EMDR Therapy Quarterly

Learning objectives

• Understand the high prevalence of PTSD in forensic populations and its association with increased risk of anger, aggression and recidivism.
• Recognise the high comorbidity between trauma, PTSD and psychosis in this group, and how these conditions relate to risk factors such as impulsivity, emotional regulation difficulties and dissociative symptoms.
• Appreciate the potential for EMDR to reduce prolonged maladaptive psychological consequences.
• Provide the reader with feedback on whether this treatment is effective in reducing overall symptoms of PTSD and related risk.

Introduction

Since its development, EMDR has established a clear evidence base in the mental health setting to reduce distress associated with trauma ‘symptomology’ and PTSD through multiple clinical domains. For instance, EMDR has been used to treat combat/war veterans’ critical incidents and current triggers, achieving symptom reductions in anger, depression, anxiety and physical pain (Silver et al., 2008). It has also shown its applicability to individuals with pre-verbal trauma or memory blocks with its use alongside art therapy (Struble & Struble, 2020; Tripp, 2023). Additionally, a systematic literature review of randomised control trials by Valiente-Gómez et al. (2017) found that EMDR improves overall trauma-associated experiences and comorbid psychiatric presentations, such as psychosis and schizophrenia. Furthermore, Shapiro (2001), the founder of EMDR, has provided evidence for the efficacy of EMDR across all clinical samples, with the original theory being consistently empirically reviewed and supported within forensic mental health services (Shapiro, 2012; Susanty et al., 2022). 

Psychosis as a trauma response

The term ‘psychosis’ still lacks a unified definition; however, it denotes a clinical construct characterised by several perceptual disturbances such as delusions, hallucinations and disordered thoughts, accompanied by the loss of an emotional connection with oneself or others and a pervasive sense of fear and lack of safety within daily life (Gaebel & Zielasek, 2015).

There is a well-established association between symptoms of trauma and psychosis (Fleurkens et al., 2018; Mueser et al., 2002). This connection is thought to stem from adverse childhood experiences, which are defined as any highly stressful and potentially traumatic event or situation that occurs during childhood and/or adolescence (McFarlane, 2013; Young Minds, n.d.). Research shows that adverse childhood experiences are thought to have a cumulative effect on the likelihood of psychosis occurring in later life. For example, between 50 and 98% of adults with severe mental disorders (such as psychosis) report at least one traumatising childhood experience, with an average of 3.5 incidents per person (Freuh et al., 2005; Goodman et al., 1997). Another study found that 69% of women and 59% of men with diagnoses of psychosis or schizophrenia were sexually or physically abused during childhood (Read et al., 2005). These studies indicate a clear relationship between severe trauma at a young age, overwhelming biological stress, and altered reality perception, leading to psychosis (Mueser et al., 2002).

This connection is believed to stem from alterations in brain structure and function, as well as the central nervous system’s response to perceived threat, resulting from the overwhelming stress and distorted reality often caused by severe trauma (Shevlin et al., 2008). The ‘dose effect’ is a concept that refers to the relationship between the severity, frequency and accumulation of traumatic experiences and the likelihood of developing psychotic symptoms (Varese et al., 2012). It recognises that with each additional traumatic experience, there is a heightened risk of developing psychosis as a response to overwhelming and prolonged stress. This additionally highlights the importance of trauma-informed approaches in both the understanding and treatment of psychotic disorders.

It is therefore important to recognise psychosis as a response to trauma, as some theories suggest that the treatment of PTSD in individuals with psychosis may lead to improvements not only in their trauma-related distress but also in associated risk behaviours. Examples include auditory verbal hallucinations, delusions, anxiety symptoms, depression symptoms and difficulties with self-esteem – all of which can influence a patient’s vulnerability to risk (Van den Berg et al., 2018).

Trauma in the secure forensic setting

Within the forensic population, mental disorder is considered in the risk assessment literature as an important consideration; for instance, the admission criteria for highly secure hospitals state that the individual must pose a grave danger to themselves or the public. Some patients will have been sentenced under a hospital order, whilst others will have been transferred from prison for assessment and/or treatment that is determined to be required under conditions within a highly secure environment (Nottinghamshire Healthcare, n.d.).

