Institutional firesetting in a forensic inpatient population

ABSTRACT

Institutional firesetting is a widespread and serious problem in custodial settings. However, there is very little research into the subject. This paper describes the characteristics and firesetting histories of 32 patients with histories of institutional firesetting in a secure psychiatric unit in the UK. Only six patients (18.8%) had a prior conviction for firesetting, while a further eight (25.0%) had a history of firesetting but no conviction). All institutional firesetting was carried out alone, most commonly in a cell or bedroom. Institutional firesetting was significantly more common among patients with a diagnosis of personality disorder than among those with a diagnosis of mental illness.

The results are consistent with the M-TTAF trajectories model of firesetting. Different patterns of firesetting may emerge in institutional settings because firesetting may be one of an extremely limited repertoire of problem solving strategies and may be inadvertently reinforced in these settings. These results suggest that there may be a subgroup of institutional firesetters with no previous history of firesetting and this is a group that merits more detailed study.

KEYWORDS:

• Institutional firesetting
• Arson
• M-TTAF

Introduction

People who set fires appear to be broadly similar to other justice-involved individuals in terms of their developmental and demographic characteristics but may differ in terms of their vulnerabilities and risk factors (T. Gannon et al., 2022). Among these vulnerabilities and risk factors, one of the most widely researched areas is that of mental health. There is a strong association between firesetting and various mental disorders including borderline and antisocial personality disorders, substance dependence, schizophrenia and affective disorders (Dickens & Sugarman, 2012; Sambrooks et al., 2021). Firesetting recidivism is associated with contact with mental health services in general (Ducat et al., 2015), and specifically with diagnoses of personality disorder (Dickens et al., 2009), psychotic disorder and substance misuse (Ducat et al., 2015; Lindberg et al., 2005), though not with affective disorders (Ducat et al., 2015).

While there have been a number of approaches to formulating firesetting behaviour, the most sophisticated and clinically useful systems have involved theory-based multi-modal frameworks that describe various pathways to firesetting behaviour (e.g. Barnoux et al., 2015; Gannon et al., 2012; Tyler et al., 2014). The Multi-Trajectory Theory of Adult Firesetting (M-TTAF; Gannon et al., 2022; Gannon et al., 2012) is one of the most-widely used frameworks for formulating firesetting behaviour. It includes developmental factors such as caregiving environment and learning experiences that lead to psychological vulnerabilities associated with firesetting context. Firesetting behaviour can be reinforced either positively (e.g. by social approval or excitement) or negatively (e.g. by providing relief from distress). Firesetting behaviour may also be reinforced if it results in the exacerbation of vulnerabilities such as social rejection.

Gannon et al. (2012) describe five patterns of vulnerabilities and their associated trajectories of firesetting behaviour; 1) those on the antisocial cognition trajectory are primarily motivated by antisocial beliefs, scripts and values and tend to offend in a number of ways to achieve antisocial goals, not just firesetting; 2) those on the grievance trajectory use fire as a tool for revenge and retribution; 3) those on the fire interest trajectory derive excitement or stimulation from fire or may use fire as a means of in times of high stress. Their offending tends to be limited to firesetting; 4) individuals on the emotionally expressive/need for recognition trajectory use fire to communicate needs and hopelessness, as a ‘cry for help’, in order to self-harm or commit suicide, or to achieve recognition by fighting or preventing a fire; 5) multi-faceted firesetters have two or more of the above motivations for firesetting.

Gannon et al. (2022) elaborated on the role of personality disorder in the M-TTAF. The authors proposed that personality disorder traits emerge from an interaction between genetic vulnerability and developmental adversity, resulting in characteristics such as affective instability, impulsivity problems and interpersonal instability that increase an individual’s vulnerability.

Given the association between serious mental disorder and firesetting, it is perhaps not surprising that firesetting in custodial settings such as a prisons or secure psychiatric hospitals, otherwise known as institutional firesetting, is a significant issue. In the year to April 2021, there were 1003 fires in prison establishments in England and Wales, of which 91% were believed to have been started deliberately (Home Office, 2022a). Possibly as a result of the fact that prisons were in a state of lockdown for much of this period due to the COVID-19 pandemic, this figure was significantly lower than in the previous year, when there were 1308 fires in prisons. To put this in perspective, the average prison population for the year to April 2021 was 78,536, so there were 12.8 deliberate fires per 1000 prisoners in 2020/2021. During the same period, in England the fire service recorded 63,715 deliberate fires (Home Office, 2022b), which equates to a rate of 1.12 fires per 1000 people. In other words, the number of deliberate fires started in prison was, pro rata, over 10 times the number in the community. Although equivalent figures for secure psychiatric hospitals are no longer collected, Grice (2012) states that during 2007 there were 489 fires involving 30 non-fatal casualties at psychiatric hospitals in the UK and that a disproportionate number of hospital-based fires occurred in psychiatric hospitals.

