Protective Factors in Forensic Practice: The Added Value of the SAPROF-Extended Version Pilot in Relation to Aggressive Incidents

Abstract

Protective factors are now commonly included in comprehensive risk assessment. This study concerns an initial validation of the new SAPROF-Extended Version (SAPROF-EV) pilot, containing modifications and additions to the original SAPROF. For 139 forensic psychiatric inpatients, assessment results with the SAPROF-EV pilot and HCR-20^V3 were compared with aggressive incidents. Results show good predictive validity for the SAPROF-EV pilot for all outcomes and incremental predictive validity for both the (modified) original SAPROF and the full SAPROF-EV pilot over the HCR-20^V3. For the outcome aggression toward others, the additional SAPROF-EV factors provide incremental predictive validity over the (modified) original SAPROF. In addition, the user feedback from clinicians highlights experienced additional value of the new factors for treatment guidance. Based on the findings from this study, the SAPROF-EV pilot will be adjusted further into an improved and enhanced version of the SAPROF.

Keywords

SAPROF protective factors risk assessment predictive validity aggression

Introduction

The first goal of forensic psychiatric treatment is to prevent recidivism by addressing factors related to future violent behavior (Nedopil, 2009). Therefore, repeated and accurate risk assessment is needed during treatment to determine which specific criminogenic needs should be addressed to protect society (Neves et al., 2019). Currently, various structured instruments for risk assessments of adults are available (Hogan & Olver, 2019). One of the most widely used risk assessment instruments in forensic clinical practice is the HCR-20^V3 (Douglas et al., 2013). Meta-analytic and systematic reviews have shown that structured risk-only assessment instruments are generally fairly good predictors of future (inpatient) violence (e.g., Fazel et al., 2022; Hogan & Ennis, 2010; Ramesh et al., 2018). This is also found for the HCR-20^V3 (Douglas & Shaffer, 2020).

In addition to targeting risk factors, forensic psychiatric treatment aims to promote secure recovery (i.e., restoration of mental health and social functioning) to enable safe reintegration into society (Nedopil, 2009; Neves et al., 2019). To address this second—but equally important to the first—goal, scholars and treatment providers have started challenging the one-sided view of altering an individual’s potential for future violence through diminishing risk factors alone (Rogers, 2000). There is a growing acknowledgment from different disciplines—such as the desistance paradigm (criminology; Maruna & LeBel, 2010) and the Good Lives Model (forensic psychology; Ward, 2002)—of the importance of reinforcing strengths in forensic psychiatric patients (Vandevelde et al., 2017).

From these two goals—prevent recidivism as much as possible and promote secure recovery as well as possible—the addition of protective factors in risk assessment appears to be able to provide a valuable contribution. Explicitly evaluating and strengthening protective factors for an individual is recognized as helpful to identify assets that reduce overall risk level and to support defensible decision-making to prevent future violence (Doyle et al., 2021). Positive, strengths-based treatment programs are considered to contribute to the prevention of future violence (Fortune et al., 2011). A profound evaluation of protective factors can help to better tailor strengths-based interventions and to offer improved opportunities for rehabilitation within a more person-oriented approach in forensic practice, even for high-risk individuals (Olver & Riemer, 2021).

Protective factors could be defined as characteristics of a person, their environment, or the situation which enable or assist desistance from offending for those who have already offended or prevent the onset of offending among at-risk populations (De Vries Robbé, Mann et al., 2015). They can be found in the person’s history but are often dynamic personal attributes, motivational factors or external environmental circumstances (Neves et al., 2019). Including the positive focus of personal and situational strengths could potentially improve the predictive validity of violence risk assessment and enhance its clinical utility. Protective factors can provide additional optimistic treatment targets and offer hope and inspiration for rehabilitation (De Vries Robbé & Willis, 2017).

One commonly used tool developed to explicitly assess protective factors is the Structured Assessment of Protective Factors for violence risk (SAPROF; De Vogel et al., 2009; 2nd Edition 2012). This tool was designed to complement risk-focused assessment tools, such as the HCR-20^V3. In the past decade, the SAPROF has been widely studied and evaluated in different settings and countries. Overall, good interrater reliability and predictive validity for desistance from violent recidivism, sexual recidivism, institutional misconduct and self-inflicted harm, as well as for positive community outcomes, have been demonstrated. Findings concern the use of the SAPROF in inpatient as well as outpatient settings and with shorter or longer follow-up times (for an overview see De Vries Robbé et al., 2020, and a meta-analysis by Natoli & Flake, 2022). In addition, several studies have pointed out the incremental predictive validity of the protective factors in the SAPROF when used in combination with risk-focused assessment tools, such as the HCR-20^V3 (e.g., De Vries Robbé et al., 2013; Kashiwagi et al., 2018; Oziel et al., 2020 see also a meta-analysis by Burghart et al., 2023). Since prevention of (physical) aggression is often the main focus of forensic treatment and supervision (Andreasson et al., 2014), especially for treatment providers and legal decision makers assessing both risk and protective factors, and deriving personalized (positive) treatment targets from this, offers additional opportunities for interventions.

Despite the overall positive findings regarding the validity of the SAPROF, clinicians from long-term inpatient settings have commented that the original SAPROF is not fully sufficient to assess all the relevant protective factors for this setting and that the clinical utility of protective factors for treatment and supervision practice could be further enhanced. Patients with a severe psychiatric disorder and a long-term care trajectory and/or patients with an intellectual disability often need intensive support, and protective factors may play an important role in their recovery (Jankovic et al., 2021; Neves et al., 2019). Therefore, it has been proposed that extending the original SAPROF with additional protective factors could possibly be valuable for risk assessment and treatment guidance, especially in inpatient forensic practice (Neves et al., 2019).

To develop more tailored assessments, suggestions for additional protective factors were gathered to complement the original list of protective factors in the SAPROF. These new protective factors came from literature reviews on strengths for individuals with intellectual disabilities as well as from literature regarding desistance for individuals in long-term psychiatric care (for a more detailed description see De Vries Robbé et al., 2018, or Neves et al., 2019). Both approaches led to overlapping insights regarding strengths for these groups of individuals and brought forth similar suggestions for additional protective factors to supplement the SAPROF. Subsequently, it was decided to integrate these into a new pilot version for an enhanced SAPROF: the SAPROF-Extended Version pilot (SAPROF-EV pilot; De Vries Robbé et al., 2018). This elaboration of the original SAPROF includes seven proposed new protective factors (see Table 1). In addition, the new pilot manual also includes revised coding suggestions for some of the original SAPROF factors. The proposed modifications include the splitting of the protective factors Self-control (2) and Social network (3) into several subfactors, making a new list of 20 protective factors (in this pilot study called the “modified original SAPROF”). The seven newly proposed protective factors together with the 20 modified original SAPROF factors make the new and improved SAPROF-EV pilot version, altogether containing 27 protective factors (see Table 1). Since health care providers in clinical practice indicated that in their view the new elements in the SAPROF-EV pilot manual contain overall valuable improvements to the tool, they suggested that the new SAPROF-Extended Version pilot might provide a general enhancement to the SAPROF that could be useful for a wide range of individuals in (inpatient) forensic care.

Table 1:

The SAPROF-Extended Version Protective Factors: Original, Modified and Additional Factors

Original SAPROF factors	Modified SAPROF factors	Additional SAPROF factors
Internal factors
Intelligence
Secure attachment in childhood
Empathy
Coping
Self-control^a	Self-control:Frustration tolerance
	Self-control:Temptation resistance
		Social competence
		Quality of life
		Self-efficacy
Motivational factors
Work
Leisure activities
Financial management
Motivation for treatment
Attitudes toward authority
Life goals
Medication
		Participation in treatment program
		Hobby’s & use of personal time
		Sleep
		Therapeutic alliance
External factors
Social network^a	Social network:Parents/guardians
	Social network:Peers
	Social network:Other supportive relationships
Intimate relationship
Professional care
Living circumstances
External control

Note. SAPROF = Structured Assessment of Protective Factors for violence risk.

