Factors influencing activity performance and participation in usual life situations among people with Parkinson’s disease: A mixed methods study

Abstract

Introduction:

Multiple factors contribute to disability in persons with Parkinson’s disease (PD). Understanding which factors most influence activity performance and participation in usual life situations may help clinicians develop personalized rehabilitation interventions.

Methods:

A triangulation mixed methods study was conducted using 137 people with PD; 13 were purposively sampled for semi-structured interviews. Two regression models examined the contributions of motor and non-motor impairments, activity limitations, personal and health factors to activity of daily living (ADL) performance and participation in usual life situations, respectively. Interviews probed participants’ life experiences with PD.

Results:

Motor and non-motor impairments, activity limitations and PD duration explained 70% of the variance in ADL performance. Non-motor impairments, activity limitations, personal factors and PD medications explained 54% of the variance in participation. Qualitative results revealed themes of “motivation to keep going” and “resilience to live well” in maintaining participation across a diverse range of life situations and highlighted the impacts of PD motor and non-motor impairments.

Discussion:

Complex interactions were revealed among individuals’ unique interests, life situations in which they participated, environments in which tasks were performed, willingness to make adaptations, impairment profile and PD severity. Motor, non-motor and contextual factors should be addressed in personalized intervention plans.

Introduction

Parkinson disease (PD) is a degenerative neurological condition associated with a variety of motor (e.g., bradykinesia, tremor, rigidity, postural instability) and non-motor (e.g., cognitive impairment, depression, anxiety, pain, sleep disturbance, constipation) impairments. At a population level, these impairments contribute to disability and negatively impact quality of life¹, which collectively result in high disease burden and healthcare costs². With PD progression, quality of life worsens³ and healthcare costs increase⁴, likely due to increasing activity limitations and participation restrictions that impact on valued life roles^5,6. Although pharmacological interventions provide some degree of symptomatic control of impairments and may partially improve activity performance and participation in usual life situations, they do not alter the disease course⁷ nor show long-term benefits on quality of life⁸.

At an individual level, activity limitations or severity of impairments do not necessarily predict level of participation in life situations⁹. Results to date also lack clarity around the key drivers of participation restrictions among individuals with PD. For example, some studies have highlighted the importance of non-motor impairments in early PD¹⁰ while other studies emphasized motor impairments¹¹. Furthermore, evidence from other populations indicated that beyond comorbidities and impairments, individuals’ beliefs and motivations are important predictors of participation^12,13. Developing a clearer understanding of what influences participation in individuals living with PD may enable clinicians to better direct their interventions, supports, and coaching/counselling to delay disability. By using a mixed methods approach to triangulate data from multiple sources including objective measures and the individual’s perspective, this study aimed to examine the interplay of factors underpinning activity performance and participation in usual life situations among individuals with mild to moderate PD.

Methods

A concurrent triangulation mixed methods study¹⁴ was conducted involving data from a randomized controlled trial based in the USA (Walking and mHealth to Increase Participation in Parkinson Disease [WHIP-PD]¹⁵, ClinicalTrials.gov registration: NCT03517371), which prescribed a three-month walking and strengthening exercise program to all participants. The quantitative portion of the mixed methods study involved secondary analyses of baseline and three-month WHIP-PD outcome data. For the qualitative component, participants who had recently completed the WHIP-PD trial were invited to participate in an interview. Ethical approval for this study was granted by the Institutional Review Board of Boston University (protocol 4854E). All data were collected prior to concurrent analyses of the quantitative and qualitative data.

Participants provided written consent prior to trial enrolment, and procedures for how baseline and three-month WHIP-PD outcome data were collected have been previously reported¹⁵. Quantitative measures pertinent to this analysis are summarised using the International Classification of Functioning, Disability, and Health¹⁶ (ICF, figure 1), with supplementary table S1 describing the instruments used to measure each predictor variable. The outcomes of interest were activities of daily living (ADL) performance as measured by the Movement Disorder Society sponsored version of the Unified Parkinson Disease Rating Scale (MDS-UPDRS) Part II¹⁷, and ability to participate in usual life situations as measured by the National Institute of Neurologic Disorders and Stroke quality of life (NeuroQOL) instrument¹⁸. Outcomes were obtained from the three-month assessment timepoint, while predictors were obtained from baseline.

Figure 1.

