Because of motor symptoms, Parkinson’s disease (PD) patients need a caregiver to assist them with their daily activities. Although research on PD caregivers is limited, data show that taking care of someone with a chronic condition is considered a chronic life stressor. ^{Reference Bevan and Sternberg1,Reference Kudlicka, Clare and Hindle2} As for mental health, depression and anxiety are approximately two and three times more frequent in PD caregivers, respectively, than in general population. ^{Reference Bevan and Sternberg1,Reference Kudlicka, Clare and Hindle2,Reference Grün, Pieri, Vaillant and Diederich3}

A study showed that, together with disease stage, behavioural problems associated with cognitive dysfunction in PD patients explained 66% of caregiver burden variance. ^{Reference Grün, Pieri, Vaillant and Diederich3} Cognitive symptoms alone are also associated with caregiver burden. ^{Reference Bevan and Sternberg1,Reference Kudlicka, Clare and Hindle2}

Among cognitive functions, theory of mind (ToM) can be defined as ‘the ability to attribute mental states (beliefs, intentions) to others as well as to understand and predict others’ behaviour based on mental states’. ^{Reference Bora, Walterfang and Velakoulis4} ToM is part of social cognition. ^{Reference Bora, Walterfang and Velakoulis4,Reference Dvash and Shamay-Tsoory5} Affective ToM, referring to empathy (i.e., attribution of feelings different from oneself), and cognitive ToM, referring to attribution of beliefs and intentions, are two components of ToM.5 In PD, both affective and cognitive ToM deteriorate overtime. ^{Reference Bora, Walterfang and Velakoulis4}

Impaired ToM has an impact on the relationship PD patients have with their social environment, especially their caregivers. ^{Reference Dvash and Shamay-Tsoory5,Reference Pomponi, Ricciardi and La Torre6} To our knowledge, few studies have investigated the association between social cognition, such as impairments in ToM, and caregiver well-being in PD. So far only a study ^{Reference Pomponi, Ricciardi and La Torre6} reported some link between patient empathy (affective ToM) and caregiver burden. However, this study has neither investigated cognitive ToM nor its association with caregiver burden or distress. Thus, the association between impairment of both components of patient’s ToM and caregiver burden and distress needs further investigation.

In a previous study (submitted for publication) conducted by our team, ^{Reference Giguère-Rancourt, Plourde and Couture7} additional data were gathered about caregiver burden and distress, as well as data of patients’ cognitive and affective ToM. The present pilot study thus performed secondary analyses on these data to verify if caregiver burden and distress are associated with patient’s cognitive and affective ToM. It was hypothesised that caregiver burden and distress would be associated with both components of patients’ ToM, that is, affective and cognitive ToM.

Participants and methods were described in another paper from our team. ^{Reference Giguère-Rancourt, Plourde and Couture7} Table 1 presents demographics and clinical characteristics of PD patients with mild cognitive impairment (PD-MCI) and their caregivers.

Table 1: Descriptive statistics at screening of participants with PD-MCI total sample (n = 12)

LED = levodopa equivalent dose; ZBI-12 = Zarit Burden Interview, 12 items; AES = apathy evaluation scale; MoCA = Montreal cognitive assessment; PDQ-39 = Parkinson disease questionnaire, 39 items.

As for this study and the present analyses, all participants and caregivers signed an informed consent before entering in the study. All nominative data were kept confidential by coding of all documents. The present analyses were approved by the Ethics and Research Committee of CHU de Québec – Université Laval.

Faux Pas recognition test (social mistaking) is a classical test of ToM and was administered to patients. In this test, participants are asked to read a story that may or may not contain a social Faux pas. After each story, participants were asked the following questions:

• ‘Did anyone say something s/he should not have said or something awkward?’ (Faux Pas detection, referred as ‘Hits’ in Table 1). If participants answered ‘yes’, they were asked the following questions:

• ‘Who said something s/he should not have said or something awkward?’ (Faux Pas comprehension).

• ‘Why shouldn’t s/he have said it or why was it awkward?’ (Faux Pas comprehension).

• ‘Why do you think s/he said it?’ (Faux Pas comprehension).

• ‘Did X know that Y (…)?’ (test for realising that the Faux pas was unintentional, referred as ‘Intentions’ in Table 1).

• ‘How did X feel?’ (test for attributing emotions to the protagonists, referred as ‘Empathy’ in Table 1).

Questions 7 and 8 were control questions to check that participants understood the details of the story: scores are adjusted if the person does not understand the story. Three composite scores were calculated for the present study, as follows:

• ‘Hits’, correct detection of Faux Pas (max/10, converted to %);

• ‘Intentions’, correct explanation of characters intentions (max/10, converted to %);

• ‘Empathy’, correct explanation of emotions (max/10, converted to %).

‘Intentions’ is considered a cognitive ToM measure, whereas ‘Empathy’ is considered an affective ToM measure. All the scores were analysed using normative data adjusted for sex and education that were developed for French-speaking elderly. The analysis provides Z-score for each composite score of the Faux Pas test. This task has validity to assess ToM. ^{Reference Bora, Walterfang and Velakoulis4}

The Zarit Burden Interview-12 items or ZBI-12 was administered to measure the caregiver burden. Items are rated on a five-point Likert scale from 0 (never) to 4 (very often), with higher scores indicating more burden.

