Eye movement desensitisation and reprocessing (EMDR) is an evidence-based intervention originally designed to address the psychological distress that arises from adverse and potentially traumatic life experiences (Kaptan et al., 2021). Twenty-four systematic reviews and at least three meta-analyses have demonstrated the efficacy of EMDR in treating symptoms of posttraumatic stress disorder (PTSD), depression, anxiety, distress due to natural disasters and chronic pain (Kaptan et al., 2021). The EMDR group traumatic episode protocol (G-TEP) was developed by Elan Shapiro and is a standardised and empirically supported intervention designed to address the effects of traumatic events in a group setting (Kaptan et al., 2021; Shapiro, 2009). The protocol is an adaptation of the individual EMDR protocol that incorporates the principles of the adaptive information processing (AIP) model and consists of eight phases. It includes the use of a worksheet to guide participants’ processing through each phase (Shapiro, 2009). The protocol acknowledges the fragmented and unconsolidated nature of recent traumatic memories, as well as the need for safety and containment during the intervention (Lehnung et al., 2017). This innovative approach combines the principles of EMDR therapy with the benefits of group therapy, making it a potentially more efficient and scalable form of treatment (Kaptan et al., 2021). This is increasingly relevant in the Aotearoa New Zealand context, which is a country with a rich history of disasters; like most counties, it has an overburdened mental health system.
Aotearoa New Zealand has faced numerous natural disasters and tragedies, including a litany of landslides, earthquakes, floods, cyclones, shipwrecks, volcano eruptions, mining accidents, fires and aeroplane crashes. The history of tragedies is well-documented, with records from the 1840s (New Zealand History, n.d.). The Ministry of Health developed a national health emergency plan in 2007, which was revised in 2015 (Ministry of Health, 2015). The objective of this document was to outline the process of psychosocial recovery from emergencies. Psychosocial recovery involves easing physical and psychological difficulties for individuals, whānau (family) and communities, and supporting social and psychological well-being (Ministry of Health, 2015). The national health emergency plan in place at the time informed the strategic framework for the Canterbury earthquake recovery.
The Joint Centre for Disaster Research was established following the Canterbury earthquakes that occurred in 2010 and 2011 (Massey University, n.d.). An advisory group of specialists with experience researching and working in psychosocial recovery from disaster was formed. This group helped develop general guidelines for responding to a disaster. The psychological response for mental health following the Christchurch earthquakes involved support through the District Health Board (DHB), general practitioners (GPs), local counselling services and not-for-profits in the area. Christchurch-based Relationship Services Whakawhanaungatanga offered free counselling services to the community (Scoop, 2010). In addition, the All Right? campaign was developed, which used social media as a means of promoting well-being messages (Calder et al., 2020). Following the Kaikoura earthquake in 2016, the government announced $1.76 million to address resulting mental health issues (RNZ, 2016). Notably, after the Whakaari volcanic eruption in 2019, families of victims struggled to receive counselling because of the lack of availability of mental health support. On the first anniversary of the eruption, local mental health professionals offered victims’ families free mental health support (Bell, 2020). After the flooding in Auckland in 2023, the Goodspace digital platform, which was developed in 2021, was used to assess mental well-being among children (Education Gazette, 2023). The pre-existing Mana Ake programme was introduced to all Hawke’s Bay and Tairāwhiti primary and intermediate schools to support children impacted by the cyclone. This programme provides support to schools and family/whanau when children are experiencing issues that affect their mental well-being (Verrall, 2023).
An extensive literature review to identify a designated mental health trauma response team in Aotearoa New Zealand showed that no entity existed that could handle that important role. The (former) DHBs, not-for-profit community agencies and GPs filled the gap by providing rudimentary mental health support to the community. However, public health services in New Zealand were already under strain and working at or near capacity. For these services, even small increases in demand may result in considerable extra burden for health workers (Beaglehole et al., 2019).