The International Classification of Functioning, Disability and Health framework¹⁶, including the specific variables in each domain that were used in the quantitative analysis. In this study ‘impairment’ refers to alterations in structure and/or functioning of body systems, ‘activity’ to specific task/s an individual executes, and ‘participation’ to the individual’s involvement in usual life situations, including their quality of life.

For the qualitative component, a purposive sample of participants who had recently (within the past three months) completed the WHIP-PD trial were invited to participate in one semi-structured interview, 45–60 minutes in duration (see supplementary material S2 for the interview prompts). The interview specifically probed domains of the ICF framework not available in the quantitative data (e.g., environment). Participants were purposively sampled across geography (i.e., both sites: New England and Midwest, USA; and urban versus less urban settings), sex (male and female), PD stages (HY 2 versus 2.5–3), employment (working and retired), and attitude toward exercise (Multidimensional Outcome Expectations for Exercise Scale¹⁹); ensuring at least two participants per characteristic were recruited for a diversity of views. While our initial intention was to invite only participants who had completed the WHIP-PD trial within the past three months, the recruitment window was extended due to the small number of participants sampled who were still working; thus, one participant was interviewed five months after finishing the trial.

A member of the research team (SP) conducted the interviews using Zoom videoconferencing software or by phone. The interviewer was an Australian physical therapist with clinical and research expertise in PD who had no prior relationship with the participants. Oral consent was obtained at interview commencement. At this time, the interviewer disclosed her background, including her lack of familiarity with the healthcare setting or environment participants engaged with, thus encouraging participants to share openly. The interviewer took field notes during the interview and made reflexive notes afterward in a log. Interviews were audio recorded then transcribed verbatim by a member of the research team or a research assistant. De-identified transcripts were offered to participants for member checking; three participants chose to check their transcripts with two making minor edits to improve wording clarity. After every 2–4 interviews, the interviewer (SP) debriefed with members of the research team (TE, JC and FP, American physical therapists with clinical and research expertise in PD) to review purposive sampling, emergent themes, refine the interview prompts and determine thematic saturation. The reflexive log was referred to in these meetings and the different perspectives of the research team were discussed.

Statistical analysis

For the quantitative component, potential predictors were examined for associations with each outcome (i.e., ADL performance as measured by the MDS-UPDRS Part II¹⁷, and ability to participate in usual life situations as measured by the NeuroQOL instrument¹⁸) using univariate linear regression. The predictors most strongly associated with each outcome (p<.20) were entered into separate multivariable models, ensuring at least 10 cases per predictor model to avoid overfitting models²⁰. Predictors that were collinear (i.e., r>.70) were not entered into the same model; instead, the predictor with the strongest relationship with the outcome was selected for entry. Two multivariable models were constructed: one for each outcome. The data distribution of all models satisfied the assumptions for regression analyses. Data were analysed using SAS Enterprise Guide v7.1 (SAS Institute, Cary NJ, USA).

For the qualitative component, inductive thematic analysis²¹ was undertaken using a social constructivist approach, whereby participants constructed meaning based on their individual experiences and contexts²². A researcher (SP) firstly familiarised herself with the data by listening to the interviews while simultaneously reading the transcripts and referring back to the field notes and reflexive log. Transcripts were then uploaded into NVivo v14 (QSR International, Denver CO, USA) and openly coded line by line by the researcher (SP), who generated a codebook during this phase of coding. Two coding meetings were held between researchers with qualitative experience (SP and SD, a British-Australian allied health researcher experienced in health services, behaviour change and qualitative research) to refine the codes, revise if necessary, and update the codebook. Visual maps were drawn in these meeting to group the codes into themes. Thematic saturation was determined when there was no change to the codebook with subsequent interview data. Saturation was apparent after 10 interviews; however, we continued to sample to ensure there were at least two participants per characteristic, this was achieved after 13 interviews. Hence, we included data from all 13 interviews.

Quantitative and qualitative analyses were undertaken in parallel. In the final stages of analysis, results from both components were integrated to form a single overall interpretation of the findings. Findings from the qualitative component prompted sensitivity analyses to examine the inter-relationship between ADL performance and participation, whereby ADL performance at baseline was omitted as a covariate in the regression model examining participation, and participation at baseline was omitted as a covariate in the model examining ADL performance.

Results

The quantitative analysis included 137 participants at baseline (table 1). Of these individuals, 117 (85%) provided three-month follow-up data. For the qualitative component, 15 participants were invited and 13 consented to be interviewed (table 2).

Table 1.

Characteristics of WHIP-PD trial participants at baseline (n=137). Data reported as mean (SD) or median (Q1-Q3), range; or n (%).

