Determinants of generic and specific health-related quality of life in patients with Parkinson’s disease

Objectives Generic and disease-specific health-related quality of life (HRQoL) instruments may reflect different aspects of lives in patients with Parkinson’s disease (PD) and thus be associated with different determinants. We used the same cluster of predictors for the generic and disease-specific HRQoL instruments to examine and compare the determinants of HRQoL. Method HRQoL was measured in 92 patients with PD by the 36-item Short-Form Health Survey (SF-36) and the 39-item Parkinson’s Disease Questionnaire (PDQ-39). The predictors included demographic and disease characteristics, and motor and non-motor symptoms. Multiple regression analyses were used to identify HRQoL determinants. Results Depressive symptoms and motor difficulties of daily living were the first two significant determinants for both instruments. The other significant determinant for the SF-36 was fatigue and non-motor difficulties of daily living, and for the PDQ-39 was motor signs of PD. Conclusions The results suggest the importance of the evaluation and intervention focused on depressive symptoms and motor difficulties of daily living in patients with PD. In addition, the SF-36 seems more related to non-motor symptoms, while the PDQ-39 appears more associated with motor symptoms. This information is important for understanding results from these two instruments and for choosing which to use.

Introduction like age, disease severity, and fatigue, were significant determinants in some, but not in all of the studies that examined these factors. A possible reason for the inconsistent findings across different studies is the different clusters of predictors analyzed and different outcome measures used.
We hypothesized that the significant determinants would depend upon the HRQoL instrument used. The purpose of this study was to identify determinants of HRQoL in patients with PD by using the same set of predictors for the SF-36 and the PDQ-39. The potential determinants examined included demographic and disease characteristics, and motor and non-motor symptoms.

Methods Participants
In this cross-sectional study, 92 participants were recruited consecutively from the neurology departments of two medical centers in southern Taiwan. The inclusion criterion was a diagnosis of PD by a neurologist according to the United Kingdom PD Society Brain Bank criteria [15]. Patients were excluded if they had dementia (Saint Louis University Mental Status Examination [SLUMS] scores < 19 for people with less than a high school education and < 20 for those with a high school education and above) [16]. This study followed the principles of the Declaration of Helsinki. The study protocol was approved by the Institutional Review Board of the hospitals (approval number: B-ER-101-171, EMRP-102-024). Written informed consent was obtained from each participant.

Procedures and measures
All participants were interviewed face-to-face by one interviewer with the following instruments in a random order. Data were collected from May 2013 to July 2014. Most participants (74/92, 80.4%) were evaluated when they were in the "on" phase of the medication cycle.
Motor and non-motor symptoms were measured using the Movement Disorder Society Revision of the Unified Parkinson's Disease Rating Scale (MDS-UPDRS) [17], which includes Part I: Non-motor difficulties of daily living (13 items); Part II: Motor difficulties of daily living (13 items); Part III: Motor signs (33 items); and Part IV: Motor complications (6 items). Each item is scored from 0 (normal) to 4 (severe) and item scores are summed for the total score of each part. The Traditional Chinese version of the MDS-UPDRS has been validated in Taiwan and Hong Kong [18].
In addition, non-motor symptoms of depression, fatigue, and cognitive dysfunction were measured using the Geriatric Depression Scale (GDS) [19], the Fatigue Severity Scale (FSS) [20], and the SLUMS, respectively. The 30-item GDS is an efficient depression screening tool for patients with PD [19]. Each item is scored as either 0 or 1. A total score of 10-19 is regarded as mild depression, and 20-30 as severe depression [19]. The Chinese version of the GDS has been reported to have adequate reliability and validity [21].
The 9-item FSS is a unidimensional generic fatigue rating scale that includes items on physical fatigue, mental fatigue, and social aspects [20]. Each item in the FSS is scored from 1 (disagree) to 7 (agree). A mean score ! 4 indicates significant fatigue. The FSS has shown good reliability, validity, and sensitivity to changes in patients with PD [20]. The Chinese version of the FSS has been reported to have acceptable reliability and validity in patients with major depressive disorder and non-depressive people [22].
The SLUMS is 30-point screening questionnaire that includes four domains: orientation, memory, attention, and executive functions. The cutoff score for people with less than a high school education is 19 for dementia and 24 for a mild neurocognitive disorder. For those with a high school or higher education, the cutoff score is 20 for dementia and 26 for a mild neurocognitive disorder [16]. The SLUMS has been culturally adapted and translated into Chinese for use in Taiwan [23].
The SF-36 and PDQ-39 were used to measure HRQoL. The SF-36 includes eight domains: physical functioning, role limitations due to physical problems, bodily pain, general health perceptions, mental health, role limitations due to emotional problems, social functioning, and vitality. Eight domain scores and a summary score can be derived (range: 0-100), with higher scores indicating better HRQoL. The Chinese version of the SF-36 has been reported to have good reliability and validity [24].
The PDQ-39 has eight domains: mobility, ADL, emotional well-being, stigma, social support, cognition, communication, and bodily discomfort. It is composed of a single index (PDQ-39 SI) score and eight domain scores (range: 0-100), with higher scores indicating worse HRQoL. Psychometric testing of the Chinese-translated version of the PDQ-39 showed satisfactory reliability and validity in Taiwanese patients with PD [14].
Statistical analysis SPSS 17.0 was used for all data analyses. Descriptive statistics were used to summarize the participants' demographic and clinical characteristics, motor and non-motor symptoms, and HRQoL. Spearman's rank-correlations were computed for associations between HRQoL and participants' characteristics and symptoms. Scores of the GDS, FSS and SLUMS were entered as continuous date. Multiple regression analyses were also done using significantly correlated variables to identify the determinants of HRQoL for the SF-36 and for the PDQ-39. Significance was set at p < 0.05 (two tailed).