Institutional firesetting is potentially extremely dangerous, not just to the person starting the fire who may well be locked in a room and unable to escape, but also to other prisoners or patients locked in the same building, and to the staff and firefighters who have to manage the fire. The number of staff and prisoners injured as a result of prison fires in 2020/21 in England and Wales was 134, up from 41 the previous year. The large increase is also believed to reflect the increase in time prisoners spent locked in their cells due to covid-19. Fortunately, there were no fatalities as a result of these fires. However, fire setting within institutional settings can result in fatal consequences; in 2012, a fire believed to have been started deliberately in a prison cell led to the deaths of 360 prisoners in Comayagua prison in Honduras, while a fire at that Oksochi Psychiatric Hospital in Russia in 2013, believed to have been deliberately started by a patient, caused 38 deaths (Tyler et al., 2019). Even where they do not result in fatalities, institutional fires can result in life-changing injuries and psychological harm (Lee et al., 1996; Warr, 2021).

While there is a considerable body of literature on firesetting in the community, there has been relatively little research into institutional firesetting. Gannon et al. (2015) in their report on a prison-based treatment group for firesetters, reported that 12% of fires set by prisoners in the treatment group, and 18% of fires set by prisoners in the control group had been started in prison. Wyatt et al. (2019) compared demographic details and dynamic risk factors of psychiatric patients who had deliberately set a fire while under NHS care, with patients matched on gender, age and location who had been involved in a non-firesetting incident. Participants were in inpatient and community services, though it is unclear how many, if any, were in secure services. Firesetters had lower levels of substance use disorder and higher rates of previous psychiatric admissions, they were more likely to express suicidal ideation and to be isolated in the month prior to the incident, but less likely to exhibit an external locus of control. Repeat firesetters were more likely to have a diagnosis of personality disorder and were more likely to demonstrate impulsivity and isolation in the month prior to the incident.

Institutional firesetting is associated with increased risk of harm to self and others. Plutchik et al. (1989) reported that psychiatric patients who engaged in both suicidal and violent behaviour were more likely to engage in firesetting, while similar patterns have been reported in both male and female prisoners (Kottler et al., 2018; Slade et al., 2020). Slade (2018) studied dual harm (histories of aggression towards self and others, not necessarily simultaneously) in two English prisons and reported that 22.9% of prisoners engaging in dual harm in one prison had engaged in institutional firesetting, while in the other prison, the proportion was 31.8%). Slade et al. (2020) also found that male prisoners who engage in dual harm use a greater variety of self-harm methods and more highly lethal methods than those who only self-harm. Pickering et al. (2022) studied the experiences of prisoners who engaged in dual harm and identified a number of factors that stemmed from early traumatic experiences and for violence and self-harm were forms of coping. These included feelings of intense distress, an incoherent sense of self and experiencing difficult and unpredictable environments. Since institutional firesetting can function as both an act of self-harm and an act of violence towards others it could be supposed that it would be commonly used by individuals who seek to express violence both towards themselves and others. Alternatively, for individuals who are experiencing intense distress which they struggle to understand or articulate, institutional firesetting may provide a form of expression that directed both towards themselves and towards others.

The present study was carried out in a high secure psychiatric hospital in the UK and was commissioned after a number of fires started by patients which highlighted the need to better understand and manage institutional firesetting behaviour in this population. Given the paucity of research on institutional firesetting, the first aim of the present study was to provide a description of the frequency and characteristics of institutional firesetting among patients in this population, their modi operandi and motivations for institutional firesetting, based on information from patient records.

The second aim was to ascertain whether, for those who had set fires in the community and in institutions, the motivations for community-based firesetting and institutional firesetting were the same or different. In order to explore this question codebook thematic analysis (Braun and Clark, 2006) was used to analyse the motivations for community-based and institutional firesetting, based on the Multi-Trajectory Theory of Adult Firesetting (M-TTAF; Gannon et al., 2012), and the motivations for community-based and institutional fires by each patient who had set fires both in the community and in institutions were compared.

