Effectiveness of a dual-task training program for fall prevention and executive functions in individuals wirk Parkinson’s disease: A quasi-experimental study

Abstract

Introduction

Parkinson’s disease (PD) is a progressive neurodegenerative disorder that affects the dopaminergic system, causing motor symptoms such as bradykinesia, rigidity, and resting tremors, as well as cognitive impairments and balance disturbances.

Objective

To evaluate the clinical effect of a dual-task training program on executive functions and fall prevention in individuals with PD.

Methods

This was a quasi-experimental study with non-random sampling. The initial sample consisted of 21 individuals with Parkinson’s disease in Hoehn and Yahr stages I–III, all with independent gait. Participants with moderate or severe cognitive impairment (MoCA), disabling musculoskeletal conditions, or other neurological disorders were excluded. A 12-week sensorimotor training program was implemented, consisting of 60-minute sessions held once per week. Pre- and post-intervention assessments were conducted to measure changes in executive functions and fall risk. Global cognitive function was assessed using the Montreal Cognitive Assessment (MoCA), frontal executive function with the Frontal Assessment Battery (FAB), and functional mobility and fall risk using the Tinetti scale and the Timed Up & Go (TUG) test.

Data were organized in Excel spreadsheets and analyzed using Python. The study was approved by the Scientific Ethics Committee of the University de Los Lagos (No. 045/2024, 08/12/2024).

Results

Of the 21 initial participants, 17 completed the 12-week sensorimotor program. Significant improvements were observed in gait items, increasing from 10 to 11 points (SD = 2) (p = 0.01), and in the total Tinetti score, which increased from 25 to 27 (p = 0.02). No significant changes were found in global cognitive function (MoCA) — Pre: 18.82 (SD = 4.23) and Post: 19.53 (SD = 4.19) — nor in specific executive functions — Pre: 13.41 (SD = 3.13) and Post: 12.71 (SD = 2.84) (p = 0.33).

Conclusions

No significant differences were observed between pre- and post-intervention measurements in MoCA, FAB, the balance subscale of the Tinetti Test, or the TUG. However, dual-task training showed potential for improving gait parameters in individuals with PD. Its implementation in community settings could expand access to effective and sustainable rehabilitation interventions.

Resumen

Introducción

La enfermedad de Parkinson (EP) es una enfermedad neurodegenerativa progresiva que afecta el sistema dopaminérgico, provocando síntomas motores como bradicinesia, rigidez y temblores en reposo, además de alteraciones cognitivas y del equilibrio.

Objetivo

Evaluar el efecto clínico de un programa de entrenamiento dual sobre las funciones ejecutivas y prevención de caídas en personas con EP.

Método

Estudio cuasiexperimental con muestreo no aleatorio. La muestra inicial estuvo compuesta por 21 personas con enfermedad de Parkinson en estadios I–III de Hoehn y Yahr, con marcha independiente. Se excluyeron participantes con deterioro cognitivo moderado o severo (MoCA), enfermedades musculoesqueléticas incapacitantes u otros trastornos neurológicos.

Se implementó un programa sensoriomotor de 12 semanas, con sesiones de 60 minutos una vez por semana. Se realizaron evaluaciones pre y postintervención para medir cambios en funciones ejecutivas y riesgo de caídas. La función cognitiva global se evaluó con la Montreal Cognitive Assessment (MoCA), la función ejecutiva frontal con la Frontal Assessment Battery (FAB), y la movilidad funcional y riesgo de caídas mediante escala de Tinetti y Timed Up & Go (TUG).

Los datos fueron organizados en planillas Excel y analizados con Python. El estudio fue aprobado por el Comité de Ética Científica de la Universidad de Los Lagos (n.º 045/2024, 12/08/2024).

Resultados

De los 21 participantes iniciales, 17 completaron el programa sensoriomotor de 12 semanas. Hubo mejoras significativas en ítems de marcha, con aumento de 10 a 11 puntos (DE = 2) (p = 0,01) y en la puntuación total de Tinetti, de 25 a 27 (p = 0,02). No se encontraron cambios significativos en función cognitiva global (MoCA) Pre: 18,82 (DE = 4,23) y Post: 19.53 (DE = 4,19), ni en funciones ejecutivas específicas Pre: 13,41 (DE = 3,13) y Post: 12,71 (DE = 2,84) (p = 0,33).