Cyclone Gabrielle
Cyclone Gabrielle struck Aotearoa New Zealand on 12–16 February, 2023. Gabrielle’s impact was compounded by Cyclone Hale and the Auckland Anniversary floods earlier in 2023, both of which caused widespread flooding and devastation during 10–17 January. A national state of emergency for Cyclone Gabrielle was declared on 14 February in the six regions most affected by the cyclone: Northland, Auckland, Tairāwhiti, Bay of Plenty, Waikato, Hawke’s Bay and Tararua District (Wilson et al., 2023). Cyclone Gabrielle caused significant damage to homes, livelihoods and infrastructure across northern and eastern regions of the North Island (Ministry of Foreign Affairs and Trade [MFAT], 2023). An estimated 10,500 people were displaced (McClure, 2023) and 11 people lost their lives (Wilson et al., 2023). Furthermore, over 70,000 residents were left without crucial services, including health services, power, road connectivity (in every direction), wastewater, drinking water, and Internet and cell phone networks (Napier City Council, n.d.). Cyclone Gabrielle also had a direct impact on the primary and exporting sectors. The cyclone hit just as the harvest was reaching its height and initial estimates of the financial loss of revenue from these farms were between 500 million and 1 billion New Zealand dollars. Estimates on farm capital losses were up to $1 billion, including damage to fencing, machinery, buildings, orchards and vines (MFAT, 2023).
Six months after these devastating events, the cost to the country was estimated to be up to $14.5 billion (Wilson et al., 2023). The economic losses are expected to exceed those attributed to the 2016 Kaikoura earthquake, which was estimated to be $2–$4 billion (MFAT, 2023). Cyclone Gabrielle has been Aotearoa New Zealand’s costliest non-earthquake natural disaster and the second costliest natural disaster in its history, second only to the Canterbury earthquakes in 2010/2011 (MFAT, 2023). These facts and figures illustrate the quantifiable damage to New Zealand as a whole and people in general; however, they do not capture the felt sense or lived experience of those who witnessed these events and bore the brunt of their impact. New Zealand’s national health agency, Te Whatu Ora, told RNZ that the number of people seeking mental health support from their GP had jumped 30% in Hawke’s Bay alone since Cyclone Gabrielle (Crimp, 2023b). The cyclone added to the already high levels of stress people were experiencing as a result of measures implemented during the COVID-19 pandemic and the spiralling cost of living (Crimp, 2023a). Around one-third of the country’s population of 5,000,000 was impacted by Cyclone Gabrielle (Center for Disaster Philanthropy, 2023). When the authors of this article went to Napier in August 2023, some participants reported they were still living in transient accommodation, without clothing or possessions save for those given to them as donations, and they still did not know how they were going to continue with their lives.
Methods
Many survivors of Cyclone Gabrielle were displaced from their homes for an extended period of time. Others were able to remain in their homes, but lost power, water and Internet connectivity for several weeks. This situation of internally displaced persons and inaccessible communities made recruitment of participants for this study a challenge. The initial plan was to hold G-TEP sessions as soon as possible after the storm. However, the logistics of the situation soon made that untenable. Instead, based on advice from Civil Defence authorities that power and Internet connectivity would be restored for most people within 3 weeks, the decision was made to hold the first group sessions approximately 4 weeks after the cyclone.
Aotearoa New Zealand is a country with a rich history of collectivism, and since the signing of The Treaty of Waitangi in 1840, the country has worked towards building a bicultural, collectivist identity (Bennett & Liu, 2017). The present authors were mindful of specific cultural considerations when developing the present study and followed both the guidelines in the Code of Ethics (New Zealand Psychologists Board, 2002) and the principles elucidated by Roberts et al. (2017). Although a full cultural analysis was beyond the scope of this study, it is an area for further research that is discussed below. Participants in the present study were recruited in an inclusive manner, which included specific outreach efforts via community organisations in areas affected by Cyclone Gabrielle. As is noted elsewhere in this article, the EMDR protocol was adapted for the culture in Aotearoa New Zealand, and further research is indicated to ensure ongoing G-TEP sessions are conducted in a culturally appropriate manner.
Participant Recruitment
Participants were recruited via an extensive campaign involving outreach to national, regional and local contacts and agencies. Organisations such as the New Zealand Red Cross, local GP practices, the New Zealand Police, Fire and Emergency services, and other community agencies were contacted. Notably, the Napier Family Centre (NFC) was integral to planning for the 6-month follow-up sessions, as discussed below. Parallel to the recruiting of participants, volunteers were sought to support the sessions. Calls were made via the EMDR Association of New Zealand and the facilitators’ extensive networks of professional contacts. Over 70 volunteers offered their support. Volunteers included clinical psychologists, psychiatrists, social workers, counsellors and various other mental health professionals from Aotearoa New Zealand, Australia and throughout the Pacific region. More discussion of the volunteers’ role is included elsewhere in this report.
