Effect of concurrent training on trainability performance factors in youth elite golf players

Study design

A parallel, 2-group, longitudinal study was designed to investigate the effects of two different approaches of training on selected golf performance factors. Selected subjects had similar handicap to avoid golf swing technical differences. We assigned volunteers to either a training group conducting a concurrent physical conditioning program (CT) before golf specific training (BG) or a group that performed CT after golf specific training (AG). After a familiarization period, laboratory tests, and a specific range of physical- performance, participants were evaluated on three occasions; 1 week before the start of the training program (T1), after 12 weeks of training (T2) and after 18 weeks of training (T3). Also, subjects reported to be free from any injuries, surgeries or sport related rehabilitation during the 12 months prior to starting the study. The flowchart for recruitment and testing is displayed in Fig. 1.

The research was conducted during the competitive season (i.e., February, March, April, May and June). Two months before the beginning of the study participants conducted the same regular golf training program. Participants were instructed for not alter their lifestyle during the investigation period in order to reduce potential interference. They were not allowed to exercise or consume stimulant drinks at least 24 h prior to test.

Participants

Based on the previous study by Alvarez et al. (2012) a priori power analysis (G*Power3) with α < 0.05 and 1–β = 80 indicated that a sample size of at least 14 was required to explore the differences between sequencing effects of neuromuscular training. A total of 16 elite right-handed youth male golfers voluntarily agreed to participate in the study and were randomly divided into two groups: before golf specific training (BG, n = 8) and after golf specific training (AG, n = 8). There were no group differences (p > .05) with regard to demographic and anthropometric data showed in Table 1. Players averaged 9.4 ± 0.9 h of training per week and completed at least one full round of golf per week.

All players involved in the study attended all the sessions. Legal guardians and all participants were provided an explanation of testing and training protocols and they gave written informed consent prior to data collection. They also completed a set of questionnaires on their health history and golf-playing history. This study was conducted in accordance with the guidelines found in the Declaration of Helsinki and all procedures involving human subjects were approved by the University of León Ethics Committee (ULE2018-2019-76).

Testing procedures

Anthropometric data

Anthropometric testing followed the International Society for the Advancement of Kinanthropometry protocols (ISAK). Fat mass, residual mass, bone mass, and muscle mass and their respective percentages were computed to estimate body composition; Deurenberg, Weststrate & Seidell, 1991.

In order to estimate the maturity status of participants, the peak-height-velocity (PHV) was calculated according to Mirwald et al. (2002). All anthropometric measures were highly reliable with intraclass correlation coefficients (ICCs) of 0.91 to 0.98 for skinfolds and 0.93 to 0.98 for diameters.

Training load quantification

During 18 in-season weeks, the perceived training load (TL) was quantified using the session rating of perceived exertion (sRPE) method (Foster et al., 2001). Ten minutes after each training and using Foster’s 0–10 scale (Foster et al., 2001), participants were asked by the same person (fitness coach) on all occasions to rate their general perception of the session difficulty (PE) (Chen, Fan & Moe, 2002). We allowed players to mark a plus sign (interpreted as 0.5 points) alongside the integer value (Otaegi & Los Arcos, 2020). All the golfers were familiarized with this method during the previous months. All golf specific training and CT PEs recorded during the study were summed separately. Then, TL was calculated by multiplying the PE value by the duration of the training. Partial 12-weeks (T2) and 18-weeks (T3) TL in golf specific training (TL-G) and TL in CT (TL-CT) were considered for each group (Otaegi & Los Arcos, 2020). The duration of a training session (training volume) was recorded for each player from the start to the end of the session, including recovery periods but excluding stretching exercises (Los Arcos et al., 2015).

Golf movement screen

We applied a specific golf movement screen (GMS) to examine the movement competency of golf players. According to Gulgin, Schulte & Crawley (2014), the subjects performed 13 different tests (movement screens). These tests, established by the Titleist Performance Institute (TPI), provide data with respect to stability, mobility, coordination of body segments, and balance. The sum of the 13 GMS was recorded. The ICC was 0.98.

Lower limbs explosive strength

Golfers performed a countermovement jump (CMJ) without arm swing on a jumping mat (SportJUMP System; DSD, Spain) according to Bosco, Luhtanen & Komi (1983). Golfers performed two maximal CMJs intercalated with 60 s of passive recovery. Only the best height for each participant was recorded. The ICC of the CMJ was 0.97 and the CV was 4.1%.

Rotational golf-specific exercise

According to Keogh et al. (2009) the golf swing-specific cable woodchop (GSCWC) is a rotational exercise that is very similar to the golf swing in terms of posture, range of motion, intended velocity, direction of force (torque) application, and coordination patterns.

A one-repetition-maximum (1-RM) test following the protocol established by the National Strength and Conditioning Association was performed to measure peak power output (highest instantaneous value) during each GSCWC exercise. The peak power outputs (Wmax) expressed in watts were measured with a pneumatic resistance device (Infinity, Keiser, Calif. USA) according to Peltonen, Häkkinen & Avela (2013). The ICC was 0.99.

Driving performance

Ball speed (Sball) was assessed using new regulation golf balls (Titlest Pro V1, USA), and tees of various heights to suit the preference of each participant. According to Alvarez et al. (2012), Sball expressed in km h-1 was measured with a Stalker’s type hyperfrequency radar (Stalker Professional Radar, Radar Sales, Plymouth, MA, USA). Each participant performed five drives at the maximum speed possible using his own club. The ICC for this test was 0.94.