Characteristic				Value
Health conditions
HY stage†	2			67 (49%)
	3			70 (51%)
Levodopa equivalent daily dose (mg)†				500 (300 – 900), 0 – 2700
PD duration (years)1				4 (2 – 8), 0 – 25
Modified Charlson Index with musculoskeletal comorbidities as a group (n, max 23)£				1 (0 – 1), 0 – 4
		0		44 (32%)
		1–2		81 (59%)
		3–5		12 (9%)
Number of musculoskeletal comorbidities				0 (0 – 1), 0 – 4
		0		72 (53%)
		1–2		52 (38%)
		3–5		13 (9%)
Body structures and functions (i.e., impairments)
Motor impairments
MDS-UPDRS-III (0 – 132)†, 1				36.9 (11.7), 8 – 64
Freezers				39 (28%)
FOG severity (NFOGQ, 0–28)†				0 (0 – 5), 0 – 20
Non-motor impairments
NeuroQOL fatigue (8 – 40)‡, †				19 (14 – 25), 8 – 39
NeuroQOL sleep disturbance (8 – 40)‡, †				14 (11 – 18), 8 – 28
NeuroQOL anxiety (8 – 40)‡, †				14 (9 – 19), 8 – 36
NeuroQOL depression (8 – 40)‡, †				10 (8 – 15), 8 – 27
Apathy rating scale (0 – 42)†, 1				10 (7 – 14), 0 – 24
NeuroQOL positive affect and well-being (9 – 45)‡				36 (33 – 40), 19 – 45
NeuroQOL emotional and behavior dyscontrol (8 – 40‡, †				12 (9 – 15), 8 – 26
Executive function (Trails B - Trails A; s)‡, 3				49.4 (29.3 – 78.2), -2.5 – 270.4
Dimensional Card Sort Test Computed score (0–10)3				7.6 (6.9 – 8.2), 2.5 – 9.6
Kings Parkinson's Pain Scale (0 – 168)†, 1				10 (5 – 16), 0 – 56
	Presence of pain			127 (93%)
Presence of disability due to low back pain§				89 (65%)
Modified Oswestry low back pain disability score (%)				8 (0 – 20), 0 – 50
Activity
MDS-UPDRS-II (0 – 52)†				9 (6 – 14), 0 – 33
NeuroQOL upper extremity function (8 – 40)‡				38 (36 – 40), 26 – 40
NeuroQOL communication (5 – 25)‡				23 (19 – 25), 9 – 25
Walking aid use1,¥				38 (28%)
Comfortable gait speed (m/s)				1.16 (0.20), 0.56 – 1.59
6-minute walk test (m)				447.71 (98.00); 126.05 – 698.74
MiniBEST score (0–28)				19 (17 – 21), 3 – 26
Participation
NeuroQOL ability to participate (8 – 40)‡, 1				37 (32 – 40), 13 – 40
Personal factors
Age (years)				67.7 (8.2), 46 – 90
Sex	F			65 (47%)
	M			72 (53%)
Living situation1		Alone		20 (15%)
		With family		115 (85%)
		Assisted living		1 (1%)
12-month fall history1		Non-faller		77 (57%)
		Single faller		27 (20%)
		Recurrent (2+) faller		32 (24%)
Fear of falls impacts ADL performance1			Never	75 (55%)
			Monthly	46 (34%)
			Weekly/always	15 (11%)
NeuroQOL stigma (8 – 40)‡, †				10 (8 – 13), 8 – 28

FOG: freezing of gait; HY: Hoehn and Yahr; MDS-UPDRS: Movement Disorder Society sponsored version of the Unified Parkinson Disease Rating Scale; MiniBEST: Mini Balance Evaluation Systems Test; NeuroQOL: National Institute of Neurologic Disorders and Stroke quality of life; NFOGQ: New Freezing of Gait Questionnaire; PD: Parkinson’s disease.

Superscript numerals indicate number of participants with missing data.

^†Higher values indicate worse function.

^£The original Charlson comorbidity index (1984) included 18 conditions; the updated version (2008) includes 22 conditions. The updated version was modified to include musculoskeletal comorbidities as a twenty-third category.

^‡Raw NeuroQOL scores are reported to better enable interpretation.

^§Where the Modified Oswestry low back pain disability score was >0.

^¥includes single point stick, quad stick, walking poles, crutches.

Table 2.

