Fear is defined as an emotional reaction to a specific, identifiable and immediate threat (McNeil and Vowles, 2004). When it is related to pain, it can evolve as a result of negative interpretations and catastrophic amplifications according to which pain is judged as equivalent to ongoing harm (Turk and Wilson, 2010). Fear falls under the primary control of the amygdala, described as the emotional center of the brain (Neugebauer, 2015). This brain area intervenes in the emotional physiological response, playing a primary role in emotional regulation following a painful stimulus, and on its related modulation (Öhman, 2005; Hulsman et al., 2021; Šimić et al., 2021). The model proposed by Vlaeyen and Linton (2000) posits that there may be two types of opposing behavioral responses: confrontation and avoidance of pain (Linton et al., 2018). They conducted an extensive analysis on the fear-avoidance model, highlighting possible precursors of pain-related fear, including the role of anxiety about pain as a limiting element to functional abilities (Lucchetti et al., 2012).

In the past it has been stated that confrontation leads to a reduction or elimination of fear, while avoidance maintains and increases fear to the point of becoming a phobia (Lethem et al., 1983; Philips and Jahanshahi, 1986). By analyzing these two responses, several factors can be associated that may go to influence this process such as negative evaluations regarding pain, avoidance and escape behavior, or hypervigilance (Leeuw et al., 2007). Some findings indicate that one of the reasons why individuals have a high fear of pain is because they are particularly susceptible to negative experiences of pain (Keogh et al., 2001).

According to current fear-avoidance models, when pain is perceived, judgments are made about the purpose of that pain (Crombez et al., 2012) and some attribute catastrophic meaning to pain, which, in turn, elicits fear of pain (Goubert et al., 2004). Since fear has a significant bearing on pain, the assessment of fear of pain turns out to be an important task for scholars to understand the mechanisms and individual differences in the very role of anxiety and fear (Ochsner et al., 2006). However, the relationship between avoidance behavior and the individual’s specific fears is very complex as it is difficult to establish the actual relationship. In this regard, McNeil and Rainwater (1998) developed the Fear of Pain Questionnaire (FPQ) with the aim of detecting by means of a questionnaire the excessive fear of pain both in individuals with chronic pain and in nonclinical individuals (Osman et al., 2002).

The most recent form of the questionnaire is the Fear of Pain Questionnaire-III (FPQ-III), a self-assessment instrument developed specifically to assess fear based on various pain stimuli. The FPQ-III includes 30 items from which three subscales (Severe Pain, Minor Pain, and Medical Pain) can be measured. Although many studies have evaluated its validity and reliability (Carleton and Asmundson, 2009; Yang et al., 2013), other studies have identified limitations in the instrument, confirmed by the same factor analysis, indicating that improvements could be applied to the instrument itself (Roelofs et al., 2005). Albaret et al. (2004) conducted a study to examine the factorial structure of the Fear of Pain questionnaire on three samples consisting of young, adult and elderly Europeans. The authors, in addition to identifying the French-language adaptation of the questionnaire, tested the inverse relationship between previous exposure to pain and fear of pain, noting significant differences in model fit based on the use of shorter versions with 15 items of the FPQ-III questionnaire. However, there are other instruments in the literature, which similar to the FPQ measure constructs associated with fear of pain. These include the Pain Anxiety Symptoms Scale (McCracken et al., 1992), the Fear Avoidance Beliefs Questionnaire (FABQ, Waddell et al., 1993; Pfingsten et al., 2000); the Dental Fear Assessment Scale (DFAS, Rowe, 1997), the Fear-Avoidance of Pain Scale (Crowley and Kendall, 1999), the Brief Behavioral Distress Scale (BBDS, Tucker et al., 2001), the Child Adult Medical Procedure Interaction Scale (CAMPISSF, Blount et al., 2001).

