Maternity Registers: An untapped source for monitoring maternal health status in Guatemala?

Mandy Rose MB.BS, MPH, MRCP¹; Pierre Buekens MD, PhD²; Fannie Fonseca Becker DrPH, MPH, RD¹; Edgar Barillas MD, MSc³; Ricardo Valladares³; Heidi Reynolds MSc, PhD candidate².

¹ MEASURE Evaluation, ORC Macro 11785 Beltsville Drive, Calverton, MD 20910.

² MEASURE Evaluation, University of North Carolina at Chapel Hill, 123 W. Franklin St, Suite 304, Chapel Hill, NC 27516-2524.

³ Gestion de Soluciones en Desarrollo (GSD), Guatemala City, Guatemala.

Background

Better tools and methods are required for monitoring progress in maternal health. Until recently most countries have relied on monitoring changes in the maternal mortality ratio, but this method of assessing progress is now recognised as being inappropriate for determining short term change. Experience with other approaches, such as the use of process indicators, is limited and there is an urgent need to find sustainable, low cost methods for monitoring and evaluating maternal health status, particularly at district level. In most countries maternity facilities routinely collect information about women and the outcome of their pregnancy in delivery or maternity log books. There is, however, little evidence that these log books are used for monitoring purposes, even in many developed countries. Although the possibility of deriving process indicators from maternity registers has previously been considered, no other study, as far as we are aware, has systematically explored the use of maternity registers for this purpose.

Methods

In late 2000 MEASURE Evaluation and Gestion de Soluciones en Desarrollo (GSD) conducted an in-depth case study of maternity registers in three states in Guatemala in collaboration with the Ministry of Health. The overall aim of the study was to determine the feasibility of using maternity registers for monitoring maternal health status at facility, district and national level. The study was designed to identify the range of process indicators that might be derived from maternity registers and to assess the quality and validity of the register data. Financial support for the study was provided by USAID-Guatemala and USAID-Washington.

In the first phase of the study a detailed census of public and private facilities was conducted. Multiple sources of ascertainment were used, including the state public health departments and project staff from the Maternal and Newborn Health Project. Each facility was then visited to determine if they conducted deliveries and if they possessed a maternity register.

Results

In addition to the 8 MOH facilities from which state statistics are compiled, the initial listing identified a further 36 facilities. Thirteen of these facilities were later found to be ineligible (either closed or did not attend deliveries) one had been open less than six months and had not managed any deliveries by the time of the census and one facility refused participation in the study. Only two delivery facilities reported that they did not routinely record information on their clients. The majority of the additional 16 study facilities performed caesarean sections.

All the remaining 27 facilities maintained a maternity register of some type. For the purposes of the study we defined "maternity register" as the record that is kept permanently in or near the delivery room and that is used for the purpose of recording all admissions and deliveries to the delivery suite. Larger facilities usually had several registers in addition to a maternity register. Smaller facilities recorded similar information but usually in one all purpose register, variously described as an admissions, discharge or procedures register. A total of 46 different variables were collected in the 27 registers. Each register contained between 2-27 variables (median 10 variables).

Discussion

Preliminary assessment suggests that maternity register data in the three states may be able to provide at least two internationally recognised indicators; proportion of skilled attendant deliveries and caesarean section rate. However, the finding of a large number of additional facilities in our census, whose information is not routinely included in national statistics, highlights the importance of identifying all delivery facilities to obtain accurate district/state rates for monitoring purposes. In a next step, in each participating facility a retrospective sample of register entries will be examined to determine the range and validity of the data.