Characteristics of WHIP-PD trial participants who were interviewed 3–6 months following trial completion (n=13).

Participant	Age group (years)	Sex	PD stage (HY)	PD duration (years)	Employment	Living situation	Geography	Region	MOEES (0 – 100)¥	Community access frequency (days per week)	Interval to interview (days)
1	65 – 69	F	2.5	10 – 14	Retired	With spouse	Suburban	NE	67	7	63
2	70 – 74	F	2.5	5 – 9	Retired	With spouse	Suburban	NE	52	7	86
3	65 – 69	M	2	0 – 4	Retired	With spouse	Semi-rural	NE	62	7	80
4	80 – 84	M	2.5	5 – 9	Retired	With spouse	Suburban	MW	57	7	88
5	65 – 69	F	2.5	0 – 4	Retired	Alone	Urban	MW	47	5	5
6	70 – 74	F	2	0 – 4	Retired	With spouse	Semi-rural	NE	66	5	53
7	65 – 69	F	2.5	0 – 4	Retired§	With spouse	Suburban	NE	73	7	77
8	85 – 89	M	2.5	5 – 9	Retired	With spouse	Suburban	MW	56	7	120
9	55 – 59	M	2	0 – 4	Works FT	With spouse	Urban	MW	67	7	119
10	65 – 69	F	2.5	0 – 4	Retired	With spouse	Urban	MW	52	3	85
11	70 – 74	M	2	5 – 9	Retired§	Alone	Suburban	MW	50	7	15
12	55 – 59	F	2	0 – 4	Retired§	With family	Suburban	NE	60	7	95
13	60 – 64	M	2	0 – 4	Works FT	With spouse	Semi-rural	NE	65	7	152
Summary ^	68.5 (8.0), 55 – 85	M: 6 (46%), F: 7 (54%)	2: 6 (46%), 2.5: 7 (54%)	4.5 (2.8), 1.5 – 11.0	Full-time: 2 (15%), Retired: 11 (85%)	Alone: 2 (15%), With spouse: 10 (77%), With family: 1 (8%)	Urban: 3 (23%), Suburban: 7 (54%), Semi-rural: 3 (23%)	MW: 6 (46%), NE: 7 (54%)	59.5 (7.6), 47 – 73	6.4 (1.2), 3 – 7	80 (39), 5 – 152

FT: full-time; HY: Hoehn and Yahr; MOEES: Multidimensional Outcome Expectations for Exercise Scale; MW: Midwest; NE: New England; WHIP-PD: Walking and mHealth to Increase Participation in Parkinson Disease trial.

^§Chose to work part-time following retirement.

^¥The MOEES score reported is from the 12-month assessment of the WHIP-PD trial. Higher scores indicate higher exercise self-efficacy.

^{^}Summary values are presented as mean (SD), range; or n (%).

Quantitative models

Predictors of ADL performance at three-months that met the univariate regression criterion for inclusion in the multivariable model were health condition (PD duration, PD stage, PD medications, comorbidities), motor impairments (MDS-UPDRS-III, freezing of gait), multiple non-motor impairments, activity limitations (communication, upper extremity function, balance) and personal factors (stigma, sex, fear of falls) (table 3). A multivariable model including motor and non-motor impairments, activity limitations and PD duration explained 70% of the variance in ADL performance (p<.001, table 4), with PD duration, motor impairments (freezing of gait, MDS-UPDRS-III), apathy, upper extremity function and communication continuing to make significant independent contributions.

Table 3.

Predictors of ADL performance and participation at 3-months from univariate linear regression.

