Abstract
Background:
Recent research suggests that a significant number of those who receive advanced treatments for Parkinson’s disease (PD) do not report improvements for some symptoms, which may relate to their pre-treatment expectations. It is important that expectations of treatment are measured and discussed prior to advanced treatment.
Objective:
The primary aim of this study was to develop a measure of treatment expectations of two advanced-stage treatments in PD, deep brain stimulation (DBS), and Levodopa/Carbidopa Intestinal Gel (LCIG). A secondary aim was to explore potential predictors of treatment expectations.
Methods:
The questionnaire-based measure was developed by researchers in conjunction with a highly experienced clinician, and evaluated treatment expectations in 189 people aged 46–91 years (M = 71.35, SD = 8.73; 61% male) with idiopathic PD.
Results:
The overall measure demonstrated excellent internal consistency (α= 0.96). Exploratory factor analysis suggested the scale was unidimensional for both DBS and LCIG. Participant expectations of the two treatments differed significantly, with expectations being higher for DBS. Perceived symptom severity was the strongest predictor of treatment expectations.
Conclusion:
This scale has potential to inform clinicians about client expectations prior to advanced stage therapy for PD, with a view to the management of these expectations. Further evaluation of the scale is required across different treatment contexts.
Keywords
INTRODUCTION
In individuals with Parkinson’s disease (PD), it is important to understand the impact of expectations regarding treatment(s), including Levodopa/Carbidopa Intestinal Gel (LCIG) and Deep Brain Stimulation (DBS) that are typically used when the symptoms of PD are advanced. Recent research suggests that individuals with PD who have unrealistic expectations (e.g., improved cognition) may not be as satisfied with DBS outcomes as those who have realistic expectations (e.g., reduced ‘off’ periods) [1–4]. Not all those who have received DBS experience significant improvement, with many reporting modest improvements or no change in symptom severity post-treatment [5]. Many of those who have received DBS report that social and non-motor symptoms have not improved in the manner or magnitude they anticipated [6]. One potential explanation for this is the influence of pre-treatment expectations, including those relating to psychological factors such as anxiety and depression [7]. Due to the potential impact of treatment expectations on perceived clinical outcomes in PD, it is important that expectations are measured and discussed prior to any treatment. A reliable and valid measure of treatment expectations in PD is therefore required.
Two studies to date have examined treatment expectations in PD, both using self-report measures [8, 9]. Nisenzon et al. (2011) [8] used a modified version of the Patient Centred Outcomes Questionnaire [10] (PCOQ) to identify which symptoms were central to perceived treatment success. They reported that reduced motor symptom severity, particularly gait, walking, and fatigue, were key outcome priorities. As the PCOQ was designed for use in chronic pain, however, many aspects of PD were omitted from the study and expectations relating to advanced treatments for PD (e.g., LCIG, DBS) were not specifically addressed. Reddy et al. (2014) [9] developed the Patient-Reported Outcome Tool for Advanced Parkinson’s Disease (PRO-APD) to assess perceived symptom severity and advanced treatment expectations in PD. The 25-item PRO-APD was administered to 22 participants with PD prior to their receiving DBS, Intrajejunal Levodopa Infusion (IJL), or Apomorphine (Apo). Individuals rated the severity of their PD symptoms across a range of domains, including motor and non-motor symptoms, cognition, and social. Higher perceived symptom severity was associated with higher expectations for improvement, with expectations being highest for motor symptoms and weakest for cognitive and social domains. Although this study provides some valuable insight into the relationships between symptoms and expectations of treatment, there are some limitations to be considered. All participants had already been scheduled to receive an advanced treatment (DBS, IJL, Apo) and were asked only to rate expectations of the therapy they were to receive, meaning that no comparisons between expectations for different therapies could be made. The small sample size, coupled with the fact that participants had already engaged extensively in discussions about the therapy they were to receive, may have impacted the findings.
