Introduction
Social anxiety is one of the most commonly reported experiences of anxiety among people with idiopathic Parkinson’s disease (PwIPD). Currently there is no specific guidance around the best clinical interventions to support PwIPD who experience social anxiety. This paper begins to explore the associated factors of social anxiety for PwIPD. It will contribute to the assessment of whether existing evidence-based cognitive behavioural approaches for social anxiety are applicable in this population. This paper will first introduce idiopathic Parkinson’s disease (IPD) and the prevalence of anxiety disorders. Existing cognitive behavioural models of social anxiety are outlined before the specific area of social anxiety in IPD is discussed.
IPD is a complex neurodegenerative condition that is not yet fully understood. Pathophysiological changes such as a loss of dopaminergic neurons in the basal ganglia, an area of the brain associated with voluntary motor movement, worsen over time (Stoker and Greenland, Reference Stoker and Greenland2018). To establish a clinical diagnosis of IPD there must be evidence of Parkinsonism. Parkinsonism is defined by the presence of bradykinesia in combination with either muscular rigidity and/or resting tremor. These symptoms are assessed against a set of exclusion criteria and supportive positive criteria before a diagnosis can be made (Postuma et al., Reference Postuma, Berg, Stern, Poewe, Olanow, Oertel, Obeso, Marek, Litvan, Lang, Halliday, Goetz, Gasser, Dubois, Chan, Bloem, Adler and Deuschl2015).
In addition to the physical presentation, there is increased recognition of non-motor symptoms (NMS). Although it is proposed that these commonly present before the onset of motor symptoms (Chaudhuri et al., Reference Chaudhuri, Healy and Schapira2006), they are often overlooked at this early stage and not recognised as part of a complex neurodegenerative condition (Klingelhoefer and Reichmann, Reference Klingelhoefer and Reichmann2017). NMS include autonomic, sensory and sleep changes (Ishihara and Brayne, Reference Ishihara and Brayne2006). Anxiety and depression are also often conceptualised as NMS of IPD. While there are undoubtedly neurobiological changes associated with anxiety and depression, conceptualising both this way risks viewing such experiences as inevitable pathological consequences of IPD. In doing so, potentially treatable components of these understandable mental health responses to a challenging physical condition could be overlooked (Stephens et al., Reference Stephens, Dysch and Gregory2018). Despite their substantial influence on disability and quality of life (Weintraub et al., Reference Weintraub, Moberg, Duda, Katz and Stern2004), NMS and mental health symptoms have received less attention than motor symptoms in IPD (Pachana et al., Reference Pachana, Egan, Laidlaw, Dissanayaka, Byrne, Brockman and Starkstein2013). There is increasing demand for a better understanding of these important aspects through research (Deane et al., Reference Deane, Flaherty, Daley, Pascoe, Penhale, Clarke, Sackley and Storey2014).
Prevalence of mental health difficulties in people with IPD
Existing research has highlighted that a range of mental health problems are frequently reported by PwIPD including generalised anxiety, low mood and social anxiety (Broen et al., Reference Broen, Narayen, Kuijf, Dissanayaka and Leentjens2016; Gultekin et al., Reference Gultekin, Ozdilek and Bestepe2014; Nègre-Pagès et al., Reference Nègre-Pagès, Grandjean, Lapeyre-Mestre, Montastruc, Fourrier and Lépine2010; Reijnders et al., Reference Reijnders, Ehrt, Weber, Aarsland and Leentjens2008). Social anxiety disorder (SAD) is among the most prevalent anxiety disorders in PwIPD (Broen et al., Reference Broen, Narayen, Kuijf, Dissanayaka and Leentjens2016). Social anxiety is highly prevalent in people with visible differences, for example people who have a scar, mark or condition that makes their face or body look different (Rumsey et al., Reference Rumsey, Clarke and White2003). IPD could arguably be percieved as a condition that causes visible differences for some people. For example, it can cause changes to posture, movement and facial expressions, all visible to others. Symptoms can also cause changes to competency with motor tasks such as walking, pouring a drink or speaking. Performance might be visibly different compared with other people and/or a person’s own previous ability in these areas. Although one could hypothesise that these visible objective differences might be related to the experience of social anxiety in some people with IPD (for example, their beliefs about their symptoms and how they are perceived by others are distorted and exaggerated), this has not been yet been explored in research.
