Phase 1: Survey data collection and analysis

A cross-sectional national survey amongst PwP and caregivers living across all regions in the United Kingdom (UK) was conducted from September 2021 to January 2022. The aim of the survey was to determine the extent of treatment burden in PD and identify key associated factors. Adult participants (age > 18 years), with a diagnosis of PD, as well as caregivers for someone with PD, were recruited via two methods: 1) via Parkinson’s UK Research Support Network and Take Part Hub, where a link to the participant information sheet and online surveys were advertised on the organisation’s website, and 2) through two PD outpatient clinics in the South of England, where interested participants were given a survey pack containing the participant information sheet, survey booklet, and prepaid return envelopes. Two separate surveys, one for PwP and one for caregivers, were available in both paper and online formats (see S1 File) with matching questions wherever feasible. PwP were eligible irrespective of whether they had a caregiver. Similarly, caregivers could participate even if the person with PD they cared for was unable or did not want to. All data were self-reported. As this was an exploratory study, formal sample size calculation was not performed as the intention was not to establish causal relationships or predictive aspects of treatment burden in PD. All paper survey responses were manually double entered into the data spreadsheet by the first author (QYT), a research administrator and a medical student.

Treatment burden measure

Treatment burden levels were measured using the MTBQ, which has been validated in older adults with multimorbidity in the UK [13]. For PwP, the 13-item MTBQ included difficulties related to taking multiple medications, self-monitoring, lifestyle changes, obtaining health information, coordinating healthcare, impact on family and friends, financial burden, access to healthcare out of hours and access to community services. A 16-item MTBQ, adapted by the original research team but not yet validated, was used following direct communication and permission from the developers. The 16-item MTBQ included additional questions regarding difficulties arranging respite care, the financial impact of being a caregiver, and lifestyle adjustments to look after the person they cared for. Each item on the MTBQ was rated on a 5-point scale from ‘0’ (not difficult/does not apply) to ‘4’ (extremely difficult). Responses with >50% missing data were excluded from analysis [13].

Other variables

Sociodemographic data collected included age, sex, ethnicity, marital status (married/civil partnership vs. single/divorced/dissolved civil partnership/widowed), living arrangements (living alone vs. living with spouse/partner/family) and employment status (employed vs. unemployed/retired). Variables related to treatment burden and capacity included the number of medications, frequency of medications, perceived difficulty getting PD-related information, and frequency of healthcare appointments for PD.

Health literacy was assessed using the single-item literacy score (SILS), scored on a 5-point Likert scale, where responses of ‘some’, ‘often’ and ‘always’ indicated limited health literacy [17]. Overall health and PD characteristics were assessed with length of PD diagnosis (years), PD severity (Hoehn and Yahr (H&Y) staging 1–5, with higher stage indicative of worse PD severity) [18], presence of non-motor symptoms (Non-Motor Symptoms Questionnaire (NMSQuest)) [19], self-reported number of other long-term conditions, frailty measure (PRISMA-7) [20], and health-related quality of life (Short-Form 12-Item Survey version 2)(SF12v2) [21] producing physical component summary (PCS) and mental component summary (MCS) scores. Caregiver burden was measured using the 12-item Zarit Burden Interview (ZBI-12) [22] with higher total scores indicative of higher caregiver burden levels. Caregiver burden is related to the emotional, social, financial, physical and spiritual impact and well-being of being a caregiver, and is a separate concept that may impact caregiver treatment burden [10,23].

Quantitative statistical analysis

Statistical analysis was conducted using the IBM Statistical Product and Service Solutions (SPSS) software. Surveys completed by PwP and caregivers were analysed independently. No comparisons were made between online and paper survey responses due to the small number of paper surveys (N = 31). Global MTBQ scores were calculated as per Duncan et al and categorised into ‘no burden’ (score = 0), ‘low burden (score <10)’, ‘medium burden’ (score 10–21) and ‘high burden’ (score ≥22) [13]. Participants with no and low treatment burdens were combined into one group, and those with medium and high treatment burdens were combined into another group for analysis. This dichotomisation was felt to be clinically appropriate given the progressive nature of PD, where early recognition of PwP and caregivers with medium and high treatment burden could allow timely interventions to prevent increasing treatment burden or reduce treatment burden. Other studies using the MTBQ have used similar statistical approaches to dichotomisation using high vs no/low/medium treatment burden levels [24,25].

