5.5% had a sick leave.

Of the 46 relatives that completed the 6-month follow-up assessment, most were also females (80.4%) and mothers who lived with the patient (89.1%). The mean age of the relatives at follow-up was 51.9 years old (S.D = 8.7).

Descriptive baseline data for all relatives’ and patients’ measures are presented in Table 1. The mean of the EE dimensions at follow-up was 20.52 (SD = 6.15) and 22.74 (SD = 5.48) for EE-criticism and EE- EOI, respectively. Dependent t-tests revealed that, on average, the levels of EE-criticism (t (45) = -0.670, p = 0.51) and EE-EOI (t (45) = 0.834, p = 0.41) did not differ significantly between baseline and follow-up assessments.

Table 1

Descriptive baseline data of early psychosis patients and their respective relatives (n = 91).

	α	Possible score range	Observed Score Range	Mean (SD)
Patients
Clinical status (PANSS)
Positive symptoms	-	7–49	7–24	12.74 (3.40)
Negative symptoms	-	7–49	7–34	17.40 (5.80)
General symptoms	-	16–112	18–66	32.93 (7.51)
Functional status (SFS)
Social Functioning	0.76	0–43	2–36	21.19 (6.12)
Relatives
Expressed Emotion (FQ)
Criticism	0.86	10–40	10–36	20.86 (6.15)
EOI	0.82	10–40	11–36	23.98 (5.82)
Distress (SCL-90-R)
Anxiety	0.90	0–40	0–34	7.13 (7.41)
Depression	0.91	0–52	0–39	14.73 (10.62)
Illness Attributions (IPQS-R)
Personal control-Patient	0.63	4–20	8–20	14.69 (2.58)
Personal control-Relative	0.61	4–20	6–20	13.45 (2.68)
Personal blame-Patient	0.81	3–15	3–15	9.93 (2.94)
Personal blame-Relative	0.83	3–15	3–15	7.45 (2.76)
Emotional representation	0.81	9–45	10–41	27.76 (7.10)

SD, Standard Deviation; PANSS, Positive and Negative Syndrome Scale; SFS, Social Functioning Scale; FQ, Family Questionnaire; EOI, Emotional Over-Involvement; SCL-90-R, Symptom Checkllist-90-Revised; IPQS-R, Illness Perception Questionnaire for Schizophrenia-Relatives' version.

Associations of EE with patients and relatives’ variables

Results of the association of patients’ and relatives’ baseline measures with relatives’ levels of criticism and EOI at both baseline and follow-up are provided in Table 2. Regarding the cross-sectional associations, only patients’ social functioning was significantly associated with relatives’ EE-criticism, such that worse social functioning in patients at baseline was related to relatives’ increased levels of criticism. No associations were found between patients’ baseline clinical status and relatives’ baseline EE indices. Furthermore, no longitudinal associations were found between patients’ variables at baseline and relatives’ EE at follow-up.

Table 2

Pearson correlations between baseline predictors (patient and relative factors) and relatives’ EE dimensions at baseline and follow-up.

	Baseline (n = 91)		Follow-up (n = 46)
	EE-Criticism	EE-EOI	EE-Criticism	EE-EOI
Patient factors
Clinical status (PANSS)
Positive symptoms	0.14	0.09	0.28	0.12
Negative symptoms	0.05	0.08	-0.08	0.06
General symptoms	0.15	0.11	0.25	0.02
PANSS total score	0.13	0.12	0.18	0.07
Functional status (SFS)
Social Functioning	-0.21*	-0.03	-0.12	-0.10
Relative factors
Distress (SCL-90-R)
Anxiety	0.60***	0.51***	0.61***	0.62***
Depression	0.59***	0.50***	0.63***	0.62***
Illness Attributions (IPQS-R)
Personal control-Patient	0.13	0.02	0.28	0.06
Personal control-Relative	-0.02	0.21*	0.09	0.29
Personal blame-Patient	0.27**	0.07	0.27	0.23
Personal blame-Relative	0.04	0.09	0.10	0.27
Emotional representation	0.44***	0.61***	0.52***	0.56***

*p<0.05

**p≤ 0.01

*** p<0.001.

EOI, Emotional Over-Involvement; PANSS, Positive and Negative Syndrome Scale; SFS, Social Functioning Scale; SCL-90-R, Symptom Checkllist-90-Revised; IPQS-R, Illness Perception Questionnaire for Schizophrenia-Relatives' version.

Medium effect sizes (r ≥ 0.30) in bold, large effect sizes (r ≥ 0.50) in bold and italics.

As for the associations between relatives’ baseline psychological distress and relatives’ EE dimensions, both relatives’ baseline levels of anxiety and depression were strongly associated with relatives’ EE dimensions at both time points. Moreover, relatives’ baseline attributions of blame toward the patient were significantly related with relatives’ EE-criticism at baseline, whereas relatives’ baseline attributions of control by the relative were significantly associated with relatives’ EE-EOI at baseline. Finally, relatives’ emotional negative representation about the disorder at baseline showed significant associations with both EE dimensions at both time points. Neither attributions of control toward the patient nor self-blame attributions were associated with EE indices.

Predictors of EE at baseline

Hierarchical regressions showed that neither patients’ baseline clinical status nor baseline psychosocial functioning accounted for significant variance in the prediction of baseline EE indices (see Table 3). Relatives’ baseline data revealed that high levels of anxiety (but not depression), attributions of blame toward the patient, emotional negative representation of the disorder as well as low scores on self-blame attributions significantly accounted for variance in relatives’ EE-criticism at baseline. In addition, high levels of anxiety (but not depression), attributions of control by the relative and emotional negative representation about the disorder significantly accounted for variance in relatives’ EE-EOI at baseline. However, ratings of relatives’ depression and anxiety were highly collinear, and they had comparable zero-order associations with EE.

Table 3

Predictors of EE at baseline (N = 91 patients and their respective relatives).