These factors have been split up into multiple new factors in the “modified original SAPROF.”

From an implementation research perspective (Peters et al., 2013), it is also important to gain insight into the user experiences of clinicians regarding the updated instrument. Gaining insight into the perspectives of those anticipated to apply innovations in practice and listening to their experienced challenges and observed benefits in clinical practice is important in any sound implementation process, also when it concerns novel risk assessment practice (Vincent et al., 2012). It is important to evaluate whether the new instrument fits with the setting and the clinicians and to indicate possible barriers and facilitators to ensure the successful adoption of a (new) assessment tool (De Beuf et al., 2020). Risk assessments will only be able to contribute to the reduction of violence if they bring forth relevant intervention targets and if assessment results are communicated and implemented in individualized risk management plans (Völlm et al., 2018).

The aim of the present study was to examine the added value of the SAPROF-EV pilot version for violence risk assessment in inpatient forensic practice. The study consisted of two parts: (a) examination of the (incremental) predictive validity of this instrument for different types of aggressive incidents during treatment and (b) evaluation of the experienced clinical utility by clinicians in the daily clinical practice of an inpatient forensic psychiatric treatment setting. For the purpose of this study, the SAPROF-EV pilot was implemented in a forensic psychiatric hospital in Belgium and subsequently tested extensively for 2 years. This hospital was chosen for the current study because of its ambition to increasingly implement strengths-based approaches in daily clinical practice and because the original SAPROF and the HCR-20^V3 have been used there routinely for many years. Furthermore, the hospital had a sound registration in place for aggressive incidents (Cappon et al., 2022) by means of the Modified Overt Aggression Scale (MOAS, Kay et al., 1988). The findings from this study will likely generate insights that guide further adjustment of the SAPROF-EV pilot.

First, the predictive validity of the HCR-20^V3 and the SAPROF-EV pilot for different types of aggression during follow-up, as administered with the MOAS, was examined. This also included an investigation of the incremental predictive validity of the SAPROF-EV pilot over the HCR-20^V3 and of the additional protective factors over the modified original SAPROF. The different types of aggression measured by the MOAS—physical aggression toward others, verbal aggression, self-directed aggression, and aggression toward objects—were all considered relevant outcome measures for this study, as previous research demonstrated good predictive validity of the SAPROF (original version) for these outcomes (e.g., Abidin et al., 2013; Neil et al., 2020). As the proposed new protective factors in the SAPROF-EV pilot stem from literature reviews regarding strengths for patients in intensive mental health care, it was hypothesized that the new factors would also show good predictive validity for the different types of inpatient aggression in the forensic psychiatric hospital. Furthermore, in line with previous findings (Burghart et al., 2023), incremental predictive validity of the SAPROF-EV protective factors over the HCR-20^V3 risk factors was expected. The incremental predictive validity of protective factors over risk factors was anticipated to further increase when the additional protective factor parameters of the extended version are added.

Second, a qualitative user evaluation of the SAPROF-EV pilot was carried out with the clinicians who applied this new assessment tool firsthand during the study period. This evaluation focused on the user-experiences with the instrument (information gathering process, clarity of scoring instructions) and on the perceived added value of the modified and additional factors (cf. implementation research, De Beuf et al., 2020; Vincent et al., 2012). It was hypothesized that the experienced clinical relevance of the extended version would bring forth added value to the hospital’s standard risk assessment practice with the HCR-20^V3 and the original SAPROF.

Method

Setting

The current study was conducted in the forensic psychiatric hospital Sint-Jan-Baptist in Zelzate (Belgium). This hospital offers forensic psychiatric treatment to both males and females (185 patients) who are declared not fully criminally responsible due to a serious mental disorder. It consists of seven mixed-gender medium-security wards, one high-security female-only ward, and one male-only ward for patients who have committed sexual offenses. The different wards vary in the level of treatment intensity. The average length of stay at the forensic psychiatric hospital is approximately 5 years.

Participants

The 162 risk assessments were rated for 139 individual patients, of which about three-quarters were male (74.8%). For 23 patients, two risk assessments were available during the study time. The mean age of the patients at the time of assessment was 44.7 years (25-70 years). Their mean IQ-score—measured with the WAIS-IV—was 73.7, ranging from 45 to 124. The most commonly diagnosed psychiatric disorders in this patient population were personality disorders (66.9%; primarily borderline [24.4%], antisocial (15.1%) and not otherwise specified (18.7%)), substance abuse disorders (40.3%), and psychotic disorders (32.4%). All patients were admitted for committing an offense. For the majority, these offenses were categorized as violent (e.g., [attempted] manslaughter, violent assault).

Instruments

HCR-20^V3

The HCR-20^V3 (Douglas et al., 2013) is the revised version of the HCR-20 containing 20 main risk factors divided into the Historical (10 items), Clinical (five items), and Risk Management (five items) scales. The risk factors were rated on a three-level response format (No, Partially, Yes; in the present study transposed to numerical scores 0, 1, and 2). The HCR-20^V3 subscales and relevance ratings were not used in the current study. Research in forensic psychiatry had shown good interrater reliability and predictive validity for the HCR-20^V3 total scores, both for inpatient violence and for violent recidivism (e.g., Brookstein et al., 2020; Douglas & Shaffer, 2020; Smith et al., 2020).

In clinical practice, the risk assessment concludes with a structured professional judgment (SPJ) conclusion regarding the overall level of violence risk. The assessor decides upon this final risk judgment based on carefully considering, weighing and integrating the most relevant assessed risk and protective factors, including an extensive contemplation regarding the most likely violence risk scenarios for the individual in the near future. In the current research study, however, the focus was on the added value of specific factors. For this reason, total scores were composed for each tool, which were subsequently used for the predictive validity analyses. In the current study, the total score of the HCR-20^V3 ranged from 8 to 33, with a mean of 22.1.

SAPROF

The SAPROF (De Vogel et al., 2012) is an instrument to assess protective factors that are usually applied in addition to risk-focused risk assessment tools, such as the HCR-20^V3. The original SAPROF contains 17 factors divided into the Internal (five items), Motivational (seven items), and External (five items) scales. Although in practice the protective factors may also be scored on a seven-point scale, in this study the factors were rated on a three-point scale (0, 1, 2). Also, the SAPROF subscales and relevance ratings (keys and goals) for the individual were not included in the present study. Research in forensic psychiatry had demonstrated good psychometric properties for the SAPROF in regard to interrater reliability and predictive validity for both inpatient aggression and community recidivism (e.g., Burghart et al., 2023; De Vries Robbé et al., 2020; Natoli & Flake, 2022).