Descriptive statistics were used and presented as median (interquartile range (IQR)), mean (standard deviation (SD)) and number (%) as appropriate. For PwP, univariable and multivariable binary logistic regression analysis were conducted using the dichotomised MTBQ scores (no/low vs medium/high burden). Variables included in the multivariable logistic regression model were decided a priori based on known associations with treatment burden from previous studies (age, number of medications, number of long-term conditions), those hypothesised to be clinically relevant after discussion within the research team, and variables that achieved p < 0.25 at univariable analysis. Univariable pre-screening is an approach that can be used to determine inclusion of variables into the regression model using a less stringent p-value (p < 0.25 in this study) [26]. This is an established approach to ensure that potentially important confounders are not excluded prematurely, thereby preserving the validity of the final model [26,27].

Due to the smaller sample size, regression analysis was not possible for caregivers. Instead, participant characteristics for caregivers in the ‘no/low’ burden vs. ‘medium/high’ burden groups were compared using the independent t-tests, Mann-Whitney U tests, Pearson Chi-square tests, Fischer’s exact tests, or likelihood ratio tests according to the distributional properties of each variable.

Phase 2: Focus group data collection and analysis

Findings from the national surveys were integrated with results from our previous qualitative studies to identify the key issues of treatment burden and capacity in PD [9,10]. These relate to attending appointments and interactions with healthcare professionals, satisfactory information provision, managing prescriptions and medications, and personal life adaptations. These key issues informed the development of the focus group guide (see S2 File), with the use of open-ended questions to facilitate discussions of potential strategies to reduce the treatment burden for PwP and caregivers. A purposive sample of key stakeholders in the South of England were invited to participate in online focus groups. These included PwP, caregivers and healthcare professionals involved in the care of PD such as PD specialists, PD nurse specialists, general practitioners, psychiatrists, pharmacists, physiotherapists, and occupational therapists. Interested PwP and caregivers from outpatient PD clinics were approached after their clinic appointment, following consent from their clinician, and provided a participant information sheet. Healthcare professionals were recruited from local services, the Parkinson’s UK Wessex regional network and the professional networks of the research team.

All consenting participants received a one-page summary of the identified key issues of treatment burden and capacity in PD prior to the focus group session. Each focus group began with an overview of the study, a reminder of the discussion aims, and agreed ground rules such as confidentiality, listening and respecting others. Three focus groups were conducted on Microsoft Teams between May and July 2022 and recorded with written consent. The discussions were moderated by QYT, a female geriatrician with a specialist interest in PD. Automatic transcription software within Microsoft Teams was used, and transcripts were anonymised and manually corrected by QYT before thematic analysis [28]. QYT took written notes, listened to the recordings, edited transcriptions, and read the transcriptions multiple times to enable immersion in the data. Inductive coding was conducted based on the issues of treatment burden discussed. Mind maps were then used to determine the connections and interlinks between the codes which led to the overall themes of recommendations for improvement. This was then reviewed and refined with the wider research team.

Patient and public involvement

The study involved a patient and public involvement group of a caregiver of someone with PD and a caregiver of someone with dementia and other long-term conditions in the study design, review of protocol and all patient-facing documents. The national surveys and focus group guide were piloted with PwP and caregivers prior to finalisation. This feedback led to revisions to ensure the survey captured all aspects of treatment burden and capacity experienced in PD, reduced survey burden and improved ease of understanding.