		EE-Criticism (at baseline)		EE-EOI (at baseline)
Step	Baseline Predictors	β	p	β	p
1	Patients’ clinical status (PANSS)
	PANSS total score	0.131	0.22	0.117	0.27
2	Patients’ functional status (SFS)
	Social Functioning	-0.188	0.10	0.012	0.92
3	Relatives’ distress (SCL-90-R)
	Anxiety	0.466	0.001	0.331	0.04
	Depression	0.196	0.17	0.223	0.17
4	Relatives’ illness attributions (IPQS-R)
	Personal control-Patient	0.116	0.15	0.048	0.58
	Personal control-Relative	0.001	0.99	0.207	0.02
	Personal blame-Patient	0.240	0.008	0.024	0.80
	Personal blame-Relative	-0.238	0.01	-0.160	0.10
	Emotional representation	0.250	0.004	0.465	0.000

EOI, Emotional Over-Involvement; PANSS, Positive and Negative Syndrome Scale; SFS, Social Functioning Scale; SCL-90-R, Symptom Checkllist-90-Revised; IPQS-R, Illness Perception Questionnaire for Schizophrenia-Relatives' version.

Significant at p<0.05 (two-tailed) are bolded.

Baseline predictors of EE at follow-up

Hierarchical regressions indicated that neither patients’ baseline clinical status nor baseline psychosocial functioning accounted for significant variance in the prediction of follow-up EE indices (see Table 4). Relatives’ high levels of anxiety (but not depression) and emotional negative representation about the disorder at baseline significantly accounted for variance in relatives’ EE-criticism at follow-up. For EE-EOI, relatives’ high levels of anxiety (but not depression), attributions of control by the relative and emotional negative representation about the disorder at baseline significantly accounted for variance at follow-up.

Table 4

Baseline predictors of EE at follow-up (N = 46 patients and their respective relatives).

		EE-Criticism (at follow-up)		EE-EOI (at follow-up)
Step	Baseline Predictors	β	p	β	p
1	Patients’ clinical status (PANSS)
	PANSS total score	0.178	0.24	0.070	0.65
2	Patients’ functional status (SFS)
	Social Functioning	-0.080	0.61	-0.082	0.60
3	Relatives’ distress (SCL-90-R)
	Anxiety	0.449	0.02	0.410	0.04
	Depression	0.279	0.144	0.326	0.10
4	Relatives’ illness attributions (IPQS-R)
	Personal control-Patient	0.185	0.13	-0.018	0.88
	Personal control-Relative	0.024	0.84	0.233	0.04
	Personal blame-Patient	0.141	0.28	0.101	0.42
	Personal blame-Relative	-0.089	0.50	-0.005	0.97
	Emotional representation	0.278	0.04	0.346	0.01

EOI, Emotional Over-Involvement; PANSS, Positive and Negative Syndrome Scale; SFS, Social Functioning Scale; SCL-90-R, Symptom Checkllist-90-Revised; IPQS-R, Illness Perception Questionnaire for Schizophrenia Relatives' version.

Significant at p<0.05 (two-tailed) are bolded.

Discussion

This study examined the association of relatives’ EE with psychological distress and illness attributions as well as with patients’ clinical features and functioning at baseline and at the 6-month follow-up. Furthermore, it was investigated whether relatives’ psychological distress and illness attributions predicted EE dimensions over-and-above patients’ baseline clinical and functional status at both time points. Overall, findings support the attributional model of EE as specific and differential attributions predicted criticism and EOI. In addition, emotional distress appears as a critical factor in the expression and maintenance of both EE dimensions in the early stages of psychosis. Evidence confirms that early family-based interventions have great potential to reduce high-EE attitudes and improve patients’ outcomes [43,44]. In general, family therapies aimed at reducing high-EE include training in family communication and problem solving in addition to psychoeducation [45,46]. However, relatives’ own needs and the emotional impact of caregiving are still a neglected intervention area in the early stages of psychosis [44]. Thus, these findings highlight that early family interventions would benefit from providing relatives with proper psychological support. This could involve helping relatives to handle harming thoughts and emotions, facilitate emotional processing, and provide specific techniques to reduce negative appraisals and stress. This, in turn, might prevent relatives’ high-EE over the psychotic process.

Consistent with some previous early psychosis studies [12,21,23], no associations were found between patients’ clinical status and relatives’ EE at baseline or follow-up. Of the patient-related variables, only patients’ poor social functioning was slightly related to relatives’ criticism at baseline, which is consistent with previous cross-sectional findings [19,23]. However, it is important to note that most early psychosis studies show inconsistent results regarding the relationship of EE with patients’ symptoms/functioning. Such differences might be related to the variability of the samples analysed (as some studies use a combined sample of early psychosis participants whereas others only examined ARMS or FEP participants separately), and to the diversity of designs employed (e.g., comparison of ARMS with FEP, ARMS and FEP vs. health controls, ARMS and FEP vs. chronic psychosis). Besides, it is likely that the heterogeneity characterizing psychosis, even in the early stages of the disorder make it difficult to find a conclusive pattern regarding the associations between patients’ illness-related characteristics and relatives’ EE.

The significant associations of relatives’ anxiety and depression symptoms with EE dimensions at both baseline and follow-up is consistent with previous cross-sectional findings [16–18, 30]. Although relatives’ distress and EE may have a complex pattern of interactions, our findings suggest that early psychosis has a profound impact on relatives’ emotional state that leads to significant levels of distress [47], which in turn may exacerbate EE attitudes.

Regarding the association between illness attributions and EE, findings indicated that attributions of blame toward the patient were significantly related with EE-criticism at baseline. This result support previous findings of cross-sectional studies [16,19] and lends further support to the attributional model [24], which states that relatives who believe that patients are guilty of their behaviors are more prone to manifest critical attitudes. Moreover, in agreement with Bolton et al. [26], it was found that relatives’ self-control attributions were significantly associated with EE-EOI at baseline. Thus, relatives’ who perceive themselves as responsible for the behaviors of their offspring are more likely to exhibit intrusive and/or self-sacrificing attitudes. Finally, strong relationships were observed between relatives’ negative emotional representation of the disorder and both EE dimensions at baseline and follow-up. This further supports the assumption that EE attitudes may be driven by relatives’ negative emotional responses to illness such as a sense of fear, frustration, anger or worry [21], which stands out as a crucial factor to consider in early family interventions.