The primary aim of the present study was to develop a measure of expectations for treatments in PD among those who had not yet received formal information about such treatments. The measure was developed to assess both the perceived severity of an individual’s PD symptoms and the individual’s expectations that LCIG/DBS could reduce these symptoms. Part of the measure (subscale A) was developed to align with aspects of the primary clinical assessment tool for PD symptoms –the Movement Disorder Society Unified Parkinson’s Disease Rating Scale [11] (MDS-UPDRS), thereby enabling the measure to be used to examine the relationships between perceived PD symptoms (subscale A), treatment expectations (subscales B and C), and future (post-treatment) clinical assessment of symptoms using the MDS-UPDRS. In line with previous research, and to provide some indication of the validity of the measure, it was predicted that higher symptom severity would be associated with poorer quality of life (for a review, see Soh et al. (2011) [12]).
This study also sought to identify psychological and psychosocial factors that may predict advanced treatment expectations in PD. In accordance with research indicating a relationship between depression and expectations in heart disease [13] and major depression [14], it was proposed that depression and anxiety would significantly predict treatment expectations for both DBS and LCIG. In accordance with recent research indicating that self-efficacy influences outcome expectations in stroke [15], it was also predicted that self-efficacy would significantly predict expectations for both therapies. This study also explored whether dysfunctional attitudes (negative thoughts) contribute to treatment expectations.
MATERIALS AND METHODS
The study was approved by the Curtin University Human Research Ethics Committee (HRE2016-0124).
Participants
Participants were 189 individuals aged 46–91 years (M = 71.35, SD = 8.73; 61% male) with a diagnosis of idiopathic PD made by a neurologist. All participants were members of Parkinson’s Western Australia, a not-for-profit organization offering support and services for those with PD in Western Australia. Participants were recruited via postal invitation through Parkinson’s Western Australia’s member database. Self-reported age at diagnosis ranged between 38 and 86 years (M = 63.26, SD = 10.16). Most participants were currently married (74.2%). Participants were excluded from the study if they had received DBS or LCIG.
Measures
The Parkinson’s disease treatment expectation scale (PTES)
The PTES was developed as a measure of treatment expectations in PD in accord with DeVellis’ scale development [16]. The content and structure of the scale was based on symptoms in the MDS-UPDRS [11]. The PTES was developed in collaboration with a clinical team comprising a PD neurologist and Psychiatrist, a PD specialist nurse, and Psychologists with a track record of research in PD. The PTES comprised 3 subscales (A, B, C). Subscale A comprised of 31 items and measured self-reported symptom severity across a range of PD symptoms. The symptoms included in Subscale A aligned with symptoms included in the MDS-UPDRS, including motor (e.g., balance, tremor, rigidity), speech, swallowing, autonomic (e.g., urinary, bowel), cognition (e.g., mood, apathy, impulsivity), dyskinesias, and sleep. Part 3 of the MDS-UPDRS comprises 33 items assessing the motor symptoms of PD and is administered by a trained researcher or clinician. Motor symptoms are rated on a 4-point Likert scale from 0 (Normal) to 4 (Severe). In the present study, participants indicated the severity of a symptom using a 5-point Likert scale ranging from slight (1) to very severe (5). If a symptom did not apply to them, participants selected ‘not applicable’, which was scored as zero. Overall scores could therefore range from 0 (none of the symptoms) to 155, with higher scores indicating more severe symptoms. Subscale B measured expectations of DBS as a treatment for the symptoms in subscale A, and subscale C the expectations of LCIG. After indicating the perceived severity of a symptom, participants were asked to indicate the extent to which they expected DBS (subscale B) and LCIG (subscale C) to improve that symptom, using a 5-point Likert scale ranging from strongly disagree (1) to strongly agree (5). If the participant had indicated in Subscale A that a given symptom did not apply to them, that item was scored as zero (as an expectation rating would be irrelevant). Expectation scores could range from 0 (not applicable) to 155, with higher scores indicating a stronger belief that the treatment (DBS, LCIG) would improve symptoms. Four further questions asked participants (i) whether they expected side effects after DBS or LCIG, and (ii) how severe they would expect any side effects to be. All four side-effect questions were rated using a 5-point Likert scale ranging from strongly disagree/slight (1) to strongly agree/severe (5).