It has been highlighted in previous research that some items on general anxiety self-report measures overlap with symptoms of IPD. For example, the Beck Anxiety Inventory (BAI) (Beck and Steer, Reference Beck and Steer1993) contains hands trembling (Salazar et al., Reference Salazar, Le, Neargarder and Cronin-Golomb2017). Such an item might reflect motor symptoms in IPD rather than, or perhaps in addition to, genuine symptoms of anxiety and therefore inflate scores. Therefore using these measures in the context of IPD might affect their validity, regardless if symptoms are objective or subjective in the context of self-report measures. This could also be true of social anxiety measures where some symptoms could be attributed to either social anxiety, IPD or some combination of the two. Clinical observations indicate that people with IPD are often aware that anxiety excaerbates their IPD symptoms but find it difficult to know how much of their experience is attributable to IPD and/or anxiety. Examples of items presenting ambiguity include references to sweating and trembling/shaking in the Social Phobia Inventory (Connor et al, Reference Connor, Davidson, Churchill, Sherwood, Foa and Weisler2000) and shaking/trembling and losing control in the Social Phobia Scale (Mattick and Clarke, Reference Mattick and Clarke1998). Performance-related items such as I would get tense if I had to carry a tray across a crowded cafeteria and I would find it difficult if I had to drink something in a group of people in the Social Phobia Scale (Mattick and Clarke, Reference Mattick and Clarke1998) could also arguably be contaminated by the physical symptoms of IPD. Some measures such as the Liebowitz Social Anxiety Scale (LSAS) (Liebowitz, Reference Liebowitz1987) do not include such items and are situational based as opposed to perceived symptom or performance based.
The numerous different anxiety measures used in the literature might contribute to the large variability in prevalence estimates across studies from 6 to 55% (Broen et al., Reference Broen, Narayen, Kuijf, Dissanayaka and Leentjens2016). In relation to social anxiety in particular, Gultekin et al. (Reference Gultekin, Ozdilek and Bestepe2014) found that 42.5% of PwIPD met diagnostic criteria for SAD. This is higher than the 12% prevalence rate estimated in the general population (National Institute for Health and Care Excellence [NICE], 2013) and closer to the 30.6% reported in multiple sclerosis (MS), another neurodegenerative condition (Poder et al., Reference Poder, Ghatavi, Fisk, Campbell, Kisely and Sarty2009). Even after accounting for possible score inflation due to symptom overlap, Salazar et al. (Reference Salazar, Le, Neargarder and Cronin-Golomb2017) demonstrated that anxiety scores remained high in PwIPD compared with the general population, indicating a need for closer clinical attention. Finding approaches to reduce stress and anxiety in IPD was identified as a top research priority by PwIPD (Deane et al., Reference Deane, Flaherty, Daley, Pascoe, Penhale, Clarke, Sackley and Storey2014). Most attention to date has been paid to depression in IPD. Although depression and anxiety are referred to in NICE clinical guidance for IPD, there is not yet acknowledgement that social anxiety is also a common problem in this population.
Social anxiety
SAD is characterised by a marked fear of one or more social situations that involves exposure to possible scrutiny from others (APA, 2013). SAD is associated with poor quality of life (QoL) (Barrera and Norton, Reference Barrera and Norton2009; Safren et al., Reference Safren, Heimberg, Brown and Holle1996). Cognitive behavioural therapy (CBT) is recommended as first-line treatment (NICE, 2013) and is based upon two empirically supported cognitive behavioural models (see Fig. 1).
Figure 1. Models of SAD. Top: a cognitive model of social phobia, Clark and Wells, from Heimberg et al. (Reference Clark, Wells, Heimberg, Liebowitz, Hope and Schneier1995); reprinted with permission of Guilford Press. Bottom: an updated cognitive behavioural model of social anxiety; Heimberg et al. (Reference Heimberg, Brozovich, Rapee, Hofmann and DiBartolo2010).
Common to both models in the maintenance of SAD are:
-
(1) Negative cognitions about oneself in social situations and the perception of the social situation as threatening.
-
(2) Safety-seeking behaviours (Salkovskis, Reference Salkovskis1991) such as total avoidance or more subtle avoidance to cope with the perceived threat. Safety-seeking behaviours are often employed to conceal symptoms that are perceived to be observable and negatively judged by others. If objective symptoms are present as well as perceived symptoms, the severity of these and the perception of negative judgement by others is often largely over-estimated.
-
(3) Attention to threat stimuli which feeds negative assumptions about how one appears to others. For example, noticing a racing heart and feeling highly anxious may lead to the conclusion that one must also appear anxious.