The third aim was to determine the extent to which risk of firesetting was assessed and managed.

The study also sought seeks to test two specific hypotheses:

Firstly, the association between firesetting and a diagnosis of personality disorder, and between institutional firesetting and both self-harm and aggression towards others, as well as the findings of Pickering et al. (2022) that dual harm is associated with childhood trauma, feelings of intense distress and an incoherent sense of self suggest that dual harm and institutional firesetting would be particularly common among individuals with a diagnosis of personality disorder, and in particular borderline personality disorder.

Secondly, the findings of Pickering et al. (2022) would also suggest that institutional firesetters would predominantly follow the emotionally expressive/need for recognition M-TTAF trajectory, using fire to communicate needs and hopelessness, and as a ‘cry for help’.

Method

Participants

Data was collected from the files of male patients with a primary diagnosis of mental illness or personality disorder detained in a high secure psychiatric hospital in the UK and who had deliberately started a fire while detained in prison or a secure psychiatric unit, regardless of whether they been prosecuted for this behaviour.

Data collection

Data collection took place during August 2022. Patients with any history of firesetting behaviour had previously been identified in an audit. For the purposes of this study, ward psychologists were asked to identify any patients on their ward with a history of firesetting in prison or secure hospital. The researchers reviewed the HCRv3 (Douglas et al., 2013), Police National Computer records and psychological assessment reports of all patients thus identified who were currently detained in the hospital. All files were reviewed independently by both researchers. The following data were collected:

Demographic information

Data was collected on psychiatric diagnoses and whether the patient had a history of institutional violence or institutional self-harm. For the purposes of this study, episodes of institutional firesetting were not included as episodes of violence or self-harm.

Firesetting history information

Data was collected on patients’ firesetting history in the community: whether they had convictions for fire setting in the community, whether they had a recorded history of fire setting in the community for which they had not been convicted, and the motives for community-based fire setting.

Institutional firesetting information

The following data was collected on patients’ firesetting history in forensic institutions: the location (prison/hospital) and number of recorded institutional fire setting incidents; whether they had been convicted for any of these offences; details of firesetting incidents (location in the prison/hospital; whether the offence was committed alone or with others; ignition source and fuel used and motivation.

Motivational thematic analysis

Codebook thematic analysis (Braun & Clarke, 2006) was carried out on the available data to classify the motivation for community and institutional firesetting offences. The authors created a priori codes, based on the Multi-Trajectory Theory of Adult Firesetting (M-TTAF; Gannon et al., 2012). Codes included a label corresponding to each of the five firesetting trajectories described in the M-TTAF, a definition of the trajectory and a description of behaviours and internal processes associated with that trajectory. Cases were rated independently by both authors. Given the paucity of information on firesetting offences in some cases, M-TTAF trajectories were only recorded where both raters agreed on the same rating.

Risk information

Information was collected on whether patients had a history of institutional aggression towards themselves or others. HCRv3 risk assessments for each patient were analysed to assess whether they contained a formulation of fire setting risk or future scenario planning relating to institutional firesetting.

Ethical considerations

Ethical approval for this study was given by the University of Lincoln Ethics Application Service, reference UoL2022_9037 and from the Service Evaluation Committee of the NHS Trust in which the research took place. Permission was granted by the Trust’s Caldicott Guardian to access patients’ records without seeking their consent on the ground of patient safety. Data was anonymised and stored on secure trust servers.

Results

Demographic details

Thirty-two patients with a history of institutional firesetting were identified. Their mean age of first institutional firesetting was 25.9 years (range 16–42). Fourteen patients had a history of firesetting in the community, mean age of first community fire was 13.2 years (range 3 to 33). Twenty eight out of 32 patients identified as White British or White Irish. The other four identified as Black or Asian.

Inter-rater reliability of M-TTAF ratings

Sufficient file information was available for the researcher to rate the M-TTAF trajectories of six community fires and 17 institutional fires. Ratings were made independently by both authors on each of these episodes. They agreed in all six community cases and in 12 of the 17 cases of institutional firesetting. Cohen’s kappa was calculated to be 0.73 for the total sample, which McHugh (2012) described as reflecting a moderate level of agreement. For the six community cases, kappa was 1.0, reflecting perfect agreement, while for the institutional firesetting cases, kappa was 0.63, which again fell within the moderate range of agreement. The difference between kappas for community and institutional firesetting probably reflects the greater amount of information that was available on offences committed in the community. Given the paucity of data in some cases, it was agreed to exclude cases where the raters disagreed rather than trying to reach a consensus. Therefore, only those cases where the authors agreed on the classification are included in the results.