	ADL performance (MDS-UPDRS-II)		Participation (NeuroQOL ability to participate)
Predictor	r2	p-value	r2	p-value
Health conditions
PD severity (HY II vs III)	.038	.04	.018	.15
Levodopa equivalent daily dose (mg)†	.066	.006	.094	<.001
PD duration (years)	.159	<.001	.068	.005
Musculoskeletal comorbidities (n)	.020	.13	.061	.008
Updated Charlson Index comorbidities (n)£	.052	.01	.044	.02
Impairments
MDS-UPDRS-III (0 – 132)†	.196	<.001	.059	.01
FOG severity (NFOGQ, 0–28)†	.190	<.001	.009	.33
NeuroQOL fatigue (8 – 40)†	.157	<.001	.223	<.001
NeuroQOL sleep disturbance (8 – 40)†	.085	.002	.125	<.001
NeuroQOL anxiety (8 – 40)†, a, b	.089	.001	.206	<.001
NeuroQOL depression (8 – 40)†, a, c, d	.107	<.001	.209	<.001
Apathy rating scale (0 – 42)†	.118	<.001	.182	<.001
NeuroQOL affect (9 – 45)c	.084	.002	.215	<.001
NeuroQOL emotional disturbance (8 – 40)†, b, d	.113	<.001	.199	<.001
Executive function (Trails B – Trails A; s)†	.057	.01	.015	.20
Dimensional Change Card Sort Test (computed score, 0–10)	.014	.23	.010	.30
Kings Parkinson’s Pain Scale (0 – 168)†	.009	.32	.044	.02
Modified Oswestry low back pain disability score (%)	.123	<.001	.177	<.001
Activity
MDS-UPDRS-II (0 – 52)†	--	--	.265	<.001
NeuroQOL upper extremity function (8 – 40)	.361	<.001	.070	.004
NeuroQOL communication (5 – 25)	.413	<.001	.162	<.001
Uses a walking aid	.081	.002	.031	.06
Comfortable 6m gait speed (m/s)e	.089	.001	.051	.02
6-minute walk test (m)e	.075	.003	.027	.08
MiniBEST score (0–28)	.124	<.001	.031	.06
Participation
NeuroQOL ability to participate (8 – 40)	.161	<.001	--	--
Personal factors
Age	.030	.07	.000	.86
Sex (M)	.106	<.001	.000	.84
Education (completed vs did not complete tertiary education)	.002	.67	.004	.53
Employment (retired, employed, disability, unemployed)	.067	.06	.049	.14
12-month fall history (none, single, recurrent)	.025	.25	.073	.01
Fear of falls impacts ADLs (no, little, sometimes/always)	.076	.03	.123	.001
NeuroQOL stigma (8 – 40)†	.182	<.001	.223	<.001

FOG: freezing of gait; MDS-UPDRS: Movement Disorder Society sponsored version of the Unified Parkinson Disease Rating Scale; MiniBEST: Mini Balance Evaluation Systems Test; NeuroQOL: National Institute of Neurologic Disorders and Stroke quality of life; NFOGQ: New Freezing of Gait Questionnaire; PD: Parkinson’s disease

Superscript letters indicate collinear variables (r≥.70).

^†Higher values indicates worse function.

^£The original Charlson comorbidity index (1984) included 18 conditions; the updated version (2008) includes 22 conditions. The updated version was modified to include musculoskeletal comorbidities as a twenty-third category.

Table 4.

Strong predictors of ADL performance and participation at 3-months from multivariable linear regression models.

Predictor		Estimate (95% CI)	Standardised β	p-value
Multivariable model of factors influencing ADL performance (n=110, r2 = 0.70, p<.001)
PD duration (years)		0.26 (0.09 to 0.44)	0.19	.003
MDS-UPDRS-III (0–132)†		0.10 (0.02 to 0.17)	0.18	.01
FOG severity (NFOGQ, 0–28)†		0.25 (0.13 to 0.38)	0.24	<.001
NeuroQOL fatigue (8–40)†		0.11 (−0.03 to 0.25)	0.12	.13
Apathy rating scale (0 – 42)†		0.19 (0.03 to 0.34)	0.15	.02
Oswestry LBP disability score (%)		0.06 (−0.01 to 0.13)	0.12	.07
NeuroQOL UE function (8–40)		−0.59 (−0.85 to −0.32)	−0.32	<.001
NeuroQOL communication (5–25)		−0.36 (−0.64 to −0.08)	−0.20	.01
MiniBEST score (0–28)		−0.03 (−0.25 to 0.18)	−0.02	.76
NeuroQOL ability to participate (8–40)		−0.03 (−0.20 to 0.15)	−0.02	.77
NeuroQOL stigma (8–40)†		−0.15 (−0.43 to 0.13)	−0.08	.29
Multivariable model of factors influencing participation (n=114, r2 = 0.54, p<.001)
Levodopa equivalent daily dose (mg)†		−0.001 (−0.003 to −0.000)	−0.15	.03
NeuroQOL fatigue (8 – 40)†		−0.11 (−0.28 to 0.07)	−0.13	.23
NeuroQOL sleep disturbance (8 – 40)†		0.14 (−0.11 to 0.39)	0.12	.27
NeuroQOL anxiety (8 – 40)†, a, b		−0.08 (−0.26 to 0.09)	−0.10	.33
Apathy rating scale (0 – 42)†		−0.21 (−0.40 to −0.01)	−0.19	.04
NeuroQOL affect (9 – 45)c		0.06 (−0.14 to 0.26)	0.07	.54
Oswestry LBP disability score (%)		−0.07 (−0.14 to 0.00)	−0.15	.06
MDS-UPDRS-II (0 – 52)†		−0.21 (−0.38 to −0.04)	−0.24	.02
NeuroQOL communication (5–25)		0.07 (−0.22 to 0.35)	0.04	.65
Fear of falls^	Little	−2.34 (−4.09 to −0.59)	−0.20	.009
	Sometimes/always	−2.38 (−4.97 to 0.20)	−0.14	.07
NeuroQOL stigma (8 – 40)†		−0.20 (−0.50 to 0.10)	−0.12	.18

