Effects of Psychological Interventions to Enhance Athletic Performance: A Systematic Review and Meta-Analysis

Abstract

Background: Psychological interventions are commonly applied in sports to help athletes enhance their performance, but the effect psychological interventions have on actual performance is unclear despite decades of research.

Objective: We conducted a systematic review with meta-analyses to investigate the effects of a wide range of psychological interventions on performance in competitive athletes.

Methods: A study protocol was preregistered in PROSPERO, and a literary search was performed in the MEDLINE, PsycINFO, Web of Science, and SPORTDiscus databases. Psychological intervention studies were eligible by using a group design and a quantitative performance outcome with athletes competing at a regional or university level or higher. Included studies were assessed regarding intervention characteristics, research methodology, and risk of bias. A multi-level meta-analysis framework with cluster robust variance estimation was used to quantitatively synthesize the results.

Results: A total of 111 studies met the inclusion criteria, and 25 of these studies (37 effects) could be synthesized into five meta-analyses in which there were similarities in the type of psychological intervention, comparator, and experimental design. Meta-analyses I (multimodal psychological skills training vs control), II (mindfulness- and acceptance-based approaches vs control), and III (imagery vs control) consisted of parallel-group studies, and random-effects models were used to calculate the standardized mean difference. Meta-analyses IV (attentional focus strategies, external vs internal) and V (regulatory focus performance instructions, prevention vs promotion) consisted of counterbalanced crossover design studies, and random-effects models were used to calculate the standardized mean change using change score standardization. Significant results were found in three of the meta-analyses (I, II, and III). Psychological skills training (g = 0.83, 95% confidence interval 0.21-1.45), mindfulness- and acceptance-based approaches (g = 0.67, 95% confidence interval 0.01-1.32), and imagery (g = 0.75, 95% confidence interval 0.14-1.36) outperformed controls with moderate effects. However, when non-randomized trials and subjective performance outcomes were removed in sensitivity analyses, the overall estimates of the effect size were no longer significant in any of the syntheses.

Conclusions: The significant moderate effects for psychological skills training, mindfulness- and acceptance-based approaches, and imagery are not stable, and further trials with robust research methodology, such as randomized controlled trials, are requested for all types of psychological interventions aiming to enhance performance in athletes. Moreover, improved reporting standards and the provision of datasets in open science repositories are important to consider in future trials in sport psychology.

Clinical trial registration: PROSPERO CRD42017056677.

Fig. 1

PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) flow chart of the eligibility process [25]. ^aSome reports (publications) included more than one study, therefore the number of studies exceeds the number of reports. PICO Participants, Interventions, Comparator, and Outcome

Fig. 2

Forest plot for meta-analysis I illustrating the effect of multimodal psychological skills training on sport performance outcomes in athletes in comparison to controls. Effect sizes are the standardized mean difference (SMD [Hedges’ g]) with a 95% confidence interval (CI). Black squares are individual effect sizes and the square size represents the relative weight. Black rhombuses are the summary effect size estimate for the random-effect (RE) model with model-based variance estimates and with robust variance estimates (RVE). a and b are separate performance outcomes in the same study

Fig. 3

Funnel plot for meta-analysis I: multimodal psychological skills training versus control. Black dots are individual effect sizes with the standardized mean difference (Hedges’ g) on the x-axis and the standard error of the effect sizes on the y-axis. The dashed vertical line represents the summary effect size estimate in the meta-analysis model. Positive standardized mean difference values favor psychological skills training and negative standardized mean differences favor controls for the performance outcome. The plot does not account for the fact that some effect sizes are dependent because of being clustered within the same study

Fig. 4

Forest plot for meta-analysis II illustrating the effect of mindfulness- and acceptance-based interventions on sport performance outcomes in athletes in comparison to controls. Effect sizes are the standardized mean difference (SMD [Hedges’ g]) with a 95% confidence interval (CI). Black squares are individual effect sizes and the square size represents the relative weight. Black rhombuses are the summary effect size estimate for the random-effect (RE) model with model-based variance estimates and with robust variance estimates (RVE). a and b are separate performance outcomes in the same study

Fig. 5

Funnel plot for meta-analysis II: mindfulness- and acceptance-based interventions versus control. Black dots are individual effect sizes with the standardized mean difference (Hedges’ g) on the x-axis and the standard error of the effect sizes on the y-axis. The dashed vertical line represents the summary effect size estimate in the meta-analysis model. Positive standardized mean difference values favor mindfulness- and acceptance-based interventions and negative standardized mean differences favor controls for the performance outcome. The plot does not account for the fact that some effects sizes are dependent because of being clustered within the same study

Fig. 6

Forest plot for meta-analysis III illustrating the effect of imagery on sport performance outcomes in athletes in comparison to controls. Effect sizes are the standardized mean difference (SMD [Hedges’ g]) with a 95% confidence interval (CI). Black squares are individual effect sizes and the square size represent the relative weight. Black rhombuses are the summary effect size estimate for the random-effect (RE) model with model-based variance estimates and with robust variance estimates (RVE). a and b are separate performance outcomes in the same study

Fig. 7

Funnel plot for meta-analysis III: imagery versus control. Black dots are individual effect sizes with the standardized mean difference (Hedges’ g) on the x-axis and the standard error of the effect sizes on the y-axis. The dashed vertical line represents the summary effect size estimate in the meta-analysis model. Positive standardized mean difference values favor imagery and negative standardized mean differences favor controls for the performance outcome. The plot does not account for the fact that some effects sizes are dependent because of being clustered within the same study

Fig. 8

Forest plot for meta-analysis IV illustrating the effect of applying either an external or internal attentional focus during sport performance in athletes. Effect sizes are the standardized mean change (SMC) using change score standardization with a 95% confidence interval (CI). Positive SMC values favor an internal focus and negative SMCs favor an external focus on the performance outcome. Black squares are individual effect sizes and the square size represent the relative weight. The black rhombus is the summary effect size estimate for the random-effect (RE) model

Fig. 9

Funnel plot for meta-analysis IV: attentional focus, external versus internal. Black dots are individual effect sizes with the standardized mean change on the x-axis and the standard error of the effect sizes on the y-axis. The dashed vertical line represents the summary effect size estimate in the meta-analysis model. Positive standardized mean change values favor an internal focus and negative standardized mean changes favor an external focus on the performance outcome

Fig. 10

Forest plot for meta-analysis V illustrating the effect of applying either a prevention- or promotion-oriented regulatory focus instruction on sport performance in athletes. Effect sizes are the standardized mean change (SMC) using change score standardization with a 95% confidence interval (CI). Positive SMC values favor a prevention instruction and negative SMCs favor a promotion instruction on the performance outcome. Black squares are individual effect sizes and the square size represent the relative weight. Black rhombuses are the summary effect size estimate for the random-effect (RE) model with model-based variance estimates and with robust variance estimates (RVE). a and b are separate performance outcomes in the same study

Fig. 11

Funnel plot for meta-analysis V: regulatory focus instructions, prevention vs promotion. Black dots are individual effect sizes with the standardized mean change on the x-axis and the standard error of the effect sizes on the y-axis. The dashed vertical line represents the summary effect size estimate in the meta-analysis model. Positive standardized mean change values favor a prevention instruction and negative standardized mean changes favor a promotion instruction on the performance outcome. The plot does not account for the fact that some effect sizes are dependent because of being clustered within the same study