The SAPROF-EV pilot includes both additional new items and modifications to the original SAPROF to increase the clinical relevance of the original 17 factors. The modifications included the subdivision of two of the original SAPROF items into subitems, based on observations in previous studies and experiences with the SAPROF-Youth Version (De Vries Robbé, Geers et al., 2015). First, the item Self-control was divided into (a) Frustration tolerance and (b) Temptation resistance. Second, the item Social network was divided into (a) Parents/guardians; (b) Peers; and (c) Other supportive relationships. The original SAPROF together with the modified factors altogether consists of 20 protective factors, which for the purpose of this study is named the “modified original SAPROF.” The seven additional factors of the SAPROF-EV pilot are: (a) Social competence (Internal); (b) Quality of life (Internal); (c) Self-efficacy (Internal); (d) Participation in treatment program (Motivational); (e) Hobbies and use of personal time (Motivational); (f) Sleep (Motivational); and (g) Therapeutic alliance (Motivational). Merging the 20 modified original SAPROF factors with the seven additional factors resulted in the SAPROF-EV pilot version (named SAPROF-EV from here on in this paper) consisting of 27 protective factors (Table 1). In practice, the SAPROF is applied as an SPJ measure, concluding with an overall final protection judgment. However, in the current study, the final judgments were not analyzed as the study explicitly concerns the added value of the revised and new factors compared with the original version. Therefore, a total score was composed. The modified original SAPROF had a mean total score of 20.9, ranging from 10 to 31. The average total score of the SAPROF-EV in this study was 30 and ranged from 13 to 45. Finally, the mean total score of the additional factors was 9.1, ranging from 3 to 14.

Modified Overt Aggression Scale

The Modified Overt Aggression Scale (MOAS; Kay et al., 1988) has been used since 2016 in the study setting to register inpatient aggressive incidents (Cappon et al., 2022). The MOAS is one of the most commonly used instruments for the administration of aggressive incidents in clinical practice and has been shown to be applicable in forensic psychiatric settings (Nijman et al., 2006; Verstegen et al., 2017). It has demonstrated good psychometric properties (Kay et al., 1988; Oliver et al., 2007). The instrument encompasses a broad definition of aggression—including physical aggression toward persons, verbal aggression, self-directed aggression, and aggression toward objects—and provides clear behavioral anchors to minimize interpretation by the rater. The severity of each type of aggression can also be rated (mild, moderate, strong and extreme). However, in the current study, the MOAS scores gathered were converted into binary responses for each type of aggression (0 = not present, 1 = present, regardless of severity) in two follow-up periods: 3 months and 6 months after the moment of risk assessment. All staff members at the different wards were trained to use the MOAS to ensure that each staff member correctly registered each aggressive incident. Therefore, the MOAS scores can be considered an appropriate measure to examine the presence of aggressive incidents during forensic psychiatric treatment.

The base rates for aggression ranged between 4% to 17% for the different types of aggression at the 3-month follow-up (7.4% physical aggression toward others; 16.7% verbal aggression; 3.7% self-directed aggression; and 4.9% aggression toward objects) and increased to 5% to 27% at the 6-month follow-up (11.6% physical aggression toward others; 27.2% verbal aggression; 4.8% self-directed aggression; and 10.2% aggression toward objects).

Procedure

At the start of the study period, the clinicians responsible for the risk assessment were trained in the use of the SAPROF-EV. They were all already experienced and trained users of the HCR-20^V3 and the original SAPROF. Their average level of experience with these tools was 6.5 years (ranging from 4.5 to 8 years). Risk assessment data (HCR-20^V3 and SAPROF-EV) were collected from the forensic psychiatric hospital between January 2019 and December 2020, resulting in 162 risk assessments for 139 individuals. These were all the individual patients who had a risk assessment during the research period and had a follow-up period of at least 3 months. The medium-security ward for persons who have committed sexual offenses (24 beds) was excluded from this study because the SAPROF-EV pilot was not used there. At this ward, protective factors were rated based on the original SAPROF supplemented by the pilot SPJ—sexual offending version (SAPROF-SO^SPJ pilot, De Vries Robbé et al., 2019).

To assess the predictive validity of the risk assessment instruments, the already registered MOAS data were used at two follow-up periods—3 months and 6 months after the administration date of the risk assessment tools. These follow-up periods were chosen for practical reasons and were in line with previous research regarding inpatient aggression (e.g., Abidin et al., 2013; Judges, 2016). MOAS data were available for all 162 risk assessments for the 3-month follow-up and 147 risk assessments for the 6-month follow-up.

To assess the usefulness of the SAPROF-EV in daily clinical practice, a focus group was conducted in June 2021 with the five criminologists responsible for the administration of the risk assessment tools in the studied setting. At the time of the focus group, they each had already scored the SAPROF-EV approximately 30 times. The main goal of the focus group was to explore whether the modified factors and the additional factors were considered of added value in daily clinical practice. The participants scored the added value of each item on a 4-point Likert-type scale: (0) no added value, (1) limited added value, (2) sufficient added value, and (3) large added value. In addition, the clarity of the scoring instructions and the ease of information gathering for the modified and additional factors were evaluated. Finally, the participants were also asked to elaborate on the importance of administering the protective factors in clinical practice. The focus group lasted for approximately 2 hr.

Data Analysis

Statistical analyses were carried out using R studio version 3.4.2 and SPSS version 25. The convergent validity of the SAPROF-EV was examined by estimating Pearson correlations between the HCR-20^V3, the modified original SAPROF and the additional factors. Second, Receiver–Operating Characteristic (ROC) analyses were conducted. ROC analyses result in area under the curve values (AUC) values. They are relatively insensitive to fluctuations in base rates. Rice and Harris (2005) provided guidelines for interpreting AUC values and facilitated comparison across studies by applying different effect sizes. AUC values between .56 and .64 are interpreted as small effect; AUC values between .64 and .71 medium effect; and AUC values of .71 and above large effect. These ROC analyses were conducted for the total scores of the different instruments and at item level for the additional and modified factors, for each type of aggression registered with the MOAS at both follow-up periods. AUC values for the SAPROF were mirrored, thus reflecting the absence of aggression. Finally, incremental predictive validity was examined through a set of logistic regression analyses for each type of aggression at both follow-up periods. These analyses were done in three steps: (1) HCR-20^V3; (2) HCR-20^V3 and modified original SAPROF; and (3) HCR-20^V3, modified original SAPROF and additional protective factors. To evaluate the overall prediction accuracy of the models, we also calculated AUC values for all logistic regression analyses.

The notes of the focus group were integrated in a report, which was sent for approval to the participants. The revised report was then thematically analyzed (Braun & Clarke, 2006). The modified and new factors were used as overarching nodes in the analysis and were further divided into nodes concerning information on the added value, the scoring instructions and the information gathering process of the discussed factors.

Results

Predictive Accuracy of the Different Measures

Risk and protective factors demonstrated to be strongly and negatively associated with each other (r[HCR-20^V3, SAPROF modified original] = −.67, r[HCR-20^V3, additional protective factors] = −.53, r[HCR-20^V3, SAPROF-EV] = −.68). The modified original SAPROF was strongly and positively associated with the additional protective factors (r[SAPROF modified original, additional factors] = .67).

Table 2 provides the predictive accuracy of the total scores of the different tools used in this study in terms of AUC analyses. For all outcomes, the AUC values varied between .68 and .91 at 3 months follow-up, or .69 and .89 at 6 months follow-up. Apart from the predictive validity of the HCR-20^V3 for object aggression, all AUC values were greater than or equal to .71 for the four types of aggressive behavior at both follow-up times, which can be considered a large effect size.