The most relevant finding of this study is that relatives’ levels of anxiety and several illness attributions accounted for significant variance over-and-above patient-related variables in the prediction of criticism and EOI both at baseline and the 6-month follow-up. These results confirm that, as in schizophrenia, relatives’ emotional state, along with the cognitive representation of psychosis, are playing an important role in the emergence of emotional attitudes toward the patient. In particular, relatives’ increased levels of anxiety, attributions of blame toward the patient, decreased levels of self-blame attributions and an emotional negative representation of the disorder predicted EE-criticism at baseline. These results are in accordance with Domínguez-Martínez et al. [16] thus suggesting that anxiety responses and attributions of blame toward the patient play an important role in predicting relatives’ levels of criticism in early psychosis. These findings are also consistent the attributional model of schizophrenia [24], which posits that relatives who attribute responsibility or blame to the patients for their behaviors are more likely to exhibit critical attitudes towards them. The fact that decreased levels of self-blame attributions were predictive of relatives’ criticism at baseline is not surprising given that critical relatives tend to place the blame or attribute the responsibility of illness behaviors mainly to patients’ internal/personal factors [48], which could reduce feelings of guilt or responsibility for patients' problems. Finally, our results pointed out the role of relatives’ negative emotional representation of patients’ disorder as an important cognitive/affective predictor of EE-criticism at baseline.

Interestingly, in the regression analyses, the only predictors of relatives’ EE-criticism at follow-up were relatives’ anxiety levels and an emotional negative representation of the disorder. Attributions of blame toward the patient and/or low levels of self-blame attributions lost its predictive value at follow-up. This may reflect that relatives receive adequate information during treatment regarding the nature and causes of the behaviors of their ill relative, which helps them to better understand the cause of the problem and thus reduce their blaming attributions. Of note, emotional negative representation of illness, the only attribution that encompasses both cognitive and emotional components (e.g., sense of fear, frustration, anger, worry), predicted relatives’ criticism both at baseline and follow-up assessments over-and-above all the predictors in the model. This suggests that negative emotional responses (i.e., anxiety) and cognitions that are clearly intertwined with emotions (i.e., emotional negative representation of the disorder) persistently promote the maintenance of EE-criticism, even when ‘core’ cognitive attributions of blame no longer seem to have a relevant role. A similar pattern of results was observed for EE-EOI. Both relatives’ anxiety levels and emotional negative representation of the disorder appeared as significant predictors of relatives’ EOI at both assessments’ points, which further indicates that the emergence of criticism and EOI attitudes stems from negative emotional responses and negative ‘hot’ cognitions. Overall, these findings indicate that emotional distress is a critical element for understanding the expression and maintenance of EE in the early phases of psychosis and suggests that therapeutic efforts should address negative affective processes in addition to the more generally implemented psychoeducational techniques that predominantly focus on cognitive restructuring.

Consistent with the results of Bolton et al. [26], and lending further support to the attributional model, EE-EOI was also predicted by relatives’ self-control attributions both at baseline and follow-up, which may indicate that, unlike attributions of blame towards the patient, parental beliefs regarding their capacity to change their offspring’s situation are more difficult to change as they tap on core aspects of the parental function (e.g., protection). Thus, overinvolved behaviors might be conceptualized as reflecting a hyperactivation of the parental system in a moment of crisis. Therefore, the malleability of this attribution and its impact on parental behaviors might require overcoming the acute crisis state characterizing early psychosis.

Consistent with previous studies, attributions of control toward the patient did not predict relatives’ criticism either at baseline or follow-up [16,19]. These findings seem to suggest that attributions of blame toward the patient, rather than attributions of control, have a significant impact on relatives’ criticism. It is thus attractive to speculate that attributions of blame toward the patient may be emotionally-driven in response to crisis situation defining the early stages of psychosis, but that in later and chronic stages they may turn into stronger beliefs of patient’s personal control over the illness. Furthermore, contrary to our hypotheses, and in contrast with some schizophrenia studies [25,27], self-blame attributions did not predict relatives’ EOI either at baseline or follow-up. It might be that self-blame attributions tend to become more evident over time when diagnoses are clearly established, as relatives might then wonder about their contribution in the development of illness.

The strengths of this study include the use of a longitudinal design, which allowed to test prospectively the assumption that relatives’ attributions and distress predict EE levels across time in both subclinical and onset stages of psychosis. Besides, the present study examined a more comprehensive predictive model of EE in early psychosis by including both patients’ illness-related variables and relatives’ emotional distress and illness attributions. A limitation of this study is that the FQ does not contemplate the positive components of the EE construct (e.g., warmth). Considering that preliminary research emphasizes the protective effect of warmth in both chronic [49] and early psychosis samples [13,14, 50], further research should examine the psychological mechanisms underlying the expression of positive components of EE and its potential buffering effect of negative components on the course of early psychosis. Another limitation is that all measures administered to relatives (e.g., EE, attributions and distress) were assessed using self-reported scales, which raises concern about the contribution of shared method variance in the findings reported. The use of additional observational and/or interview ratings of these family constructs would strengthen the validity of these findings and should be pursued in future studies. Finally, it is important to note that content overlap among measures (and even constructs) of EE, anxiety and relatives’ emotional representation of the patients’ disorder may have partly contributed to the observed relationships. That is, some of the items of measures that assess different constructs present some overlap (i.e., they tap comparable content). A specific example is that both the EE-EOI subscale of the FQ [38] and the relatives’ emotional representation of the disorder subscale of the IPQS-R [42] contained items assessing relatives’ worry about the patient (e.g., an item of the EE-EOI subscale reads ‘I’m very worried about him/her’, and an item of the relatives’ emotional representation of the disorder subscale reads ‘Their mental health problems do not worry me’). However, it is important to note that this measurement overlap probably reflects to some extent a certain degree of construct overlap. EE, anxiety and relatives’ emotional representation of the disorder probably all partly draw on a general factor of relatives’ emotional distress.

In conclusion, this study shows that relatives’ psychological distress and negative illness attributions accounted for significant variance over-and-above patients’ clinical and functional status in the prediction of EE attitudes both at baseline and 6-month follow-up assessments. This highlights the importance of considering how relatives’ emotional states and the early formation of cognitive representations of psychosis can affect the attitudes they take towards the disorder at the at-risk and recent onset stages of psychosis. Moreover, these results underscore the need for early family interventions to go beyond educating relatives about psychosis [e.g., 51] to attend to their specific psychological needs. The findings of the present study suggest that providing relatives with stress management and problem-solving techniques [52] along with the use of reattribution techniques [24] might be beneficial to acquire a more balanced emotional and/or attributional stance. This, in turn, might help relatives to gain insight over their EE attitudes and manage them more effectively.

