School achievement and adolescent self-harm: methodological issues may have led to misleading conclusions in a highly powered national study