By the way, McNeil and Rainwater’s (1998) instrument remains the most studied and used internationally, even in its short versions such as the FPQ-Short Form consisting of 20 items (FPQ-SF) (Asmundson et al., 2008) and the FPQ-9 consisting of 9 items (McNeil et al., 2018).

Some research that has employed this questionnaire has identified gender differences on fear of pain, that confirms women’s tendency to react more fearfully to painful stimuli (Thibodeau et al., 2013; Vambheim and Øien, 2017). Other scholars have pointed out that cultural and linguistic factors may also play an important role in the assessment of fear of pain (Orhan et al., 2018).

In the international context, Spanish (Solé et al., 2019), French (Albaret et al., 2004), Dutch (Van Wijk and Hoogstraten, 2006), Portuguese (Cardoso et al., 2016), and Norwegian (Vambheim et al., 2017) validations are currently available. Di Tella et al. (2019) first proposed an Italian study of the FPQ-III, comparing the 30-item version with the short 20-item and 9-item versions. They employed a heterogeneous sample of Italian adults and showed the psychometric properties of the three models exhibited acceptable fit values overall, however better in the short versions of the instrument. They reported divergent validity with anxiety and depression measures, test–retest reliability, and measurement invariance across gender and different age groups.

The purpose of our work is to propose a new validation study oriented to confirm the good psychometric properties of a short version of the FPQ-III for the Italian context. Since among the limitations of the study by Di Tella et al. (2019) was mentioned the disproportion in the gender distribution among participants, in order to be able to exhibit robust results with respect to the measurement invariance, in our survey we recruited a large number of participants with substantially homogeneous distribution with respect to gender. To assess the convergent validity of the instrument, we combined the administration of the short version with three additional measures, different from the two used in the previous study by Di Tella et al. (2019) that were related to anxiety and depression: the Pain Catastrophing Scale (PCS, Sullivan et al., 1995; Italian validation Monticone et al., 2012), the Locus of Control of Behavior Scale (LCBS, Craig et al., 1984; Italian validation Farma and Cortivonis, 2000), and the Coping Orientations to Problems Experienced (COPE, Carver et al., 1989; Italian validation COPE-NVI, Sica et al., 2008). Considering the literature evidence, we therefore hypothesized to find significant associations with PCS, both total but particularly with the brooding subscale (Burri et al., 2018; Priore et al., 2019; Diotaiuti et al., 2021; Nunziato et al., 2021), with the external locus of control of the LBCS (Seville and Robinson, 2000; Torres et al., 2009; Sweeney et al., 2018), with the avoidance coping style (Benoit-Piau et al., 2018; Du et al., 2018; Ibrahim et al., 2020; Lentz et al., 2022) and transcendent reliance strategies of the COPE-NVI (Flanigan et al., 2019; Ferreira-Valente et al., 2020).

In order to lend support to the discriminant validity of the instrument, we added an additional laboratory study using three short 20-s videos as pain fear manipulation tools with the aim of eliciting by their presentation to participants an internal activation with respect to the sources of severe, medical, and minor pain. We hypothesized that with respect to a control group, the measurement using the FPQ-III short form would have revealed, in the comparison before and after the presentation of the videos, a significant increase in the scores related to the subscales of fear of severe, medical, or minor pain, confirming that the scale can adequately be used for both dispositional measurement of fear of pain but also for the assessment of a state condition.

The calculation of the Mardia Index (average of the squares of the Mahalanobis Distances) produced a coefficient (976.23) lower than the limit value (1.123), confirming the assumptions of multivariate normality. Low co-linearity was indicated by the low VIF values (Variance Inflation Factor) < 2 and high tolerance values > 0.60. For verification of the assumptions on the residuals, the average between the standardized and raw residuals was equal to 0; the Durbin–Watson test had a value of 1.756 and was therefore indicative of the absence of autocorrelation.

The evaluation of the metric properties of the scale was conducted through a confirming analysis (CFA) designed to test the goodness of the three-dimensional model adopted by McNeil and Rainwater (1998). The averages and standard deviations for the single items are reported in the following Table 1.