Although both models are conceptually similar, Clark and Wells (Reference Clark, Wells, Heimberg, Liebowitz, Hope and Schneier1995) place an emphasis on an internal, self-focus of attention that is interpreted as threatening whereas Heimberg et al. (Reference Heimberg, Brozovich, Rapee, Hofmann and DiBartolo2010) include attention to environmental cues that are interpreted as negative in addition. This study will focus on the Clark and Wells (Reference Clark, Wells, Heimberg, Liebowitz, Hope and Schneier1995) model primarily due to the operationalised measures available for the processes outlined above. Support for these models have been demonstrated in the general population (McManus et al., Reference McManus, Sacadura and Clark2008; Spurr and Stopa, Reference Spurr and Stopa2002; Woody et al., Reference Woody, Chambless and Glass1997); young people (Ranta et al., Reference Ranta, Tuomisto, Kaltiala-Heino, Rantanen and Marttunen2014; Schreiber et al., Reference Schreiber, Höfling, Stangier, Bohn and Steil2012) and support for CBT interventions based on these models have been demonstrated in the context of autism (Cardaciotto and Herbert, Reference Cardaciotto and Herbert2004); visible differences (Kleve et al., Reference Kleve, Rumsey, Wyn-Williams and White2002) and a single case study of social anxiety in IPD (Heinrichs et al., Reference Heinrichs, Hoffman and Hoffman2001). These findings indicate that cognitive behavioural approaches to SAD may be widely generalisable to populations including PwIPD.
Social anxiety in people with IPD
SAD in the context of a medical condition was excluded from the Diagnostical and Statistical Manual (DSM) until its most recent revision (DSM-IV; APA, 2013). Social anxiety in a medical context was regarded as having a different symptom profile, pathophysiology, demographic, course and treatment outcome compared with the general population (Heimberg et al., Reference Heimberg, Hofmann, Liebowitz, Schneier, Smits, Stein and Craske2014). Therefore, existing cognitive behavioural interventions for SAD were deemed potentially inappropriate, leading to a lack of understanding and support for social anxiety in these groups.
Some have suggested that an association between anxiety and IPD severity indicates that anxiety in PwIPD is fundamentally different from that in the general population (e.g. Sagna et al., Reference Sagna, Gallo and Pontone2014). Theories have explained the association as a result of pathological dopamine dysfunction (Moriyama et al., Reference Moriyama, Felicio, Chagas, Tardelli, Ferraz, Tumas and Bressan2011; Weintraub et al., Reference Weintraub, Newberg, Cary, Siderowf, Moberg, Kleiner-Fisman and Katz2005), i.e. that anxiety results from organic pathology. However, not all studies have replicated this finding and some have found anxiety in PwIPD to be independent of disease severity (Brown and Fernie, Reference Brown and Fernie2015). As seen in MS (Wood et al., Reference Wood, van der Mei, Ponsonby, Pittas, Quinn, Dwyer and Taylor2013), Brown and Fernie (Reference Brown and Fernie2015) attributed anxiety in PwIPD largely to psychological variables. Some qualitative research findings are consistent with psychological explanations of anxiety in PwIPD. One finding showed that some PwIPD described feeling trapped in an unpredictable, unreliable body (Goddard, Reference Goddard2014). Another study showed that some PwIPD reported anxiety around others and a sense of being visibly different (Chen and Marsh, Reference Chen and Marsh2014) while other PwIPD reported fear and embarrassment regarding visible symptoms. For example, some participants described worries that their tremor will cause them to spill food when eating (Goddard, Reference Goddard2014; Heinrichs et al., Reference Heinrichs, Hoffman and Hoffman2001). The way PwIPD thought about their symptoms (their metacognitions) was found to be a better predictor of anxiety than IPD symptom severity (Brown and Fernie, Reference Brown and Fernie2015; Fernie et al., Reference Fernie, Spada, Ray Chaudhuri, Klingelhoefer and Brown2015). This finding indicates that a purely biomedical view of anxiety in PwIPD is inadequate. It suggests that psychosocial factors, particularly how symptoms are interpreted by an individual, are influential in an individual’s level of distress. A recent generic cognitive behavioural model has attempted to describe how disease-related and psychosocial-related vulnerabilities interact in the development and maintenance of psychological distress including social anxiety in PwIPD (Stephens et al., Reference Stephens, Dysch and Gregory2018). It posits that the way PwIPD think, interpret and react to IPD influences anxiety. Although conceptually appealing, this model remains to be empirically tested, possibly due to a lack of available operationalised constructs.