FOG: freezing of gait, LBP: low back pain, MDS-UPDRS: Movement Society sponsored version of the Unified Parkinson Disease Rating Scale, MiniBEST: Mini Balance Evaluation Systems Test, NeuroQOL: National Institute of Neurologic Disorders and Stroke quality of life instrument, PD: Parkinson’s disease; UE: upper extremity.

Bolded values indicate variables that maintained an independent contribution to the outcome (p<.05).

^†Higher values indicates worse function.

^{^}Fear of falls: reference level = none.

Predictors of participation in usual life situations at three-months that met the univariate regression criterion for inclusion in the multivariable model were health condition (PD medications, PD duration, comorbidities), motor impairment severity (i.e., MDS-UPDRS-III), multiple non-motor impairments, activity limitations (ADL performance [i.e., MDS-UPDRS-II], communication, upper extremity function, gait speed) and personal factors (stigma, fear of falls, fall history) (table 3). A multivariable model including PD medications, non-motor impairments, activity limitations (ADL performance, communication), and personal factors (fear of falling, stigma) explained 54% of the variance in participation (p<.001, table 4), with PD medications, apathy, ADL performance and fear of falling continuing to make significant independent contributions.

Sensitivity analyses revealed no change in the proportion of variance explained in ADL performance when participation was omitted from the multivariable model, and a 2% reduction in the variance explained in participation when ADL performance was omitted from that model (supplementary table S3). Although ADL performance made an independent contribution to participation in multivariable analyses, when this variable was omitted from the model low back pain disability and communication instead made independent contributions, reiterating the combined contributions of the health condition, non-motor impairments, activity limitations and personal factors to participation.

Qualitative results

Two themes relating to participation were identified from the interviews: motivation to keep going, and resilience to live well.

Motivation to keep going

All interviewed participants were highly motivated to maintain participation. Participants did not distinguish between participation and activity performance; rather, they outlined participation across a variety of essential activities, leisure activities, social activities and work. They described the leisure, social and even work activities as being of interest to them, and therefore they were enthusiastic and motivated to participate. In contrast, essential activities (e.g., chores such as grocery shopping, attending medical appointments) were often described as mundane, though caring responsibilities were a priority: “I’m still able to support them [children with disability]… that trip to the doctor, or…needs to have somebody go along to interpret, something like that” (P10, F, HY 2.5).

Participants engaged in a range of leisure and social activities, from physically active (e.g., playing sports, hiking, gardening): “I also play golf, and I ski and I want to keep being able to do that forever” (P06, F, HY 2); to sedentary (e.g., reading, doing word puzzles): “Now my pastime … To keep my cognitive skills going, I play Bridge and Gin Rummy. And I found two games… I can play them on my phone or on my computer” (P04, M, HY 2.5).

Activities often reflected participants’ personalities and interests: “Oh, I love being creative. I love working with my hands, working with my head, seeing things. I love the decorating, and I’m a knitter” (P05, F, HY 2.5). Some activities involved participants going out into their community, while others were performed at home.

For the most part, participants were inherently motivated to participate in the various activities and roles in which they engaged (i.e., internal motivation): “Because even if you’re slow, you can still do your stuff. It just takes longer. It’s not a big deal” (P01, F, HY 2.5). However, some participants described using external motivators to assist them with participation: “I got started by doing a thing for the trip…But I realize now how important it is to continue to stay on” (P11, M, HY 2). Some participants highlighted the need for external motivators to maintain physical activity: “Left to my own devices I tend to get more lazy” (P03, M, HY 2).