Informed Consent
All participants completed the assessments along with an informed consent form, which the researchers held on file. Participants were invited to submit their data for inclusion in the present study anonymously, and no names or identifying information were used in the analysis. The project was identified as ‘out of scope’ by the New Zealand Health and Disability Ethics Committees (HDECs) online scope of review form, and therefore did not require review by the HDECs. The evaluation met HDECs criteria for an audit study of usual practice (HDECs Standard Operating Procedures, Section 3). The authors do not have access to an alternative ethics or research and development committee. All participants provided written informed consent for their outcome data and feedback to be included anonymously in this evaluation study.
Group EMDR Treatment
Initial online group sessions
Based on the logistical considerations noted previously, the initial group sessions were held online via a secure webinar format. This method of delivery ensured that participants could interact with the facilitators, but not with each other, to ensure maximum privacy as many people experience shame when seeking psychological help. Participants were led through the standard G-TEP protocol, which includes grounding and container exercises at the beginning and end of the session, respectively. Participants were instructed to respond to prompts using the ‘thumbs up’ in the Zoom platform to check on their understanding at various times during the session. They were also invited to submit their subjective units of disturbance (SUDs) scores via the chat function; their responses were only visible to the facilitators.
During the online sessions, volunteers were on call to handle any abreactions or requests for individual support. These volunteers were available via Zoom breakout rooms, and were available to provide individual support to participants with the goal of re-integrating the individual into the group. During the course of five online sessions, none of the participants required individual support beyond brief redirection and via chat. The most common issues raised were technical (i.e. ‘I can’t hear or see the facilitator’) or procedural (i.e. ‘I forgot what step we are on’).
In-person sessions
Many participants reported significant treatment effects from the online G-TEP sessions, but there was still a call from the Hawke’s Bay community to access in-person support. The local DHB was consulted, but were unable to provide for the community as their service delivery model did not include group therapy for survivors of a natural disaster. As Hawke’s Bay was the site of the majority of the damage and fatalities from the cyclone, the decision was made to offer in-person support in partnership with the NFC. As a not-for-profit community-based social services organisation, the NFC has served the Hawke’s Bay for over 40 years (Napier Family Centre, 2023). The Centre, located in a city that was significantly impacted by Cyclone Gabrielle, was a natural partner for the present study. The NFC generously donated the use of their community centre to the authors and assisted in local outreach efforts focused on marginalised communities impacted by Cyclone Gabrielle. Participants were recruited for the two in-person sessions in the same manner as the online sessions, with an additional focus on high-need communities such as Marewa, Puketapu, Esk Valley and others. Local volunteers were again recruited to assist the authors with facilitating the sessions and to help with any abreactions; no significant abreactions were noted during the two in-person sessions. All sessions were offered free of charge to all participants. Table 1 presents a summary of the sessions.
Assessment measures
Participants were asked to complete a standardised battery of self-report assessments (see Appendix A) a well as a brief demographic questionnaire. Pre-treatment assessment data were collected by the authors both digitally and via traditional administration of assessments. Pre-treatment screenings were conducted both before and on the day of the sessions. Participants were encouraged to complete these assessments, but there was no requirement to do so.
Post-treatment screenings were sent to participants in the weeks following their sessions. Because of the dispersed nature of participants, all post-treatment screening data were collected electronically. Participants were sent a total of three reminder emails at 2, 4 and 6 weeks post-treatment. Participants who consented to being contacted by phone were called approximately 2 weeks post-treatment to ensure they were able to complete the follow-up assessments. A small number of participants completed the post-treatment assessments (N=14), which is discussed further below.
Results
Participants
In total, 62 participants completed the assessments pre-treatment. As these assessments were not a requirement for participation in the therapy, a significantly greater number of people were involved in the sessions. The second online session that was held in March 2023 had over 160 people logged in during the session. Unfortunately, data collection was not robust at that time, and apart from standard consent and confidentiality forms, no information about those participants is available. For participants who provided demographic information (N=76), the average age was 43 years, 65% identified as European New Zealander, 9.3% identified as Māori and 1.3% identified as Pacific Islander; 81% of participants identified as female.