In order to examine the validity of a 30-item scale, a confirmatory factor analysis was performed. The results obtained by considering three factors and 30 items did not show a good fit to the data: CFI = 0.756; TLI = 0.736; NNFI = 0.736; NFI = 0.724; RMSEA = 0.093; RMSEA 90% 0.90–0.0597; value of p = 4.38e − 9. Therefore, the existence of a lower number of items or different factor structure was verified by performing an EFA with ML on half of the sample (selected by randomization) and CFA on the remaining portion. Considering Promax as rotation method and Cattell’s scree test indications (that three main factors lay above the debris), 26 items resulted in the factor loadings Structure Matrix. Factor loadings indicated the elimination of the items 9, 16, 27, 29 arriving at a final number of 26 items. Kaiser–Meyer–Olkin (KMO) index score was 0.928, Bartlett’s test p < 0.001; Chi-squared Test < 0.001; RMSEA = 0.055, RMSEA 90% confidence 0.04–na; TLI = 0.66. Figure 1 shows the scree plot while loadings pattern matrix was reported in Table 2.

Subsequently, the items 1, 5, 12, 14, 15, 18, 19, 20, 22, 23, 25, 26, and 30 were removed because they were found to damage the fit between the model and the covariance structure. Through the omission of these 13 items the following fit values were reached: Kaiser–Meyer–Olkin (KMO) index score was 0.858, Chi-squared Test < 0.001; RMSEA = 0.063; RMSEA 90% 0.063–0.084; TLI = 0.950.

Table 3 shows the final pattern matrix with saturations on the three identified factors, McDonald’s ω and Cronbach’s Alpha values, Guttman Split-Half Coefficients, Corrected item/total correlations, factor intercorrelations; while in Table 4 the factorial interrelationships are reported.

The confirmatory factor analysis performed with the split sample consisting of 354 participants confirmed that the model with three related factors and 13 items presented overall good indices of fit to data (see Figure 2): Chi-square = 148.092; χ²/df = 2.51; CFI = 0.971; TLI = 0.962; RMSEA = 0.046 and RMSEA 90% CI [0.037–0.056]. The first factor measures Severe Pain Jealousy (4 items); the second factor measures Medical Pain (4 items); the third factor measures Minor Pain (5 items).

The following Table 5 shows item statistics. All of the items showed some ceiling effects, and these ranged from 11.4% (item 10) to 58.3% (item 13). Item 13, which is the one that aroused the most fear of pain, had a mean score of 4.29, whereas item 24, which is the one that aroused the least fear of pain, had a mean score of 1.77. The overall mean score was 35.71 (13–65) with a SD of 8.46.

Furthermore, the measurement invariance of the factorial structure of the FPQ-III by gender was assessed. Four nested models with increasing degrees of restriction were tested. Table 6 shows the goodness-of-fit indices of the multidimensional model by gender and nested models of invariance in ascending order of restriction level. Results showed that the FPQ-III had strict invariance across gender and that the fit of the three-dimensional model for males and females was excellent.

These results indicate that the latent means can be compared by gender. The latent mean values were fixed to zero for females and, as can be seen in the following Table 7, males showed lower latent means of fear of pain in this study.

A further sample was used for convergent validity testing: correlations between the short form of FPQ-III, PCS, LBCS, and COPE-NVI were performed by administering a sample of 292 people, 113 males (38.7%) and 179 females (61.3%), M_age 24.46, and SD = 7.39. In relation to the hypotheses stated with regard to these associations, as shown in the following Table 8, the results have substantially confirmed the directions assumed. More specifically, the robust association with catastrophism (0.38^**) and brooding thinking (0.39**) was found. Correlations with the other scales were congruent with what was expected, but still negligible in terms of effect size (<0.20). The association with external locus of control was positive (0.22^**) while it was negative with respect to internally projected behavioral locus of control (−0.12^*). Avoidance coping and transcendent reliance coping showed positive (albeit slight) relationships with fear of pain, 0.19^** and 0.13^*, respectively.