The current study aims to test the applicability of a specific cognitive behavioural conceptualisation of social anxiety in PwIPD. Negative social cognitions, safety-seeking behaviours and focus of attention were chosen as key variables because they are common to both NICE-recommended cognitive behavioural models and they have been operationalised. In line with the Clark and Wells (Reference Clark, Wells, Heimberg, Liebowitz, Hope and Schneier1995) model, the presence of internally focused attention in PwIPD during social interactions was included. Based on current literature, it was hypothesized that:
-
(1a) Negative social cognitions, safety-seeking behaviours and internally focused attention (cognitive behavioural variables) would all be positively associated with social anxiety in PwIPD.
-
(1b) These cognitive behavioural variables will account for a unique portion of variance in social anxiety in PwIPD even after accounting for demographic and IPD related variables.
-
(2c) Social anxiety will be negatively associated with QoL in PwIPD.
-
(2d) Social anxiety will account for a unique portion of variance in QoL in PwIPD.
Method
Participants and procedure
Recruitment was supported by an NHS Movement Disorders Clinic in the South West and the charity Parkinson’s Disease UK (PDUK). Both organisations shared study advertisements with potential participants. Interested participants contacted the researcher who conducted a telephone screen to establish eligibility. Table 1 outlines exclusion criteria. Eligible participants completed informed consent and received their preference of either a paper questionnaire pack or link to an electronic version. If they were recruited via the clinic, they also had the option of face-to-face support. Participants were given researcher contact details for any clarification or technical support during participation. Some participants received practical support from a relative or carer to complete the study; however, they were instructed to provide their own answers to the questions. Participants were given the option to have £2 donated to a relevant charity for their participation. The triple data collection strategy (online, post or in-person) maximised recruitment in a potentially hard-to-reach population. The opportunity to take part remotely facilitated accessibility for participants across the UK who might otherwise be unable to attend face-to-face (e.g. due to high levels of social anxiety, physical impairment or resource limitations).
Table 1. Exclusion criteria
Design
This study used a cross-sectional design. Participants were asked to complete a battery of self-report measures at one time point.
Sample size
G*Power statistical software (version 3.41) was used to estimate sample size. The following assumptions were made: linear multiple regression; fixed model R 2; deviation from 0 alpha of 0.05; power of 0.80; nine predictors and a medium effect size (0.15) (based on Hodson et al., Reference Hodson, McManus, Clark and Doll2008). The estimated required sample was 114 participants for the primary analysis.
Demographics
One hundred and twenty-four PwIPD took part; 52% were male. Participants were predominantly White British (95.9%). Most participants were aged between 51 and 79 years (90.3%); 44.3% had been living with IPD for <5 years, 40.3% for 5–10 years, 12% for 11–15 years and 0.8% >15 years. Over half (52.4%) reported mild to moderate IPD, 37.9% reported moderate IPD and 9.7% moderate to severe IPD; 54.8% reported fluctuating IPD symptoms. Sixty per cent of the sample scored above the recommended clinical cut-off for social anxiety.
Primary outcome variable
The Liebowitz Social Anxiety Scale (LSAS) (Liebowitz, Reference Liebowitz1987) is a NICE-recommended screening tool for social anxiety. Cronbach’s alpha is reported as 0.95, indicating good internal consistency with test–re-test reliability 0.83 at 12 weeks (National Institute for Health and Care Excellence [NICE], 2020). The self-report version (used in this study) correlates well with the original clinician reported version (Fresco et al., Reference Fresco, Coles, Heimberg, Liebowitz, Hami, Stein and Goetz2001). A cut-off of 41/42 is reportedly the best balance between sensitivity (77.8%) and specificity (64.3%) for PwIPD (Kummer et al., Reference Kummer, Cardoso and Teixeira2008) and was used to determine the prevalence of clinically relevant social anxiety in this sample.
Secondary outcome variable
The Quality of Life Index generic version (QLI) (Ferrans and Powers, Reference Ferrans and Powers1985) measures quality of life and is reported to have good psychometric properties with Cronbach’s α ranging from 0.89 to 0.96 (Ferrans and Powers, Reference Ferrans and Powers1992).
Key predictor variables
All key predictor variables have been used to test the applicability of a cognitive behavioural understanding of social anxiety in other populations (Hodson et al., Reference Hodson, McMannus, Clark and Doll2008; Schreiber et al., Reference Schreiber, Höfling, Stangier, Bohn and Steil2012).