All participants reported that PD impacted their ability to participate. Participants highlighted the contributions of motor impairments: “I have a lot of trouble with balance… I think it’s my rigidity and my balance” (P12, F, HY 2); non-motor impairments: “I’m really struggling with the apathy…and anxiety” (P05, F, HY 2.5); or both (e.g., muscle weakness and fatigue): “I would like to do more, but my legs get really, sort of dead weight easily. So it’s hard for me to do, oh you know, consecutive days for example…So I have to, I tend to do things every other day…and sometimes every third” (P08, M, HY 2.5). The presence of other comorbidities also negatively impacted participation, regardless of whether it was a chronic disease: “I’m not sure it [osteoporosis] affects how I go about my day…Or if it does, I’m so used to it I don’t notice it. It certainly affects what I’m willing to take on by my challenge in terms of balance” (P02, F, HY 2.5); or acute injury: “I’ve had some injuries over the past six to eight months which have prevented me from exercising to a level that I would want to” (P13, M, HY 2).

Resilience to live well

This theme was encapsulated by the following model: when PD symptoms (and other comorbidities) affected participants’ capability to meet the demands of the task, they found workarounds to maintain participation. To accommodate their altered capability, participants modified their behaviour : “Like maybe I would have a slice of pizza instead of a bowl of spaghetti when I was out because it’s easier to eat a piece of pizza in front of people than it is to manage twirling on a fork and that type of stuff… but it just makes you more aware of your situation and causes you to make little adjustments like that” (P09, M, HY 2). Others modified the demands of the activity and/or environment in addition to modifying behaviour: “You know, we get there [the symphony] relatively early. And we stay till the crowd is gone…and where we park the police direct traffic. So it’s easy to walk across the street to the car or parking lot” (P08, M, HY 2.5). Sometimes the workaround involved social support: “My husband takes on a fair amount of what feels to me like more than his share of responsibilities, so that my slowness doesn’t prevent things from happening entirely…I would probably be a lot more isolated without him” (P02, F, HY 2.5).

Many participants highlighted the need for social support when driving was required to access the community: “There’s some limits I started putting on it [driving] …if it’s not difficult for them I ask the other person to pick me up” (P10, F, HY 2.5). Social support was particularly pronounced when driving occurred under challenging conditions, such as inclement weather or when visibility was low: “I’m not as willing or happy to drive at night as I used to be […] although my husband will, because we’re gonna meet up with [friends]…But if he couldn’t do that, I would do a Saturday or a Sunday [brunch], but I don’t like to drive at night” (P07, F, HY 2.5).

Participants’ workarounds were varied and reflected adaptations in accordance with their PD stage and the severity of their motor and/or non-motor impairments. Those with reduced mobility and balance often reported using mobility aids to maintain participation: “And I can do better actually in the stores because I can hold on to the cart and so that really helps with balance” (P10, F, HY 2.5); particularly in community environments: “Like we’re going to a museum. We’ll take a wheelchair” (P01, F, HY 2.5). However, some participants with balance and mobility problems were also fearful of falling and used strategies to mitigate this risk: “I always park in the handicap. I’m not walking it, I’m not going to take a chance and have to walk half a mile across the parking lot when I can park up front” (P04, M, HY 2.5). For some, experiencing falls resulted in curtailment of activity: “I was playing pickleball, but I fell and I hurt my knee. So I kind of curtailed that activity, and I decided that I probably shouldn’t be doing that… but I liked playing. It was fun” (P12, F, HY 2).

Participants cycled through the process of finding workarounds until they could no longer find a suitable alternative, at which point they gave up participating in that activity: “Other than not being able to continue with the flying…You have to take a medical exam, and as soon as I said Parkinson’s, that was kind of a deal killer. And it’s not because of my physical health, but it’s just the symptoms that come along with it” (P03, M, HY 2). Some participants described less satisfaction with the workarounds than the original activity: “I have always loved making miniatures and doll houses... And now I’m retired and I have plenty of time, but I shake and I can’t, it’s just so frustrating… The thing I spend most of my day doing is you know, what every other person does, and that’s on my cell phone. I do word games a lot… And I figure well, at least it’s something that’s good for me.” (P10, F, HY 2.5).

Discussion

This triangulation mixed methods study demonstrated that both motor and non-motor impairments along with other comorbidities affected ADL performance and participation in usual life situations in people with mild to moderate PD. The quantitative models demonstrating bidirectional relationships between ADL performance and participation, together with the lack of differentiation between activity performance and participation by interviewed participants²³, reinforced the conceptual bidirectionality of relationships within the ICF model¹⁶ (see figure 1). The results of sensitivity analyses further confirmed these findings, with a variety of physical activities, communication, non-motor and motor impairment contributors identified.