Considering psychosis as a trauma response is crucial, as it may increase the risk of violent behaviour when it’s related to the individual’s perception of threats (McFarlane, 2013; Sweeney et al., 2018). This criterion is supported by the literature, which indicates a complex relationship between initial traumatic experiences and presenting violence (Ardino, 2011; Bentall et al., 2014). Violence is described as a way for a patient to mobilise against unresolved trauma, whereby an individual may re-enact an innate ‘blueprint’ that is protective (e.g., hurting someone else to prevent themselves from becoming hurt) (Porges, 2022). Treating trauma in patients who experience psychosis can reduce symptoms such as hallucinations, delusions, anxiety and depression (Van den Berg et al., 2018). Such treatment can further reduce the likelihood of these individuals engaging in criminogenic behaviours, including violence as a method of coping (The Good Lives Model, Ward et al., 2007).

Considering trauma in risk formulation within high secure forensic populations is therefore critical for understanding the distress and coping mechanisms that contribute to mental health disorders. These mechanisms can include addiction, self-harm and suicidal ideation; understanding these links can guide appropriate interventions and help to minimise the risk of re-traumatisation and re-offending (Cleary et al., 2020).

Furthermore, trauma-informed care posits that this risk factor must be embedded within the relationship and ethos of service delivery, as by acknowledging trauma, risk can be appropriately managed while providing patients with the best possible opportunity for safe and effective care (GOV, 2022; Harris & Fallot, 2001). Trauma and PTSD are essential factors when working and formulating risk in a forensic population, not just to reduce levels of risk to others, but also to reduce the impact of re-traumatisation during hospitalisation and treatment of the patient (Goff et al., 2007; Mental Health Act, 1983).

Rationale

There is limited published data on EMDR therapy within this complex forensic population. Most EMDR evaluations have used self-report measures of satisfaction and progress, semi-structured interviews, and thematic analysis to determine service users’ experiences in NHS high secure services (Every-Palmer et al., 2019; 2023). However, these are not behavioural measures, nor are they reflective of the service from a non-biased perspective. For instance, a robust behavioural measure is important for accurately measuring risk, and, although in most services, patients’ self-reports can be relied upon, those in secure mental health settings may be motivated to under- or over-report the severity of their symptoms for a variety of reasons. Therefore, this evaluation was designed to consider the impact of EMDR on the severity of symptoms of PTSD (as linked to psychosis) and related risk, as recorded by staff in the patients’ notes. This approach allows for the clear exploration of day-to-day stressors and symptoms systematically through the analysis of records relating directly to each patient’s behaviours. Our hypothesis here expects to see a reduction in symptoms following EMDR.

The effect EMDR therapy has on high secure patients remains unclear regarding related risk and practice efficacy. ‘Least restrictive practice’ is a core clinical and ethical principle that underpins guidelines by the National Institute for Health and Care Excellence (NICE, 2005), ensuring that therapy is respectful, person-centred and supportive of patient recovery and progression to less secure environments. This evaluation was therefore in both the service users’ and service providers’ best interest to better understand the effectiveness of EMDR within this population.

Method

This service evaluation consisted of a multiple single-case design using content analysis to compare the case notes of four forensic patients recruited within a high secure hospital. Each patient had completed EMDR at least six months before the start of data collection and had the capacity to consent to participate.

Demographic information, including age, gender, ethnicity and diagnosis of the patients was extracted (see Table 1). No personally identifiable information was included in the data.

Table 1

Demographic information.

Patient	Age	Gender	Ethnicity	Psychiatric diagnosis
1	48	Male	White British	Paranoid schizophrenia. Antisocial personality disorder.
2	45	Male	White British	Paranoid schizophrenia. Antisocial personality disorder. Emotionally unstable personality disorder.
3	31	Male	White British	Schizophrenia.
4	54	Male	White British	Treatment-resistant schizoaffective disorder. Chronic hypomania and psychosis.

This service evaluation was approved by the University of Lincoln School of Psychology ethics committee (2024_17513) and the NHS Trust’s Research and Ethics team. Patients provided informed consent via a form that outlined the usage of EMDR data, their voluntary participation, anonymity and withdrawal rights. No deception occurred, and a support section was included.

Any names or initials within case notes and reports were also anonymised by replacing them with non-related characters (e.g., ‘X’).

The DSM-5 (American Psychiatric Association, 2013; SAMHSA, 2014) diagnostic criteria for PTSD was used as a pre-determined symptom checklist to identify symptoms in each patient’s RiO nursing notes.