Table 2:

AUC Values for Predicting Aggressive Behavior at 3- and 6-Month Follow-Up

SPJ instruments	Physical aggression		Verbal aggression		Self-directed aggression		Aggression toward objects
SPJ instruments	AUC	95% CI	AUC	95% CI	AUC	95% CI	AUC	95% CI
3-month follow-up (N = 162)
HCR-20^V3	.78***	[.64, .91]	.77***	[.66, .87]	.91***	[.86, .97]	.68	[.47, .88]
SAPROF modified original	.86***	[.77, .96]	.83***	[.71, .92]	.91***	[.83, .99]	.83***	[.73, .94]
Additional factors	.84***	[.72, .96]	.79***	[.65, .93]	.78***	[.66, .90]	.73*	[.51, .94]
SAPROF-EV	.88***	[.80, .97]	.83***	[.73, .93]	.89***	[.82, .97]	.82***	[.69, .95]
6-month follow-up (N = 147)
HCR-20^V3	.73**	[.59, .87]	.76***	[.67, .85]	.89***	[.83, .95]	.69*	[.54, .83]
SAPROF modified original	.81***	[.70, .93]	.78***	[.69, .87]	.89***	[.81, .97]	.77***	[.64, .90]
Additional factors	.80***	[.69, .91]	.71***	[.61, .80]	.79***	[.69, .90]	.74***	[.60, .87]
SAPROF-EV	.83***	[.72, .93]	.77***	[.68, .86]	.88***	[.82, .95]	.78***	[.67, .90]

Note. AUC = area under the ROC curve; SPJ = structured professional judgment; 95% CI = 95% confidence interval for AUC; “HCR-20V3” = sum score of the 20 HCR-20V3 risk factors; SAPROF = Structured Assessment of Protective Factors for violence risk; SAPROF-EV = Structured Assessment of Protective Factors for violence risk —Extended Version; “SAPROF-EV” = sum score of the 27 modified original SAPROF and additional factors; “SAPROF modified original” = sum score of the 20 modified original SAPROF protective factors; “Additional factors”: sum score of the 7 additional protective factors; . Reference category is “0” (no aggressive behavior) for the HCR-20V3 and “1” (aggressive behavior) for all SAPROF measures.

p < .05. ** p < .01. *** p < .001.

Item-Level Descriptives and Predictive Validity of the Modified and Additional Factors of the SAPROF-EV

Table 3 presents item score percentages and AUC values for the modified and additional factors of the SAPROF-EV. The descriptives show large variability across these protective factors. In general, most individual modified and additional SAPROF-EV factors demonstrated AUC values <.64 for at least one of the outcomes, some were as high as .88, which can be considered a medium to large effect. For most factors, predictive accuracy was best for not showing physical aggression and slightly lower for no verbal aggression. However, the factors Sleep and Parents/guardians (Social network) yielded poor AUC values overall.

Table 3:

Item Scores and AUC Values at 3 and 6 Months Follow-Up for Modified and Additional SAPROF-EV Factors

Modified and additional factors of the SAPROF-EV	Item scores (N = 162)			AUC values at 3-month follow-up (N = 162)				AUC values at 6-month follow-up (N = 147)
Modified and additional factors of the SAPROF-EV	0(%)	1(%)	2(%)	Physical Aggression	Verbal Aggression	Self-directed aggression	Aggression toward objects	Physical Aggression	Verbal Aggression	Self-directed aggression	Aggression toward objects
Frustration tolerance	20.3	55.1	24.7	.82	.69	.83	.78	.73	.73	.75	.78
Temptation resistance	17.6	43.1	39.2	.81	.73	.80	.70	.83	.72	.82	.72
Social competence	17.3	59.3	23.5	.74	.63	.69	.73	.70	.62	.71	.73
Quality of life	13.0	54.3	32.7	.81	.64	.79	.88	.74	.65	.69	.71
Self-efficacy	19.9	60.2	19.9	.72	.58	.67	.71	.72	.59	.62	.65
Participation in treatment program	5.6	39.8	54.7	.66	.58	.55	.54	.70	.61	.62	.56
Hobbies & use of personal time	13.0	43.2	43.8	.74	.67	.69	.79	.69	.58	.72	.61
Sleep	2.5	22.2	75.3	.43	.56	.45	.46	.43	.53	.51	.48
Therapeutic alliance	9.3	46.3	44.4	.79	.60	.56	.71	.73	.64	.63	.62
Parents/guardians	42.2	23.0	34.8	.52	.57	.54	.47	.53	.54	.62	.45
Peers	48.1	38.9	13.0	.66	.56	.77	.69	.62	.64	.64	.66
Other supportive relationships	62.7	29.2	8.1	.71	.59	.70	.62	.58	.56	.60	.61

Note. Reference category is “1” (aggressive behavior) for all SAPROF-EV factors. AUC = area under the ROC curve; SAPROF-EV = Structured Assessment of Protective Factors for violence risk—extended version.

Incremental Predictive Validity of the Modified Original SAPROF and the Additional Factors

In Table 4, the logistic regression analyses concerning incremental predictive validity are presented. Across both follow-up periods and the four types of aggressive behavior, it was found that risk factors were statistically significantly associated with an 11% to 49% increase in odds except for object aggression at 3 months (see the first blocks in Table 4). Similarly, the modified original SAPROF provided additional value over the HCR-20^V3 for all eight outcomes. There was a 13-28% decrease in odds for violent behavior if the modified original SAPROF score increased by one point (see the second blocks in Table 4). The decrease in odds was statistically significant in all cases. However, further including the additional protective factors lead to mixed findings. For verbal aggression, aggression toward objects and self-directed aggression at 3- and 6-month of follow-up, the results were consistent with the previous step, that is., the modified original SAPROF was a statistically significant predictor of violent behavior in Step 2, but the additional protective factors did not provide significant additional predictive value in Step 3. On the contrary, for physical aggression there was a statistically significant effect of the additional protective factors over the other measures, such that the odds dropped by 33% to 37% for both follow-up periods, meaning that the likelihood of physical aggression was further reduced by the presence of the additional protective factors. The accuracy of the overall prediction model was high for the various outcomes at both follow-up periods (AUC for physical aggression = .90/.85; verbal aggression = .82/.80; self-directed aggression = .93/.92; object aggression = .83/.79; see Table 4).

Table 4:

Logistic Regression Analyses Predicting Aggressive Incidents During Treatment at 3- and 6-Month Follow-Up

SPJ instruments	Physical aggression		Verbal aggression		Self-directed aggression		Aggression toward objects
SPJ instruments	OR	95% CI	OR	95% CI	OR	95% CI	OR	95% CI
3 months follow-up (N = 162)
Block 1
HCR-20^V3	1.28**	[1.09, 1.50]	1.28***	[1.13, 1.44]	1.49***	[1.27, 1.76]	1.18	[0.97, 1.44]
AUC	.76**		.75***		.89***		.67
Block 2
HCR-20^V3	1.03	[0.78, 1.35]	1.10	[0.95, 1.27]	1.42*	[1.03, 1.49]	0.96	[0.69, 1.34]
SAPROF modified original	0.72*	[0.56, 0.93]	0.81***	[0.71, 0.92]	0.73*	[0.57, 0.95]	0.75*	[0.58, 0.97]
AUC	.87***		.81***		.89***		.83**
Block 3
HCR-20^V3	0.97	[0.75, 1.24]	1.11	[0.95, 1.29]	1.30*	[1.01, 1.67]	0.93	[0.69, 1.26]
SAPROF modified original	0.79	[0.60, 1.03]	0.79**	[0.68, 0.92]	0.69*	[0.49, 0.96]	0.80	[0.59, 1.08]
Additional factors	0.63*	[0.43, 0.92]	1.08	[0.85, 1.36]	1.29	[0.79, 2.09]	0.79	[0.47, 1.31]
AUC	.90***		.82***		.93***		.83**
6 months follow-up (N = 147)
Block 1
HCR-20^V3	1.12**	[1.03, 1.21]	1.14***	[1.07, 1.21]	1.24***	[1.15, 1.34]	1.11**	[1.03, 1.19]
AUC	.73**		.76***		.89***		.70*
Block 2
HCR-20^V3	1.03	[0.91, 1.16]	1.08*	[1.01, 1.15]	1.14*	[1.02, 1.27]	1.02	[0.93, 1.15]
SAPROF modified original	0.79*	[0.65, 0.95]	0.87*	[0.77, 0.99]	0.79*	[0.62, 0.99]	0.82*	[0.70, 0.99]
AUC	.81***		.79***		.92***		.78***
Block 3
HCR-20^V3	0.99	[0.88, 1.13]	1.08*	[1.01, 1.16]	1.16*	[1.04, 1.31]	1.02	[0.91, 1.13]
SAPROF modified original	0.87	[0.71, 1.05]	0.85*	[0.74, 0.98]	0.77	[0.58, 1.02]	0.86	[0.71, 1.04]
Additional factors	0.67**	[0.48, 0.89]	1.02	[0.83, 1.25]	1.14	[0.77, 1.72]	0.85	[0.63, 1.12]
AUC	.85***		.80***		.92***		.79***

