Note: In hospitals in developing countries, treatment of abortion complications may be performed in many dif­ferent locations within the facility, such as the gyneco­logical ward, emergency room or operating room; data collection should therefore include admissions from all locations.

Peru provides an example of current efforts to improve the quality of information about PAC caseloads at pub­lic sector health centers and hospitals in one state. The process involves completion of a standardized clinical history form for each postabortion patient receiving treatment in the facility.  The form requests informa­tion about a limited number of key indicators, such as diagnosis, patient age, evacuation technique used, and duration of pregnancy.  Providers in these facilities are accustomed to completing a similar form for obstetric deliveries, so they have easily adopted the form. Staff are responsible for entering the information into a data­base at each facility, and the Ministry of Health makes the information available for use at the facility, state, and national levels.

This indicator can estimate the extent of induced abor­tion in countries where abortion is restricted. Research­ers have utilized data on abortion-related hospital ad­missions to construct such estimates. Evaluators can extrapolate the number of abortion-related hospital ad­missions to estimate the number of abortions in the population by using a variety of multipliers. These multipliers will be region and country specific. They will vary by the degree of restrictiveness of the legal and social climate, the availability of induced abortion performed by trained providers, the procedures used by clandestine providers, the availability of antibiotics, and the socioeconomic status of the women who undergo abortions. To address the uncertainty related to these multipliers, researchers have suggested using a range of estimates with several different multipliers (Singh and Wulf, 1994).

The best way to collect data for this indicator may be to conduct special studies at specific facilities (e.g., hos­pitals in urban areas). A hospital-based study in Nige­ria indicated that over 75 percent of gynecological ad­missions to hospitals were due to abortion-related causes (Rogo, Lema, and Rae, 1999).

A possible alternative indicator is Percent of women treated for PAC at service facilities who die.

Evaluators have used several approaches in attempting to distinguish between spontaneous and induced abor­tions. These approaches range from a series of ques­tions asked of the patient to multipliers based on the biological occurrence of spontaneous abortion to mul­tipliers based on expert opinion of the proportion of hospitalizations due to complications of induced abor­tion.