Frequency and characteristics of institutional firesetting in this population

Out of a total population of 153 patients, 32 (18.3%) had a history of institutional firesetting. Of these, 14 (43.8%) had a history of firesetting in the community.

Six patients (18.8%) had been convicted for firesetting offences in the community. A further eight (25.0%) had a history of firesetting on their records but no convictions. One patient had both firesetting convictions and a history of unconvicted firesetting.

The 32 patients in this study had started a total of 57 recorded institutional fires (43 in prison, 14 in hospital), though this number will be an underestimate because several patients were simply recorded as having started ‘several’ or ‘numerous’ fires in prison. Most (19 patient, 59.4%) had started one recorded fire, six had started two fires, four patients had started three fires, one patient was responsible for four fires and two patients had started five fires. Seven patients had only started fires in hospital, 22 had only started fires in prison and three had set fires in both hospital and prison.

Information about incidents was incomplete, particularly for incidents in prison. In every case the perpetrator appeared to have been acting alone, rather than in collaboration with others. In the majority of cases where information was recorded (23 out of 28 incidents, 82.1%), fires were started in a prison cell or hospital bedroom. Other fires were started in a day area, in an office, on an exercise yard and outside a cell. Information on the method of starting the fire was often not recorded. In five cases the perpetrator used a lighter or matches as their ignition source, while in another five cases they had generated a spark from a battery or power socket. In all cases where information was available, the perpetrator used available items, such as clothing, bedding or paper as fuel. Of the 57 incidents, only six (10.5%) resulted in the perpetrator being convicted of a criminal offence; four out of 14 fires in hospital (28.5%) and two out of 43 in prison (4.6%), though it was not clear how many offences in prison had resulted in an adjudication.

Comparison of community and institutional firesetting trajectories

There was insufficient information on the previous firesetting by the nine patients with histories of unconvicted firesetting only. Therefore, only the six patients with previous firesetting convictions were included in this part of the analysis. Their M-TTAF trajectories are summarised in Table 1. For three of these patients, their community and institutional M-TTAF trajectories were similar, while for the other three, their trajectories appeared different.

Table 1. Community and institutional M-TTAF trajectories.

Patient	Community M-TTAF Trajectory	Institutional M-TTAF Trajectory
A	Antisocial Cognitions	Emotional Expression
B	Grievance	Multi-faceted
C	Fire Interest	Fire Interest
D	Antisocial Cognitions	Emotional Expression
E	Antisocial Cognitions	Antisocial Cognitions
F	Emotional Expression	Emotional Expression

Firesetting risk assessment

An assessment of fire setting risk was included in the HCRv3 of 10 patients in the sample (31.3%), including all six of those with firesetting convictions. There was no mention of fire setting risk for any of the other patients.

Institutional firesetting and borderline personality disorder

Of the 32 patients with a history of institutional firesetting, 13 had a primary diagnosis of mental illness (14.4% of those with a primary diagnosis of mental illness) and 19 had a primary diagnosis of personality disorder (30.1% of those with a primary diagnosis of personality disorder). The proportions of patients with a history of institutional firesetting was significantly greater among those with a primary diagnosis of personality disorder (Fisher’s exact test, 2-tailed p = .026).

Perhaps not surprisingly for a high secure forensic population, 30 out of 32 patients had a history of institutional violence. Most patients with a primary diagnosis of personality disorder (18 out of 19; 94.7%) but only 4 out of 13 (30.8%) of those with a primary diagnosis of mental illness had a history of institutional self-harm. A history of self-harm was significantly more common among patients with a primary diagnosis of personality disorder (Fisher’s exact test, 2-tailed p = .0002). Most of those with a primary diagnosis of personality disorder had a history of aggression both to self and others (16 out of 19; 84.2%), whereas only four out of thirteen patients (30.8%) with a primary diagnosis of mental illness had a record of dual harm. A history of dual harm was significantly more common among patients with a primary diagnosis of personality disorder (Fisher’s exact test, 2-tailed p = .0035). These results are summarised in Table 2.

Table 2. Firesetting history and risk information.