12 Feb 2020

PONE-D-19-32832

Predictors of criticism and emotional over-involvement in relatives of early psychosis patients

PLOS ONE

Dear Dr. Barrantes-Vidal,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by Mar 28 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Sarah Hope Lincoln

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors frame this study as examining the “mechanisms underlying the ontogenesis of expressed emotion (EE) in the early stages of psychosis.” This is clearly an important area of research as the findings have implications for intervention, particularly family interventions. The authors examine the role carers’ psychological distress and subjective appraisals of the illness relate to their level of criticism and emotional overinvolvement, two important predictors of illness course. The major strength of this study is the prospective design. They assessed not only the interrelation of the variables of interest at baseline but also at a six month follow-up. The authors found that family carers’ psychological distress (anxiety) and “negative emotional representation of the disorder” at baseline predicted 6 months later the carers’ level of criticism directed at the person with early psychosis. In addition, they observed that together a set of baseline indicators, specifically the relatives’ level of anxiety, their perceptions of their ability to control their loved ones mental health problems, and again the “negative emotional representation of the disorder” predicted their level of emotional overinvolvement at follow-up. The importance of the carers’ psychological distress in predicting both criticism and emotional overinvolvement calls for attending to caregivers’ distress in the early stages of the illness course of psychotic disorders. This strikes me as an important point.

Despite the strengths of the paper and its potential contribution, my enthusiasm for the paper is dampened by a number of weaknesses. Except for the symptomatology and social functioning scales, all the measures are based on the carers’ self report. This becomes a problem as there is some overlap between the measures, particularly family carers’ level of anxiety, “negative emotional representation of the disorder”, and emotional overinvolvement. The SCL-90R is a well-validated measure of anxiety symptoms. A look at the Illness perception questionnaire for schizophrenia (Lobban et al., 2005) reveals that of the nine items that assess “Emotional representation”, three include anxiety related items: “Their mental health problems do not worry me, their mental health problems make me feel anxious, and their mental health problems make me feel afraid.” A look at the 10 items to assess emotional overinvolvement (Wiedemann et al., 2002) also reveals anxiety related items. “I’m very worried about him/her” strikes me as the most obvious. Another is “I often think about what is to become of him/her,” although this is less direct. Clearly the measures of emotional representation and EOI include a number of what appear to be nonanxiety related items (e.g., the EOI scale has a number of items measuring self-sacrifice). The measure of emotional representation also overlaps with the measure of criticism. Two items stand out: “Their mental health problems make me feel angry” and “I get very frustrated by their mental health problems.” One item of the measure of criticism is “I’m often angry with him/her.” Two others seem related to frustration, those that refer to annoyance and irritation. My concern is that at least some of the observed associations may be a reflection of common method variance, not just regarding the use of self-report measures, but also regarding some of the measures content.

I found the speculation about the shift of attributions of blame to control over the illness course as problematic. First of all, the measures of controllability attributions in some of the prior studies of schizophrenia (e.g., Weisman et al, 1998, and Brewin et al., 1991) are quite different (based on coders’ ratings of statements generated from the Camberwell Family Interview). The authors would be on stronger grounds if the presuimed difference was observed across the illness course using similar methods. Second, I find the cognition-emotion distinction muddled in this speculation, and to some extent throughout the paper. Weiner’s attribution of responsibility model, which underlies some of the attribution-EE literature, makes a clear distinction between the appraiser’s cognitive construal and the subsequent affective reaction. To refer to emotional negative representation of illness as “the only attribution that encompasses both cognitive and emotional components (e.g., sense of fear, frustration, anger, worry) blurs the conceptual distinction between cognition and affect.

On a few occasions the authors point out that their findings are consistent with their prior work (Dominguez-Martinez et al., 2014; Dominguez-Martinez et al., 2017) and other’s work as well. For example they wrote on p. 13 that “The significant associations of relatives’ anxiety and depression symptoms with EE dimensions at both baseline and follow-up is consistent with previous cross-sectional findings [14-16, 18]. Also on page 14 they noted, “Regarding the association between illness attributions and EE, findings indicated that attributions of blame toward the patient were significantly related with EE-criticism at baseline. This result support (sic) previous findings of cross-sectional studies [14,17] and lends further support to the attributional model [22], which states that relatives who believe that patients are guilty of their behaviors are more prone to manifest critical attitudes. What concerns me is that earlier in the paper they report that the findings of the previous two studies were based on 78 relatives and 44 patient-relative dyads from the current study. That turns out to be 86% of the relatives (78 of 91) and 95% of the dyads (44 of 46). To refer to the findings from previous studies based on nearly the same sample as being supportive of the current findings is potentially misleading.

The last major point is that the authors refer to the ontogenesis and origin of expressed emotion in the first sentence of the abstract and the last sentence in the first paragraph of the discussion section. Indeed the researchers are studying these processes early in the illness course. The authors provide no evidence, however, that they captured the actual beginning of the caregivers’ attitudes and emotional reactions that reflect expressed emotion. They might want to be more circumspect in describing the study and interpreting their findings as an examination of the period of early psychosis, which they do throughout most of the paper.

Minor concerns

It would be helpful to briefly mention the ARMS criteria and the criteria they used to identify first episode psychosis. For example, was there any consideration of time since illness onset as seen in some first episode studies (Srihari et al., 2014)?

In assessing the carers’ level of anxiety, the authors report using SCL-90 R, yet they cite a much earlier paper that refers to the SCL-90 not the SCL-90R.

One limitation that the authors point out is that they did not consider the role of carers’ positive affect. The authors might consider other research that points out that carers’ supportive stance is associated with positive clinical outcomes for early stages of psychosis (Lee et al., 2014).

Reviewer #2: The present study is examining potential predictors of expressed emotion (EE), specifically emotional over involvement (EOI) and criticism in relatives of early psychosis patients. Overall, this is an interesting study and has important implications for early psychosis interventions. The manuscript is well-written, the justification for the study is clear, and is overall an interesting contribution to the growing field of understanding how family dynamics may interact with an existing vulnerability to severe mental illness. Nevertheless, the methodological concerns stated below should be emphasized. In addition, a few suggestions/comments:

Introduction

• In the first paragraph, please specify or give examples of the "macroenvironmental factors" referenced. Aside from EE, which is the focus of this paper, what else might be at play? Or is this primarily what the authors mean by macroenvironment?

• At the end of the first paragraph of the introduction, and throughout the paper, be careful not to imply blame on the family system for exacerbating symptoms or causing relapse, but rather clarifying that the family can play an important role in the course of the patient's illness.