Note. SPJ = structured professional judgment; OR = odds ratio; 95% CI = 95% confidence interval for odds ratio; AUC = area under the ROC curve. “HCR-20V3” = sum score of the 20 HCR-20V3 risk factors; SAPROF = Structured Assessment of Protective Factors for violence risk; “SAPROF modified original” = sum score of the 20 modified original SAPROF protective factors; “Additional factors” = sum score of the 7 additional protective factors. Reference category is “0” (no aggressive behavior).

p < .05. ** p < .01. *** p < .001.

Incremental Predictive Validity of the SAPROF-EV Over the HCR-20^V3

To investigate the overall incremental predictive validity of the total SAPROF-EV over risk factors alone, another set of logistic regressions was performed (Step 1: HCR-20^V3; Step 2: HCR-20^V3 and SAPROF-EV). As hypothesized, for both follow-up periods, SAPROF-EV scores were associated with a statistically significant decrease in the odds of all four types of aggressive behavior, controlling for the baseline HCR-20^V3 score. For aggressive behavior at 3 months, the results were: physical aggression (odds ratio [OR] = 0.74, 95% confidence interval [CI] = [0.61, 0.89], p = .002, AUC = .90), verbal aggression (OR = 0.88, 95% CI = [0.80, 0.95], p = .003, AUC = .80), self-directed aggression (OR = 0.85, 95% CI = [0.74, 0.98], p = .024, AUC = .92), and object aggression (OR = 0.80, 95% CI = [0.68, 0.94], p = .006, AUC = .83). For aggressive behavior at 6 months, the results were: physical aggression (OR = 0.80, 95% CI = [0.68, 0.92], p = .003, AUC = .84), verbal aggression (OR = 0.91, 95% CI = [0.84, 0.99], p = .033, AUC = .79), self-directed aggression (OR = 0.87, 95% CI = [0.77, 0.98], p = .024, AUC = .90), and object aggression (OR = 0.86, 95% CI = [0.75, 0.96], p = .009, AUC = .79).

User-Feedback Regarding the Usefulness of the Modified and Additional SAPROF-EV Factors in Clinical Practice

First, all five participants highlighted that in their opinion protective factors should not be ignored in risk assessment in forensic psychiatric treatment. These factors were considered by the clinicians as an important addition to risk factors in a comprehensive risk assessment. Using a structured tool such as the SAPROF helped them to focus on protective factors for which there is enough scientific evidence. This way, a more positive strengths-based approach was utilized in the decision-making process regarding future treatment goals. According to the clinicians, the administration of risk and protective factors should be combined simultaneously to come to an integrated judgment concerning the feasibility of a next treatment step.

Second, the participants overall indicated that the scoring instructions of the modified and new factors were sufficiently clear. Specific comments on the scoring instructions are addressed below when the added value of each item is discussed. The information-gathering process for the new factors was also considered to be simple and time-efficient enough to make these factors easy to use in clinical practice. The participants highlighted that discussing the different topics from these factors with the patients themselves and even with their social network members often helped to more accurately rate each item. The extra time needed to rate the additional factors was considered worthwhile by the participants, as they expressed that the new factors provided additional value for generating and prioritizing positive treatment goals.

Third, the participants scored their experienced added value of the modified and additional factors of the SAPROF-EV on a four-point scale (0–3). All modified factors were considered to provide sufficient to large added value (score range: 2.2–2.8). The largest added value was pointed out for the division of the item Self-control in Frustration tolerance and Temptation resistance. According to the participants, this division helped to see self-control from a broader perspective and not only in relation to substance (ab)use. They also indicated that each of these subfactors often received a different score for the individual. The added value of the division of the item Social network was also considered sufficient to large. Within the social network factors, the lowest added value was seen for Other supportive relationships. Given the inpatient context of the assessed patients and given their overall limited social network, the participants indicated that it was quite difficult to denote these additional supportive relationships. Further clarification regarding what type of relationships can be considered here was deemed helpful to be able to rate this item more accurately and to translate this into a clear treatment goal.

Of the seven additional protective factors, five were considered of at least sufficient added value (score range: 2.0-2.8). The largest added value was allocated to the Therapeutic alliance and to Participation in the treatment program. The participants found it important to elaborate on the quality of this alliance and participation, given that the observed quality of treatment participation is seen to be an important factor for aggression prevention. The addition of the item Hobbies and use of personal time (unstructured use of leisure time) next to the item Leisure activities (structured social leisure time activities) was also considered highly useful in clinical practice, especially at (the beginning of) an inpatient treatment. This division gives the rater the opportunity to see whether patients are able to occupy themselves in their free time and to determine future possibilities of attending more structured leisure activities. Further in the trajectory of the patient, some of the participants stated that these 2 factors begin to show an increasing overlap in awarded score.

Only two of the additional protective factors were considered of limited added value: Self-efficacy (1.6) and Sleep (1.4). The participants found Self-efficacy a complicated factor to rate as they felt it was quite often difficult for them to decide whether there is a healthy belief in one’s own capabilities, especially when the patient is diagnosed with a (narcissistic) personality disorder. They stated that clearly specifying the different life domains for which one’s belief in his or her own capabilities should be addressed, would help to more accurately score this item. Sleep was not considered to be a key protective factor by the participating clinicians in the current setting. They commented that sleep might be protective in very specific cases (e.g., psychosis) or settings (e.g., perhaps in community care); however, they did not generally find it of added value in the inpatient setting. It was suggested that when specific importance of sleep is deemed relevant for the case at hand, the item could be added as case-specific extra protective factor but should not be considered a general additional factor in the SAPROF-EV.

Discussion

From the two-sided perspective on forensic psychiatric treatment—prevention of recidivism and promotion of secure recovery—structured and periodical assessment of risk and protective factors with for example the HCR-20^V3 and the SAPROF is recommended. This way, an accurate and up-to-date judgment of the current risk level and the most promising future treatment steps for an individual can be made (Doyle et al., 2021; Neves et al., 2019). However, clinicians from long-term inpatient settings often feel that the current SAPROF is not fully sufficient to assess all the relevant protective factors and that its clinical utility can be enhanced (Neves et al., 2019). This concern resulted in the development of a pilot extension of the SAPROF (i.e., SAPROF-EV), especially applicable for inpatient care, consisting of seven proposed additional protective factors and some modifications to existing SAPROF factors. The aim of the current study was to examine the added value of this pilot version by analyzing its predictive and incremental validity for inpatient aggression and by evaluating the experienced clinical utility in daily clinical practice.

Main Findings

Overall, the statistical analyses and observed user experiences indicated that the SAPROF-EV pilot provided a valuable elaboration of the original SAPROF, resulting in a more accurate assessment of future inpatient (physical) aggression. The study also found incremental predictive validity of protective factors over risk factors when assessing the likelihood of future inpatient violence (Burghart et al., 2023), especially for physical aggression. This is an important finding since institutional aggressive behavior is considered a good predictor of future aggression and hinders resocialization (Andreasson et al., 2014). Preventing institutional aggression by addressing protective factors may also be helpful to achieve a more successful resocialization trajectory.