In Table 9 the internal reliabilities of the two samples are shown comparatively with their confidence intervals. McDonald’s ω and Alpha coefficients for these convergent administrations ranged from 0.77 to 0.89 (Severe Pain), from 0.80 to 0.83 (Medical Pain), from 0.70 to 0.79 (Minor Pain), respectively.

The following Table 10 reports the English and Italian versions of the short form of FPQ-III, and the grouping of the items on respective factors.

As far as the scoring of the instrument is concerned, the 13 items in total are distributed over three factors. Every item has a scoring range from 1 to 5. The scoring calculation produces, through a summation of the scores of the component items, separate measurements for each factor, and a total score that ranges from 13 to 65.

The aim of this work was to present a validation study for an Italian short form of the Fear of Pain Questionnaire (FPQ-III). The analyses carried out led to the definition of a scale composed of a total of 13 items that converge separately on three factors (fear of severe pain, medical pain, and minor pain), showing an overall good fit of the model.

The first factor measures the person’s fear of receiving major damage on some part of the body as a result of events such as a disastrous fall downstairs, a broken limb, a head injury, or a broken neck. The convergent validity analysis indicated the significant positive associations with the sub-scales of catastrophism, namely brooding, magnification, helplessness, and with external causal attribution. The first relationship between brooding thought and fear of major painful harm is a confirmation of one of the main convergence hypotheses posed by this study. There is both a strand of studies in the literature that predictively links remorseful amplification to more intense pain perception (Burri et al., 2018; Priore et al., 2019; Diotaiuti et al., 2021; Häggman-Henrikson et al., 2021; Nunziato et al., 2021), and further empirical evidence reporting a direct effect of fear of pain on the level of actually perceived pain (Markfelder and Pauli, 2020; Luo et al., 2022). According to some scholars, fear of pain constitutes a stronger predictor of actually perceived pain than depressive catastrophic amplification (George et al., 2006; Hirsh et al., 2008). In each case, amplification orientation reported a significant association with the level of fear of pain in our study. An additional association was reported with the helplessness measure that is characterized by negative outcome expectancies and general, stable negative attributions ascribed to the actual pain condition. This is consistent with what has already been reported in the study by Samwel et al. (2006) where there was a robust correlation between helplessness and fear of pain. Feelings of helplessness may be an important consequence of the learning history of chronic pain patients, especially those with many years of pain. It refers to an attributional style, explaining negative events such as chronic pain and its consequences as uncontrollable, unpredictable, and unchangeable and generalizing these consequences to daily functioning (Samwel et al., 2007; Craner et al., 2016; Burland et al., 2019; Stensland, 2021). The association between fear of severe pain and external locus of control confirms findings in studies where lack of control over a totally external and unpredictable source of danger has been found to be associated with higher estimated potential harm (Heinze and Sleigh, 2003; Guszkowska, 2014; Musich et al., 2020).