Primary Diagnosis	Firesetting conviction	Unconvicted firesetting history	Convicted for institutional firesetting	History of institutional violence	History of institutional self-harm	History of dual harm
Mental Illness (n = 13)	2/13 (15.4%)	0/13 (0%)	1/13 (7.7%)	13/13 (100%)	4/14 (30.8%)	4/13 (30.8%)
Personality Disorder (n = 19)	4/19 (21.1%)	7/19 (36.8%)	5/19 (26.3%)	17/19 (89.5%)	18/19 (94.7%)	16/19 (84.2%)
Total (N = 32)	6/32 (18.8%)	7/32* (21.9%)	6/32 (18.8%)	30/32 (93.8%)	22/32*** (68.8%)	20/32** (62.5%)

Fisher exact 2-tailed test value.

*p<.05.

**p<.005.

***p<.0005.

Braun and Clarke (2006).

Slade (2018).

The most common diagnoses among the sample were antisocial personality disorder (19 cases, 59.4%) and borderline personality disorder (15 cases, 46.9%). Among those with a primary diagnosis of personality disorder, 17 out of 19 (89.5%) had a diagnosis of antisocial personality disorder while 14 (73.7%) had a diagnosis of borderline personality disorder. Thirteen (68.4%) of those with a primary diagnosis of personality disorder were diagnosed with both antisocial and borderline personality disorders. The base rate of these diagnoses in this population is not known.

M-TTAF trajectories of institutional firesetters

For 17 out of 32 cases there was sufficient information about institutional firesetting to classify the patient’s M-TTAF trajectory. Of these, the raters agreed on 12 cases. Three were classified as fitting the antisocial cognitions trajectory; one was classified as fitting the grievance trajectory; one was classified as fitting the fire interest trajectory; three were classified as fitting the emotional expression trajectory and four were classified as fitting the multi-faceted trajectory.

Discussion

One noticeable finding from this study is that only a small minority of institutional firesetters (6 out of 32, 15.6%) had convictions for firesetting in the community. Data on previous convictions from the Police National Computer was available for all patients in this study, so we can be confident in the reliability of this finding. A quarter (8 out of 32) were reported to have a history of firesetting on their record, although no conviction. Assessors are generally reliant on prisoners’ or patients’ self-report for this information, so this figure is less reliable and is likely to be an underestimate. It is recognised that detection rates for firesetting in the community are low. For example, the Arson Prevention Forum (2017), quoted by T. Gannon et al. (2022) refers to 21,961 cases or arson recorded by the police in England and Wales in 2015 and 2016, and only 1,242 successful prosecutions. It is therefore likely that many of those with no previous convictions for firesetting had simply managed to avoid detection in the community. However, it is also possible that at least some of the individuals in this study have only started fires in institutions or have started firesetting while detained.

The results provide some information about the nature of institutional firesetting behaviour. For example, it would appear from this sample that in every case, institutional firesetting was carried out alone rather than with others. This contrasts with community firesetting, which is more often carried out with others (Barnoux et al., 2015; Dolan et al., 2011). Institutional fires are most likely to be started in a prison cell or hospital bedroom, which probably reflects opportunities; these are the locations where prisoners and patients spend most of their time, and where they are least likely to be observed.

Another striking result was that, despite its seriousness, the likelihood of being prosecuted for institutional firesetting is low; only six out of 57 incidents in this study (10.5%) resulted in a criminal conviction. The likelihood of being convicted for firesetting in prison is even lower (4.6%), though it is likely that a larger proportion received a formal sanction through the prison adjudication system. In hospital, the rates of conviction are higher (28.5%). However, since hospitals have no sanctions equivalent to prison adjudication, this still means that nearly three quarters of fires started in hospitals resulted in no formal sanction. It is not entirely clear why the rates of prosecution for institutional firesetting are so low. In the experience of the lead author, police and prosecutors are often reluctant to prosecute all but the most serious offences committed by patients detained indefinitely under the Mental Health Act, on the grounds that the patient is already detained and so prosecution is not in the public interest. Similar considerations may well apply to the prosecution of long-term prisoners. An additional factor in prisons is the Crown Prosecution Service guidance on the prosecution of prison-related offences, which states that ‘Prosecutors should note that a cell fire may be an attempt to commit self-harm – these cases should not normally be referred by the prison to the police’ (Crown Prosecution Service, n.d.).

Of course, there may be other, less direct consequences to institutional firesetting, such as increased risk rating or delayed transfer or discharge. Examination of the HCRv3 risk assessments of patients in this study found that risk of firesetting was incorporated into the HCRv3 risk assessments of all those patients with a conviction for firesetting. However, among the rest, it was only considered in four out of 26 cases (15.4%). This probably reflects the lack of information on institutional firesetting, particularly in prison. However, it does suggest that for many institutional firesetters, there is no clear impact of their firesetting on their risk assessment or management.