The patient health questionnaire 9 (PHQ-9) [17]
The PHQ-9 measures depression over the previous two weeks. The scale includes nine items and participants respond using a 4-point Likert scale, ranging from ‘not at all’ to ‘nearly every day’. Overall scores range from 0 to 27, with higher scores indicating increased symptoms of depression. The PHQ-9 demonstrated good reliability in the current study (α= 0.83).
The Parkinson anxiety scale (PAS) [18]
The PAS measures anxiety over the previous month. Three subscales assess persistent anxiety, episodic anxiety, and avoidance behaviors. Participants respond using a 5-point Likert scale ranging from ‘never/not at all’ to ‘severe/nearly always’. Overall scores range from 0 to 48, with higher scores indicating higher levels of anxiety. The measure demonstrated excellent internal consistency in the present study (α= 0.90).
The Parkinson’s disease questionnaire (PDQ-39) [19]
The PDQ-39 comprises 39 items measuring the impact of PD symptoms on quality of life. Participants respond using a 5-point Likert scale to indicate the impact of specific symptoms in the last month. Responses range from ‘never’ to ‘not at all/cannot do at all’. Overall scores range from 0 to 195, with higher scores indicative of poorer quality of life. The measure demonstrates good construct and convergent validity. In the current study, the scale demonstrated excellent reliability (α= 0.93).
The general self-efficacy scale (GSES) [20]
The GSES measures how well individuals cope with unexpected and/or stressful life events. The measure contains 10 items, which participants respond to using a 4-point Likert scale ranging from ‘not at all true’ to ‘exactly true’. Overall scores range from 10 to 40, with higher scores indicating increased belief in the ability to successfully complete tasks (self-efficacy). The measure demonstrated sound internal consistency in the present study (α= 0.73)
The dysfunctional attitudes scale (DAS) [21]
The DAS comprises 18 items measuring the intensity of dysfunctional attitudes and beliefs. Participants respond using a 4-point Likert scale ranging from ‘totally agree’ to ‘totally disagree’. Overall scores range from 18 to 72, with higher scores indicative of increased dysfunctional attitudes. The measure demonstrated excellent reliability in the present study (α= 0.90).
Procedure
Members of Parkinson’s Western Australia, a not-for-profit support organization for individuals with PD, were invited to participate via postal invitation. Members received an information pack containing a paper copy of the questionnaire, Information and Consent form, and a reply-paid envelope to return should they wish to participate. There were 740 information packs distributed and 206 were returned (28% return rate). Unfortunately, a number of those returned questionnaires were missing significant amounts of information that could not be imputed. Hence, the final sample for analysis included 189 completed questionnaires.
RESULTS
There were 189 participants in total, of which 60% were male. The youngest participant was 46 years old and the oldest was 91 (M = 71.31, SD = 8.73). Average disease duration was 8.20 years (SD = 6.09), minimum duration was less than one year and maximum was 32 years. Of the 189 participants, none were scheduled to receive either DBS or LCIG treatment. Average self-reported disease severity was 58.13, minimum severity was 14 and maximum was 138 (out of 155). Using Kroenke and Spitzer’s (2002) [17] cut-offs, 44% of participants were not depressed, 29% had mild depression, 17% had moderate depression, 4% had moderately severe depression, and 2% had severe depression. The means and standard deviations for all measures (including age and disease duration) are presented in Table 3.
Overall internal reliability of the PTES was excellent (α= 0.96). Cronbach’s alpha for subscale A (symptom severity) was 0.95 and was 0.96 for both subscales B (DBS) and C (LCIG). A paired-samples t-test revealed that expectations of DBS and expectations of LCIG were significantly different, t(188) = 12.42, p = 0.001, d = 0.31. Participants expected DBS to be significantly more effective at reducing symptom severity than LCIG. A paired samples t-test revealed that expectations of side effects for each treatment did not significantly differ, t(188) = –1.11, p = 0.27, and that severity of side-effects for each treatment did not significantly differ, t(188) = –0.13, p = 0.90.