• In the introduction, family therapy is briefly mentioned, and a specific study (O'Brien et al) is referenced at the very end of the discussion. Although the focus of this paper is not intervention-based, there are important intervention implications. It would be helpful to spend 1-2 lines (earlier than the last paragraph of the discussion) speaking more specifically about how EE is currently addressed in family therapy and how it could be improved with the findings from the current study.

Methods

• It was surprising that there was no inclusion of positive emotion or protective family factors (e.g. warmth, constructive communication) to use as a differential comparison and no mention of these factors until almost the very end of the discussion. These seems like a major limitation of the study (as is briefly mentioned) and should perhaps be addressed earlier on as to why they were not included. For example, I would be interested to know not only whether higher levels of distress and attributions of illness control predict EE but whether, differentially, lower levels of warmth and constructive communication also predict this. And, conversely, how levels of distress and illness attributions differentially predict factors like warmth and constructive communication. This double dissociation would be crucial to improving the impact of this finding.

• Why weren't demographic variables such as age and sex included as covariates in the hierarchical regression models? There seems to be some variance at least in age and it would be interesting and important to see whether this interacts with the variables of interest to predict EE.

Results

• In table 1, it would be helpful and more clear for the reader to include 6 month observations side-by-side with the baseline assessments for comparison, in addition to t-tests to compare measures across the two time points.

General

It would be helpful for the reader throughout the manuscript to make it more clear which measures refer to the patient, and which refer to the relatives. For example, in the abstract, the first line refers to the "ontogenesis of expressed emotion in the early stages of psychosis" but it could be made more clear that this is EE in relatives, not patients.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

30 Mar 2020

Summary of revisions to the manuscript, Predictors of criticism and emotional over-involvement in relatives of early psychosis patients (PONE-D-19-32832)

Response to Academic Editor

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Thank you for the constructive reviews of our manuscript, “Predictors of criticism and emotional over-involvement in relatives of early psychosis patients”. We are delighted to have the opportunity to submit a revised manuscript, and we have revised the manuscript according to the Reviewers’ useful recommendations and provide a detailed summary of the revisions below.

Response to Reviewer 1

1. The authors frame this study as examining the “mechanisms underlying the ontogenesis of expressed emotion (EE) in the early stages of psychosis.” This is clearly an important area of research as the findings have implications for intervention, particularly family interventions. The authors examine the role carers’ psychological distress and subjective appraisals of the illness relate to their level of criticism and emotional overinvolvement, two important predictors of illness course. The major strength of this study is the prospective design. They assessed not only the interrelation of the variables of interest at baseline but also at a six-month follow-up. The authors found that family carers’ psychological distress (anxiety) and “negative emotional representation of the disorder” at baseline predicted 6 months later the carers’ level of criticism directed at the person with early psychosis. In addition, they observed that together a set of baseline indicators, specifically the relatives’ level of anxiety, their perceptions of their ability to control their loved ones mental health problems, and again the “negative emotional representation of the disorder” predicted their level of emotional overinvolvement at follow-up. The importance of the carers’ psychological distress in predicting both criticism and emotional overinvolvement calls for attending to caregivers’ distress in the early stages of the illness course of psychotic disorders. This strikes me as an important point.

We thank the Reviewer for the positive feedback on the study and manuscript.

2. Despite the strengths of the paper and its potential contribution, my enthusiasm for the paper is dampened by a number of weaknesses. Except for the symptomatology and social functioning scales, all the measures are based on the carers’ self-report. This becomes a problem as there is some overlap between the measures, particularly family carers’ level of anxiety, “negative emotional representation of the disorder”, and emotional overinvolvement. The SCL-90R is a well-validated measure of anxiety symptoms. A look at the Illness perception questionnaire for schizophrenia (Lobban et al., 2005) reveals that of the nine items that assess “Emotional representation”, three include anxiety related items: “Their mental health problems do not worry me, their mental health problems make me feel anxious, and their mental health problems make me feel afraid.” A look at the 10 items to assess emotional overinvolvement (Wiedemann et al., 2002) also reveals anxiety related items. “I’m very worried about him/her” strikes me as the most obvious. Another is “I often think about what is to become of him/her,” although this is less direct. Clearly the measures of emotional representation and EOI include a number of what appear to be nonanxiety related items (e.g., the EOI scale has a number of items measuring self-sacrifice). The measure of emotional representation also overlaps with the measure of criticism. Two items stand out: “Their mental health problems make me feel angry” and “I get very frustrated by their mental health problems.” One item of the measure of criticism is “I’m often angry with him/her.” Two others seem related to frustration, those that refer to annoyance and irritation. My concern is that at least some of the observed associations may be a reflection of common method variance, not just regarding the use of self-report measures, but also regarding some of the measures content.

The reviewer points out very relevant issues. We completely concur with the Reviewer that findings may partly reflect shared method variance given that all measures administered to relatives had a self-report format. Also, content overlap among measures may have partly contributed to the observed relationships—and it could be argued that the boundaries among these constructs are themselves a challenging issue. Thus, we now make reference to these points in our discussion of the limitations of the study (page 17, lines 501 to 516) as follows:

“Another limitation is that all measures administered to relatives (e.g., EE, attributions and distress) were assessed using self-reported scales, which raises concern about the contribution of shared method variance in the findings reported. The use of additional observational and/or interview ratings of these family constructs would strengthen the validity of these findings and should be pursued in future studies. Finally, it is important to note that content overlap among measures (and even constructs) of EE, anxiety and relatives’ emotional representation of the patients’ disorder may have partly contributed to the observed relationships. That is, some of the items of measures that assess different constructs present some overlap (i.e., they tap comparable content). A specific example is that both the EE-EOI subscale of the FQ [38] and the relatives’ emotional representation of the disorder subscale of the IPQS-R [42] contained items assessing relatives’ worry about the patient (e.g., an item of the EE-EOI subscale reads ‘I’m very worried about him/her’, and an item of the relatives’ emotional representation of the disorder subscale reads ‘Their mental health problems do not worry me’). However, it is important to note that this measurement overlap probably reflects to some extent a certain degree of construct overlap. EE, anxiety and relatives’ emotional representation of the disorder probably all partly draw on a general factor of relatives’ emotional distress.”