Given the small number of institutional firesetters with previous firesetting convictions, it is not possible to draw many firm conclusions about how similar the community and institutional M-TTAF trajectories individuals are. However, since they appear to differ in three of the six cases for which data was available, we can conclude that the trajectories will not always be the same. This is not incompatible with the M-TTAF framework (T. A. Gannon et al., 2012) which includes proximal factors and triggers such as life events and environmental factors that interact with psychological vulnerabilities such as poor problem solving skills and moderators such as mental health. For many people, being incarcerated will negatively impact both psychological vulnerabilities and moderating factors. In environments such as prisons or secure hospitals where people’s freedom and sense of agency are severely limited, the likelihood of firesetting as a coping strategy may increase as the number of feasible alternatives decreases.

There was insufficient data to draw any clear conclusions about the M-TTAF trajectories of institutional firesetters in this sample, but all five trajectories appeared to be represented in this sample. Perhaps unsurprisingly in a high secure hospital population the multi-faceted trajectory was the most common. As previously discussed, individuals detained in a high secure setting may have limited coping and problem solving resources, and little sense of agency. For such individuals, firesetting may one of a highly restricted repertoire of problem solving strategies. For individuals experiencing intolerable distress, firesetting in one’s cell or room is likely to result in a rapid change in internal states (causing intense pain, fear or excitement) and external circumstances (change in location) that provides at least temporary relief from distress and so be negatively reinforced. Many such individuals may struggle with consequential thinking in general, but particularly when in acute distress, so they may well not consider the longer-term consequences of their actions. However, as we have seen the likelihood of serious direct sanctions, or of changes to the individual’s risk assessment appear to be low, in which case they may consider that firesetting is an effective problem solving strategy.

Institutional firesetting was significantly more common among those with a primary diagnosis of personality disorder than among those with a primary diagnosis of mental illness. Among those with a diagnosis of personality disorder, most were diagnosed with both antisocial and borderline personality disorders. However, the baseline rate of both diagnoses is likely to be very high in a high secure personality disorder population, so this finding may not be generalisable to other populations.

This study confirms the findings of previous studies (Kottler et al., 2018; Slade, 2018; Slade et al., 2020) of a correlation between institutional firesetting and dual harm. This association was particularly strong among those with a primary diagnosis of personality disorder, though the baseline rate of dual harm in a high secure personality disorder population is also likely to be high.

Limitations

This study is based on a relatively small sample and caution should be exercised in drawing conclusions about offenders in other settings. Patients are often referred to high secure psychiatric services because of their challenging behaviour in other settings, which may include institutional firesetting. Institutional firesetting is therefore likely to be more prevalent in this sample than in the general forensic population.

The main limitation of this study has been the quality of the data. Information on fires that had been started in the hospital where the study took place, or in other hospitals was relatively good. Data from prisons was patchy and largely incomplete. This reflects the fact that complete records are not shared between prison and forensic mental health establishments and staff must rely on summary reports, meaning that many details of behaviour are lost between services. Moreover, it is likely that the vast majority of firesetting incidents in prison were dealt with, if at all, by prison adjudication, for which the amount of information recorded would be much less than for a criminal prosecution.

First-hand accounts from patients of institutional firesetting were noticeably rare, meaning that judgements about motivation generally had to be inferred from other sources, meaning that conclusions about motivations should be treated as tentative. In the experience of the lead author, staff in hospitals are often discouraged from discussing the motivations for firesetting while offences are under investigation by the police or awaiting trial, a process that can take many months or even years. This can result in patients never being asked about their institutional firesetting.

Future directions

The results of this study suggest that there may be a subgroup of institutional firesetters with no previous history of firesetting and this is a group that merits more detailed investigation. First-hand accounts of institutional firesetters were rare in this study but would be essential for any further study in understanding the firesetting histories and trajectories of these individuals, as well as the actual and perceived consequences of their firesetting. This would help to improve our understanding of the phenomenon of institutional firesetting and improve risk assessment processes.

As well as the institutional firesetters who were the focus of this study, there are also a significant number of prisoners and patients with histories of firesetting in the community who have not engaged in institutional firesetting. A related study would be to investigate the differences between these community firesetters and institutional firesetters and the protective factors that keep them from offending while detained.

Disclosure statement

No potential conflict of interest was reported by the author(s).