This is a cross-sectional survey with data collected between February and March 2019. This study is part of a larger research entitled: Voice Hearers - New Approaches to Mental Health. In the first stage of this research, which took place between September 2017 and May 2018, all the medical records of the users of a Psychosocial Care Center II in the city of southern Brazil were read. In this period there were 400 active users’ records in the service and of these, 11 presented incomplete information. For the second stage, among the remaining 389, only those who had a voice hearing record were identified, totaling a sample of 172 users. 

The users diagnosed with mental retardation and those who were no longer attending the service at the time of the interview were excluded, who totaled 46 users and 14 refusals. From the 112 who had a record of hearing voices in their medical records, 66 reported still hearing voices during the period of application of the questionnaires. Interviewers were then trained to apply a questionnaire to the voice hearers included in the survey, at CAPS or at the user’s home.

To compose the sociodemographic profile of the hearers the following variables were analyzed: gender; age; skin color; educational level - knowing how to read and write, schooling by years of study; marital status/situation; and work situation. In collecting the characteristics of the voices, the following variables were used: voices manifest themselves: inside your head; through your ears; another place inside your body; a place outside your body; you can indicate to whom the voices belong; the voices come from you or from another person; the number of voices you hear today; the voices report any kind of experiences, bizarre or extraordinary, recent or past; the number of voices has ever changed; the number of voices you heard at first; the voices are named; the gender of the voices; the age of the voices; manner or tone of voices resembles someone known; how many times a day you hear the voices; the voices resemble someone; time of day the voices consume; when the voices appear; the voices appear when you are feeling emotions; the voices say positive things, negative things; fear of the voices; the voices confuse; they hinder your daily activities; they give orders; they annoy them; type of orders, if the hearer fulfills the orders and when they fulfill it.

 After the collection, a database was built, and double typing was performed in the Epidata software. A descriptive analysis was carried out obtaining the proportions in Stata 11. The research was submitted to and approved by the Research Ethics Committee. There was no funding from a public or private institution for the development of the study.

As can be seen in Table 1, among the 112 users interviewed, 60% (n=66) reported hearing voices at the time of the interview. Among them (66), most were female (67%), aged between 43 and 64 years (60%), white skin color (62%), who can read and write (95,5%), single (51,5%) and unemployed (62%).

Table 1

Baseline characteristic	(n/%)
Gender
Male	22 (33%)
Female	44 (67%)
Age group
21 to 31	6 (10%)
32 to 42	16 (24%)
43 to 53	20 (30%)
54 to 64	20 (30%)
65 to 75	4 (6%)
Skin Color
White	41(62%)
Black	9 (14%)
Brown	11 (17%)
Yellow	3 (4%)
Indigenous	2 (3%)
Can read and write
No	3 (4.5%)
Yes	63 (95.5%)
Marital Status
Single	34 (51.5%)
Married	15 (23%)
Separated / Divorced	12 (18%)
Widowed	5 (7.5%)
Work
No	42 (62%)
No, but is retired	10 (15.5%)
No, but receives benefit	9 (14.5%)
Yes	5 (8%)

More than half of hearers (n= 35/55%) related hearing the voices with some experience of their lives. Among these, 26 reported the type of experience: (a) as a result of a physical injury or serious illness (15%); (b) when you went to live alone for the first time (4%); (c) divorce or the end of a close/near relationship (15%); (d) pregnancy or abortion (4%); (e) the death of a close friend or family member (38%); (f) fell in love and was rejected (4%); (g) attending a session, satanic ritual, spiritual event (4%); (h) during great tension in your home or relationship (16%).

Regarding the characteristics of the voices, 92% reported hearing the voices through the ears, 69% inside the head, 75% from somewhere outside their body and 17% from somewhere else inside their body. Regarding the origin of voices, 69% said that the voices come from themselves or someone else and 46% can indicate to whom the voices belong. As for the number of voices they hear, 39% can’t identify at present, 47% has never been able to identify and 52% has already had a change in the number of voices they hear. About 80% do not give names to the voices. For 52% the voices are similar to the parents and 42% believes they are similar to the brother or sister. For more than 60% of hearers the voices are male and they are on average 40 years old. Most hear the voices once a day (34%) for up to 30 minutes (46%). For 80%, voices usually appear when they are alone (73%) and when they feel depressed and insecure (74%). For 81% of hearers the voices say negative things. Over 70% of hearers reported that voices confuse them, 69% that hinder their daily activities, 73% that annoy them and 59% that give orders. Regarding the orders given by the voices, more than half of the hearers comply with what is ordered to them, with the highest frequency (37%) (Figure 1). The action most carried out after the orders are the attempted suicide.