3. I found the speculation about the shift of attributions of blame to control over the illness course as problematic. First of all, the measures of controllability attributions in some of the prior studies of schizophrenia (e.g., Weisman et al, 1998, and Brewin et al., 1991) are quite different (based on coders’ ratings of statements generated from the Camberwell Family Interview). The authors would be on stronger grounds if the presumed difference was observed across the illness course using similar methods. Second, I find the cognition-emotion distinction muddled in this speculation, and to some extent throughout the paper. Weiner’s attribution of responsibility model, which underlies some of the attribution-EE literature, makes a clear distinction between the appraiser’s cognitive construal and the subsequent affective reaction. To refer to emotional negative representation of illness as “the only attribution that encompasses both cognitive and emotional components (e.g., sense of fear, frustration, anger, worry) blurs the conceptual distinction between cognition and affect.

We concur with the Reviewer that our speculation about the possibility of a shift from attributions of blame to attributions of control over the illness course would be more grounded if there were previous findings supporting this notion that used similar methods to ours. However, please note that we make it very clear that this idea is offered as a speculation: “It is thus attractive to speculate that attributions of blame toward the patient may be emotionally-driven in response to the crisis situation defining the early stages of psychosis, but that in later and chronic stages they may turn into stronger beliefs of patient’s personal control over the illness”; page 16, lines 481 to 484). To our knowledge, there is only study measuring the change and/or stability of relatives’ illness attributions over time in early psychosis using the same method for the assessment of relatives’ attributions (Barrowclough et al., 2014). Specifically, this study found that most relatives’ illness attributions (including blame and control toward the patient) were stable over a period of 6 months in a sample of first episode of psychosis patients.

The cognition-emotion distinction presents a very interesting and very challenging issue. Indeed, Weiner’s attributional model (Weiner, 1986) assumes that attributions lead to emotions and this theory greatly influenced the development of the attributional model of EE (Barrowclough & Hooley, 2003). Nevertheless, it is important to note that the conceptualization of the attributional model of EE (Barrowclough & Hooley, 2003) was based on the subjective experience of schizophrenia caregivers (which could be significantly different from that of early psychosis caregivers). Relatives in at-risk and onset stages of the disorder are exposed for the first time to the early signs of psychosis. These potent signs of threat and biographical disruption trigger the development of cognitive appraisals about the causes of the disorder and, possibly at the same time, raise highly intense affective reactions. In fact, it is well established that early psychosis caregivers often report a heightened risk of psychological distress compared to family members of patients with chronic schizophrenia (Martins & Addington, 2001). Thus, it is likely that in the early stages of psychosis (when caregivers are dealing with abrupt disruptions and a peak of emotional distress), it is more challenging to establish a clear-cut unidirectional arrow from cognition (attributions) to emotions, and this possibly unique and intense emotional moment may be co-contributing to the shaping of attributions. in terms. This is indeed an interesting and unresolved debate, as it has been previously suggested that the attributional model of EE based on schizophrenia samples should be tailored to the developmental specificities of early psychosis (Domínguez-Martínez et al., 2017). Finally, we referred to emotional negative representation of illness as: “the only attribution that encompasses both cognitive and emotional components” because the 9 items encompassed in this subscale tap both cognitive and emotional content (i.e., I get depressed when I think about their mental health problems; When I think about their mental health problems I get upset; Their mental health problems make me feel angry; Their mental health problems do not worry me; Their mental health problems make me feel anxious; Their mental health problems make me feel afraid; Their mental health problems make me feel worthless; I get very frustrated by their mental health problems; I feel a sense of loss due to their mental health problems).

4. On a few occasions the authors point out that their findings are consistent with their prior work (Domínguez-Martínez et al., 2014; Dominguez-Martinez et al., 2017) and other’s work as well. For example, they wrote on p. 13 that “The significant associations of relatives’ anxiety and depression symptoms with EE dimensions at both baseline and follow-up is consistent with previous cross-sectional findings [14-16, 18]. Also, on page 14 they noted, “Regarding the association between illness attributions and EE, findings indicated that attributions of blame toward the patient were significantly related with EE-criticism at baseline. This result support (sic) previous findings of cross-sectional studies [14,17] and lends further support to the attributional model [22], which states that relatives who believe that patients are guilty of their behaviors are more prone to manifest critical attitudes. What concerns me is that earlier in the paper they report that the findings of the previous two studies were based on 78 relatives and 44 patient-relative dyads from the current study. That turns out to be 86% of the relatives (78 of 91) and 95% of the dyads (44 of 46). To refer to the findings from previous studies based on nearly the same sample as being supportive of the current findings is potentially misleading.

We apologize if we did not clearly describe the difference between previous cross-sectional reports and the current study. Thus, we summarize below the differences between the papers published by Domínguez-Martínez et al. (2014, 2017) and the present manuscript:

o In the present manuscript we used an extended sample of early psychosis patients and their respective relatives at baseline (i.e., 91 dyads) as compared to our previous cross-sectional studies (n=78 relatives in Domínguez-Martínez et al., 2017; n=44 dyads in Domínguez-Martínez et al., 2014). As mentioned, the study of Domínguez-Martínez et al. (2017) included only 78 relatives (not dyads). In contrast, the present manuscript included 91 dyads at baseline. Thus, the present manuscript not only includes more relatives at baseline (78 instead of 91 relatives), but also an extra sample of 91 early psychosis patients.

o With all due respect, we think that the Reviewer may have been confused when comparing the dyads included by Domínguez-Martínez et al. (2014) with the dyads included in the present study. Specifically, the reviewer has compared the 44 dyads included in the cross-sectional study of Domínguez-Martinez et al. (2014) with the 46 family members included in the follow-up assessment of the present study. Actually, the cross-sectional study of Domínguez-Martínez et al. (2014) included 44 dyads. In contrast, the present study included 91 dyads at baseline. Therefore, that turns out to be 48% of the dyads (44 instead of 91 dyads), not the 95% of the dyads referred by the Reviewer.

o Finally, this manuscript presents, for the first time, longitudinal data. We have never reported any data or findings for the 46 relatives who were reassessed regarding EE dimensions at a 6-month follow-up assessment.

5. The last major point is that the authors refer to the ontogenesis and origin of expressed emotion in the first sentence of the abstract and the last sentence in the first paragraph of the discussion section. Indeed, the researchers are studying these processes early in the illness course. The authors provide no evidence, however, that they captured the actual beginning of the caregivers’ attitudes and emotional reactions that reflect expressed emotion. They might want to be more circumspect in describing the study and interpreting their findings as an examination of the period of early psychosis, which they do throughout most of the paper.

