Case-control study of GP attendance rates by suicide cases with or without a psychiatric history.