The medical pain fear subscale reported significant associations not only with the catastrophizing measure but also with avoidance coping and external social support seeking. The first association is congruent with the literature where there is a recurrent association between fear of pain and avoidance coping (Fischerauer et al., 2018; Meulders, 2019; Buchmann et al., 2021). As suggested by van Vliet et al. (2018), the ability to avoid painful situations may paradoxically even increase the levels of fear elicited by such situations. The Fear Avoidance Model (FAM) currently represents a solid theoretical reference for understanding fear response behaviors generated by potentially painful situations (Crombez et al., 2012; Wideman et al., 2013). Maladaptive fear-avoidance behaviors include catastrophizing, anxiety/depression, and avoiding physical and social activities. These behaviors are linked to increased pain, disability, and sadness and are thought to be caused by a disorder that resembles kinesiophobia (Vlaeyen and Linton, 2000; Peeters et al., 2020). As a result, people perceive painful experiences as a warning of impending danger, injury, or a catastrophic medical ailment (catastrophizing), which causes them to dread pain and make actions associated to it. Increases in pain perception (hypervigilance) and the avoidance of pain-related movements are both consequences of elevated pain-related anxieties. Short-term reductions in pain and mental suffering may encourage avoidance behavior. Long-term avoidance and restriction of physical and social activity, on the other hand, may encourage physical deconditioning, accrue loss of social reinforcement, and eventually lead to a disuse syndrome (Nees and Becker, 2018). Contrarily, it is believed that no-fear confrontational pain reactions without catastrophizing and conditioned avoidance behavior help to prevent chronic pain and speed up healing, considering that physical exercise is considered an excellent tool for the treatment of chronic pain (Verbunt et al., 2003; Diotaiuti et al., 2022a). Coping oriented toward seeking external support also registered a positive association with fear of pain. The result finds an explanation in the disposition with low autonomy and limited action orientation of the person who preferentially seeks comfort and emotional support from others to cope with the strong fear of pain he or she is experiencing. The significant protective role of social support in mitigating the pressure of stress and perceived distress had already been emphasized in previous studies (Koivula et al., 2002; Zaza and Baine, 2002; Azimi et al., 2018). Complete and passive reliance on external support, however, can generate increased fear of pain when people who should provide emotional containment instead indirectly defer their anxieties and concerns about the unfolding of their relative or caretaker’s delicate situation.

In this regard, Neville et al. (2021) illustrated the situation of parents who had ended up increasing (by failing to contain their own anxieties and fears) in their children the fear of pain related to the medical care they were to receive. Beyond the non-functional response that support figures may have, evidence has emerged that illustrates a remarkable inhibitory effect of the presence of support figures on the formation of new fear associations for other cues, allowing the person to move with fewer learned fears, thus decreasing the activation of the threat response (Hornstein and Eisenberger, 2017). More recent models of pain coping emphasize the positive function of actively building individual resilience resources, where the focus of intervention should be on behavior change and sustaining, restoring or promoting the undertaking of personal goal-directed actions in the presence of pain (Goubert and Trompetter, 2017). Petter et al. (2014) demonstrated that changes in pain catastrophizing, which were indirectly connected to an induction of state mindfulness, were associated with lower levels of pain and improved pain tolerance. According to research, positive attitudes like optimism and acceptance coping may independently predict positive results from unfavorable personal traits and events (Smith and Zautra, 2008; Jackson et al., 2014; McCracken and Vowles, 2014).

The third measure of FPQ-III, related to fear of mild painful sources and conditions without major physical consequences, in addition to the correlations highlighted above in relation to medical pain, showed a significant inverse association with internal locus of control. In contrast to the fear of dangers and events relevant to physical safety or the outcome of medical manipulations or interventions, in the case of minor pain, the person perceives a wider sphere of control, which allows him or her a greater range of action or choice: to avoid certain situations, to better control processes, to take action on possible preventive measures. In any case, awareness of the limited duration of the condition predisposes the person to greater tolerance and forbearance, limiting the intensity of anticipatory fear associated with such conditions. However, in the literature, the role of internal locus of control in mitigating the experience of mild pain is reported in several studies (Bonafé et al., 2018; Löffler et al., 2018; Musich et al., 2020; González-Roldán et al., 2021).