Assessment measures
Data were collated and analysed by the lead author using standard statistical software. The sample sizes significantly differed between the pre- and post-treatment groups, which is discussed in detail below. Based on this significant difference in sample sizes, Welch’s unequal variances t-test was determined to be the most appropriate statistical test (West, 2021). Cohen’s d was used to determine effect sizes (Gignac & Szodorai, 2016). The significance level and effect sizes (Magnusson, 2023) are included in Table 2. Individual measures are discussed in the following paragraphs.
Depression
Depression was assessed using the PROMIS Short Form - Emotional Distress - Depression questionnaire (American Psychiatric Association, 2022b). The results from the pre-test (mean [M] = 16.98, standard deviation [SD] = 7.08) and post-test (M = 13.14, SD = 4.81) indicated a significant improvement in clinical symptoms (t(27) = 2.44; p = 0.01).
Anxiety
Anxiety was assessed using the PROMIS Short Form - Emotional Distress - Anxiety questionnaire (American Psychiatric Association, 2022b). The results from the pre-test (M = 19.80, SD = 6.56) and post-test (M = 14.78, SD = 5.52) indicated a significant improvement in clinical symptoms (t(30) = 2.65, p < 0.01).
Sleep disturbance
Sleep disturbance was assessed using the PROMIS Short Form - Sleep Disturbance questionnaire (American Psychiatric Association, 2022b). The results from the pre-test (M = 20.07, SD = 4.81) and post-test (M = 23.14, SD = 9.59) indicated a significant improvement in clinical symptoms (t(19) = 3.56, p < 0.01).
PTSD
PTSD was assessed using the Posttraumatic Stress Disorder Checklist for DSM-5 (PCL-5), which is a 20-item self-report measure for symptoms of PTSD (National Centre for PTSD, 2022). The results from the pre-test (M = 43.71, SD = 25.01) and post-test (M = 29.43, SD = 15.17) indicated a significant improvement in clinical symptoms (t(31) = 2.78, p = 0.01).
Dissociation
Dissociation was assessed using the Brief Dissociative Experiences Scale (DES-B) - Modified (American Psychiatric Association, 2022b). The results from the pre-test (M = 0.35, SD = 0.49) and post-test (M = 0.08, SD = 0.26) indicated a significant improvement in clinical symptoms (t(21) = 1.74, p = 0.05).
Suicidal behaviour
Suicidal ideation and behaviour were assessed using the Suicidal Behaviour Questionnaire - Revised (SBQ-R), which is a brief self-report measure of past suicidal ideation and behaviour (Osman et al., 2001). The results from the pre-test (M = 5.93, SD = 2.46) and post-test (M = 5.00, SD = 2.94) indicated a moderate improvement in clinical symptoms (t(16) = 1.06, p = 0.15).
Social skills
Social skills were assessed using the Scale of Protective Factors, which offers a multidimensional measure of factors that contribute to resilience and recovery (Ponce-Garcia et al., 2015). The results from the pre-test (M = 15.64, SD = 5.37) and post-test (M = 18.00, SD = 3.67) indicated a moderate improvement in this domain of resilience (t(11) = −1.08, p = 0.15).
Social support
Social support was assessed using the Scale of Protective Factors, which is a multidimensional measure of factors that contribute to resilience and recovery (Ponce-Garcia et al., 2015). The results from the pre-test (M = 15.14, SD = 3.80) and post-test (M = 18.16, SD = 2.64) indicated a significant improvement in this domain of resilience (t(14) = −2.04, p = 0.03).
Planning and prioritising behaviour
Executive functions were assessed using the Scale of Protective Factors, which offers a multidimensional measure of factors that contribute to resilience and recovery (Ponce-Garcia et al., 2015). The results from the pre-test (M = 15.07, SD = 2.78) and post-test (M = 16.50, SD = 3.39) indicated a slight improvement in this domain of resilience (t(8) = −0.91, p = 0.19).