Despite the preponderance of literature describing the negative impacts of motor impairments^11,24 or non-motor impairments^10,25–27 alone on participation and quality of life, our findings highlighted that both motor and non-motor impairments impact activity and participation in people with PD. On an individual level, participation resulted from interactions among the nature of chosen life activities, environment in which tasks were performed, predominant and/or troubling impairments and disease stage. The quantitative analysis of participation, where non-motor impairments (including fatigue, mood and pain) formed the majority of predictors entered into the multivariable model, may reflect the lack of clinical scales examining individual motor impairments, even though many of our interviewed participants described impacts of specific motor impairments (particularly bradykinesia and tremor¹¹) on their activity and participation.

Participants reported that balance and mobility problems²⁸, falls⁶ and fear of falling^6,29 also substantially impacted participation, with fear of falling similarly found to be an independent predictor of participation. However, neither mobility nor balance were found to be significant contributors to ADL performance or participation in quantitative models, which could be due to the sample comprising participants of a walking intervention trial, many with reasonable mobility. Consistent with prior research, participants also outlined changes in mobility arising from PD²⁹, with the interaction between mobility limitations and the physical environment presenting challenges to participation³⁰. This result highlights the ongoing need for inclusive environments and for clinicians such as physical therapists to ensure that interventions delivered to people with PD address not only the individual’s mobility challenges but also the contexts in which these tasks are performed³¹.

Unsurprisingly, quantitative models confirmed the contribution of PD progression to ADL performance and participation, as indicated by longer PD duration²⁶ and greater LEDD dose^32,33, respectively. Taken together with interview results, in which participants with more long-standing disease often described more severe impairments and detailed workarounds to maintain participation, these findings highlight the critical need for rehabilitation clinicians to support people with PD to develop workaround strategies to maintain participation. Such a proactive approach to delivering common rehabilitation interventions^34,35 would minimise progression of disability and enable people with PD to maintain participation and improve quality of life. Sadly, most people with PD are only referred to rehabilitation in later stages of disease in response to symptom deterioration^36,37. Our findings reiterate calls by others for proactive input during early stages of disease from a multidisciplinary team of healthcare professionals with relevant expertise across the various dimensions of health¹⁶ to optimise participation outcomes for people with PD³⁸, which should be trialled in future studies.

We acknowledge some limitations of the current study. The sample comprised individuals who participated in a year-long walking intervention study, who may represent the more motivated portion of the overall population of people with PD and therefore affect the generalisability of the findings. The correlational study design also did not allow causal inferences to be made. The overall low severity of impairments in the sample may have obscured the contribution of some non-motor impairments such as cognitive impairment and pain to participation, despite prior research describing negative effects of these impairments^29,39,40. Nonetheless, the impact of other non-motor impairments (e.g., apathy) on participation aligned with prior findings^6,26. Although communication was found to impact participation^41,42 in the quantitative models, all interviewed participants had good communication skills, and the interviewer adjusted her interview style to accommodate individual participants⁴¹, which may have limited our ability to further probe this issue.

Another limitation was that all included data were obtained in the USA, so cultural and environmental predilections may not apply to other settings. For example, the USA is particularly reliant on driving with relatively few people having easy access to public transport⁴³. However, even in countries where public transport is more readily available, people living with chronic disease including PD are still reliant on private means of transport for community access⁴⁴, with changes to driving ability impacting participation⁴⁵. And while prior studies described the need for driving assistance and sometimes family and friends’ attitude as a barrier to participation among people with PD³⁰, in line with recent evidence⁴⁶ our sample viewed such assistance as necessary and appreciated this workaround to enable ongoing participation. The strength of this study was the mixed methods approach that triangulated quantitative findings from a large sample with interview data that elucidated the perspectives of people living with PD.

In summary, this triangulation mixed methods study clarified that non-motor, motor, and contextual factors appear to most strongly influence activity performance and participation among people with mild to moderate PD. Activities of interest and participation are diverse across individuals, and both are impacted by the nature of chosen activities and the individual’s non-motor and motor impairments. Although included participants were highly motivated to maintain participation, this may differ across the population of people with PD and highlights the need for early intervention and/or education by healthcare providers to ensure that participation in usual life situations is optimised from the outset of disease. Future work should elucidate models of care with the appropriate mix of disciplines to maximise participation across the spectrum of PD severity.