Results
Of the 92 patients with PD, most (77.2%) were at Hoehn and Yahr (H&Y) stages I and II. Many (65.2%) of them had a mild neurocognitive disorder, and some had depression (23.9%) and fatigue (33.7%) ( Table 1). There was no significant difference in the HRQoL scores between those in the "on" and "off" phases of the medication cycle (S1 Table).
Employment status, the H&Y stage, the MDS-UPDRS Parts I-III, the GDS, and the FSS were significantly correlated with both the SF-36 and the PDQ-39 (Table 2). In addition, age and the SLUMS were significantly correlated only with the SF-36, but not with the PDQ-39, while disease duration and the MDS-UPDRS Part IV were significantly correlated only with the PDQ-39 but not with the SF-36.
To use the same set of predictors for the multiple regression analyses, we selected the variables that were significantly correlated with both the SF-36 and the PDQ-39 as predictors, including employment status, the H&Y stage, the MDS-UPDRS Parts I-III, the GDS, and the FSS. The SF-36 and the PDQ-39 were separately entered as the outcome variable. The Kolmogorov-Smirnov test was used to check the normality of the SF-36 total scores and the PDQ-39 SI. The test results indicated non-normality for the PDQ-39 SI and thus the logarithm transformation was used.
The results of regression analyses indicated that significant determinants for the SF-36 were the MDS-UPDRS Parts I and II, the GDS, and the FSS, all of which accounted for 61.8% of the variance in the SF-36 (Table 3). The significant determinants for the PDQ-39 were the MDS-UPDRS Parts II and III, and the GDS, which accounted for 54.0% of the variance in the PDQ-39.
We further conducted correlational analyses to understand the relationship between the HRQoL domains and the significant determinants (i.e., the MDS-UPDRS Parts I-III, the GDS, and the FSS) ( Table 4). The GDS was significantly correlated with all domains in each instrument. The MDS-UPDRS Parts I and II and the FSS were correlated with all domains in the SF-36 and most in the PDQ-39. The MDS-UPDRS Part III was correlated with the physical functioning, role-physical, bodily pain, general health, and vitality domains in the SF-36 and the mobility, ADL, and bodily discomfort domains of the PDQ-39.

Discussion
By using the same cluster of predictors for the SF-36 and the PDQ-39, we found that depressive symptoms and motor difficulties of daily living were the first two significant determinants for both instruments. The other significant determinant for the SF-36 was fatigue and nonmotor difficulties of daily living and for the PDQ-39 was motor signs. Our finding that depression was a major determinant in the SF-36 and the PDQ-39 are consistent with other studies and reviews [5,[25][26][27][28][29][30]. We found that depression was significantly correlated with not only psychosocial but also physical functioning domains in the SF-36 and the PDQ-39. Similar results were also reported in a recent study [31] that depression was significantly correlated with most domains of the PDQ-39 and was a significant determinant for the mobility, emotional well-being, stigma, and communication domains. Using mediation analysis, research [32] reported that depression might contribute to ADL difficulties, which in turn might contribute to decreased HRQoL in patients with PD. The interrelationships among depression, physical function, and HRQoL suggest the need for an interdisciplinary approach to alleviate the effects of psychological and physical dysfunction on the HRQoL of patients with PD. The MDS-UPDRS Part II was another significant determinant in both the SF-36 and the PDQ-39. Motor difficulties of daily living were associated with poor HRQoL, which is consistent with other studies using the PDQ-39 [28] and other HRQoL tools (e.g., EQ-5D, PDQ-8) [33,34]. The MDS-UPDRS Part II asks questions directly related to daily activities with patients' views of primary disease symptoms, and it has been widely used to assess motor symptoms and disabilities in PD [33][34][35]. Our finding that the MDS-UPDRS Part II was more strongly associated with the HRQoL instruments than were the MDS-UPDRS Part III and the H&Y stage is also consistent with other research [28], which suggests that the role of the motor difficulties of daily living outweighs the roles of motor symptoms and disease severity in HRQoL. The difficulties completing simple daily tasks (e.g., using eating utensils, dressing, and walking) adversely affect HRQoL in patients with PD [5,32].