We thank the Reviewer for pointing this out and we completely concur with the Reviewer’s opinion. Following the Reviewer’s observation, we have rewritten the first line of the abstract (page 2, line 43) and the line that the Reviewer has indicated in the discussion (page 13, line 375) in order to be more precise.

Please note that, we have also replaced the word origins by expression in another sentence of the discussion (page 16, line 466).

6. It would be helpful to briefly mention the ARMS criteria and the criteria they used to identify first episode psychosis. For example, was there any consideration of time since illness onset as seen in some first episode studies (Srihari et al., 2014)?

Both ARMS and FEP inclusion/exclusion criteria were already reported in the manuscript’s method section (page 6, lines 222 to 230). Following the Reviewer’s recommendation, we have now added the amount of time considered since illness onset to include FEP patients in the study (page 6, lines 226 to 228) as follows:

“FEP patients met DSM-IV-TR criteria [35] for any psychotic disorder or affective disorder with psychotic symptoms as established by the Structured Clinical Interview for DSM-IV (SCID-I) [36] and presented a first-episode of psychosis within the past two years. Mean duration of illness was 14 months (SD=9.8), although 4 patients slightly exceed the 24-month period (range 1 to 29 months) and 2 reached a length of 33 and 34 months.”

7. In assessing the carers’ level of anxiety, the authors report using SCL-90 R, yet they cite a much earlier paper that refers to the SCL-90 not the SCL-90R.

We thank the Reviewer for pointing this out. It has now been corrected in the revised manuscript.

8. One limitation that the authors point out is that they did not consider the role of carers’ positive affect. The authors might consider other research that points out that carers’ supportive stance is associated with positive clinical outcomes for early stages of psychosis (Lee et al., 2014).

We appreciate the Reviewer’s suggestion to consider other research examining how family positive attitudes relate with positive clinical outcomes in early psychosis. Following the Reviewer’s recommendation, we have now included the research of Lee et al. (2014) in the revised manuscript (page 17, line 499).

Response to Reviewer 2

1. The present study is examining potential predictors of expressed emotion (EE), specifically emotional over involvement (EOI) and criticism in relatives of early psychosis patients. Overall, this is an interesting study and has important implications for early psychosis interventions. The manuscript is well-written, the justification for the study is clear, and is overall an interesting contribution to the growing field of understanding how family dynamics may interact with an existing vulnerability to severe mental illness. Nevertheless, the methodological concerns stated below should be emphasized. In addition, a few suggestions/comments.

We thank the Reviewer for his/her positive comments on the manuscript.

2. Introduction. In the first paragraph, please specify or give examples of the "macroenvironmental factors" referenced. Aside from EE, which is the focus of this paper, what else might be at play? Or is this primarily what the authors mean by macroenvironment?

We thank the Reviewer for pointing this out. We have now provided a specific example of macroenvironmental factors (i.e., urbanicity) in the manuscript (page 3, line 72). In answer to the Reviewer’s question, we would like to clarify that we were not referring to EE as a macroenvironmental factor. Actually, EE is conceptualized as a microenvironmental risk factor, as it pertains to the person rather than social level of the environment (e.g., Sheinbaum & Barrantes-Vidal, 2015). We have now specified this in the revised manuscript (page 3, line 73) to make the text clearer.

3. Introduction. At the end of the first paragraph of the introduction, and throughout the paper, be careful not to imply blame on the family system for exacerbating symptoms or causing relapse, but rather clarifying that the family can play an important role in the course of the patient's illness.

We agree with the Reviewer that it is extremely important to clarify this aspect. We have now addressed this issue in the Introduction section (page 3, lines 79 to 82). The text now explicitly states that families are not guilty or are not the primary causal agent of exacerbating symptoms or causing patient’s relapse in the following way:

“However, it is important to note that the EE construct does not aim to blame families for contributing unidirectionally to the patient’s clinical worsening. Instead, EE is best regarded as the product of a negative dynamic interaction between patients and their families that can play a central role in the course of the patient’s disorder [6,7].

Since the main goal of the manuscript was to examine the mechanisms underlying the manifestation of EE in early psychosis, rather than analyzing the effect of EE on patients’ clinical outcomes, we consider that it is not necessary to include more statements about this topic in the paper (as the Reviewer suggested).We think that we have repeatedly emphasized the critical importance of considering caregivers’ psychological needs in the early stages of psychosis and that we have been careful about not laying the blame on the family system.

4. Introduction. In the introduction, family therapy is briefly mentioned, and a specific study (O'Brien et al) is referenced at the very end of the discussion. Although the focus of this paper is not intervention-based, there are important intervention implications. It would be helpful to spend 1-2 lines (earlier than the last paragraph of the discussion) speaking more specifically about how EE is currently addressed in family therapy and how it could be improved with the findings from the current study.

We completely concur with the point made by the Reviewer. We have now included the following text in the Discussion section (page 13, lines 375 to 382):

“Evidence confirms that early family-based interventions have great potential to reduce high-EE attitudes and improve patients’ outcomes [43,44]. In general, family therapies aimed at reducing high-EE include training in family communication and problem solving in addition to psychoeducation [45,46]. However, relatives’ own needs and the emotional impact of caregiving are still a neglected intervention area in the early stages of psychosis [44]. Thus, these findings highlight that early family interventions would benefit from providing relatives with proper psychological support. This could involve helping relatives to handle harming thoughts and emotions, facilitate emotional processing, and provide specific techniques to reduce negative appraisals and stress. This, in turn, might prevent relatives’ high-EE over the psychotic process.”

5. Methods. It was surprising that there was no inclusion of positive emotion or protective family factors (e.g. warmth, constructive communication) to use as a differential comparison and no mention of these factors until almost the very end of the discussion. These seems like a major limitation of the study (as is briefly mentioned) and should perhaps be addressed earlier on as to why they were not included. For example, I would be interested to know not only whether higher levels of distress and attributions of illness control predict EE but whether, differentially, lower levels of warmth and constructive communication also predict this. And, conversely, how levels of distress and illness attributions differentially predict factors like warmth and constructive communication. This double dissociation would be crucial to improving the impact of this finding.