Conclusiones

No se observaron diferencias significativas entre las mediciones pre y postintervención en MoCA, FAB, la subescala de equilibrio del Test de Tinetti ni en el TUG. Sin embargo, el entrenamiento dual mostró potencial para optimizar parámetros de la marcha en personas con EP. Su implementación en contextos comunitarios podría ampliar el acceso a intervenciones de rehabilitación eficaces y sostenibles.

Introduction

Parkinson's disease (PD) is one of the most common neurodegenerative diseases¹ and in Latin America, Chile stands out for a 16.5% increase in deaths attributed to this disease and a 19.9% increase in its prevalence,² highlighting the need to advance public health strategies. PD is characterised by the progressive loss of dopaminergic neurons in the substantia nigra, resulting in motor impairments such as bradykinesia, rigidity, and tremors at rest, as well as non-motor symptoms including cognitive dysfunction and balance disorders.¹ This combination of symptoms increases the risk of falls and impairs quality of life, making it essential to implement integrated interventions that address motor and cognitive deficits.³

In this context, dual-task training, which combines physical exercises with cognitive tasks, has emerged as a promising strategy for improving executive function and reducing the risk of falls in people with PD.⁴ Studies conducted in controlled clinical settings have shown improvements in postural stability, gait speed and variability, coordination, and balance,^5,6 as well as benefits in executive processing, working memory, and attention.⁷

Recent meta-analyses confirm that dual-task training produces significant improvements in spatiotemporal gait parameters, balance, functional mobility, and certain cognitive domains under dual-task conditions, although they highlight the heterogeneity of protocols and the need for standardisation.^8–10

However, most of this research has been conducted under experimental conditions that minimise the influence of external factors, with samples selected using strict inclusion criteria to favour homogeneity and control of variables.¹¹ In these settings, sessions are usually supervised by health professionals in specialised facilities with access to advanced monitoring equipment.¹² While this design has been fundamental in establishing the effectiveness of dual-task training, its applicability in community settings is unclear.^5,13

Transferring these interventions to community settings can present challenges, including variability in implementation, reduced professional supervision, difficulties in adherence, and greater heterogeneity in the PD population in terms of disease severity and comorbidities.^14–16 These factors can affect the feasibility and outcomes of dual-task training outside of a clinical setting. However, developing effective community programmes could provide accessible, sustainable alternatives that have the potential to benefit a greater number of patients.¹⁷

Consequently, it is necessary to evaluate not only the clinical effects of dual-task training in less controlled settings, but also its feasibility and adherence.¹⁸ The present study aims to examine the clinical effect of a dual-task training programme on executive functions and fall prevention in people with PD in a community setting.

Methods

Study design

Pilot study with a quasi-experimental design and pre- and post-intervention assessments (12 weeks). This project was reviewed and approved by the accredited scientific ethics committee of the University of Los Lagos in Chile, and all participants signed an informed consent form prior to the assessment.

Population and sample

Participants were selected non-randomly, recruiting 21 people diagnosed with PD and registered at family health centres in the municipality. They were invited to participate in May 2024. The inclusion criteria included being at stage I-III on the Hoehn & Yahr scale,¹⁹ being able to walk independently or with technical aids such as a cane or walker, absence of cognitive impairment or mild cognitive impairment according to the Montreal Cognitive Assessment (MoCA) (score ≥18 points, with <18 considered moderate or severe cognitive impairment) and not undergoing rehabilitation treatment at the time of the study. Those with other neurological diagnoses, musculoskeletal conditions limiting independent movement, or language disorders preventing them from responding to cognitive questionnaires were excluded. All participants were undergoing anti-Parkinson's treatment with levodopa, with no changes in medication during the 12-week study period, thereby reducing the possibility of bias arising from pharmacological adjustments during the intervention period.