Phase 1: Survey findings

The full characteristics of survey participants are presented in Table 1. A total of 162 PwP (143 online, 19 paper) and 30 caregiver (18 online, 12 paper) surveys were completed. Two PwP survey responses were excluded due to >50% missing data on the MTBQ, leaving 160 valid responses. The mean ages of PwP and caregivers were 67.6 ± 9.2 and 68.7 ± 8.9 years, respectively. Among PwP, 52% (N = 83) of respondents were females, whilst 73% (N = 22) of caregivers were females. Most participants were married, cohabiting with a spouse/partner/family member, living in their own home, retired, and had at least GCSE-level education. PwP reported a median disease duration of 5 (IQR 3–8) years, whilst caregivers looked after someone with a median disease duration of 10 (IQR 6–15) years. Using the H&Y staging severity, most PwP (94%) had early to mid stages of PD (stages 1–3), compared to most caregivers (70%) who reported caring for someone with mid to late stages of PD (stages 3–5). PwP reported a median number of 9 (IQR 6–13) non-motor symptoms on the NMSQuest. Most caregivers reported that the person with PD they cared for had problems with mood (73%), memory (73%) and hallucinations (50%) in the last 12 months. Nearly half of PwP had frailty (46%) and multimorbidity (46%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of survey participants.

https://doi.org/10.1371/journal.pone.0338620.t001

High treatment burden was reported by over one-fifth of PwP (21%) and half of caregivers (50%). For PwP, aspects of treatment burden on the MTBQ reported as most difficult (responses ‘a little to extremely difficult’) were making recommended lifestyle changes (51%), remembering how and when to take medication (49%), obtaining clear and up to date information (49%) and arranging appointments with health professionals (48%). For caregivers, the most ‘difficult’ aspects on the MTBQ were adjusting their own lifestyle to look after the person they cared for (69%), making recommended changes to their lifestyle (59%), seeing lots of different health professionals (53%), arranging appointments with health professionals (50%) and getting help from community services (50%). Full data can be seen in S3 File.

Associations with medium/high treatment burden

Univariate analysis found that PwP with higher H&Y stages (vs H&Y Stage 1) (H&Y stage 2 (odds ratio (OR) 1.45; 95% confidence interval (CI) 0.54–3.90); H&Y stage 3 (OR 3.54; 95% CI 1.69–7.42); H&Y stages 4 and 5 (OR 5.08; 95% CI 0.99–25.11), higher NMS scores (OR 1.13; 95% CI 1.06–1.21), frailty (OR 3.08; 95% CI 1.61–5.92), and those who took medications more than three times a day (OR 3.42; 95% CI 1.68–6.95) had significantly (p < 0.05) higher odds ratios of medium/high treatment burden. Multivariate analysis adjusting for age, gender, living property and employment for PwP found that PD severity (vs H&Y stage 1) (H&Y stage 2 (OR 1.56; 95% CI 0.56–14.36); H&Y stage 3 (OR 3.60; 95% CI 1.63–7.499); H&Y stages 4 and 5 (OR 3.93; 95% CI 0.72–21.48)), PD NMSQuest score (OR 1.12; 95% CI 1.04–1.21), frailty (OR 3.12; 95% CI 1.46–6.67), and frequency of medications (>3 times a day) (OR 3.01; 95% CI 1.44–6.30) had significant, independent associations (p < 0.05) with medium/high treatment burden. PwP with multimorbidity had higher odds of medium/high treatment burden (OR 1.45; 95% CI 0.75–2.93) although this was not statistically significant. These results are shown in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Associations of treatment burden in people with Parkinson’s.

https://doi.org/10.1371/journal.pone.0338620.t002

Descriptive analyses comparing those with no/low and medium/high treatment burden for caregivers are shown in Table 3. Caregivers who reported medium/high treatment burden were predominantly female, younger age, more likely to care for someone with PD H&Y stages 4–5, more likely to report memory issues in the person with PD they cared for, had lower mean mental component scores and higher caregiver burden levels compared to caregivers with no/low treatment burden.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Comparison between caregivers of people with Parkinson’s with no/low and medium/high burden.

https://doi.org/10.1371/journal.pone.0338620.t003

Phase 2: Focus group findings

Three online focus groups were conducted with a total of 11 participants, comprising three PwP, one caregiver and seven healthcare professionals (see Table 4). The recommendations of ways to reduce treatment burden and/or increase patient and caregiver capacity were categorised into four main themes which are summarised in Table 5, along with descriptions of how each recommendation may improve specific aspects of treatment burden and capacity in PD.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Focus group participants.

