Prospective (Cohort) Studies
Prospective Studies: BACKGROUND
Prospective studies follow a group of individuals (cohort or selected population) over a period of time to observe health outcomes, including mortality. Within the context of maternal mortality, a prospective study could involve following a population of pregnant women or women of reproductive age. When an entire population is followed over time, this is usually termed a Demographic Surveillance System (DSS). Prospective studies seek to test for an association between a certain exposure and the outcome of interest. Randomized controlled trials are a special form of prospective study in that the exposure (treatment) is deliberately administered at random to individuals in the cohort.
Identification of death
Active surveillance is used to identify deaths usually by asking about enumerated members of the cohort.
Ascertainment of pregnancy status
If cohort is already comprised of pregnant or recently delivered women, they are by definition pregnancy related deaths. Cause of death will need to be established to determine whether it is a maternal death (See Verbal Autopsy and ICD Coding).
If the cohort is women of reproductive age, then pregnancy status will either be ascertained from the woman herself while she was alive or from relatives.
Advantages:
- as a source of mortality data active surveillance makes it less likely that deaths are missed
- pregnancy status can be confirmed
- many exposures and risk factors can be collected in advance of the outcome, including from the woman herself, thus potentially improving data quality
Limitations:
- large sample sizes are needed - for example, even in setting where the MMR is 1000/100000, one hundred pregnant women would need to be followed for each expected death; if the general fertility rate is 75/1000 then 1333 women of reproductive age would need to be followed for each maternal death
- may be complicated as early pregnancies may be hidden by women
- there may be false reports of pregnancy or early pregnancy loss where pregnancy status is not possible to confirm
- pregnant women may move out of the study area to deliver and so may be lost to follow-up when at highest risk
- pregnant women may move into a study area to deliver and not be counted
- the fact of surveillance or enrolment in a trial with careful follow-up may reduce levels of mortality, making the cohort un-representative
- by it very nature, the cohort is unlikely to represent a national population
Back to Top
Prospective Studies: REFERENCES
 |
Bennett, DA (2003) Review of analytical methods for prospective cohort studies using time to event data: single studies and implications for meta-analysis. Statistical Methods in Medical Research; 12: 297-319. |
 |
Fenton, PM; Whitty, CJM; Reynolds, F (2003) Caesarean section in Malawi: prospective study of early maternal and perinatal mortality. BMJ; 327: 587. |
 |
Greenwood AM; Greenwood BM; Bradley AK; Williams K; Shenton FC; Tulloch S; Bypass P; Oldfield FS (1987) Prospective survey of the outcome of pregnancy in a rural area of the Gambia. Bulletin of the World Health Organization; 65(5): 635-43. |
|
Pradhan EK, West KP Jr, Katz J, Christian P, Khatry SK, Leclerq SC, Dali SM, Shrestha SR (2002) Risk of death following pregnancy in rural Nepal. Bulletin of the World Health Organization; 80(11): 887-91. |
 |
Thonneau PF, Matsudai T, Alihonou E, De Souza J, Faye O, Moreau JC, Djanhan Y, Welffens-Ekra C, Goyaux N. (2004) Distribution of causes of maternal mortality during delivery and post-partum: results of an African multicentre hospital-based study. European Journal of Obstetrics, Gynecology and Reproductive Biology, 114(2): 150-4. |
Back to Top