The proposed additional protective factors provided further incremental predictive validity over the HCR-20^V3 risk factors and the modified original SAPROF protective factors for the outcome of physical aggression. For the other types of aggression, predictive validities of the additional factors were also good, but significant incremental validity of these factors over the other tools was not found. It may be the case that for specific types of aggression, other additional factors are valuable. For example, for self-directed aggression, the predictive validity of the protective factors in the SAPROF-EV pilot overall was good and comparable to previous studies with the SAPROF (e.g., Abidin et al., 2013). However, further incremental predictive validity could possibly be obtained by including additional factors specifically important for the prevention of self-harm (e.g., Nock, 2009). Future studies should address these specific outcomes more thoroughly.

The value of the SAPROF-EV pilot was supported by the user feedback, which stated that the modified factors and most of the new factors were considered of sufficient to large added value for daily clinical practice. The information-gathering process for the additional factors was considered simple and time-efficient. Some factors need further clarification of the coding instructions (e.g., self-efficacy) to ensure better clinical relevance. In addition, the participants stated that the SAPROF-EV helped to bring forth and prioritize further positive treatment goals (e.g., use of free time, improvement of therapeutic alliance). The clinicians also experienced that discussing these protective factors with the patient was valuable to hear the individual’s own opinion regarding protective factors that are present or those that could potentially become intervention targets. In this respect, it is noteworthy to mention that an easy-to-use self-assessment version of the SAPROF-Extended Version pilot has recently also been developed. Asking about protective factors may motivate patients and may help to develop an adequate therapeutic relationship (Doyle et al., 2021; Völlm et al., 2018).

When looking more specifically at the individual modified and additional protective factors, the majority had moderate to high predictive accuracy. However, the additional factor Sleep and the modified network factor Parents/guardians overall demonstrated poor predictive validity. The factor Parents/guardians was adopted from the Youth Version but appears to be too narrow to demonstrate solid validity for adults in an inpatient setting. Nevertheless, including parental support was deemed valuable by clinicians. Based on these mixed findings, modification of this factor will carefully be considered. To improve the empirical relevance, it has been suggested to widen the scope of this factor to include all direct family members. Future studies will have to further investigate the predictive validity of this wider Family factor for aggressive behavior. Based on the clinicians’ feedback, the coding definitions regarding the network factor Other supportive relationships will also be revised. The factor Sleep was found of limited clinical value by the clinicians, in addition to its poor predictive validity. Therefore, when further improving the SAPROF-EV, it is suggested to remove this factor from the tool.

The overall positive statistical and user-feedback findings regarding the modified and additional protective factors of the SAROF-EV pilot offer new potential for intervention planning, especially in inpatient forensic care. Current ongoing studies at inpatient forensic psychiatry in Canada are further investigating the clinical utility and predictive validity of the SAPROF-EV pilot. Future studies will also need to investigate whether the additional factors are also valuable in outpatient or community supervision settings. In particular, the new factors Social competence, Quality of life, Self-efficacy and Therapeutic alliance are anticipated to provide valuable new potential for a wide range of patients and clients in varying treatment and supervision settings.

Limitations

Several limitations to this study should be acknowledged. First, data were gathered directly from clinical practice. Therefore, interrater reliability could not be calculated and the three-point rating scale instead of the commonly applied seven-point rating scale was used. Also, relevance ratings and structured professional judgments were not included. Given the good predictive validity of the new and modified factors and general findings regarding the good interrater reliability of the original SAPROF factors from previous research (e.g., De Vries Robbé et al., 2020; Natoli & Flake, 2022) it is hypothesized that the interrater reliability for the SAPROF-EV pilot factors is likely sound as well, however this needs to be studied further. The use of the three-point rating scale may have had a dampening effect on the findings as generally the seven-point scale offers slightly more accurate ratings. Unfortunately, in the present study it was not possible to analyze the impact of the SAPROF-EV assessment on the overall final risk judgments and to investigate the degree to which the modified and new protective factors altered the final conclusions of the assessment. For use of the tool in clinical practice this remains an important question that should be addressed in future research: how do the protective factors of the SAPROF-EV impact assessments performed by frontline staff and how does the addition of protective factors impact decision-making by administrators? In future studies, a methodology involving multiple independent raters for each case, could further explore the empirical and clinical advantages of incorporating protective factors in risk assessment, and more specifically of the additional SAPROF-EV factors, for formulating the final structured professional judgments. This is especially interesting given the also observed overlap between different factors in this study.

A further possible limitation concerns the study samples. Some patients were present more than once in the dataset as they had multiple risk assessment measurements at different time points. However, when the analyses were performed for only one assessment per patient, the results were virtually the same. This was also found in a previous inpatient study by De Vries Robbé and colleagues (2016) in The Netherlands. The focus group of clinicians commenting on their user-experiences with the SAPROF-EV pilot in clinical practice was rather small, which may have influenced the findings and limits generalizability. For future research examining user-feedback, it is recommended to involve a wider group of clinicians. The valuable insights gained from the focus group do highlight the importance of paying attention to implementation research when examining risk assessment tools in future studies (De Beuf et al., 2020; Vincent et al., 2012).

Third, the follow-up periods used in the current study were fairly short and only focused on the presence of aggressive incidents during hospitalization as outcome measure. Despite the short follow-up, overall base rates for the different types of investigated aggression were still sufficiently high, except for self-directed aggression. Therefore, longer follow-up periods as well as additional outcome variables, such as formal recidivism after discharge, treatment interruptions (crisis), other types of incidents (e.g., sexual aggression, general misconduct, substance abuse, absconding), as well as positive (community) outcomes (see Coupland & Olver, 2020), could be addressed in future research. Also, the severity of aggressive incidents was not considered in the present study and could be included in future studies.

Finally, the current study did not specifically differentiate between subgroups, such as patients with an intellectual disability or patients with a long treatment trajectory, although many of the new SAPROF-EV factors originated from research on these particular subgroups. However, when a comparison was made within the current sample for patients with and without intellectual disability, no notable differences were found (Jentsch, 2021). The user feedback also indicated that clinicians viewed the additional factors as equally valuable for patients with and without intellectual disability or long-term trajectory. Further research could focus more particularly on the value of different protective factors for different subgroups of individuals. In addition, in future studies, it could be useful to monitor which protective factors are most frequently highlighted as treatment goals for different subgroups of individuals in different settings.

Future Research and Clinical Implications

Based on the findings from this study, the SAPROF-EV pilot will be adjusted further into an improved and enhanced version of the SAPROF. In line with the results from this pilot study and the acknowledged limitations, future research should further investigate the (incremental) predictive validity of the new SAPROF-EV for institutional misconduct as well as negative (recidivism) and positive (successful functioning) community outcomes. This should be done for various inpatient and outpatient settings, for a wide variety of subgroups of individuals, and for different follow-up periods ranging in duration. Testing the empirical validity is after all important to assure that guidelines are evidence-based and credible (Doyle et al., 2021).

The SAPROF aims to offer a positive additional evidence-based decision support component of risk assessment, which ensures a balanced evaluation of both assets and risks. Future research should focus more on how these additional positive components of risk assessments are explicitly translated into intervention and risk management plans and how this formulation can help to prevent recidivism. Since the SAPROF and its successor the SAPROF-EV are specifically intended to provide positive guidance to treatment and intervention practice and offer new insights regarding strengths-based intervention goals, it is important that future studies also focus on the clinical utility and user-experiences of this new extended version of protective factors assessment in treatment practice.

Footnotes

AUTHORS’ NOTE:

We would like to thank the criminologists of Psychiatric Center Sint-Jan-Baptist for their efforts in scoring the pilot version of the SAPROF-EV and their participation in the focus group. Without their efforts, we would not have been able to complete this research.

