Denominator: number of women admitted to the health facility with obstetric complications.

Information required to construct the indicator should be available directly from facility registers and case notes. The lack of certain information may signal suboptimal care/management. The feasibility of the indicator also critically depends on standard definitions of admission-to-treatment-time-interval (ATTI) for each obstetric complication. For example, does admission time mean time of first arrival at the hospital, time seen by the admissions clerk, or something else? What is the treatment time for an obstetric hemorrhage? Is it when an intravenous drip is inserted, when a blood transfusion starts, when an oxytoxic is given, or when the hemorrhage stops?

This indicator should respond rapidly to changes in staff or facility administration practice. Early rapid response may simply be due to a Hawthorne effect (i.e., apparent improvement simply because of an observation taking place), but this type of improvement will not be sustained. To maintain improvement, programs should regularly audit ATTI. Better still, ATTI could be incorporated into routine management practice (e.g., discussed routinely in all case reviews). (See indicator, Percent of Facilities that Conduct Case Reviews/ Audits into Maternal Death/Near Miss).

This type of indicator has several ambiguities and deficiencies, as follows:

• The recording of the actual admission time may be seriously delayed if the facility lacks a triage system, and women must wait a long time on first arriving at the facility;
• The ATTI should be reviewed for all women developing obstetric complications. If the admissions
  register serves as the sampling frame to identify women with complications, then the register will exclude those women who had a complication but whose diagnosis was missed at the time of admission;
• The severity of the complication and hence the need for rapid treatment may be difficult to ascertain retrospectively for all cases (a further reason for facilities to set their own realistic standards);
• The indicator fails to capture the timeliness of treatment for those women who develop complications while in the hospital;
• Women who arrive at the hospital, but who are not admitted for a variety of reasons (e.g., they
  cannot afford the treatment), will be omitted from both the numerator and denominator. This omission may be a potentially important source of bias if poorer patients have a disproportionately high complication rate; and
• The indicator does not measure the appropriateness of the type of treatment given.

Because of these limitations, evaluators can best measure ATTI in combination with other indicators or approaches that measure complementary aspects of the quality of care, for example in the context of care reviews or near-miss audits.