Consistent with evidence reposted in the literature (Horn et al., 2014; Ramírez-Maestre and Esteve, 2014), analysis of invariance and related measures of latent averages, showed significantly higher levels of fear of pain in females, with particular note for the subscale of fear of pain attributable to medical treatment. However, this result differs from what was found in Vambheim and Øien (2017)’s previous study, in which the greatest gender difference was found in the severe pain factor, with the highest fear manifested by the female group. The authors attributed this difference to a higher level of fear of dying, fear of being disabled, overall anxiety, or existential anxiety, related to the consequences and implications that serious happenings might have effect on, for example, familial factors and caregiver responsibilities. Another explanation by the authors was related to gender role expectations of pain that influence males and females differently when confronted with reporting fear of pain. In this regard, the authors refer to the work of Wise et al. (2002) and Aslaksen et al. (2007) where it is shown that pain behavior is influenced by contextual factors, such as socialization, cultural expectations, and feminine and masculine gender role expectations. Therefore gender role stereotypes might have contributed to the findings of the higher fear of pain in females than in males. In relation to measures of fear of medical pain and minor pain in females, Fredrikson et al. (1996) had already reported that they have higher fear of injections, dentists, and injuries, compared to males; they worry, ruminate, and display more negative affectivity, and are more afraid of burning their fingers, cutting themselves while shaving, having a muscle cramp (minor pain), and having a wart removed by a doctor, compared to males. Courbalay et al. (2016) hypothesize a significant mediating role of neuroticism on the level of female fear of severe and medical pain. The increased sensitivity to medical pain in females is also reported by more recent studies such as Heft et al. (2007), Katanec et al. (2018), McLenon and Rogers (2019), and Labrenz et al. (2020). In addition, female seem to have a significantly higher perceived pain intensity than men, which could also explain these findings (Diotaiuti et al., 2022b). The test of measurement invariance with respect to gender in our study showed that the short version of the FPQ-III is an appropriate instrument for assessing fear of pain in males and females in the Italian context.

Discriminant assessment using the videos confirmed the adequacy of the scale in detecting changes in fear of pain in the groups obtained through the experimental manipulation. The results showed that watching the videos with severe pain and medical pain content induced greater fear in the respective groups compared to the baseline measurement and compared to a control group; on the other hand, watching a video with minor pain content induced (unexpectedly) a subsequent lowering of the level of fear in the group, both compared to the other groups and to the preliminary baseline assessment. This last result appears interesting and could be interpreted as an identificatory response with situations where the pain shown was not associated for the person with major or potentially irreversible consequences, such as in the case of serious accident or surgery. Observing others involved in painful but occasional, temporary and minor events may have elicited a feeling of potential self-efficacy and confidence in managing and overcoming the unpleasant situation, lowering the level of anticipatory fear associated with such situations.

In the present study, the Italian Short Version of the Fear of Pain Questionnaire (FPQ-III) showed good psychometric properties. The three-factor model to measure the fear of severe, medical and minor dimensions of pain was confirmed. In addition, the measurement invariance of the scale across gender was established with an extended sample of participants. The values of internal consistency and convergent validity were good as well. A further discriminant assessment using experimental manipulation through administration of fear eliciting videos demonstrated that the instrument is suitable in detecting changes in fear of pain induced by changes in situational conditions. The scale in this short version is particularly suitable to quickly and efficiently collect information on the perceived intensity of such anticipatory fears that could also dysfunctionally affect the person. High medical fear, for example, can activate avoidance responses with negative repercussions for diagnostic monitoring, delaying any necessary interventions and therapeutic treatments. Excessive fear of severe pain (coupled with continued fears of accidents, serious injuries, disabling injuries, etc.) can lead the person to progressive self-limitation in movement and normal interaction with his or her surroundings, lowering his or her overall quality of life. Finally, excessive fear of frequent minor painful events closely associated with the sphere of daily life (cutting oneself, tripping, hitting an edge, having dust or stinging liquid in one’s eyes, etc.) may portend states of anxiety and difficulty in focusing on coordinating actions in personal space. It therefore appears important to assess the intensity and frequency of such anticipatory fears in order to put in place restraining interventions and functional support for the person. The Italian validation of the short version of the FPQ-III is therefore a useful and original contribution to the conduct of such an assessment, possibly associated with other more in-depth psychometric instruments.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Institutional Review Board (IRB) of the University of Cassino and Southern Lazio. The participants provided their written informed consent to participate in this study.

PD, SC, and SM designed the study, analyzed the data, and discussed the results. PD, AC, and EC drafted the manuscript. TS and AA revised the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.