Table 1

Summary of the evidence for hamstrings strength, hip ROM, previous injury of the same type, and age as risk factors for hamstring strains. The table also shows whether a multivariate approach was used to analyse the data

Study	Sport (n)	Total sample   size (n)	No. of injured athletes   (n, % per season)	Strength*	ROM*	Previous injury*	Other internal risk factors studied	Design (study type, statistical approach)†
*+, significant association; −, no significant association observed; NA, not applicable (case-control studies). †Univariate indicates that t test or similar were used for continuous variable, and χ²-tests or similar were used for categorical variables. Multivariate indicates that logistic regression analysis or discriminant analysis were used.
Burkett (1970)³⁹	Athletics & soccer	30 + 37		+		NA		Case-control
Liemohn (1978)⁵⁸	Athletics	27	7 (26%)	+	+			Cohort, univariate
Worrell et al. (1991)⁴⁰	Football, soccer, lacrosse & athletics	32	16 cases &   16 controls	−	+	NA		Case-control, univariate
Yamamoto (1993)⁵⁹	Athletics	64	26 (20%)	+				Cohort, univariate
Jönhagen et al. (1994)⁴¹	Athletics (sprinters)	20	11 cases &   9 controls	+	+	NA		Case-control, univariate
Orchard et al. (1997)⁶⁰	Australian rules football	37	6 (16%)	+	−		Age, height, weight, body composition, aerobic & anaerobic fitness, etc.	Cohort, uni- & multivariate
Bennell et al. (1998)⁴²	Australian rules football	102	12 (12%)	−		+		Cohort, univariate
Fredriksen (1998)⁶¹	Athletics (sprinters)	19	6 (32%)	+	−			Cohort, univariate
Verrall et al. (2001)⁵⁰	Australian rules football	114	26 (23%)			+	Age, height, weight, race, previous knee, back or groin injury	Cohort, univariate
Orchard (2001)⁵¹	Australian rules football	83503 player-games	672 injuries			+	Age, height, weight, body mass index, race	Cohort, uni- & multivariate
Arnason et al. (2003)⁴⁴	Soccer	306	31 (10%)		−	+	Age, height, weight, body composition, aerobic fitness, jumping ability, etc.	Cohort, uni- & multivariate