We completely concur with the point made by the Reviewer about the huge interest of including protective family factors. Indeed, the assessment of positive family environment variables would have provided a complementary, enriched understanding of the research question. However, as mentioned in the limitations of the study, the only available validated measure to examine relatives’ EE [i.e., Family Questionnaire (FQ; Wiedemann et al., 2002)] at the outset of the study does not assess the positive components of the EE construct (i.e., warmth, positive comments, constructive communication). Please note that these data were collected over a longitudinal study and that, unfortunately, the critical importance of positive/protective variables at the time when this study started was much less realized than what it is now—and, consequently, the number of available measures was also scarce.

We appreciate the suggestion of addressing why positive family factors were not included earlier in the text. However, we honestly think that it is very clear since we define the goals and hypotheses that the study is exclusively focused on the negative components of the EE construct. Finally, as the Reviewer has indicated in detail (and we also globally suggested in the discussion: page 17, lines 499-501), future follow-up studies will be able to provide relevant information on a range of positive family factors (e.g., warmth, positive comments, constructive communication) that might affect the expression of the negative attitudes of the EE construct (i.e., criticism and emotional over-involvement).

6. Methods. Why weren't demographic variables such as age and sex included as covariates in the hierarchical regression models? There seems to be some variance at least in age and it would be interesting and important to see whether this interacts with the variables of interest to predict EE.

Please note that we did not include any demographic variables (e.g., sex or age) in our analyses because we did not have any specific a priori hypotheses about these variables. Given the number of analyses performed to address our specific research questions, we limited our analyses only to testing a priori hypotheses. Given that the inclusion of sex or age would have been exploratory and given that the study was not designed to sample and test hypotheses about these variables, we did not believe it would have been appropriate to include them in the analyses.

It is also worth noting that the issue of “adjusting” for confounders presents some conceptual and methodological challenges. As noted, given that we did not have hypotheses about sex or age, it would not be entirely clear why we would adjust for them. Secondly, the practice of examining covariates or “adjusting” for them is not completely unequivocal. Partialling out these variables from our hierarchical linear analyses would indeed test whether significant findings occurred over-and-above the effects of sex or age. However, as noted in the literature, such analyses (partialling or covarying) do not “adjust” or “correct” for third variables (in the sense that it does not equate for them). Rather such analyses examine partialled or residualized effects. However, such partialling can fundamentally change the psychometric properties and the conceptual nature of variables. Such analyses also do not provide any information about whether the effects of interest differ between males and females. This would require examining the interaction of sex and age on outcome measures (over and above the component main effects). Again, we did not have any hypotheses about these relationships and did not design the study to test such effects. Therefore, we felt it was important to avoid conducting unhypothesized, exploratory analyses.

7. Results. In table 1, it would be helpful and more clear for the reader to include 6 month observations side-by-side with the baseline assessments for comparison, in addition to t-tests to compare measures across the two time points.

Please note that, as mentioned in the manuscript, relatives’ EE dimensions (i.e., criticism and EOI) were the only variables reassessed at the 6-month follow-up. The rest of the study variables were only assessed at baseline (i.e., patients’ clinical/functional status, relatives’ distress and relatives’ illness attributions). Thus, we think that it is not necessary to include an extra column in Table 1 for reporting only the descriptive statistics of EE dimensions at follow-up. On the other hand, and following the Reviewer’s recommendation, we have now provided in the text the descriptive statistics (Mean, SD) of the EE values at follow-up (page 8, lines 300 to 301). The comparison of EE levels across the two time points was already informed in the text (page 8, lines 301 to 303) by including the results obtained with the Dependent t-test statistics.

8. General. It would be helpful for the reader throughout the manuscript to make it more clear which measures refer to the patient, and which refer to the relatives. For example, in the abstract, the first line refers to the "ontogenesis of expressed emotion in the early stages of psychosis" but it could be made more clear that this is EE in relatives, not patients.

Thank you for raising this opportunity to present findings is a more precise way. Following the Reviewer’s recommendation, we have now specified in parts of text that might read ambiguous when we were referring to the relatives' variables in order to avoid any confusion (page 2, line 43; page 5, line 187; page 5, line 190; page 15, line 446)

Cited references:

o Barrowclough C, Gooding P, Hartley S, Lee G, Lobban F. Factors associated with distress in relatives of a family member experiencing recent-onset psychosis. J Nerv Ment Dis. 2014;202: 40–46.

o Barrowclough C, Hooley JM. Attributions and expressed emotion: A review. Clin Psychol Rev. 2003; 23:849–880.

o Martens L, Addington, J. The psychological well-being of family members of individuals with schizophrenia. Soc Psychiatry Psychiatr Epidemiol. 2001; 36:128–133.

o Hinojosa-Marqués L, Domínguez-Martínez T, Kwapil TR, Barrantes-Vidal N. Ecological Validity of Expressed Emotion in Early Psychosis. Front. Psychiatry. 2019; 10:854.

o Domínguez-Martínez T, Medina-Pradas C, Kwapil TR, Barrantes-Vidal N. Relatives’ illness attributions mediate the association of expressed emotion with early psychosis symptoms and functioning. Psychiatry Res. 2014; 218:48–53.

o Domínguez-Martínez T, Medina-Pradas C, Kwapil TR, Barrantes-Vidal N. Relatives’ expressed emotion, distress and attributions in clinical high-risk and recent onset of psychosis. Psychiatry Res. 2017; 247:323–329.

o Lee G, Barrowclough C, Lobban F. Positive affect in the family environment protects against relapse in first-episode psychosis. Soc Psychiatry Psychiatr Epidemiol. 2014; 49: 367–376.

o Sheinbaum T, Barrantes-Vidal N. Mechanisms mediating the pathway from environmental adversity to psychosis proneness. In: Mason O, Claridge G. editors. Schizotypy: New dimensions. Oxford: Routledge; 2015. pp. 116-131.

o Wiedemann G, Rayki O, Feinstein E, Hahlweg K. The Family Questionnaire: Development and validation of a new self-report scale for assessing expressed emotion. 2002;109: 265–279.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

26 May 2020

Predictors of criticism and emotional over-involvement in relatives of early psychosis patients

PONE-D-19-32832R1

Dear Dr. Barrantes-Vidal,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Sinan Guloksuz, M.D., Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #2: No

5 Jun 2020

PONE-D-19-32832R1

Predictors of criticism and emotional over-involvement in relatives of early psychosis patients

Dear Dr. Barrantes-Vidal:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Sinan Guloksuz

Academic Editor

PLOS ONE