Figure 1

As noticed in other research on voice hearers (Beavan, 2011; Couto & Kantorski, 2020; Shevlin et al., 2007; Sidgwick, 1894), in this study, the majority (67%) of users who heard voices were female. Sidgwick's study (1894) analyzed a sample of 17.000 people and showed a 12% prevalence of voice hearing in women compared to 7.8% in men. Similarly, in the study by Shevlin et al. (2007), in a sample of 5.893 people, a prevalence of 59% of voice hearing in women was found. Although on a smaller scale, the studies by Beavan (2011) and Couto and Kantorski (2020) also met these results. The former found a 60% prevalence of hearing voices in women in a sample of 50 people, while the latter, when conducting semi-structured interviews with 16 voice hearers using a Psychosocial Care Center (CAPS), found that 11 were women, and only five were men. It is believed that these data may be related to the fact that women have been assuming an increasing number of roles in society, accumulating tasks and responsibilities that often lead to physical and psychological burden, which can culminate in mental illness. In addition, according to the latest data from the National Health Survey (Instituto Brasileiro de Geografia e Estatística [IBGE], 2020), women are the ones who most seek health services, which influences the number that shows that women hear more voices than men.

In contrast, in the study by Connor and Birchwood (2011), which analyzed a sample of 74 UK clinical voice hearers, it was found that the majority (60%) of participants were male. The study also assessed the age, marital status and employment status of the participants, and found an average age of 43, with more than half being single and 87% unemployed, data that are in line with those found in this study. Similar results were presented by Peters et al. (2016), when comparing clinical hearers (n=84), non-clinicians (n=92), and control group (n=83), showed that 91% clinicians had never worked or were unemployed for a long period, while for the non-clinicians, this percentage dropped to 32%, and reached 23% among the control group. This consistency between the data reinforces the idea that the illness of subjects who, due to age, would be economically active, as well as unemployed, are factors that worsen the prognosis for producing feelings of incapacity, unproductiveness, and dependence.

Regarding schooling, it was found in this survey that the majority (97%) knew how to read and write and had from one to eight years of study (67%), which is equivalent to incomplete elementary education. No studies were found with clinical hearers who analyzed this variable reinforcing the importance of including the variable education in research, since it is closely linked to the subjects’ socioeconomic conditions. Finally, regarding the variable skin color, in the study by Peters et al. (2016), the prevalence of white-skinned individuals in the three groups analyzed was verified: 90.4% in the control group, 87% in the non-clinical hearers and 65.5% in the clinical hearers, in line with the results of this research.

In view of the above, these characteristics (gender, age, work/income, education, skin color) should be observed and analyzed - in addition to their compulsory completion in medical records - when it is intended to identify factors associated with listening to voices in users of mental health services, since inadequate living conditions, resulting from social inequality, contribute to the emergence of psychic suffering in general.

Regarding the origin of voices, there is currently a great deal of research showing that there is an association between traumatic experiences and hearing voices (Andrew et al., 2008; Daalman et al., 2013; Kantorski et al., 2018; Varese et al., 2012). The study by Varese et al. (2012) consisted of a meta-analysis of 41 works that found evidence that adversities experienced in childhood are associated with a higher risk of developing some type of psychosis. The types of traumatic experiences identified in these studies were: sexual abuse, physical abuse, emotional abuse, bullying, death of parents and neglect, the former being the most prevalent. Similarly, Andrew et al. (2008), when comparing two groups of hearers, one consisted of people who used mental health services and had predominantly negative interpretations of voices, and the other of people who did not use the services and had positive experiences with voices, found that both groups had high rates of traumatic events, with reports of sexual abuse prevalent in the first group.

Unlike these results, in this study, the death of a friend or relative proved to be the triggering event for the most frequent hearing of voices, as well as in the study by Kantorski et al. (2018). The authors analyzed 21 interviews conducted with Italian voice hearers with the aim of identifying, in their life history, the situations related to the origin of voices. There were situations in which the origin of these was linked to work issues, such as the case of one of the interviewees who reported being overburdened when one of his colleagues left maternity leave. Situations of illness were also identified, such as the case of two interviewees who began to hear voices after some period of coma and situations of loss of close relatives and friends. In this case, the authors obtained the same number of reports referring to situations of illness and loss as the origin of voices, and in this study the results indicate a prevalence of situations of loss.