The Social Cognitions Questionnaire (SCQ) (Wells et al., Reference Wells, Stopa and Clark1993) measures the frequency and belief in negative social cognitions (only frequency was used in this study). This decision was based on limiting the number of variables in this study to essential core variables to reduce loss of power. If frequency of SCQ is found to be a relevant variable, then further studies may wish to include percentage belief ratings. This would advance the research question from are negative social cognitions present?, to what influences the impact of negative social cognitions? Stopa is reported to have factor analysed data from 335 participants and established good internal consistency of this measure across three factors: negative self-beliefs α = .72, fear of performance failure α = .84 and fear of negative evaluation α = .81 (Stopa, 1995; unpublished data referred to in (Tanner et al., Reference Tanner, Stopa and De Houwer2006); (Calvete et al., Reference Calvete, Orue and Hankin2013). For the purposes of this study, mean scores were used as recommended (Clark, Reference Clark2005).
The Social Behaviour Questionnaire (SBQ) (Clark et al., Reference Clark, Wells, Salkovskis and Hackmann1995) measures safety-seeking behaviours in social situations. It has good internal consistency (α = 0.8) and good discrimination between anxiety disorders (Clark, Reference Clark2005). Mean scores were used as recommended (Clark, Reference Clark2005).
The internal subscale of the Focus of Attention Questionnaire (FAQ) (Woody, Reference Woody1996) measures internally focused attention. It has acceptable internal consistency (Cronbach’s α = .76; Woody et al., Reference Woody, Chambless and Glass1997). The FAQ was modified by asking participants where they ‘usually’ focused their attention during a social interaction to fit the study design.
Control variables
The Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) was used as a measure of IPD severity and is considered to have good reliability and validity (Goetz et al., Reference Goetz, Tilley, Shaftman, Stebbins, Fahn and Martinez-Martin2008). It has a reliable factor structure supporting the use of each part separately with the lowest Cronbach’s α reported to be .79 (Goetz et al., Reference Goetz, Tilley, Shaftman, Stebbins, Fahn and Martinez-Martin2008). Only self-report items in parts I and II were used in this study to fit the study design. Self-report items of NMS severity (part I) and motor symptom severity (part II) significantly model clinician rated reports, suggesting that patient perceptions of disability represent valid estimates of symptom severity (Goetz et al., Reference Goetz, Luo and Stebbins2019). Recommended motor symptom severity cut-off scores are: mild/moderate, 12/13 and moderate/severe, 28/29 (Martínez-Martín et al., Reference Martínez-Martín, Rodríguez-Blázquez, Mario, Arakaki, Arillo, Chaná and Merello2015). These cut-offs were used to establish IPD severity in the sample. Part II has appropriate performance in assessing IPD disability compared with other measures (Rodríguez-Blázquez et al., Reference Rodríguez-Blázquez, Alvarez, Arakaki, Campos Arillo, Chaná, Fernández and Martínez-Martín2017).
The Hospital Anxiety and Depression Scale (HADS) (Zigmond and Snaith, Reference Zigmond and Snaith1983) was used to assess anxiety and depression. It is shown to have satisfactory psychometric properties in PwIPD (Anxiety subscale Cronbach’s α = 0.86; depression subscale α = 0.78) (Marinus et al., Reference Marinus, Leentjens, Visser, Stiggelbout and van Hilten2002).
Statistical analysis
IBM SPSS statistics version 25 was the statistical software package used to perform a 3-step hierarchical linear regression with social anxiety as the dependent variable (Table 3). In step 1 the effects of age, sex, depression and anxiety were controlled for due to associations with anxiety in previous literature (Bolluk et al., Reference Bolluk, Ozel-Kizil, Akbostanci and Atbasoglu2010). Age was not entered into the model after observing insufficient evidence for a linear relationship with social anxiety (see Table 2). In step 2, NMS and motor symptom severity were controlled for due to their possible associations with anxiety (Sagna et al., Reference Sagna, Gallo and Pontone2014). In step 3 cognitive behavioural variables were entered. A second hierarchical regression was conducted with QoL as the dependent variable: anxiety and depression were entered in step 1, IPD variables were entered in step 2 and social anxiety was entered in the final step. The final model is reported (Table 4).
Table 2. Spearman’s Rho Correlation Matrix
*Correlation is significant at the 0.05 level (2-tailed).