https://doi.org/10.1371/journal.pone.0338620.t004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Focus group themes and suggested recommendations.

https://doi.org/10.1371/journal.pone.0338620.t005

Implications for clinical practice

This study highlights the importance for healthcare professionals involved in PD care to understand the concepts of treatment burden and capacity, recognise their potential impact on health outcomes and address these in clinical settings. The MTBQ be a practical tool to help identify PwP and/or caregivers at risk of high treatment burden, enabling timely and proactive interventions. The brevity of the questionnaire makes it feasible for use in clinical settings, although further validation is needed. Attributes such as frailty, PD severity, number of non-motor symptoms and medication frequency should be routinely assessed in PD reviews. Healthcare professionals should consider using patient-centred questions such as those proposed by Mair and May (“Can you really do what I’m asking you to do?”) [40], and by Linzer et al (“What challenges do you experience in your treatment and self-management?”) to facilitate shared-decision making conversations towards achieving “Minimally Disruptive Medicine”, which contain important principles towards the implementation of change at a system level [12,40–42].

Wider utilisation of valuable existing resources including Parkinson’s UK, PD support groups, and social prescribing link workers can reduce treatment burden by enabling PwP and caregivers to access tailored information, resources and services. Lower levels of self-management adherence are associated with higher treatment burden levels [43,44]. Furthermore, healthcare professionals have an important role in building trust, increasing motivation and empowering individuals to self-manage their health [45]. A study in patients with end-stage renal failure reported differences in self-management practices, which may be modifiable between those with high and low treatment burden levels [46]. Key self-management components in PD include medication management, completing exercises, symptom monitoring, psychological coping strategies, maintaining independence and autonomy, engaging with social networks, and obtaining knowledge and information [47]. Capacity coaching, a strategy shown to be feasible within a primary care setting in the USA, could help healthcare professionals co-create strategies with PwP and caregivers to bolster existing sources of capacity or cultivate new capacity for self-management of PD [48,49].

Policy changes to raise awareness and education about treatment burden and capacity in PD amongst healthcare professionals could lead to meaningful change. For example, the 2016 NICE UK multimorbidity guidelines recommended assessing treatment burden alongside an individualised management plan [50]. Yet, a systematic review in 2022 identified only one multi-centre randomised control trial in the UK that measured the effect of primary care service changes across multiple providers on treatment burden using the MTBQ for patients with multimorbidity [51,52]. Our study provides recommendations for interventions in future research studies that could address treatment burden and capacity not just in PD, but also in other long-term conditions.

Strengths and limitations

The use of the MTBQ to assess the extent and key drivers of treatment burden in PD is an important and novel contribution of this study. However, the MTBQ has not been validated for use amongst caregivers and may be a limitation. This mixed-methods approach enabled the identification of the key drivers of treatment burden through a cross-sectional survey followed by focus groups to develop relevant recommendations for improvement. The survey included a wide sociodemographic range of PwP and caregivers across the UK, with varying lengths of PD diagnosis and severity. Multistakeholder involvement, particularly the inclusion of three PwP and a caregiver in two of the three focus groups alongside healthcare professionals is seen as a strength, allowing discussion from differing perspectives. Regardless, the small number of PwP and caregivers in the focus groups are limitations. The lack of ethnic diversity, the potential for recruitment bias as participants were a self-selected population recruited through Parkinson’s UK who had expressed an active interest in participating in research studies and have access to technology, and the small number of caregiver survey respondents may limit the generalisability of findings. Recruitment from PD clinics with the distribution of paper surveys attempted to reduce this bias. Additionally, self-reported data relies heavily on participant recall, and a review of healthcare records should be considered in future studies. Data collection during the COVID-19 pandemic may have had an impact on healthcare service provision and access for PwP and caregivers in the UK. Challenges recruiting PD nurse specialists and general practitioners are a limitation of the focus groups, but perhaps reflect the broader healthcare service constraints in the UK.