Due to the exploratory nature of the study, no prior sample size calculation was performed, and all eligible subjects were included during the study period. Each participant was asked to complete an initial questionnaire collecting sociodemographic data, medical history, comorbidities, medication use, and history of falls in the last 6 and 12 months.

A healthcare professional measured each participant's vital signs (heart rate, blood pressure, respiratory rate, and oxygen saturation), and their subjective perception of exertion was recorded using the Borg scale, with scores ranging from 0 to 10.

Variables

Following the initial assessment, validated tools were used to evaluate cognitive and executive function, balance, and risk of falls. The MoCA was used to assess overall cognitive function in domains such as memory, attention, language, abstraction, and executive functions. A total score of 0–30 points was possible, with a score ≥26 considered to indicate normal, 18–25 to indicate mild cognitive impairment, and <18 moderate or severe impairment.²⁰ The Frontal Assessment Battery (FAB) assessed frontal executive function using six items: similarities, fluency, motor planning, interference, inhibitory control, and autonomy. A total score of 0–18 points was possible, with lower scores indicating greater impairment.²¹ Functional mobility and risk of falls were assessed using the Timed Up & Go (TUG) test, which involves getting up from a chair, walking 3 m, turning around, walking back and sitting down again. Times >13.5 s indicate a higher risk of falls.²² Finally, the Tinetti test assessed balance (0–16 points) and gait (0–12 points), with a total of 28 points. Scores <19 were considered to indicate a high risk of falls, scores 19−23 moderate risk, and scores ≥24 as low risk.²³ These tools have been shown to be valid and reliable in populations with PD and older adults, which facilitates the interpretation of results. The assessments were repeated at the end of the intervention to analyse changes in the variables of interest. These assessments were carried out by an independent assessor, who was not the professional delivering the intervention. However, due to the pilot nature of the study, there was no formal blinding. Post-assessments were carried out immediately after week 12 of the programme ended.

A control group was not considered due to the progression of PD itself, but a pre-intervention baseline was established to evaluate longitudinal changes in the aforementioned variables.

Intervention

The intervention consisted of a standardised therapeutic programme based on the NM-FITT® protocol.²⁴ This programme was designed to improve participants' physical and cognitive function. The programme lasted 12 weeks. One session per week was defined as the optimal frequency, taking into account logistical feasibility, resource availability, and the need to avoid physical and cognitive overload for participants in this community setting. The intervention was delivered by a physiotherapist/kinesiologist who had received specialised training in neurorehabilitation. Prior to the start of the study, they reviewed the NM-FITT® protocol manual in detail and received direct guidance from the local health team to resolve operational queries and standardise procedures. This preparation process ensured consistent implementation in line with the original protocol guidelines. In terms of adherence, participants attended an average of 10 out of 12 sessions.

The protocol combined physical exercises with cognitive tasks (sequences, memory, calculation, abstraction) using a progressive 3-phase scheme:

• 1

  Static phase: postural control in standing position with variations in the base of support, eyes open/closed and simple cognitive tasks;

• 2

  Dynamic phase: movements, changes of direction and use of unstable surfaces, combined with memory and attention tasks;

• 3

  Functional phase: simulation of activities of daily living with dual-tasking and variations in speed.

Cognitive tasks were integrated with physical exercises simultaneously in each phase of the protocol,⁹ constituting true dual-task training designed to challenge motor and cognitive function together.

Each exercise was designed to provoke automatic and reflex muscle destabilisation, thereby challenging postural control in various conditions.²⁵ The intensity and complexity of the exercises were adjusted according to individual tolerance and disease stage (Hoehn & Yahr I–III). The complete protocol, including examples of exercises, progressions and adaptations, is described in the freely available manual “Entrenando en casa en tiempos de COVID-19” (Training at home during the COVID-19 pandemic),²⁴ which ensures its replicability in different clinical and community settings. Previous studies have shown that this strategy can positively affect walking speed and step symmetry, as well as reducing freezing episodes in people with PD.^4,25

Statistical analysis

The collected data were organised in Excel® spreadsheets and processed using the Python programming language for statistical analysis. The normal distribution of the data was verified using the Shapiro-Wilk test. For variables that followed a normal distribution, descriptive analyses were performed using the mean and standard deviation. For data that did not follow a normal distribution, associated with the Tinetti test, the median and interquartile range (IQR) were used. It should be noted that the Wilcoxon signed-rank test evaluates the distribution of the ranks of paired differences, so it can detect statistical significance even when the medians appear unchanged.