**Correlation is significant at the 0.01 level (2-tailed).
Missing data
Guidance on handling missing items was followed for each measure where available. SCQ, SBQ and QLI mean scores were calculated, which automatically deals with missing items. As recommended, no more than one item from part I or two items from part II of the MDS-UPDRS was pro-rated (Goetz et al., Reference Goetz, Luo, Wang, Tilley, LaPelle and Stebbins2015). Mean imputation was applied when no more than 20% of items were missing (Enders, Reference Enders2003). Where more than 20% of items were missing, data were removed. See Table 5 for a breakdown of missing data across each measure. To maximise use of data, cases were excluded pairwise during statistical analysis (Pallant, Reference Pallant2010).
Results
Parametric assumptions
Spearman’s rho correlations (Table 2) were used to assess all bivariate relationships. Tolerance (0.409–0.936) and variance inflation factor (VIF) values (1.069–2.445) for all predictors fell within acceptable thresholds, suggesting no evidence of multicollinearity (Pallant, Reference Pallant2010). Visual inspection of a histogram, normal P-P plot and scatter plot of standardised residuals indicated no evidence that parametric assumptions were violated (Field, Reference Field2013).
Hypothesis 1a
Spearman’s rho correlations indicated moderate to strong positive linear relationships between each cognitive behavioural variable and social anxiety. Results indicated support for hypothesis 1a. The association between social anxiety and negative social cognitions was strong (R s = .772, n = 122, p<0.001), moderate with safety-seeking behaviours (R s = .694, n = 123, p<0.001) and moderate with an internal focus of attention (R s = .511, n = 119, p<0.001).
Hypothesis 1b
The hypothesis that each cognitive behavioural factor would explain a unique portion of variance in social anxiety was partly supported. Hierarchical linear regression (Table 3) indicated that cognitive behavioural factors accounted for 43.5% of variance in social anxiety in PwIPD. The final model accounted for 71.6% of variance overall and was statistically significant (F = 32.10, d.f. = 8, 102, p<0.001). Negative social cognitions (t 102 = 7.30, p<.001) and safety-seeking behaviours (t 102 = 4.19, p<.001) were statistically significant predictors; however, an internal focus of attention (t 102 = –.70, p = .677) was not.
Table 3. Hierarchical regression model with social anxiety (LSAS) as the outcome variable*
* p < 0.05, ** p < 0.001
IPD NMS severity (t 102 = –2.47, p = .015) and depression (t 102 = 2.25, p = .026) were significant predictors of social anxiety; however, motor symptoms (t 102 = 1.97, p = .051), sex (t 102==1.26, p = .21) and anxiety (t 102 = –.53, p = .60) were not.
Hypothesis 2a
Spearman’s rho correlations indicated a weak negative linear association between social anxiety and QoL providing some support for this hypothesis (R s = –.39, n = 116, p<.01).
Hypothesis 2b
The hypothesis that social anxiety would explain a unique portion of variance in QoL was not supported. Hierarchical regression indicated that social anxiety was not predictive of QoL in PwIPD (t 106 = –1.65, p = .10). The final model (Table 4), which accounted for 38.8% of the variance in QoL, indicated that only depression and NMS made significant contributions to the model. Anxiety, motor symptoms and social anxiety did not. This model was statistically significant (F = 13.45, d.f. = 5, 106, p<.01).
Table 4. Final model with QoL as the outcome
** p < 0.1
Table 5. Missing data
Discussion
This is the first study to demonstrate that negative social cognitions and safety-seeking behaviours are strongly associated with social anxiety in PwIPD. Social anxiety was more strongly correlated with negative social cognitions and safety-seeking behaviours than symptoms of IPD severity. When the influence of other relevant factors were statistically controlled for, both negative social cognitions and safety-seeking behaviours predicted unique variance in social anxiety in PwIPD. Internally focused attention did not. Social anxiety was not predicted by motor symptoms and NMS were a very weak predictor.
These findings are consistent with the theoretical position that psychological factors are influential to the experience of social anxiety in PwIPD. Results indicate that a cognitive behavioural understanding of social anxiety may be a useful conceptual framework in PwIPD. This contrasts with a purely biomedical disease model and is consistent with existing evidence that psychological factors play a role in influencing distress in PwIPD (Brown and Fernie, Reference Brown and Fernie2015; Fernie et al., Reference Fernie, Spada, Ray Chaudhuri, Klingelhoefer and Brown2015).
