Purpose: To determine whether activity-modifying behaviour mediates the relationship between the severity of knee pain and each of physical function and knee-related quality of life. Methods: A total of 105 participants with medial knee pain and no diagnosis of knee osteoarthritis (mean age 52.2 [SD 6.7] y) completed two self-report questionnaires. The Questionnaire to Identify Knee Symptoms assessed activity-modifying behaviour; the Knee injury and Osteoarthritis Outcome Score assessed pain severity, physical function, and knee-related quality of life. Simple mediation analysis was performed using linear regression. Results: The unstandardized regression coefficient for activity-modifying behaviour revealed partial mediation of the effect of pain severity on physical function (0.31 (SE 0.09), p<0.001) and on knee-related quality of life (0.24 (SE 0.07), p<0.001). After accounting for activity-modifying behaviour, the variance in physical function that was explained by pain decreased from 45% to 15%, and the variance in knee-related quality of life that was explained by pain decreased from 64% to 25%. Conclusion: Activity-modifying behaviour partially mediates the relationship between pain severity and physical function and between pain severity and knee-related quality of life. Activity-modifying behaviour may thus counteract the impact of knee pain on physical function and knee-related quality of life, which explains why it is used by people with emergent knee pain.