Training intervention

During the 18-week intervention, golfers carried out four training sessions per week: two CT sessions (on Wednesday and Friday), one putter-and-approach session on Monday and one full round of golf (on Saturday or Sunday).

The CT program based on a mix of golf-specific functional movement training and neuromuscular training program (Table 2) was undertaken at an indoor facility (high performance sports center) on average 66.70 ± 3.1 min per session. The regular golf training took place at an outdoor facility (sport golf club center) located 30 min from each other. According to Fernandez-Fernandez et al. (2018) during recovery period between training bouts, all participants could ingest water and carbohydrate/electrolyte drink. Regular golf training lasted on average 83.2 ± 9.6 min and was characterized by a ∼10-minute specific warm-up (i.e., general mobility and low-intensity golf shots), ∼30 min of technical swing adjustments, and ∼40 min of specific drills (i.e., mixed iron-drives-putter drills).

All players had previous experience of this type of training. Prior to starting CT, participants performed a standardized 10-minute warm-up protocol. After the warm-up, each golfer developed a 30-minutes personal golf-specific functional movement training program with conditioning exercises designed to enhance the lower body stability and the upper body mobility (Lephart et al., 2007). Lastly, participants proceeded with the neuromuscular training program divided into three parts (maximal strength; explosive strength and golf-specific strength training), each six weeks long (Alvarez et al., 2012). Details are given in Table 2.

Data analyses

The data were checked for normality using the Shapiro–Wilk test and found to be suitable for parametric testing. Student’s t-tests were performed to determine differences between groups at baseline. A 2x3 repeated measures ANOVA was used to explore the effects of group (AG and BG), and time (one week before training, 12 weeks and 18 weeks after training). When a significant F value was achieved by means of Wilks’ lambda, Scheffe’s post hoc procedures were performed to locate the pairwise differences. In addition, partial eta squared (${\eta }_p^2$) was computed to determine the effect size which was interpreted as small 0.1, medium 0.3, and large 0.5. The percentage difference between groups was assessed using one-way ANOVA by comparing T1–T2, T1–T3 and T2–T3 and the Cohen’s d (Cohen, 1988) was calculated to determine the magnitude of differences between experimental conditions for each variable. The significance level was set at p ≤ .05. Statistical analysis was performed with SPSS 24.0 (IBM® SPSS Statistics 24, IBM GmbH).

Anthropometric data

ANOVA revealed no significant time x group interaction for anthropometric measures, although significant improvements were seen between the time points for both groups. Further post hoc analysis showed significant increase of body mass between T1 and T2 (p < .001; d = 0.14), T2 and T3 (p = .03; d = 0.04) and T1 and T3 (p = .002; d = 0.18) in BG, and between T2 and T3 (p = .04; d = 0.06) in AG. Related to percent muscle mass, a significant increase was observed in BG between T1 and T3 (p = .04; d = 0.34).

Perceived training load

Data analysis revealed significant time x group interaction effects just for TL-CT (p = .005; ${\eta }_p^2=0.45$). Differences between T2 and T3 were dismissed as non-logic (neuromuscular training load in T2 not comparable with golf-specific training load in T3).

Performance variables

Analysis of variance located significant time x group interaction effects for GMS (p = .02; ${\eta }_p^2=0.25$) and Wmax (p < .001; ${\eta }_p^2=0.53$). Additionally, ANOVA revealed a significant effect for time in all the performance variables. Regarding to BG, Scheffe’s post hoc tests located the differences between T1 and T2 differences were located in GMS (p < .002; d = 1.06), CMJ (p < .001; d = 0.34), Wmax (p < .001; d = 0.67) and Sball (p < .01; d = 0.24), between T1 and T3 in GMS (p < .001; d = 1.81), CMJ (p < .001; d = 0.78), Wmax (p < .001; d = 1.21) and Sball (p < .001; d = 0.31), and between T2 and T3 differences were located in GMS (p = .003; d = 0.97), CMJ (p = .03; d = 0.44) and Wmax (p < .001; d = 0.56). Furthermore, post hoc analysis for AG located the differences between T1 and T2 differences were located in CMJ (p < .001; d = 0.40) and Wmax (p < .001; d = 0.22), between T1 and T3 in CMJ (p = .02; d = 0.43), Wmax (p = .002; d = 0.41) and Sball (p = .02; d = 0.34), and between T2 and T3 differences were located in CMJ (p = .02; d = 0.03) and Wmax (p = .03; d = 0.18).

With regard to comparison of the percentage of change between evaluations (T1, T2 and T3) in association with the TL data are represented in Fig. 2. Concerning TL-CT, one-way ANOVA revealed a significant effect between AG and BG in T2 (p < .001; d = 3.43) and T3 (p = .018; d = 1.34). T1–T3 comparison between groups shows that BG obtains higher percentages of change in all performance variables: CMJ (AG +37.15%; BG +50.52%; p = .041; d = 0.91), GMS (AG +5.08%; BG +9.38%; p = .165; d = 0.73), Wmax (AG +8.03%; BG +16.96%; p = .001; d = 2.02) and Sball (AG +1.03%; BG +1.82%; p = .018; d = 0.92).