Exploratory Factor Analysis (EFA) with principal axis factoring with promax rotation was used to determine whether subscales B (DBS) and C (LCIG) were driven by the same underlying variable(s). Five factors were extracted with Eigenvalues > 1 for subscale B. Visual examination of the scree plot revealed that the inflection point occurred after the first factor, with an eigenvalue of 14.97. This single factor accounted for 48.31% of the total variance. Item 25 was the only item to cross-load onto two or more factors, with both primary and secondary factor loadings < 0.5, indicating uni-dimensionality. The factor loadings for each of the items in subscale B are in Table 1.
Promax Rotated Factor Structure of the PTES for Subscale B (DBS)
Seven factors were extracted with Eigenvalues > 1 for subscale C. Visual examination of the scree plot revealed that the inflection point occurred after the first factor, with an eigenvalue of 10.33. This single factor solution accounted for 33% of the total variance. Some items cross- loaded on two or more factors, with all but one item (22) loading onto Factor 1. Internal clustering of those items suggests that Factors 1, 2, 3, and 4 represent the following categories of PD symptoms; general PD symptoms, psychological factors, autonomic symptoms, and tremor-related symptoms. The factor loadings for each of the items in subscale C are presented in Table 2. Overall EFA findings indicated that subscales B and C shared similar factor structures and were uni-dimensional.
Bivariate correlations examined the relationships between study measures and are presented in Table 3. As predicted, scores on subscale A (perceived symptom severity) and the full PDQ-39 were strongly, positively correlated, r(204) = 0.799, p < 0.001, r2 =0.638, demonstrating concurrent validity. More severe self-reported symptoms were associated with lower quality of life.
Two separate hierarchical regression analyses then examined the predictive ability of self-reported severity and psychological measures for (i) DBS and (ii) LCIG expectations.
The first hierarchical regression included subscale B (DBS) as the outcome variable. Disease duration did not correlate with any of the psychological measures, so was not retained as a control variable in the analyses. As age correlated with some psychological measures, it was retained as a control variable at step 1 in both regressions. Age accounted for a non-significant 0.30% of the variance in DBS expectation scores, R2 = 0.003, F(1, 186) = 0.47, p = 0.49. At step 2, self-reported severity (subscale A) was added and significantly predicted 36% of the variance, R2 = 0.36, F(1, 185) = 51.13, p = 0.000. At step 3, depression was added and non-significantly accounted for an additional 0.6% of the variance, R2 = 0.36, F(1, 184) = 34.80, p = 0.19. At step 4, anxiety was added and did not account for any additional variance, R2 = 0.36, F(1, 183) = 26.01, p = 0.72. At step 5, dysfunctional attitudes was added and did not account for any additional variance, R2 = 0.36, F(1, 182) = 20.76, p = 0.63, At step 6, self-efficacy was added and did not account for any additional variance, R2 = 0.36, F(1, 181) = 17.22, p = 0.87. At each step, self-reported severity remained a significant predictor of DBS expectation scores. Regression statistics for subscale B are presented in Table 4.
Promax Rotated Factor Structure of the PTES for Subscale C (LCIG)
The second hierarchical regression included subscale C (LCIG) as the outcome variable. As for the previous regression, age was retained as a control variable at step 1 and accounted for none of the variance in LCIG expectation scores, R2 = 0.000, F(1, 186) = 0.004, p = 0.95. At step 2, self-reported severity (subscale A) was added and significantly predicted 36% of the variance, R2 = 0.36, F(1, 185) = 52.01, p = 0.000. At step 3, depression was added and non-significantly accounted for an additional 0.2% of the variance, R2 = 0.36, F(1, 184) = 34.78, p = 0.45. At step 4, anxiety was added and did not account for any additional variance, R2 = 0.36, F(1, 183) = 26.14, p = 0.49. At step 5, dysfunctional attitudes was added and did not account for any additional variance, R2 = 0.36, F(1, 182) = 20.81, p = 0.80, At step 6, self-efficacy was added and did not account for any additional variance, R2 = 0.36, F(1, 181) = 17.25, p = 0.99. At each step, self-reported severity remained a significant predictor of LCIG expectation scores. Regression statistics for subscale C are presented in Table 5.