To collect the data, the researcher considered a total of 12 months’ worth of nursing notes per patient (six months pre- and post-EMDR) and used the pre-determined symptom checklist to identify the presence of PTSD symptoms in each patient. Whenever the researcher encountered a behaviour or statement in the nursing notes that matched any PTSD symptoms, it was recorded. Additionally, the frequency of incident reports (IR1s) was collected to indicate the presence of aggressive behaviours – verbal, physical or otherwise – that pose a risk to the safety and well-being of patients, staff or others within the hospital environment.

All four patients had the same EMDR-practitioner-accredited therapist who only engaged with these individuals for EMDR therapy when it was deemed appropriate by a multidisciplinary care team and with consideration to their diagnosis. The length of EMDR therapy across the four patients ranged from four to 17 months.

Analytical strategy

Following the extraction of data, an inter-rater reliability check was conducted to confirm the classification of each identified symptom. The total number of reports of symptoms per patient before and after EMDR was then generated to indicate a total quantitative trauma characteristic pre- and post-EMDR for each patient. Additionally, the symptoms with the highest counts and the most prevalence across all four patients were identified.

The extracted numerical data was imported and analysed using non-parametric Wilcoxon tests in R Studio (Version 2024.04.2-764).

Results

There was a variation in the number of symptoms reported across all four patients, with Patient 4 having many more symptoms than the other three. For example, Patient 1 had a total number of 16 identified symptoms of PTSD before EMDR and five after, which was the lowest number of symptoms identified, whereas Patient 4 had 163 symptoms before EMDR and 170 after, indicating that Patient 4 had the highest number of identified symptoms. This is illustrated in Figure 1.

Figure 1

The number of identified PTSD symptoms before and after EMDR for each patient.

On average, patients displayed a higher frequency of symptoms before EMDR (M = 80.5) than after EMDR (M = 62.25), indicating that there was an overall higher level of symptoms of PTSD before EMDR treatment, which is clinically meaningful.

Of the 13 identified symptoms, only four had enough statistically meaningful data to make additional comparisons. This was determined by identifying any symptom that had a frequency count higher than 10 either before or after EMDR.

The four symptoms that met this criterion were ‘hypervigilance,’ ‘impairment in areas of social functioning,’ ‘irritable behaviour and angry outbursts,’ and ‘sleep disturbance.’ Wilcoxon tests revealed no significant differences (Hypervigilance: V = 5, p = 1;Impairment in areas of social functioning: V = 0, p = 0.18;Irritable behaviour and angry outbursts: V = 3, p = 0.37; Sleep disturbance: V = 0, p = 0.18).

These results are shown graphically for each patient in Figure 2:

Figure 2

The total number of symptoms across each patient and each of the four identified symptoms.

Finally, the number of incident reports for each patient before and after EMDR was analysed (see Table 2). These included behaviours that demonstrated recurrent, persistent negative emotional states (e.g., fear, anger and shame), reckless or self-destructive behaviour and/or irritable behaviour resulting in physical or verbal aggression toward other people, and physical arousal or reactivity (e.g., hypervigilance). This number further demonstrated the individualistic needs and characteristics of each patient.

Table 2

Discussion

It was posited that following EMDR, a reduction in PTSD symptoms would be evident due to the well-established connections between trauma, psychosis and PTSD. However, this was not the case across the whole sample, nor when the patients were split according to the four most prevalent symptoms – the influence of EMDR on each of these symptoms was also statistically non-significant. Although these results appear less consistent with the existing research, the trends in the frequencies of the data per patient can suggest that further research may reveal meaningful and clinical implications. For instance, from Figure 1, it appears that Patients 1 and 3 improved, while Patients 2 and 4 did not. In contrast, Figure 2 suggests that although the overall symptom count for Patient 4 stayed consistent, his hypervigilance and irritability got worse, and his sleep and social interactions improved. Additionally, although the p-value was above the conventional threshold, the median number of symptoms before (71.5) and after (37) EMDR indicates there was a clinically meaningful moderate relationship on an ‘individualistic’ level that warrants further investigation and discussion.