ORCID iDs

Leen Cappon

Michiel De Vries Robbé

Leen Cappon, PhD, is a researcher in the psychiatric center Sint-Jan-Baptist in Belgium, co-researcher at the ScienForCare research group, and associated researcher at Department of Special Needs Education, Ghent University.

Armin Jentsch, PhD, is a researcher interested in quantitative psychology and educational measurement. He recently obtained a postdoctoral fellowship at the University of Oslo.

Saskia Roggeman, PhD, is a researcher in the psychiatric center Sint-Jan Baptist in Zelzate, Belgium, co-researcher at the ScienceForCare research group, and associated researcher at the Department of Head and Skin, Ghent Experimental Psychiatry (GHEP) Lab, Ghent University, Ghent, Belgium.

Michiel de Vries Robbé, PhD, is a senior researcher at the Department of Child and Adolescent Psychiatry & Psychosocial Care of Amsterdam UMC and a visiting professor at the Forensic Psychiatry Department of St. Joseph’s Hospital and McMaster University in Canada. He is a co-author of the SAPROF (adult, youth, and child versions) and the Risk Screener (adult and youth).

References

Abidin

Davoren

Naughton

Gibbons

Nulty

Kennedy

H. G.

(2013). Susceptibility (risk and protective) factors for in-patient violence and self-harm: Prospective study of structured professional judgement instruments START and SAPROF, DUNDRUM-3 and DUNDRUM-4 in forensic mental health services. BMC Psychiatry, 13, 1–18. https://doi.org/10.1186/1471-244X-13-197

Andreasson

Nijman

Krona

Meyer

Anckarsäter

Nilson

Hofvander

(2014). Predictors of length of stay in forensic psychiatry: The influence of perceived risk of violence. International Journal of Law and Psychiatry, 37, 635–642. https://doi.org/10.1016/j.ijlp.2014.02.038

Braun

Clarke

(2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3, 77–101. https://doi.org/10.1191/1478088706qp063oa

Brookstein

D. M.

Daffern

Oglff

J. R. P.

Campbell

R. E.

Chu

C. M.

(2020). Predictive validity of the HCR-20V3 in a sample of Australian forensic psychiatric patients. Psychiatry, Psychology and Law, 28, 325–342. https://doi.org/10.1080/13218719.2020.1775152

Burghart

De Ruiter

Hynes

S. E.

Krishnan

Levtova

Uyar

(2023). The Structured Assessment of Protective Factors for violence risk (SAPROF): A meta-analysis of its predictive and incremental validity. Psychological Assessment, 35, 56–67. https://doi.org/10.1037/pas0001184

Cappon

Heyndrickx

Rowaert

Grootaert

De Decker

Tremmery

Vandevelde

(2022). Systematic aggression registration in forensic psychiatric care: A qualitative study on preconditions for successful implementation. International Journal of Offender Therapy and Comparative Criminology. Advance online publication. https://doi.org/10.1177/0306624X221102850

Coupland

Olver

M. E.

(2020). Assessing protective factors in treated violent offenders: Associations with recidivism reduction and positive community outcomes. Psychological Assessment, 32, 493–508. https://doi.org/10.1037/pas0000807

De Beuf

T. L. F.

De Ruiter

De Vogel

(2020). Staff perceptions on the implementation of structured risk assessment with the START-AV: Identifying barriers and facilitators in a residential youth care setting. International Journal of Forensic Mental Health, 19, 297–314. https://doi.org/10.1080/14999013.2020.1756994

De Vogel

De Ruiter

Bouman

De Vries Robbé

(2009). SAPROF: Guidelines for the assessment of protective factors for violence risk. Forum Educatief.

10.

De Vogel

De Ruiter

Bouman

De Vries Robbé

(2012). SAPROF: Guidelines for the assessment of protective factors for violence risk (2nd ed.). Van der Hoeven Stichting.

11.

De Vries Robbé

De Vogel

Douglas

K. S.

(2013). Risk factors and protective factors: A two-sided dynamic approach to violence risk assessment. The Journal of Forensic Psychiatry & Psychology, 24, 440–457. https://doi.org/10.1080/14789949.2013.818162

12.

De Vries Robbé

De Vogel

Veldhuizen

(2020). The Structured Assessment of Protective Factors for violence risk (SAPROF). In Douglas

K. S.

Otto

R. K.

(Eds.), Handbook of violence risk assessment (2nd ed., pp. 410–437). Routledge.

13.

De Vries Robbé

De Vogel

Wever

E. C.

Douglas

K. S.

Nijman

H. L. I.

(2016). Risk and protective factors for inpatient aggression. Criminal Justice and Behavior, 43, 1364–1385. https://doi.org/10.1177/0093854816637889

14.

De Vries Robbé

Geers

M. C. K.

Stapel

Hilterman

E. L. B.

De Vogel

(2015). SAPROF-Youth Version: Structured assessment of protective factors for violence risk-youth version. Guidelines for the assessment of protective factors for violence risk in juveniles. Van der Hoeven Kliniek.

15.

De Vries Robbé

Hounsome

Carey

Jones

Brown

Whittington

Van den Nagel

Bohle

Veldhuizen

(2018). Structured assessment of protective factors for violence risk—Intellectual disabilities and long—term care addition (SAPROF—ID / LC)—Pilot manual. SAPROF International.

16.

De Vries Robbé

Mann

R. E.

Maruna

Thornton

(2015). An exploration of protective factors supporting desistance from sexual offending. Sexual Abuse: A Journal of Research and Treatment, 27, 16–33. https://doi.org/10.1177/1079063214547582

17.

De Vries Robbé

Willis

(2017). Assessment of protective factors in clinical practice. Aggression and Violent Behavior, 32, 55–63. https://doi.org/10.1016/j.avb.2016.12.006

18.

De Vries Robbé

Willis

Thornton

Kelley

(2019). Structured Assessment of Protective Factors for violence risk—Sexual offense addition, structured professional judgement version (SAPROF-SOSPJ pilot)—Pilot manual. SAPROF International.

19.

Douglas

K. S.

Hart

S. D.

Webster

C. D.

Belfrage

(2013). HCR-20V3: Assessing risk of violence—User guide. Mental Health, Law and Policy Institute, Simon Fraser University.

20.

Douglas

K. S.

Shaffer

C. S.

(2020). The science of and practice with the HCR-20 V3. In Douglas

K. S.

Otto

R. K.

(Eds.), Handbook of violence risk assessment (2nd ed., pp. 253–293). Routledge.

21.

Doyle

Coid

Ullrich

Shaw

(2021). Assessing protective factors to prevent post-discharge violent behavior: Towards an assets-based approach for clinical risk management. Journal of Psychiatric Research, 143, 416–421. https://doi.org/10.1016/j.jpsychires.2021.10.010

22.

Fazel

Burghart

Fanshawe

Gil

Monahan

(2022). The predictive performance of criminal risk assessment tools used at sentencing: Systematic review of validation studies. Journal of Criminal Justice, 81, 101902. https://doi.org/10.1016/j.jcrimjus.2022.101902

23.

Fortune

C. A.

Ward

Willis

(2011). The rehabilitation of offenders: Reducing risk and promoting better lives. Psychiatry, Psychology and Law, 19, 1–16. https://doi.org/10.1080/132187199.2011.615809

24.

Hogan

N. R.

Ennis

(2010). Assessing risk for forensic inpatient violence: A meta-analysis. Open Access Journal of Forensic Psychology, 2, 137–147.

25.

Hogan

N. R.

Olver

M. E.

(2019). Static and dynamic assessment of violence risk among discharged forensic patients. Criminal Justice and Behavior, 46, 923–938. https://doi.org/10.1177/0093854819846526

26.

