The reason pre-eclampsia matters so much to maternity care providers is that adverse maternal and perinatal events cluster around the diagnosis of proteinuric gestational hypertension. While that is true, most pre-eclampsia is mild and evanescent, resolving rapidly postpartum. Therefore, every effort must be made to identify those women at greatest personal risk, and those bearing fetuses at greatest risk, so that they can be offered closer surveillance and lower thresholds for the use of effective interventions, such as delivery and the use of MgSO(4). Conversely, as delivery remote from term can increase perinatal risks and as liberal MgSO(4) use is associated with maternal morbidity, it may be as important to identify those women who have "mild" disease and bear little personal and/or fetal actuarial risk. For women with "mild" disease at presentation, expectant management remote from term or nonuse of MgSO(4) would be appropriate. Through the PIERS (Pre-eclampsia Integrated Estimate of RiSk) model research program, we have determined that most criteria for "severe" disease perform poorly when operationalized to predict adverse maternal and/or perinatal outcomes. However, with standardized assessment and surveillance of women with suspected and confirmed pre-eclampsia it is possible to lower maternal risks both within individual institutions and across regions. In addition, the PIERS group developed, and is currently validating, 2 outcome prediction models (full-PIERS and mini-PIERS) that we hope will provide an evidence base for the definition of "severe" disease and guide clinical decision-making, especially remote from term when potential perinatal gains are so great.