Goal efficacy
Goal efficacy was assessed using the Scale of Protective Factors, which is a multidimensional measure of factors that contribute to resilience and recovery (Ponce-Garcia et al., 2015). The results from the pre-test (M = 17.79, SD = 6.09) and post-test (M = 19.34, SD = 3.61) indicated no significant improvement in this domain of resilience (t(16) = −0.70, p = 0.25).
Discussion
Survivors of natural disasters are faced with countless challenges, from a lack of housing to food insecurity and a loss of community (New Zealand Red Cross, 2023). Among the devastation caused by storms such as Cyclone Gabrielle, constant re-exposure to reminders of the trauma is considered to be a significant risk factor for developing PTSD (American Psychiatric Association, 2022a). Early intervention following a trauma and timely treatment for any emerging PTSD symptoms is considered best practice (Shapiro & Maxfield, 2019) and can aid in ensuring the recovery of an individual is maintained. By treating groups of survivors, the present study addressed the psychological needs of these survivors without straining the already overburdened mental healthcare system with a flood of individual therapy clients.
The psychological symptoms assessed in the present study were varied and numerous. The results of those assessments were mixed; some areas showed significant improvements and large effect sizes, whereas others showed minimal improvement and smaller effect sizes. Many internalising conditions, including depression, anxiety, sleep disturbance and PTSD, showed significant improvement. This was consistent with much of the literature focused on group EMDR therapy (Kaptan et al., 2021), which showed group EMDR was an effective therapy for many common psychological conditions.
These results were limited by several important factors, including a small number of responses to the follow-up assessments. Although some level of drop out is expected in any psychological research, the present study lost contact with 48 participants. This was likely because of a multitude of factors, including the early nature of the interventions which occurred at a time when many participants were focused on physical safety and recovery. In addition, the decentralised nature of the recovery efforts made data collection difficult, as this depended on the authors to contact participants by phone or email to follow-up on missing assessments. This labour-intensive approach to data collection did not obtain the expected level of responses. The disparities in response rates were statistically controlled for using Welch’s t-test (West, 2021) and by comparing effect sizes; however, it remains a consideration for further research that an increased post-treatment response rate would possibly impact clinical outcomes.
Another limitation of this study was the self-report nature of the assessments. No embedded validity scales were used in any of the scales administered. This was intended to reduce the response burden on participants. However, a lack of scale validity combined with a small total sample may have contributed to the chance of type 1 errors (Banerjee et al., 2009). Although efforts were made to select assessments with the highest level of reliability and validity for the populations affected by the cyclone, further deliberation on the test selection could have resulted in a markedly different set of outcome measures. Furthermore, the outcome measures were limited by the route of administration. All follow-up data were collected electronically, which eased the analysis efforts by the authors but also potentially biased the pool of respondents as those who were uncomfortable completing the assessments online were unable to provide quantitative responses post-treatment.
A final limitation of this study concerned the G-TEP, which was developed in a markedly different cultural context than what exists in Aotearoa New Zealand. G-TEP is recognised as an effective treatment approach in many cultures, but the present study is the first that focused on its use in a New Zealand context following a natural disaster. This is an area of which the authors are conscious, and efforts were made to balance traditional Māori tikanga (customs) with adherence to the G-TEP. At the beginning of every session, a facilitator introduced the concept of G-TEP in accordance with the written protocol, then explained that in a breach of tikanga, participants would not have a chance to participate in the korero (discussion). It was explained that there would not be an opportunity for sharing of pepeha (personal introductions), as is customary in most meetings in Aotearoa New Zealand. Instead, participants were invited to focus on their own experiences and to notice the collective healing that occurs when we whakamene (come together). Following that diversion from the standard G-TEP procedure, the session returned to the script. Further research that the authors intend to conduct will focus on this cultural aspect of group EMDR, and on methods for further adapting the G-TEP for participants in Aotearoa New Zealand.
Implications for Practice
-
Cyclone Gabrielle was a massively destructive storm that exacerbated the existing strain on the mental health system in Aotearoa New Zealand. Group EMDR was shown to be a cost and resource effective method of helping large numbers of survivors following the cyclone.
-
Participants reported reductions in anxiety, depression, sleep disturbance and PTSD symptoms following group EMDR sessions, which were delivered both online and in-person.
-
Further research is needed to determine what adjustments to the G-TEP are needed to ensure psychological treatments are culturally appropriate in Aotearoa New Zealand.