It is widely accepted that substantial reductions in maternal mortality and severe morbidity are impossible to achieve without an effective referral system for complicated cases. Early detection and referral to higher levels of care might also substantially reduce neonatal deaths due to the complications of childbirth. The general goal of such a referral system is that patients are dealt with in the right place with effective treatment provided at the minimum of cost. There are real challenges, however, in monitoring the effectiveness of such referral systems once put in place. This paper describes some of the tools used to review pregnancy-related referrals in Lusaka, Zambia. The tool-mix used provided information for monitoring five different aspects of the referral system: the distribution of births across levels of facility and population coverage; the use of essential obstetric care (EOC) level facilities by women with complications; the progress towards a reduction of maternal mortality at referral facility level; inappropriate use of EOC level facility; and perinatal outcomes at peripheral facility level. Apart from the information on coverage, the data came from routinely collected facility statistics, registers and medical notes. Findings for Lusaka are reported. Consideration is given to issues of interpretation of specific indicators, and to how such tools might be used in conjunction with others, in order to help district managers to monitor the effectiveness of district maternity referral systems.