It could be posited that the results would support the need for an idiosyncratic approach to EMDR treatment and evaluation. Trauma-informed care already places great emphasis on the importance of trauma in mental health and risk (GOV, 2022), but this does not eliminate the possibility that additional/other factors may have influenced behavioural response (Malvaso et al., 2016). As shown in Figure 2, Patients 2 and 4 saw the largest increase in ‘hypervigilance’ and ‘irritable behaviour and angry outbursts’. However, Patients 2 and 4 moved to medium secure units for a three-month trial leave shortly after the completion of EMDR. This makes the results harder to interpret, as the staff in medium secure units may have very different thresholds for what they regard as ‘worthy’ of recording in a patient’s notes. Moreover, this move to a new facility will likely have increased hypervigilance due to the disrupted feelings of attachment and reinstated feelings of insecurity, abandonment or rejection (McMurran, 2002). It is important to consider protective factors, such as security and safety during the stages of change in a forensic setting, as these factors can be used to help minimise a patient’s fluctuations in hypervigilance and paranoia.

Expanding on this potential alternate factor, there are a multitude of confounding factors why the move to a medium secure unit may have had a destabilising effect on the patient’s rehabilitation. As explained by Attachment Theory (Bowlby, 1982), the consistency of a therapeutic relationship between patient and practitioner is invaluable to the reduction of maladaptive coping mechanisms, aggressive outbursts, and violent behaviours (Alshahrani et al., 2022) as it may have been the first opportunity for a corrective emotional experience in relation to patterns of attachment. Therefore, the move to a medium secure unit may have disrupted feelings of attachment and reinstated feelings of insecurity, abandonment or rejection (McMurran, 2002) and acted as a trigger for a multitude of harmful/maladaptive coping mechanisms. This may explain why Patients 2 and 4 appeared to have deteriorated after treatment, as evidenced by their increase in symptoms and behavioural incidents; Patient 2’s symptoms increased by 68%, and Patient 4’s increased by 83%.

It could also be argued that the transition to a lower-security hospital resulted from a reduction in risk and symptomology, apparent after EMDR treatment, that may not have been captured using the method of data collection. What was clinically evident was a reduction in risk-related treatment needs associated with trauma response, alongside more adaptive functioning in daily living, such as less conflict with peers, adherence to ward rules, greater distress tolerance and emotional regulation. This suggests a positive clinical impact of EMDR, supporting its role in promoting adaptive functioning in individuals with complex trauma histories (Shapiro, 2001; Porges & Dana, 2018).

The use of psychiatric diagnoses may also influence the difference in symptom frequencies. Patients had multiple diagnoses, many with overlapping symptomology; for example, symptoms such as hypervigilance could be identified both as being part of PTSD diagnostic criteria as well as Schizophrenia. This identifies the limitations of using a diagnostic and categorical approach to understanding presenting problems. As such, how staff choose to record behaviours under these categories may vary and therefore influence the data.

It is essential to address the confounding factor of comorbidity when working with a high secure population, due to the complexity of serious mental disorders and the idiosyncratic challenges in treatment. Trauma-informed care and thinking provide a greater understanding of the root causes of problematic behaviours and their links to both mental disorder and risk. Focusing on these aspects, rather than the psychiatric diagnosis, could better capture symptom expression and change (Sweeney et al., 2018).

One key finding from this service evaluation is that it demonstrates a methodology that is both practical and efficient. It accounts for significant variables before and after EMDR and utilises existing data, thereby easing the workload typically required for data collection in inpatient settings. Additionally, the methodology can be used to portray the practitioner’s capacity to tailor treatment to the individuality of the patient. For instance, it identified treatments that took place before EMDR, such as dialectical behavioural therapy (Patient 1) and schema-informed sessions (Patients 2 and 3). This retrospective method supports the evaluation of EMDR’s effectiveness across diverse patient backgrounds and offers insight into the varying lengths of EMDR treatment, hospital stay duration and differences in presentations/levels of dissociation.

Additionally, this methodology highlights the complexity of conducting research in high secure settings and the many confounding factors that are difficult to control when examining pre-recorded data. For instance, moving patients between wards and facilities can destabilise the provision of treatment; changes in psychologists can cause setbacks due to a need for attachment and a reduction in their sense of safety; and it can be problematic to use patient notes and records retrospectively, as patients have already moved on and confounding factors cannot be controlled (Silver et al., 2008; Rodenburg et al., 2009; Porges & Buczyński, 2011).