To compare pre- and post-intervention assessments, a t-test for paired data was used, with a significance level of p ≤ .05, to determine significant differences between assessments. Where the data did not follow a normal distribution, the Wilcoxon signed-rank test was used, with the same significance level. Additionally, for each main comparison, 95% confidence intervals (95% CI) of the differences were calculated to provide a measure of precision and effect size beyond the p-value. Furthermore, to enrich the data analysis, the population was subdivided by sex and age. Due to the low number of participants, the non-parametric Mann-Whitney U test was used to analyse whether the effect of the intervention differed between subgroups: sex (women vs. men) and age range (≤65 vs. >65 years).

Sample size and statistical power

Due to the exploratory and pilot nature of the study, the sample size was determined by the available population in the community. A post hoc power analysis was performed, considering pre-post comparisons with α = .05 (two-sided). With 17 participants, the statistical power to detect an effect size of d^z = .61 was 80%. For moderate effects (d^z = .40) a sample size of approximately 50 participants would be necessary to achieve the same power, which informs the planning of future confirmatory studies.

Ethical aspects

This study was approved by the Scientific Ethics Committee of the University of Los Lagos (approval number: 045/2024, date: 12/08/2024). All participants signed an informed consent form prior to their inclusion in the study. The ethical principles of the Declaration of Helsinki and current national regulations for research involving human subjects were respected.

Results

A total of 21 participants with PD were evaluated in the initial stage. Of these, 13 were men (61.9%) and 8 were women (38.1%). The mean age of the population was 63.09 ± 9.68 years. The characteristics of the study population are shown in detail in Table 1.

Table 1.

General characteristics of the study population.

Subjects	Age (years)	Sex (F/M)	Marital status	N of years since PD diagnosis	Hoehn and Yahr stage	N of falls
						Last 6 m	Last 12 m
S1	74	F	Single	1	2	2	1
S2	64	F	Married	8	1	1	0
S3	61	M	Married	15	3	2	1
S4	59	M	Single	8	3	12	6
S5	69	M	Married	7	2	0	0
S6	63	F	Divorced	1	1	1	0
S7	75	M	Widowed	6	2	0	0
S8	73	M	Married	10	2	0	0
S9	43	M	Married	5	2	0	0
S10	62	M	Divorced	6	1	4	1
S11	46	M	Single	3	3	4	2
S12	72	M	Widowed	3	3	10	6
S13	65	F	Married	1	2	3	1
S14	54	M	Married	8	3	0	0
S15	56	M	Married	7	2	0	0
S16	63	F	Married	15	2	0	0
S17	73	M	Married	1	1	0	0
S18	62	M	Married	0	2	4	3
S19	49	F	Single	3	3	1	0
S20	65	M	Married	4	3	0	0
S21	81	F	Widowed	2	3	1	0

PD: Parkinson’s disease; F: female; M: male; N: number.

Of the population evaluated, 95.5% reported having another chronic disease in addition to PD. Chronic low back pain appeared to be the most prevalent condition, affecting 63.6% of the population. Of the participants, 59% had had one or more falls over the past year. When the time window was reduced to the last 6 months, 36% of participants reported having fallen during that period.

Four of the initial 21 participants dropped out of the programme for personal reasons unrelated to the training, mainly due to difficulties travelling to the training location. This left 17 participants for the pre- and post-intervention analysis.

The results of this research study provide a detailed assessment of cognitive and physical function in PD patients before and after an intervention, using standardised tools such as MOCA, FAB, TUG, and the Tinetti test, both in terms of gait and balance. The comparison between pre- and post-intervention measurements is detailed in Table 2.

Table 2.

Changes pre- and post-intervention.