SF-36
Physical Fatigue was also a significant determinant for the SF-36, which is congruent with one review [5] and two studies using the SF-36 [36,37]. Many domains of the SF-36 encompass attributes of energy, strength, and endurance. For example, the domain of physical functioning asks questions about vigorous and moderate activities, and the domain of role limitations due to physical problems asks about overall activity performance in daily life in terms of time and types. Therefore, the association between the SF-36 and the FSS might be attributable to the similar properties that underlie the questionnaire items.
In addition, the non-motor difficulties of daily living (the MDS-UPDRS Part I) were also a significant determinant for the SF-36. The MDS-UPDRS Part I evaluates a broad range of non-motor symptoms, such as anxious mood, apathy, features of dopamine dysregulation syndrome, daytime sleepiness, and urinary problems. Our results are consistent with recent findings [38][39][40] that the overall burden of non-motor symptoms was a significant contributor to HRQoL in PD, suggesting the importance of active screening and managing non-motor symptoms.
In agreement with other research [41], we found that motor signs (the MDS-UPDRS Part III) were also a significant determinant for the PDQ-39. The results suggest that the mobility, ADL, and bodily discomfort domains of the PDQ-39 are also related to motor signs of PD in addition to motor difficulties of daily living. PD motor features are frequently evaluated to help healthcare professionals understand the response to dopaminergic therapy [42]. Our findings suggest that the PDQ-39 may be a more responsive tool than the SF-36 for reflecting HRQoL changes related to motor signs.
Our findings show that the SF-36 and the PDQ-39 reflect somewhat different aspects of the lives of patients with PD. Specifically, the SF-36 is more associated with non-motor symptoms, and the PDQ-39 is more related to the motor symptoms of PD. Therefore, it is important to know what HRQoL instrument is used to analyze the contributing factors of HRQoL, rather than assuming that all HRQoL instruments reflect the same construct. The choice of which instrument to use may depend upon how much the area of interest is related to general functioning or to PD-related symptoms. It was recently suggested [5,12] that both generic and disease-specific measures be used to capture unexpected improvement. However, considering the extended time and effort that would be needed with more evaluations, we suggest that an understanding of the HRQoL instrument's characteristics is important.
In addition, we noticed that the stigma domain in the PDQ-39 was significantly associated with only depression and motor difficulties of daily living, while most of the other HRQoL domains were associated with more than three of the significant determinants. The PDQ-39 was developed from in-depth interviews with patients with PD to learn about the areas of their lives that had been adversely affected by their PD [10]. Therefore, the domains in the PDQ-39 reflect the concerns of patients with PD and should be carefully examined. The finding that the stigma domain had only scant correlations with PD symptoms suggests that merely addressing motor and non-motor symptoms may not be enough, and that a broader consideration from the social context is needed to understand stigma in PD.
This study has some limitations. Our participants were recruited from medical centers, and thus the generalizability may be limited to clinic-based patients with PD. In addition, we did not collect information on medications or levodopa equivalent daily lose, or control for the medication cycle phase during the evaluation. Future research should include medication information and be consistent on the phase of medication cycle during the evaluation. This is a cross-sectional study in which all assessments were administered at one time point. Followup of the participants would be helpful for understanding the longitudinal change and to explore possible causal relationships between participants' characteristics and HRQoL. Finally, the determinants we examined were factors related to the individual. Future work should include physical and social environmental factors to broaden our understanding of the determinants of HRQoL in patients with PD.

Conclusions
This is the first study that uses the same set of predictors for both commonly used generic and disease-specific HRQoL instruments in PD. The common determinants in both the SF-36 and PDQ-39 suggest the importance of evaluating and intervening on depressive symptoms and motor difficulties of daily living in patients with PD. In addition, we found that the SF-36 seems to be more related to non-motor symptoms, while the PDQ-39 appears more related to motor symptoms. This information is important for choosing which instrument to use in clinical practice and for research purposes.
Supporting information S1 Table. Comparison of HRQoL scores between patients with PD in the "on" and "off" phases of the medication cycle. (PDF) S1 Appendix. Raw data of all relevant data in all participants. (XLSX)