Jankovic

Masthoff

Spreen

de Looff

Bogaerts

(2021). A latent class analysis of forensic psychiatric patients in relation to risk and protective factors. Frontiers in Psychology, 12, 695354. https://doi.org/10.3389/fpsyg.2021.695354

27.

Jentsch

(2021). Predictive validity of the SAPROF-Long care version in forensic psychiatry in Belgium [Unpublished master’s thesis, University of Hasselt]. University of Hasselt.

28.

Judges

R. C.

(2016). An exploration into the value of protective factors in violence risk assessment of psychiatric inpatients [Doctoral dissertation, University of Nottingham]. https://eprints.nottingham.ac.uk/32959/1/JUDGES%20R%20RESEARCH%20THESIS%20AMENDED.pdf

29.

Kashiwagi

Kikuchi

Koyama

Saito

Hirabayashi

(2018). Strength-based assessment for future violence risk: A retrospective validation study of the Structured Assessment of Protective Factors for violence risk (SAPROF) Japanese version in forensic psychiatric inpatients. Annals of General Psychiatry, 17, 51–58. https://doi.org/10.1186/s12991-018-0175-5

30.

Kay

S. R.

Wolkenfeld

Murrill

L. M.

(1988). Profiles of aggression among psychiatric patients: I. Nature and prevalence. The Journal of Nervous and Mental Disease, 176, 539–546. https://doi.org/10.1097/00005053-198809000-00007

31.

Maruna

LeBel

T. P.

(2010). The desistance paradigm in correctional practice: From programmes to lives. In McNeill

Raynor

Trotter

(Eds.), Offender supervision: New directions in theory, research and practice (pp. 65–89). Willan.

32.

Natoli

A. P.

Flake

(2022). Predictive validity of the Structured Assessment of Protective Factors for violence risk (SAPROF): A meta-analytic review. PsyArXiv Preprint. https://doi.org/10.31234/osf.io/d3nj4

33.

Nedopil

(2009). The role of forensic psychiatry in mental health systems in Europe. Criminal Behavior and Mental Health, 19, 224–234. https://doi.org/10.1002/cbm.719

34.

Neil

O’Rourke

Ferreira

Flynn

(2020). Protective factors in violence risk assessment: Predictive validity of the SAPROF and HCR-20V3. International Journal of Forensic Mental Health, 19, 84–102. https://doi.org/10.1080/14999013.2019.1643811

35.

Neves

A. C.

De Vries Robbé

Van den Nagel

Bohle

Veldhuizen

(2019). Enhancing protective factors in the management of violence risk in long-term psychiatric care. In Völlm

Braun

(Eds.), Long-term forensic psychiatric care (pp. 123–137). Springer.

36.

Nijman

Bjørkly

Palmstierna

Almvik

(2006). Assessing aggression of psychiatric patients: Methods of measurement and its prevalence. In Richter

Whittington

(Eds.), Violence in mental health settings (pp. 11–23). Springer.

37.

Nock

M. K.

(2009). Why do people hurt themselves? New insights into the nature and functions of self-injury. Current Directions in Psychological Science, 18, 78–83. https://doi.org/10.1111/j.1467-8721.2009.01613.x

38.

Oliver

P. C.

Crawford

M. J.

Rao

Reece

Tyrer

(2007). Modified Overt Aggression Scale (MOAS) for people with intellectual disability and aggressive challenging behavior: A reliability study. Journal of Applied Research in Intellectual Disabilities, 20, 368–372. https://doi.org/10.1111/j.1468-3148.2006.00346.x

39.

Olver

M. E.

Riemer

E. K.

(2021). High-psychopathy men with a history of sexual offending have protective factors too: But are these risk relevant and can they change in treatment? Journal of Consulting and Clinical Psychology, 89, 406–420. https://doi.org/10.1037/ccp0000638

40.

Oziel

Marshall

L. A.

Day

D. M.

(2020). Validating the SAPROF with forensic mental health patients. The Journal of Forensic Psychiatry and Psychology, 31, 667–698. https://doi.org/10.1080/14789949.2020.1785525

41.

Peters

D. H.

Adam

Alonge

Agyepong

I. A.

Tran

(2013). Implementation research: What it is and how to do it. BMJ, 347, 1–7. https://doi.org/10.1136/bmj.f6753

42.

Ramesh

Igoumenou

Vazquez Montes

Fazel

(2018). Use of risk assessment instruments to predict violence in forensic psychiatric hospitals: A systematic review and meta-analysis. European Psychiatry, 52, 47–53. https://doi.org/10.1016/j.eurpsy.2018.02.007

43.

Rice

M. E.

Harris

G. T.

(2005). Comparing effect sizes in follow-up studies: ROC Area, Cohen’s d, and r. Law and Human Behavior, 29, 615–620. https://doi.org/10.1007/s10979-005-6832-7

44.

Rogers

(2000). The uncritical acceptance of risk assessment in forensic practice. Law and Human Behavior, 24, 595–605. https://doi.org/10/b4q64t

45.

Smith

K. J.

O’Rourke

Mcpherson

(2020). The predictive validity of the HCR-20V3 within Scottish forensic inpatient facilities: A closer look at key dynamic variables. International Journal of Forensic Mental Health, 19, 1–17. https://doi.org/10.1080/14999013.2019.1618999

46.

Vandevelde

Vander Laenen

Van Damme

Vanderplasschen

Audenaert

Broekaert

Vander Beken

(2017). Dilemmas in applying strength-based approaches in working with offenders with mental illness: A critical multidisciplinary review. Aggression and Violent Behavior, 32, 71–79. https://doi.org/10.1016/j.avb.2016.11.008

47.

Verstegen

De Vogel

De Vries Robbé

Helmerhorst

(2017). Inpatient violence in a Dutch forensic psychiatric hospital. Journal of Forensic Practice, 19, 102–114. https://doi.org/10.1108/JFP-04-2016-0020

48.

Vincent

G. M.

Guy

L. S.

Grisso

(2012). Risk assessment in juvenile justice: A guidebook for implementation. Implementation Science and Practice Advances Research Center Publications. https://escholarship.umassmed.edu/psych_cmhsr/573

49.

Völlm

B. A.

Clarke

Herrando

V. T.

Seppänen

A. O.

Gosek

Heitzman

Bulten

(2018). European Psychiatric Association (EPA) guidance on forensic psychiatry: Evidence based assessment and treatment of mentally disordered offenders. European Psychiatry, 51, 58–73. https://doi.org/10.1016.j.eurpsy.2017.12.007

50.

Ward

(2002). The management of risk and the design of good lives. Australian Psychologist, 37, 172–179. https://doi.org/10.1080/00050060210001706846

Protective Factors in Forensic Practice: The Added Value of the SAPROF-Extended Version Pilot in Relation to Aggressive Incidents

Abstract

Keywords

Introduction

Method

Setting

Participants

Instruments

HCR-20V3

SAPROF

Modified Overt Aggression Scale

Procedure

Data Analysis

Results

Predictive Accuracy of the Different Measures

Item-Level Descriptives and Predictive Validity of the Modified and Additional Factors of the SAPROF-EV

Incremental Predictive Validity of the Modified Original SAPROF and the Additional Factors

Incremental Predictive Validity of the SAPROF-EV Over the HCR-20V3

User-Feedback Regarding the Usefulness of the Modified and Additional SAPROF-EV Factors in Clinical Practice

Discussion

Main Findings

Limitations

Future Research and Clinical Implications

Footnotes

AUTHORS’ NOTE:

ORCID iDs

References

HCR-20^V3

Incremental Predictive Validity of the SAPROF-EV Over the HCR-20^V3