The Neuropsychiatric Inventory-12 items or NPI-12 was conducted with the caregiver who had to answer questions about psychiatric symptoms of PD patients and about their own distress. For each symptom, frequency is rated on a three-point scale (max. 3), impact is rated on a four-point scale (max. 4) and distress of the caregiver is rated on a six-point scale (max. 5). Impact scores of 12 items are multiplied by frequency scores, resulting in a severity score (possible maximum score = 144). A total distress score can be obtained with a possible maximum score of 60.

Because of small sample size, Spearman’s rho was applied to calculate non-parametric correlations between different variables at baseline. To add corrections for multiple comparisons, p values were adjusted with false discovery rate correction. ^{Reference Benjamini and Hochberg8}

All statistical analyses were performed with SAS software (version 9.4 for Windows) at the α = 0.1 level of significance.

Table 1 shows descriptive statistics of the total sample. The Schapiro–Wilks test confirmed that most variables were not normally distributed. Thereafter, non-parametric tests were conducted.

Table 2 shows non-parametric correlations between variables. There is a significant correlation between Hits and Empathy scores (r = 0.96, p < 0.0001). This means that Empathy and Hits scores measure a related construct, that is, ToM. Patient symptom severity and caregiver distress scores of the NPI-12 are correlated (r = 0.86, p = 0.011), meaning that the more patients present with psychological and behavioural problems, the more the caregivers are in distress. There is a significant correlation between Faux Pas Intentions and distress score of NPI-12 (r = −0.76, p = 0.008). The less skillful patients are at sensing another person’s intentions, the more distressed the caregivers are. There is also a significant correlation between ZBI-12 and severity score of NPI (r = 0.75, p = 0.0853). The more patients present with psychological and behavioural problems, greater the burden on caregivers (Table 2).

Table 2: Spearman correlations between variables in a PD-MCI sample

Numbers in the table are interpreted as follows: (1) Hits (number of Faux Pas detected correctly); (2) Intentions (correct explication of intentions in the Faux Pas stories); (3) Empathy (correct explication of character’s feeling in the Faux Pas stories); (4) Caregiver burden (ZBI-12); (5) caregiver distress (distress subscale of NPI-12); (6) Severity of neuropsychiatric symptoms (severity subscale of NPI-12); and (7) Disease stage (Hoehn and Yahr).

p < 0.1.

* p < 0.05.

** p < 0.01.

Interestingly, the more difficult it is for patients to understand and guess the intentions (cognitive ToM) of their interlocutor, the greater the distress (NPI-12) experienced by caregivers. In mental health disorder literature, ToM is generally linked to social and occupational functioning of patients, ^{Reference Dvash and Shamay-Tsoory5,Reference Pomponi, Ricciardi and La Torre6} but to date, no study has investigated the impact of ToM on the mental health of their caregivers. In PD, the relationship with the caregiver might be considered as a social functioning outcome. Caregiver distress is therefore an important variable to consider, especially because it is directly related to institutionalisation of PD patients. ^{Reference Goldman, Vernaleo and Camicioli9,Reference Carrilho, Rodrigues and Oliveira10} Given the frequency of caregivers’ psychological symptoms, a future target of PD-MCI cognitive training might be ToM in order to help caregivers in their daily tasks and prevent caregiver distress.

Contrary to our hypotheses, the caregiver burden (ZBI-12) was not associated with the ToM in this PD-MCI sample. The small range of ZBI scores in this sample might have decreased correlations. However, to our knowledge, no study investigated the link between caregiver burden and both components (cognitive and affective) of ToM; therefore, this result cannot be compared with previous findings.

There was a positive correlation between ZBI-12 score and NPI-12 severity of neuropsychiatric symptoms subscale, supporting the results of previous studies indicating that psychological symptoms of PD are associated with caregiver burden. ^{Reference Kudlicka, Clare and Hindle2,Reference Grün, Pieri, Vaillant and Diederich3,Reference Goldman, Vernaleo and Camicioli9,Reference Carrilho, Rodrigues and Oliveira10} For the majority of PD patients included in this sample, depression or anxiety symptoms were identified by the caregiver, but there was also apathy, irritability and sleep problems identified for some patients. Those symptoms may also have contributed to the caregiver burden.

The caregiver burden and distress reported by the caregivers were not associated. Since there was no caregiver that scored higher than moderate burden on the ZBI-12, it is possible that caregivers in this sample might not have reported enough burden to find a correlation between the ZBI score and the NPI distress subscale score.

The small number of participants is the biggest limit of this study; it prevents generalisation to all PD-MCI. Another limitation is inherent to Spearman rho correlations, which have a tendency to inflate the effect sizes of the associations. Indeed, the rho in the present study was elevated for a small sample. Therefore, these effect sizes, although significant, must be interpreted with caution. This type of correlation, with the FDR correction, is frequently used in exploratory studies with small samples not distributed normally.

In conclusion, PD-MCI’s cognitive ToM difficulties seem to be linked to caregiver feelings of distress. However, PD-MCI’s ToM and caregiver burden must be investigated in larger samples in the future.