To mitigate this complexity and confounding factors, a working model of EMDR integrated with the psychological model of understanding is useful to facilitate understanding the individual and unique presentation of each patient. As described by Mueser et al., (2010), patients with PTSD, schizophrenia and psychosis are all likely to have different presentations and triggers based on their individual trauma histories and adverse childhood experiences. However, current NICE guidance focuses on the least restrictive settings and does not seemingly fully consider the additional complexities of patients in high secure services. This emphasises the requirement for EMDR practitioners to act responsively to presenting patient needs rather than adhering to a standard protocol across all cases (Chadwick & Billings, 2022; NICE, 2005).

It has been established that there is a need to adapt common approaches to trauma using EMDR. Given that EMDR is considered a non-directive exposure therapy, as it allows patients to have free associations and quickly move through traumatic memories in a “patient-led” manner (Shapiro, 2001), the need for preparatory work is extensive. Each practitioner should take an individualised approach to formulation to understand the presenting difficulties, resulting coping strategies and threat responses of each patient to appropriately inform treatment goals and approaches (Chadwick & Billings, 2022; Johnstone & Boyle, 2018).

Limitations

Due to the nature of working with vulnerable adults and high secure data, patients had to be selected according to the inclusion/exclusion criteria and give formal consent, which led to a lower participation number. The therapist was a white, British female, and the population in the recruited sample was all white, British and male, thus limiting the generalisability of these results to females and more diverse demographics. Despite a notable timeframe for collecting data, the study was only able to recruit four patients. This was due to the limited number of individuals who had completed EMDR therapy within the service across the six-month pre- and post-timeframe. As noted, a single-case design may have been more informative and reduced the number of confounding factors; however, this would have had implications for the overall generalisability of the results.

Although content analysis has demonstrated rigour among qualitative studies (Strijbos et al., 2006), it requires rigid adherence to predetermined themes – in this case, the DSM-5 diagnostic criteria for PTSD – which can limit the researcher’s ability to respond to unexpected insights or novel aspects of the data (Kalpokas & Hecker, 2024). This approach may have led the researcher to overlook the nuances of individuality between patients or symptoms of trauma that are not conventionally recognised. Furthermore, the use of nursing notes brings a subjective limitation to the collected data. The quality and consistency of nursing notes are likely to vary across different wards, hospitals, and over time due to individual staff variables; for instance, the experiences and subsequent recordings of one staff member may differ significantly from those of another (Font-Jimenez et al., 2020). This could also be an explanation for the increase in symptoms noted when patients have moved facilities, which could be attributed to a difficulty in validating the collected data that is inherently opinion-based and time-relevant.

Summary

This service evaluation adds to the growing literature and evidence base for the application of EMDR in treating trauma, psychosis and related risks. It demonstrates a successful methodology that provides some insight into how patients have changed pre- and post-EMDR and draws attention to the individuality and complexity of high secure patients. Furthermore, with the various models, associated mechanisms and connections from the literature, it is reasonably assumed that due to a reduction in traumatic distress, there has been a reduction in risk following EMDR. This is evident in the increase in adaptive functioning demonstrated by each of the patients moving on from a high secure to a medium secure unit.

It is also recognised that EMDR can clinically influence the frequency of specific symptoms of PTSD in some patients on a single-case basis and, thus, has clinical utility in reducing traumatic distress when therapy is patient-led. Exposure to previous trauma that occurs during EMDR therapy should follow the least direct route possible, addressing original traumatic memories whilst minimising overall distress. This approach supports the effectiveness of EMDR by helping individuals to start to heal unresolved trauma in line with EMDR’s underlying Adaptive Information Processing (AIP) model, reducing the emotional charge of traumatic memories and weakening harmful psychotic experiences and associated behavioural responses. Continued investigation in both clinical practice and research would allow clinicians and services to have the best understanding of how EMDR can help reduce distress that is representative of the manifestations of trauma and associated risk.

Further reading

Every-Palmer, S., Ross, B., Flewett, T., Rutledge, E., Hansby, O., & Bell, E. (2023).  Eye movement desensitisation and reprocessing (EMDR) therapy in prison and forensic services: A qualitative study of lived experience. European Journal of Psychotraumatology, 14(2). 

https://doi.org/10.1080/20008066.2023.2282029

Shapiro, F. (2012). EMDR therapy: An overview of current and future research. European Review of Applied Psychology, 62(4), 193–195.

https://doi.org/10.1016/j.erap.2012.09.005