Regarding the characteristics of voices, these have been widely studied in different social and cultural contexts, due to the current understanding that the contents, the tone, the moments when they normally appear, gender, among other characteristics of the voices, can be beneficial in helping the hearer to deal with the experience (Andrew et al., 2008; Beavan & Read, 2010; Beavan, 2011; Couto & Kantorski, 2020; Daalman et al., 2013; Rosen et al., 2015). As previously explained, Andrew et al. (2008) conducted a comparative study between clinical and non-clinical hearers. Regarding characteristics of voices, the authors found that out of 22 clinical hearers, 20 heard the voices every day, 15 heard for hours, 19 heard voices with negative content, most (20) were moderately or severely bothered by the experience, 20 reported rarely or never having control over the voices, and 19 said that the voices originated internally.

Similarly, Daalman et al. (2013) also conducted a comparative study between 72 clinical and 72 non-clinical hearers, and among the former they obtained the following descriptions of the most prevalent voices: the duration of the voices used to be up to one hour; they came either from inside or outside the head, or from a place close to the ears; the height of the voices was the same as that of the voice itself; less than 50% of individuals believed that the voices had an external cause; most voices were unpleasant or negative and had as content verbal abuses related to self-concept; most voices were distressing, related to a moderate amount of life disturbance, and the hearers had only occasional control over them.

As in these studies, Beavan and Read (2010) also found a prevalence of negative content among the 50 hearers interviewed, which totaled 76%. Of these, it is important to point out that the majority (50%) criticized the hearers, and that 32% ordered them to hurt themselves or other people. As explained, this survey also found a prevalence of negative voices (81%), which is corroborated by the surveys presented here, and is in line with population data that show that hearers who use mental health services hear more negative voices than non-users (León-Palacios et al. 2015; Mackinnon et al., 2004; Varese et al. 2016). In the study by Couto and Kantorski (2020), conducted with 16 clinical voice hearers, voices with negative content also predominated (13), the command voices were the most common, since they were heard by eight subjects. Almost half of respondents reported hearing more than one type of voice, such as derogatory content, call, debauchery, premonitory and threatening voices.

From the above, it can be observed that the events that lead to hearing voices are related to the life story, to the social, family, and relational context of the subjects. In addition, the experiences are strongly heterogeneous. Therefore, such aspects must compose the listening and practice of health professionals, especially in Brazil, which despite having undergone a Psychiatric Reform, still maintains psychotropic drugs as the main treatment, leaving the singularity of life experiences and stories in the background. It should be noticed that countries such as the United Kingdom, Finland, Italy, among others, already adopt practices aimed at new approaches in mental health, such as Autonomous Medication Management (GAM), groups of voice hearers and the use of Maastricht Interview, proposed by the international Hearing Voices Movement and the Open Dialogue approach.

Regarding identity and gender of voices, there are the studies of Beavan (2011), Faccio et al. (2013), and Rosen et al. (2015). After interviewing 50 voice hearers, analysis of Beavan’s (2011) data showed that there are five essential characteristics of voice hearing, one of which is the identity of the voices. Of the 50 interviewees, only five did not fit into this model, and to illustrate this, the author cites four examples that she considers representative of the sample. Of these four hearers, two are men and two are women, and all listen to at least one male voice, like of a deceased father, the voice of God, of relatives and ancestors, and even of an “old Shakespearean”. Similarly, in the study of Faccio et al. (2013), of the ten interviewee’s voices, eight heard male voices. Of these, five also heard female voices, and three heard only male voices. In this study, there was also a prevalence of male voices (55%), as most hearers associated them with parents (52%) or siblings (42%).

In the study by Rosen et al. (2015), male voices were also very present. The authors conducted a mixed quantitative and qualitative study with twenty voice hearers from the United States, and the qualitative stage included 14 subjects. Regarding the characteristics of the voices, the authors found that of the 14 hearers interviewed, 11 heard male and female voices from adults and children; one reported only male voices from adults and children; one reported female and male voices from adults only; and only one hearer reported male and adult voices. Therefore, even if male voices, for many hearers, came with female voices, as in the study by Faccio et al. (2013) none reported hearing only female voices. It can be thought that the higher prevalence of male voices is associated with the type of patriarchal society that has been built, in which men have a privileged position and power in relation to women.