Negative social cognitions were the strongest predictor of social anxiety, while motor symptom severity did not predict any variance in the final model. This study did not specifically look at individual interpretations of IPD symptoms. However, given the initial finding that negative social cognitions are associated with social anxiety in PwIPD, it might be a reasonable next step to tentatively hypothesise that the way someone interprets their IPD symptoms has an impact on their behaviour and experience of social anxiety. It would be useful to gain a fuller understanding on what the content of their negative social cognitions are and what specific safety-seeking behaviours PwIPD are using to cope. For future research to test this hypothesis adequately, the subjective interpretations of IPD symptoms from self-report measures would have to be compared against an objective measure of IPD symptom severity. Completing the full MDS-UPDRS alongside a gold standard measure of social anxiety based on the DSM-IV or ICD criteria, as well as a semi-structured interview schedule to capture a detailed account of how each participant interprets their symptoms of IPD, would be one way to achieve this. It would enable better understanding of common cognitive biases that may be present in this population that would be useful for clinicians working with PwIPD to be mindful of.
This is the first study to demonstrate that the degree to which a PwIPD engages in safety-seeking behaviours (e.g. putting one’s hand in a pocket to hide a tremor) is moderately predictive of social anxiety. This is consistent with a cognitive behavioural understanding of SAD, where thoughts about oneself can drive behavioural responses intended to reduce social anxiety but actually exacerbate it. Future research would also benefit from measuring a more detailed account of the safety-seeking behaviours PwIPD use and whether these are carried out in response to particular IPD symptoms. This would allow a richer understanding of how PwIPD use safety-seeking behaviours and help inform clinicians working with PwIPD of what to be mindful of in this context.
The finding that depression and NMS were predictive of social anxiety in IPD was consistent with previous literature (Bolluk et al., Reference Bolluk, Ozel-Kizil, Akbostanci and Atbasoglu2010). The finding that motor symptom severity was not predictive of social anxiety was also consistent with previous literature in PwIPD (Brown and Fernie, Reference Brown and Fernie2015).
The association between NMS and social anxiety in this study was not predicted. One possible explanation is symptom overlap. For example, question 1.9 in the MDS-UPDRS part I asks about pain and other sensations such as uncomfortable feelings in the body like tingling. Another explanation is that NMS such as urinary problems or fainting are interpreted as more socially threatening than motor symptoms. Future qualitative studies could explore PwIPD interpretations of NMS and motor symptoms to gain insight into this finding and develop a more IPD-specific understanding of social anxiety.
The finding that social anxiety shared a negative linear relationship with QoL was consistent with previous literature (Barrera and Norton, Reference Barrera and Norton2009; Safren et al., Reference Safren, Heimberg, Brown and Holle1996). Although this was predicted, the association was weak and social anxiety did not predict QoL. Findings therefore suggest that depression and NMS might be more important factors when considering QoL in PwIPD. It is also possible that variance shared by depression and social anxiety might have masked a more complex relationship reflecting a limitation of the sensitivity and specificity of the measures used in this study.
As expected, the prevalence of social anxiety in this sample was high, with 60% scoring above the clinical cut-off. This is higher than the prevalence in the general population, other samples of PwIPD and people with other movement disorders such as MS (Gultekin et al., Reference Gultekin, Ozdilek and Bestepe2014; Poder et al., Reference Poder, Ghatavi, Fisk, Campbell, Kisely and Sarty2009). Sampling bias to a study designed to explore SAD in PwIPD and a screening tool to assess social anxiety could have contributed to this high prevalence. Even so, such high prevalence in this sample supports the argument that SAD is a clinically relevant problem for PwIPD warranting further attention.
Contrary to Schreiber et al. (Reference Schreiber, Höfling, Stangier, Bohn and Steil2012) and Hodson et al. (Reference Hodson, McMannus, Clark and Doll2008), an internal focus of attention did not predict social anxiety in this study. While this might represent a genuine finding it might also be due to methodology. This study adapted the FAQ from situation-specific to generic questions to fit the non-experimental design. This may have impacted recall accuracy. The importance of an internal focus of attention to social anxiety in PwIPD could be clarified by using the original FAQs and directing participants to focus on a specific recent social interaction via semi-structured interview in future research. Another possible explanation could be the regression analysis which involved entering all cognitive behavioural variables during the same step. This could have led to an overlap of variance accounted for between the three predictors masking the unique contributions of self-focused attention. One way to test this in future studies would be to enter cognitive behavioural variables in a hypothesis led in a step-wise way in line with cognitive models.
