Ground condition as a risk factor in sports injury aetiology studies: the level of concordance between objective and subjective measures


Extremes in climatic conditions, particularly over the past decade, have resulted
in an increasing amount of literature being published on the association between ground
conditions and injury risk and the need to reduce/prevent such injuries (Ramirez et
al. 2006]; Takemura et al. 2007]; Twomey et al. 2012a]). Subjectively rated ground hardness has been linked to injury risk in many sports
injury epidemiology studies; however, the strength of the potential injury risk factors
is dependent on the quality and accuracy of the measures used, and the validity of
subjective assessment in this context is yet to be reported (Petrass and Twomey 2013]). This is the first study to compare subjective and objective ground assessments,
and the findings suggest that the ability to accurately rate ground hardness and moisture
level subjectively may be difficult. Overall, the results show a greater level of
agreement between objective and subjectively rated ground hardness compared to soil
moisture levels. However, for both ground conditions, when the objective and subjective
assessments did not agree on scales needing a high level of differentiation (e.g.
soft vs. very soft), they were just as likely to under- or overestimate the ground
condition under review. This finding has implications for future recommendations and
the use of subjective assessment measures over more accurate but costly objective
ground measures in sports injury epidemiology studies.

Ground hardness refers to the ability of the surface to absorb impact forces (Orchard
et al. 1999]) and has been reported in previous sports injury studies using a variety of subjective
rating scales. The most common has been a simple dichotomous scale of ‘hard’ or ‘soft’
(Ryan, McQullian 1992]; Inglis, Stewart 1979]) to more complex scales including aspects of traction as well as hardness, for example,
‘hard’, ‘firm’, ‘yielding’, ‘slippery’ or ‘heavy’ (Lee, Garraway 2000]). In this study, the four-point scale of ‘very hard’, ‘hard’, ‘soft’ or ‘very soft’
was selected for reasons of simplicity and its use in previous work (Alsop et al.
2005]). The results show that the subjective assessments of a ground rated as soft correlated
well with the combined average objective ratings as either ‘soft’ or ‘very soft’ and
a hard ground as ‘hard’ or ‘very hard’. While this is a pleasing result, perfect agreement
was only reached in approximately half of the 36 assessments when categories were
not combined. Given the inability of the subjective raters to distinguish between
‘soft’ and ‘very soft’ and ‘hard’ and ‘very hard’, replacing ‘very’ with ‘unacceptably’
might be valuable in future assessments. To ensure there is accuracy in the association
between subjectively rated ground conditions and sports injury, it is critical that
the reliability and validity of any subjective scale are established and reported
in future studies or, at the very least, the limitations of the subjective scale acknowledged.
Ideally, the standardisation of ground condition measurement scales/categories would
be valuable for comparisons between findings of different studies. It is also essential
that the descriptors represent the particular ground condition under review, for example,
hardness or traction, to ensure that injury risk is accurately associated with the
specific ground property.

Even though a subjective ground assessment is generally based on the average condition
of the whole ground, this study also investigated whether the level of agreement was
stronger with extremes of objective measures such as the maximum or minimum hardness
values of each ground. A slightly higher level of agreement was found between the
subjective ratings and the minimum hardness objective values than with the average
or maximum values. This result was based on the increased agreement of 9/13 for the
‘very soft’ ratings with minimum hardness objective values compared to 7/13 for the
same category with the average or maximum objective values. Regardless of which objective
measure was used, the findings show that the subjective and objective ratings only
perfectly matched on approximately 50% to 60% of the occasions. It is questionable
if this is a sufficient level of agreement to have confidence in reports linking subjective
ratings of ground hardness to injury risk and indicates that, where possible, objective
measures or proven valid and reliable subjective measures should be used in future
sports injury epidemiology studies.

In the agronomic literature, the level of soil moisture has been linked with ground
hardness (Baker 1991]), and it has also been used in subjective ratings of ground conditions in sports
injury epidemiology studies (Hagel et al. 2003]). It has even been suggested that soil moisture could be used as a proxy measure
for ground hardness, given the expensive and time-consuming nature of objectively
measuring ground hardness. The findings of this study show a very low level of concordance
between subjectively rated and objectively measured soil moisture levels and therefore
adds doubt to some of the previous studies purporting the link between sports injury
risk and soil moisture (Cromwell et al. 2000]; Hagel et al. 2003]; Andresen et al. 1989]). It was anticipated that it would be easier to accurately assess the moisture level
on the more worn, bare ground areas. However, there was slightly higher agreement
for the grassed areas than the bare areas. These soil moisture findings have implications
for the use of subjective soil moisture ratings in future injury-related research,
and it is recommended that hardness rather than soil moisture ratings be used in such
studies.

The strength of this study is that it was conducted over a full playing season which
resulted in a range of sporting ground conditions. It is acknowledged, however, that
differences in rainfall and more extreme conditions could be experienced in other
playing seasons or other regions. As this study was embedded in the PAFIX project,
there were multiple subjective raters involved. The results presented do not discern
between the raters and so it is possible that some raters were more accurate at assessing
grounds than others. Having said that, the data collection process was deliberately
chosen to mirror what happens in the normal sports injury studies context where, before
a given game, different raters would assess and record the condition of the ground
they are assigned to determine if a given playing surface was safe for play.

A possible limitation of this study is that it was not feasible to independently observe
each subjective assessment to ensure the instructions were being followed correctly.
Hence, it relied on the integrity of the PDCs to undertake a thorough walk around
the ground prior to making the assessment each time. The instructions and training
provided to the PDCs included going through each term in the subjective assessment
and providing the recommended pathway to walk around the ground. In this study, no
psychometric testing of the data collectors was undertaken. The inclusion of such
tests to assess the suitability of data collectors may lead to increased reliability
and validity of subjective assessment of ground conditions in future studies. Training
the data collectors on grounds of varying conditions might also prove valuable in
future work. While the use of objective measures is the most ideal option, where it
is not possible, some standardisation of subjective category labels in future sports
injury epidemiology studies will help to improve the accuracy of results provided
by subjective ground assessments and lead to greater confidence in using the injury
risk results to inform injury prevention strategies.