On the content, most (10) reported hearing both positive and negative voices, with four reporting hearing only negative ones (Rosen et al., 2015). The authors also evaluated the number of voices that each subject heard, and obtained very varied numbers, with twelve being the highest number of voices reported by a hearer (Rosen et al., 2015). The study by Faccio et al. (2013) also investigated the number of voices of each hearer, and the results also showed no pattern: two hearers reported hearing more than one voice; three reported hearing only one; two could not quantify; one said hearing two; one said hearing more than ten; and one said hearing three or four voices. This variation in the number of voices seems to be in line with the result of this study, which showed that 52% of hearers have already had a change in the number of voices they heard.

This study also investigated the frequency (34% hears once a day), duration (46% hears up to 30 minutes), location (92% hears through the ears), origin (69% say voices come from themselves), and whether hearers named the voices (79% do not name the voices). This data is also discussed by Faccio et al. (2013), Nayani and David (1996), Daalman et al. (2011), and Garrett and Silva (2003). In the study by Faccio et al. (2013), already described above, the frequency and location of voices were analyzed. Regarding the first, very diverse data was found, as of ten hearers interviewed, two reported hearing almost every day, two heard every day, one heard once a week, one heard once a month, one heard a few times a month, one heard rarely, one heard four to five times a year, and one heard only during stressful periods. Regarding the location of voices, the authors identified, as in this study, a prevalence of subjects hearing them through the ears (six).

In the study by Nayani and David (1996), the frequency of voices was measured based on the period of one day, and the authors found, from a sample of 100 clinical hearers, that 12% heard once or twice during the day, 36% heard during a limited part of the day, 37% heard during most of the day, and 15% heard voices throughout the day. Such data differs from that found in this study, where most reported hearing voices only once a day. On duration, the authors identified that for 33% of hearers voices last a few seconds or minutes, for 25% voices last less than an hour, and for 42% voices last more than an hour, with subjects who hear voices more often tending to experience a longer duration of them. The same study also showed that 72% of the subjects analyzed attributed their voices to external factors, such as divine or diabolic forces, the Central Intelligence Agency (CIA), ghosts, spirits, or aliens.

Similarly, the study by Daalman et al. (2011) also found a prevalence of voices with external origin among the subjects. The authors compared two groups, 118 clinical and 111 non-clinical hearers, and the results showed that 41% of the subjects believed that the voices had an internal origin, 42% believed that they came from other people, and 17% did not know if the sounds were real or imaginary. However, it is noticeable that the difference between the number of hearers who assigned voices to internal and external sources was only 1%, making the result closer to that found in this research. Finally, regarding the name of the voices, the study by Garrett and Silva (2003) corroborates the findings of this study, since they found that only a minority of the subjects give names to the voices. The authors worked with a sample of 41 hearers, and of these only 22% reported naming the voices.

From the above, it can be seen that most of the data was corroborated by other studies, such as the fact that the origin of voices is very related to traumatic events, such as the loss of close people; the prevalence of male voices and negative content; the prevalence of subjects who report hearing voices through the ears, as well as the fact that most of them do not give names to voices. While data related to frequency, duration and origin of voices showed greater variation from those found by this study. Thus, the need to take into account the singularities of the experiences of each hearer in the planning of their care practices is reinforced, since these can serve as guides, indicating the best approaches for subjects to develop a better relationship with their voices and, consequently, have a reduction in the suffering caused by them.

Regarding the content of the voices, it was found in this survey that 58.5% of the participants reported hearing commanding voices, and in 37% of the cases the voices ordered the subjects to kill themselves, and in 24% ordered the subjects to assault themselves or other people. In addition, the most accomplished action after the orders is the suicide attempt. This information is also investigated in the studies by Nayani and David (1996), Birchwood et al. (2014), Trower et al. (2004), Fox et al. (2004), Connor and Birchwood (2011), and Mackinnon et al. (2004).

In the study by Nayani and David (1996), which included a sample of 100 clinical hearers, the data showed that most of the voices reported by the participants were negative and, as in this study, most were commanding (84%). The research developed by Birchwood et al. (2014), had only hearers who presented command voices, and investigated the content of the voices, as well as the response of 197 participants to these commands. The three most commonly reported types of commands were: to hurt yourself, such as "cut yourself", "drink bleach" and "don't take your medication"; to kill yourself, such as "overdose and kill yourself", "jump off the bridge" and "commit suicide"; and to hurt others, such as "attack someone", "hurt your children", and "smash his head in".

