Expressed Emotion (EE) is a construct that describes how a family member or caregiver, when speaking, expresses emotions about another family member with a mental disorder. Numerous definitions of EE have been proposed: Vaughn [1] defined it as “an index of the emotional temperature in the family environment, an indicator of the intensity of the emotional response of the family member, which reveals a lack of affection or his overly intrusive concern”. Bertrando [2] described it as “the measurement of certain characteristics of the family emotional environment in the course of various pathologies, disorders or problems, generally—but not exclusively—psychiatric”. Finally, Peters, Calam, and Harrington [3] proposed it as an “umbrella” comprising a set of behaviors, attitudes, and emotions within familial relationships. In short, EE is the “expressive way of reaction of family members” [4].

The idea that EE might predict mental disorder relapse originated in the 1950s, from the observations of George Brown at the Medical Research Council Social Psychiatry unit in London [5]. Following the introduction of first-generation anti-psychotics (i.e., chlorpromazine), schizophrenic patients began to be discharged from hospitals. Brown noted that patients who were reintegrated into the family context were significantly more likely to present an aggravation of symptoms—sometimes requiring re-hospitalization—relative to patients who were reintegrated into other public structures. This observation led him to hypothesize that patients’ family environment could contribute to predicting a return of their symptoms. EE was thus born as an empirical, predictive index, and the need to measure it led to the development of rating scales and interviews. It was Brown himself who developed the first coding scheme, the Camberwell Family Interview (CFI; [6], later modified by Vaughn & Leff ^[7][8][7,8]), which is still considered the conventional model for measuring EE.

The CFI is administered to the caregiver after the patient’s first hospitalization. The semi-structured interview, which generally takes 30 min to administer, does not include pre-established questions, but merely guidance on eliciting information from the caregiver about the patient’s onset of disease symptoms in the months preceding the worsening of their disease or their re-hospitalization. The interview is audio-recorded and later coded—a process that generally takes 45 min. EE is then estimated on the basis of five coded dimensions: Emotional Overinvolvement, Criticism, and Hostility contribute to raising EE; and Positive Comments and Warmth contribute to reducing EE. Prior to administering the measure, interviewers must complete 1–2 weeks of training.

A second measure of EE, Level of Expressed Emotion (LEE), is a self-report questionnaire comprising 60 items that measure the emotional environment of family relationships. Four parameters are analyzed with reference to the caregiver, with each assessed using 15 items: Intrusiveness, Emotional Response, Emotional Attitude, and Tolerance/Expectations. All items are rated as true or false. Two versions of the LEE are available: one that is administered to caregivers and one that is administered to patients.

The Perceived Criticism (PC) scale is the fastest measure of EE (1 min), but it only assesses a single dimension of EE: Criticism. The measure is administered via two questions—one directed to the caregiver (1), and one directed to the patient (2):

(1)

How critical do you think you are to (name of caregiver)?

(2)

How critical do you think (name of caregiver) is toward you?

Answers are coded on a 10-point Likert scale ranging from 1 (not at all critical) to 10 (very critical).

The Five-Minute Speech Sample (FMSS) asks the caregiver to talk about their thoughts and feelings about the patient for 5 min, without interruption. Originally developed by Gottschalk and Gleser [9] to assess anxiety, hostility, and hope, it was later proposed by Magana et al. [10] as a specific measure of EE, for which they designed a coding scheme. The FMSS interview is audio-recorded and coded according to Magana’s [11] scoring system. The coding generally takes 20 min to complete. The measure analyzes five parameters of EE: Initial Statement, Relationship, Criticism, Dissatisfaction, and Emotional Overinvolvement (and thereby only assesses the Criticism and Emotional Overinvolvement dimensions). A scale of values is established for each parameter, and scores determine whether respondents exhibit either low (LEE) or high EE (HEE). Training is mandatory to code the FMSS, but not to administer it. The training consists of approximately 20 h of classroom teaching and 15 h of individual study (coding speech samples). Training seminars take place over 5 days and are usually completed within 1–2 weeks.

2. FMSS

Several studies have shown that the FMSS is a more convenient measure of EE than the CFI. Training in coding the FMSS to measure EE takes approximately 20 h, as opposed to 70 h for the CFI. Furthermore, the actual practice of FMSS coding takes approximately 20 min, whereas CFI coding requires 4–5 h ^[12][43]. Concerning administration time, the FMSS takes 5 min, and the CFI involves approximately 2 h. However, FMSS scores for EE correspond to those produced by the CFI, although caregivers classified as HEE by the CFI tend to be under-identified by the FMSS, thus, producing false negatives ^[13][44]. The rapid administration of the interview and coding, and thus, the easy accessibility (without the sacrifice of the interview method) of the FMSS make it a convenient measure for EE. Compared to the CFI, which represents the conventional instrument, the FMSS assesses the dimensions of Criticism and Emotional Overinvolvement, only, and the categories of Initial Statement, Relationship, and Dissatisfaction. It does not consider the dimensions of Warmth and Positive Remarks, which are considered irrelevant for the assessment of EE, as well as Hostility, given that the brevity of the interview results in a relative infrequency of hostile comments ^[14][45]. Moreover, in light of the FMSS interview structure, based on a free association of ideas, the measure may not only assess EE, but it may also analyze the psychopathological structure of caregivers themselves. As a consequence, it may be possible to apply family psychoeducation interventions, jointly with patient’s pharmacological and psychological treatments, to manage emotional reactions, both to minimize relapse in patients and to benefit caregivers. Such interventions might reinforce dimensions of EE that are configured as protective factors and reduce domains described as risk factors, with the ultimate goal of effecting a renewed therapeutic relationship in the family context. Psychoeducation consists of the provision of information (about the disease, including its possible course and the consequences of non-adherence to pharmacological therapy) and instructions to prevent and manage mental and psychological disorders. This is particularly relevant, as poor compliance with pharmacological therapy has been identified as the primary cause of relapse ^[15][46]. Over the past 10 years, several studies have been conducted with the aim of reducing detrimental EE domains (i.e., Emotional Overinvolvement, Criticism, Hostility) through family psychoeducation, and the effect on global EE has been equivocal. However, a very recent meta-analysis ^[16][12] showed that, with schizophrenic patients, reduction in Criticism was associated with a reduction in global EE and a lower risk of relapse. Furthermore, psycho-educated families could benefit from cognitive-behavioral family psychotherapy, as illustrated by two recent reviews showing that this form of therapy is effective in promoting and improving family dynamics. Finally, it would be helpful to evaluate EE as a predictor of not only relapse, but also psychosis onset.