Correlation matrix for all measures
DURATION, disease duration in years; PTES_A, Symptom severity subscale; PTES_B, DBS expectations subscale; PTES_C, LCIG expectations subscale; PDQ-39, Parkinson’s Disease Questionnaire-39; PHQ-9, Patient Health Questionnaire; DAS, Dysfunctional Attitudes Scale; GSES, General Self Efficacy Scale; PAS, Parkinson’s Anxiety Scale; M, Mean; SD, Standard Deviation. *p < 0.05; **p < 0.01.
Regression statistics for hierarchical regression with DBS expectations score as the criterion (outcome) variable
CI, confidence intervals. *p < 0.001.
Regression statistics for hierarchical regression with LCIG expectations score as the criterion (outcome) variable
CI, confidence intervals. *p < 0.001.
DISCUSSION
The present study developed the Parkinson’s Disease Treatment Expectation Scale (PTES) to measure expectations of advanced treatments in PD. For both DBS and LCIG expectations, a single factor explained most of the variance. This factor related to general PD symptoms and suggests that the measure is generalizable across PD treatment contexts. Future research should explore the factor structure of the measure with a range of different PD treatments and consider different factor-analytic solutions (such as bi-factor).
Perceived symptom severity was the only statistically significant predictor of both DBS and LCIG expectations. This is in line with Reddy et al.’s findings [9] and indicates the importance of subjective PD symptom severity for treatment expectations and, potentially, post-treatment satisfaction. PD symptom severity is typically assessed by a clinician using the MDS-UPDRS, but it is unclear whether those who participated in this study (and previous expectation studies) accurately self-reported their PD symptoms. This is important, as treatment satisfaction requires the individual to (i) subjectively assess their symptom severity and (ii) compare their pre and post subjective assessments, in order to determine whether they think the treatment has led to any improvement in symptoms. If self-reported symptoms do not align with the (objective) clinician’s assessment of symptoms, then perceived improvement of symptoms will most likely also be disparate. Multiple studies indicate that self-awareness of symptoms, both motor and non-motor, may be impaired in PD [22–25]. Given that self-awareness of symptoms is integral to the process of evaluating symptom improvement, it is important that an individual’s own assessment of their symptoms is realistic both pre and post treatment. An overly negative or positive subjective assessment of symptoms may impact treatment expectation and, in turn, satisfaction. As this study did not include an objective (MDS-UPDRS) assessment of participant’s symptoms, it was not possible to examine the relationships between subjective symptom severity, objective symptom severity, and treatment expectations. Future research should examine whether (subjective) self-reported PD symptom severity aligns with (objective) clinical assessment, and whether the two differentially predict treatment expectations.
Another key finding was that expectations of DBS and LCIG were significantly different, but the expectations of side effects were not. Participants had much higher outcome expectations for DBS compared to LCIG. This is important to consider in clinical practice, as those with lower expectations of LCIG may be less inclined to consider it as a treatment option when it could be highly effective for them. This finding may also indicate that participants could have unrealistically high outcome expectations relating to DBS. LCIG can be highly effective at reducing PD symptoms and improving quality of life where traditional oral or transdermal therapy has become ineffective [26]. DBS is also highly effective for specifically targeted PD patients and has been demonstrated to improve a range of motor symptoms [27, 28], speech [29, 30], and activities of daily living [31]. However, research concerning changes relating to neurocognition, behavior, and mood after DBS are limited and have yielded mixed results that are difficult to interpret (for a review, see Moldovan et al. (2015) [32]). Some research suggests an overall improvement in neuropsychiatric symptoms following DBS, but there is also some research to suggest that DBS is associated with neuropsychological decline and increased negative affect [33–35]. Overall, a lack of randomized control trials (for obvious ethical reasons) has made comprehensive assessment of outcomes relating to LCIG and DBS difficult; larger, controlled, multi-center studies using standardized assessment tools are required.
