A typical cohort study should include the components of:
1)A study population.
For the cohort study, the study population was chosen from Matlab, southeast from Dhaka, the capital of Bangladesh. It is one of the areas registering the highest contamination of tube-well water by arsenic. A total of 29, 134 pregnancies were used as the study population. The study population comes from two areas in Matlab 1)The International Center For Diarrhoeal Disease Research area, where a Maternal, Child Health, and Family Planning Program is in place 2) a government health service area where citizens receive government subsidized health care.
2)The incidence of those exposed to the harmful element compared to those not exposed to the harmful element (this is the comparison group).
In the cohort study, 29, 134 pregnancies qualified for analysis out of 51, 500 pregnancies which resulted in fetal deaths or singleton births. Of the 51,500 number, 46, 441 used tube-water during their pregnancies, but 5, 059 did not. Of the 46, 441 women who used tube-well water during their pregnancies, 17, 307 women either had non-functioning tube-wells or used tube-wells from another area outside of Matlab. The remainder 29, 134 drank tube-well water during their pregnancies.
Comparison groups may be taken from the general population (in this case, the Matlab area), lower exposed group, or even from those pregnancies dependent on tube-well water outside the study area. The study shows that location of the tube-well water did not modify the risk of infant deaths from arsenic exposure. For the first comparison group, researchers collected data on drinking sources for all inhabitants of the study area above four years of age.
3)The formulation of a clear hypothesis for the study including the definition of the exposure to the harmful element and clear outcomes of interest.
Researchers established the hypothesis that drinking water contaminated with arsenic increased the risk of infant death and/or fetal loss during pregnancies. The study noted an increased risk of infant deaths and fetal loss at higher quintiles of water arsenic concentration. The highest infant death risk was sustained by pregnancies exposed to arsenic concentrations of 276-408 micrograms per liter (compared to the lowest exposure group).
The Bangladeshi local drinking standard is capped at 50 micrograms/liter. For those mothers exposed to arsenic levels above this figure, there was a 17 percent increase in infant death and a 14 percent increase in fetal loss. So, it has been established that the level of risk increases from the 50 microgram/liter figure. This has established a need to strengthen arsenic mitigation programs in areas with high arsenic concentrations in the drinking water.
4)A surveillance system to ascertain outcomes of interest such as sending regular questionnaires to study subjects, regular physical exams of study subjects and/or regular interviews.
Monthly household surveillances are conducted in Matlab. The researchers have found this to be more effective in amassing accurate figures for pregnancy losses than the retrospective interview. A quality control team regularly cross-references test area drinking water sources from census figures collected in 174, 1982, and 1996. Water samples were also analyzed in duplicate for arsenic concentrations. Water from 61 randomly selected tube-wells in the study area were also analyzed three times a year for a three year period.