Limitations
The cross-sectional design means that causality cannot be inferred, limiting the conclusions that can be drawn about associations in this study. All data in this study were self-reported, a necessary strategy to maximise recruitment in a hard-to-reach population. Truly objective estimates of IPD symptom severity were thus not achieved, limiting conclusions about the relationship between social anxiety, NMS and motor symptoms of IPD. Self-report alone prevents elimination of the possible contamination of cognitive biases that could have otherwise been ruled out in contrast with clinician ratings. Although Goetz et al. (Reference Goetz, Luo and Stebbins2019) indicate that self-report items on the MDS-UPDRS correlate well with clinician-reported items, future research would benefit from controlling more robustly for the potential cognitive bias of symptoms of IPD by using the full MDS-UPDRS.
This sample consisted largely of PwIPD who had an interest in research. Evidence suggests that volunteer bias can influence outcomes that are not representative more generally (Jordan et al., Reference Jordan, Watkins, Storey, Allen, Brooks, Garaiova and Morgan2013). Therefore caution should be used about generalising the findings of this study. While the sample was geographically diverse, non-White British and severe levels of IPD were under-represented. This could be related to the online recruitment strategy and inclusion criteria that attracted younger participants, who were therefore on average in earlier stages of disease. PwIPD in earlier stages are more likely to be adjusting to their diagnosis and the challenges IPD brings to their identity in different social circumstances. For example, they may still be working and have an active role in larger social networks that might challenge their functional abilities. Arguably they might therefore experience a heightened level of self-consciousness compared with people in latter stages, who may be more adjusted, making social anxiety less relevant. Assessment of social anxiety experiences among people with severe IPD would deepen our understanding of social anxiety across the course of IPD.
Clinical implications
These findings are consistent with the revision in the DSM-5 which now accepts that people within a medical context may also meet criteria for SAD (Heimberg et al., Reference Heimberg, Hofmann, Liebowitz, Schneier, Smits, Stein and Craske2014). As we know from other medical conditions, how an individual interprets their disability is a critical part of their distress and experience, regardless of how their disability may ‘objectively’ appear when measured clinically or using standardised tools.
This study thus partially supports the applicability of a cognitive behavioural understanding of social anxiety in PwIPD. In practice, assessment of a PwIPD’s interpretation of their IPD should be regarded as equally important to the assessment of disability or disease ‘severity’. Anxiety cannot be assumed to be linearly associated with disease progression or ‘organic pathology’. Clinicians should pay attention to any such incongruence and consider referral for psychological intervention when social anxiety levels are high.
Such interventions may include cognitive behavioural therapy (CBT) particularly if there are clear cognitive biases and safety-seeking behaviours likely to maintain anxiety. However, further research must clarify which processes from the available models would be most applicable for PwIPD. For example, combining the key processes from disorder-specific models such as Clark and Wells (Reference Clark, Wells, Heimberg, Liebowitz, Hope and Schneier1995) and Rapee and Heimberg (Reference Rapee and Heimberg1997) with a more generic context that acknowledges the vulnerabilities arising due to IPD such as those proposed by Stephens et al. (Reference Stephens, Dysch and Gregory2018) might offer more tailored interventions for PwIPD and social anxiety. Such an approach offers great hope as such psychological interventions have been shown effective in other populations.
Conclusion
This novel study provides evidence that negative social cognitions and safety-seeking behaviours are cognitive behavioural variables that are associated with and predictive of social anxiety in PwIPD. Results are similar to what we might expect to see in the general population. Interpretations and reactions to IPD symptoms and the relationship of these factors with social anxiety are important areas for further research.
Acknowledgements
The authors would like to thank all participants with Parkinson’s disease who showed interest in this study and those who took the time to talk with the researcher and complete the measures. We extend our thanks to PDUK for supporting the researchers with all aspects of this study including enabling contact with people with experience to pilot measures and circulating the study details to many people affected by Parkinson’s disease. Without the help of PDUK this study might not have met its recruitment target. Thanks to all those the staff at the Clara Cross Centre, St Martins Hospital in Bath for their consultation and support regarding recruitment.
Financial support
This research was completed at the University of Bath Department of Psychology and did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Conflicts of interest
The authors have no conflicts of interest to report.
Ethics statement
Authors have abided by the Ethical Principles of Psychologists and Code of Conduct as set out by the BABCP and BPS. Ethical approval was granted by the East of Scotland Research Ethics Service (ESRES), Health and Care Research Wales (HCRW) committee and Health Research Authority (HRA), Ref: 18/ES/0140, the University of Bath Psychology Department ethics committee, PREC 18/19 030 and Research Governance at VirginCare, BRD Ref 2019_006.
Comments
No Comments have been published for this article.