Although psychological and psychosocial factors in this study did not significantly predict treatment expectations, it is worth examining the relationships. Those participants who reported higher anxiety and depression demonstrated higher expectations for both DBS and LCIG treatments. Given the prevalence rates of depression and anxiety in PD, with some studies indicating a co-morbidity rate as high as 76% [36], this finding suggests it is important to consider anxiety and depression at a pre-treatment stage. Cognitive Behavior Therapy (CBT) and Mindfulness-based Cognitive Therapy (MBCT) have demonstrated efficacy in reducing both anxiety and depression in PD [37–39] and may help to address potentially unrealistically high treatment expectations. Future research should examine if psychological treatment prior to advanced treatment may impact on the perceived efficacy and satisfaction with treatment outcomes. If expectations are very high prior to treatment, and the treatment fails to meet the individuals’ expectations regarding improvement of their PD symptoms, this could exacerbate existing symptoms of anxiety and depression and further negatively impact upon quality of life.
Although general self-efficacy did not predict treatment expectations, correlational analysis revealed significant negative relationships, where higher self-efficacy was associated with lower expectations for both therapies. This finding is consistent with previous research demonstrating self-efficacy interacts with outcome expectations [40] and suggests that self-efficacy is associated with expectations for advanced PD treatments. Lower self-efficacy is associated with poorer self-management of PD symptoms [41]. Future research should examine whether psychological therapies tailored for PD, including MBCT and CBT [37, 42], administered prior to treatment may improve self-efficacy associated with reductions in anxiety and depression from psychological treatment. Contrary to predictions, negative cognitions on the Dysfunctional Attitudes Scale showed no significant relationship with expectations in the current study. Future research may want to consider more thorough examination of the role of repetitive negative thinking, which has an established relationship with anxiety and depression [43] and was related to anxiety and depression in the current study.
There are a number of limitations of the current study to be considered. One limitation of the current study is that an individual’s prior exposure to treatment interventions, or how this exposure may influence treatment expectations, was not measured. Future research should consider the moderating influence of prior treatment exposure on treatment expectations and outcomes in PD [44]. Another limitation of the current study is that participants were not scheduled to receive either DBS or LCIG, meaning that they would not yet have received any formal information about the potential of either treatment. It is difficult to determine the source(s) of their knowledge concerning both treatments. The media, internet, support groups, and general PD clinics are all potential sources of information (and misinformation) regarding treatments. Future studies should ask explicitly about the source(s) of knowledge concerning such treatments and to establish any potential sources of misinformation and whether this information is related to expectations. A final limitation of the study is that clinical assessment of disease severity was not included, future research examining the scale should include objective assessment of PD symptoms.
Whilst we acknowledge that there is much research yet to be done and that the present measure requires further development, the PTES can be used to further explore the nature of treatment expectations in PD. Future studies should examine whether the PTES can identify individuals with unrealistic treatment expectations and those factors that may influence expectations. With a greater understanding of an individual’s pre-treatment expectations, even prior to any intention to explore those treatments, health care providers can implement strategies to moderate expectations prior to the intervention.
CONFLICT OF INTEREST
There are no conflicts of interest for any author.
Footnotes
ACKNOWLEDGMENTS
This work was partially supported by Parkinson’s Western Australia Zrinski Grant awarded to principal investigator Sergio Starkstein and co-investigator Sarah Egan. The funder of the study, Parkinson’s Western Australia, had no role in study design, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.