	Pre-intervention	Post-intervention	Difference (95% CI)	p-value
MOCA, mean (SD)	18.82 (4.23)	19.53 (4.19)	+.71 (–2.82 to +1.40)	.49
FAB, mean (SD)	13.41 (3.13)	12.71 (2.84)	–.70 (–.79 to +2.20)	.33
TUG, mean (SD)	10.72 (3.42)	11.73 (3.0)	+1.01 (–3.14 to +1.12)	.33
Tinetti-gait median [IQR]	10 [9–11]	11 [10–12]	+1.00 (–2.50 to .00)	.01
Tinetti-balance, median [IQR]	15 [14–16]	16 [15–16]	+1.00 (–3.50 to +.00)	.12
Tinetti-total, median [IQR]	25 [23–27]	27 [25–28]	+2.00 (–4.49 to .00)	.024

95% CI: 95% confidence interval; FAB: Frontal Assessment Battery; MOCA: Montreal Cognitive Assessment; TUG: Timed Up & Go.

Values are expressed as mean ± SD for normally distributed variables (MOCA, FAB, and TUG), and median [IQR] for non-normally distributed variables (Tinetti).

In the MOCA, which assesses overall cognitive function, no significant differences were observed between pre- and post-intervention measurements (p = .49), with a mean difference of +.71 points (95% CI: −2.82 to +1.40) (Table 2). No differences were identified when analysed by sex (p = .77) or age (p = .21). Taken together, these results suggest that the intervention did not produce relevant changes in the participants' overall cognitive function (Fig. 1).

Figure 1.

Cognitive changes by sex and age sub-analysis.

In the FAB, which specifically assesses executive functions, no significant differences were observed between the pre- and post-intervention measurements (p = .33), with a mean difference of –.70 points (95% CI: –.79 to +2.20) (Table 2). No differences were found when analysed by sex (p = .69) or age (p = .62). These findings suggest that the intervention did not produce significant changes in the participants' executive function, regardless of sex or age (Fig. 2).

Figure 2.

Changes in frontal assessment by sex and age sub-analysis.

In the TUG test, which assesses functional mobility and risk of falls, no significant differences were observed between pre- and post-intervention measurements (p = .33), with a mean difference of +1.01 s (95% CI: –3.14 to +1.12) (Table 2). Similarly, analyses by sex (p = .19) and age group (p = .31) also showed no significant differences. These findings suggest that the intervention had no clear impact on functional mobility, regardless of sex or age (Fig. 3).

Figure 3.

Changes in mobility by sex and age sub-analysis.

In the Tinetti test, which assesses gait and balance, significant improvements were observed only in the gait subscale. The score increased after the intervention with a mean difference of +1.00 points (95% CI: −2.50 to .00; p = .01) (Table 2). When analysing the results by sex (p = .54) and age (p = .44), no significant differences were found, suggesting that the improvement in gait was consistent across the different groups.

No significant differences were observed between pre- and post-intervention measurements on the Tinetti test balance subscale (p = .12), with a mean difference of +1.00 point (95% CI: –3.50 to .00) (Table 2). Similarly, analyses by sex (p = .36) and age (p = .79) also showed no significant differences, indicating that the intervention had no clear effect on the participants' balance.

Finally, in the total score of the Tinetti test, which integrates gait and balance, a significant improvement was observed after the intervention (p = .024), with a mean difference of +2.00 points (95% CI: –4.49 to .00) (Table 2). This result indicates an overall positive effect of the intervention on participants' functionality. However, when analysed by sex (p = .80) and age (p = .84), no significant differences were found, suggesting that the improvement was consistent across the different subgroups.

Discussion

The results of this study suggest that the intervention had no significant impact on cognitive function, as there were no statistically significant changes in the Montreal Cognitive Assessment (MOCA) and the Frontal Assessment Battery (FAB). Similarly, no significant differences were observed in physical performance measures related to balance, specifically in the balance subscale of the Tinetti or in the Timed Up and Go (TUG) test. In contrast, significant improvements were detected in the Tinetti test's gait subscale (p = .01) and in the total scale score (p = .02), suggesting that the intervention effectively enhanced the participants' walking ability and motor function in general. These effects were independent of age and sex, indicating a homogeneous impact on the study population. However, the balance subscale of the Tinetti test did not show a significant improvement, showing only a slight trend towards an increase (median of 15–16 points; p = .12), consistent with the reported results. Furthermore, the statistical significance observed in the gait subscale using the Wilcoxon test reflects changes in the distribution of scores rather than in the median, which justifies the presentation of the results as median and interquartile range (IQR). The magnitude of the improvement in gait (an increase in the median from 10 to 11 points) is clinically relevant, reflecting an increase in agility and motor control, key factors in fall prevention.