In cases where there was greater submission of subjects to voice commands, the following actions were reported as consequences of the commands: cutting the wrists, not taking medication, hitting other people, attempting suicide by putting a rope around the neck or jumping in front of moving cars. And in cases where subjects were able to calm voice commands a little more, actions were reported such as: making superficial cuts in wrists, threatening to attack others, planning how to execute the attack and mentally rehearsing ways to commit suicide (Birchwood et al., 2014).

Similar results were found by Trower et al. (2004), who investigated 38 users of mental health services in England who heard commanding voices, and demonstrated the higher prevalence of voices commanding suicide, actions harmful to other people, homicide, and actions harmful to oneself, respectively. Among the participants, there was a high risk of command compliance, as 79% (30) reported having executed what the voice commanded and 76% (29) expressed fear of reprisal in the event of non-compliance with the command. Of the 25 respondents who were commanded to commit suicide, nine had already tried, and seven had found ways to soften their voices, such as holding a knife close to their fist, taking a razor to the bath, or imagining the execution of the suicide. Of the 14 respondents who were given command to harm another person, seven reported assaults or attacks with cold weapons and two reported calming behavior, such as assault with minimal force or the thought of postponing voice commanded action.

The study by Fox et al. (2004) also investigated the type of commands that hearers received from voices and whether or not they complied with these orders. A sample of 32 participants who reported hearing violent command voices was selected and divided into two groups: command compliant and non-compliant. The types of commands have also been divided into two groups: commands to hurt themselves and commands to hurt others. As a result, the authors found that 75% (24) of the participants who heard violent command voices had complied with the commands. When considering the types of commands, they showed that of the 14 participants who received commands to cause harm to themselves, 79% (11) complied with the command, and of the 18 participants who received commands to cause harm to another person, 72% (13) complied with it. An important finding of this study, therefore, was that compliance with voice commands is more common among clinical voice hearers than non-compliance with commands, which is in line with the results found in this research.

Finally, there are the studies of Connor and Birchwood (2011) and Mackinnon et al. (2004) on the theme. The first, already described above, analyzed a sample of 74 UK clinical voice hearers, found that approximately 70% of participants were diagnosed with schizophrenia, and more than half (58%) of the participants heard five or more voices, with dominance voices predominating. Most participants (40.5%) presented severe depression, about 30% presented ideation or attempted suicide and 18.9% presented moderate or severe hopelessness levels. The study found that the greater the power attributed to voices by the hearer, the higher the levels of depression and suicidal thinking.

In the study of Mackinnon et al. (2004), 199 clinical voice hearers were investigated, with more than two thirds of the sample reporting hearing command voices, and a quarter of these subjects reporting not being able to resist commands. Regarding the strategies used by the subjects, among those who heard command voices 61% said shout or respond verbally to the commands of the voices, 56% were looking for someone to talk to, and 52.3% listened to the radio. By comparing the strategies used by resilient hearers and those who followed the commands of the voices, the authors identified that the first group listened to more music, talked more about the voices and slept, as ways to deal with the commands. Finally, the authors identified that, participants who said they were unable to resist voice commands, described them as more intrusive, had fewer strategies for dealing with them, and took higher doses of medication.

Therefore, it is perceived that hearing more negative and commanding voices increases the odds of hearers having a negative emotional response to the experience. The limitation of this study is that it was conducted with hearers users of only one health service. However, despite the sample size, it was possible to corroborate the results of several research that also aimed to investigate the experiences of clinical hearers. In view of these results, it is important to invest in research on the theme in Brazil, since only two studies were found that addressed the characteristics and origin of voices.

This study highlighted important factors that must be considered in the care for people who hear voices that other people do not. Paying attention to the socioeconomic and demographic profile of hearers who use mental health services, allows us to think about how these characteristics influence the hearing of voices. In this study, for example, it was identified that single women, with low education and unemployed are more susceptible to this experience. In addition, it was found that most participants heard negative, commanding voices, and that the most performed action in the face of the experience was the suicide attempt.

Therefore, it is believed that the characterization of voices is of paramount importance to help making the relationship with voices positive, since investing only in drug treatments has proven to be ineffective for many subjects. This study contributes to the care for voice hearers in the Psychosocial Care Centers, services in the territory, constituted in the process of Brazilian psychiatric reform, with a view to constituting new approaches in mental health in the field of psychosocial care.