This finding is consistent with previous studies demonstrating the effectiveness of physical rehabilitation programmes in enhancing motor performance in individuals with chronic diseases.^9,10,26,27 In terms of cognition, the absence of significant changes is consistent with research indicating that interventions with low-intensity or insufficiently specific cognitive components may not produce measurable improvements in standardised tests. The literature also suggests that more intensive interventions combining physical exercise and cognitive stimulation tend to be more effective in producing overall improvements in functionality and cognition in patients with neurodegenerative diseases.^28,29

The lack of effects on cognitive function suggests that, although the programme was designed as a 'dual' training programme, the cognitive component may have been insufficient or not specific enough to induce measurable changes. Similarly, the duration and intensity of the programme (one session per week for 12 weeks) may not have been sufficient to generate cognitive adaptations in individuals with chronic comorbidities.³⁰ In contrast, the improvement observed in gait can be attributed to the fact that the physical component of the programme was adequate to stimulate this domain, generating clinically relevant increases in motor control.

A critical limitation of the study is the absence of a control group, which restricts the ability to attribute the observed changes exclusively to the intervention. While the pilot nature of the study and the ethical and logistical conditions of the community setting made implementing a control group difficult, the literature has highlighted the importance of controlled clinical trials in isolating the effects of physical and cognitive interventions on motor function and fall risk in PD.^13,16 The small sample size and lack of an a priori calculation also restrict the generalisation of the results. While the post hoc analysis suggests the ability to detect moderate-to-high effects (d^z≈.61), confirmatory studies with larger sample sizes are needed to for a more accurate assessment of the changes observed. Furthermore, the lack of longitudinal follow-up means that the sustainability of improvements in motor function cannot be determined, highlighting the need to incorporate 3–6 month measurements in future studies.^13,18 Likewise, potentially influential variables such as previous physical activity level, sleep quality, mood, or the use of medication other than levodopa, which may affect motor and cognitive performance, were not collected.^28–30 The absence of these data limits the detailed interpretation of the results.

The positive results regarding walking suggest that interventions with well-targeted physical components can improve relevant aspects of motor function in populations with chronic diseases, which could help to prevent falls. However, these findings should be interpreted with caution due to the aforementioned methodological limitations. Future studies should incorporate control groups, for example, conventional physical activity interventions without a dual component, to more rigorously evaluate the differential efficacy of dual training in this population.^12,17 It is also recommended to increase the frequency and intensity of the sessions, given that previous research has shown that more intensive protocols (2–3 sessions per week) with a progressive increase in physical and cognitive load are associated with more pronounced improvements, especially in cognitive domains.^26,27 Similarly, the cognitive and physical components should be integrated more robustly and additional variables be included that allow for better characterisation of the sample and a more accurate understanding of the observed effects.

Conclusion

The dual-task training programme produced significant improvements in gait parameters, without evidence of changes in cognitive function or balance measures, suggesting that it could be a complementary intervention in community settings for people with PD. However, since the incidence of falls was not directly recorded during the 3-month follow-up period, the reported risk reduction is based on indirect indicators and cannot be confirmed by the occurrence of actual falls.

In terms of cognitive function, no significant changes were observed in participants, suggesting that the effects of these programmes on cognition were limited under the study conditions.

It is recommended that future research incorporate more intensive and multidisciplinary intervention programmes that integrate specific physical and cognitive components to more accurately examine their combined effects on functionality and cognition. Increasing the sample size and conducting longitudinal follow-ups is also necessary to determine the stability and magnitude of long-term effects.

Funding

This work was funded by the Research Department of the Universidad de Los Lagos (Regular Internal Project IVM 04-23).