Below are some strengths and weaknesses for this cohort study. There will also be two links to help you explore further questions regarding the strengths or weaknesses in the arsenic study. Basically, a bias is a weakness in any study and you will see that I address bias in the section on weaknesses.
Quality Assessment Tools A
Quality Assessment Tools B
1)Was selection of exposed or non-exposed cohorts drawn from the same population? In other words, was information such as demographics, time period, and location explicitly stated?
Yes. The abstract at the top of the page clearly states that a cohort population of 29, 134 pregnancies was selected from Matlab, Bangladesh in 1991-2000. The question the study is trying to determine is whether exposure to arsenic in drinking water increases the risk of fetal loss and infant deaths and if so, at what levels. This fact is established in the abstract, a strength. The target population here refers to the pregnancies which resulted in fetal losses or singleton births.
According to the study, this number is 51, 500. 5, 059 pregnancies were not exposed to tube-well water, but 46, 441 pregnancies were. Of the 46, 441 number, 17, 037 pregnancies were either exposed to tube-well water outside the study area or had no access to functional tube-wells. The remaining 29, 134 number has been determined as relevant to the analysis. This figure is more than 50% of the 51, 500 number; a figure lower than 50% would increase the risk of bias.
2)Are different levels of the exposure of interest chronicled?
Yes. The study states that the local Bangladesh drinking standard lists the supposedly safe presence of arsenic at 50 micrograms/liter of water. The study finds that pregnancies exposed to levels even slightly above this level incurred a 17 % increased risk of infant deaths and a 14% increased risk of fetal losses. The highest risks are incurred by pregnancies exposed to tube-well water with arsenic concentrations of 276-408 micrograms/liter. The mean concentration is at 239 micrograms/liter with less than 1 microgram/liter exposure at the 10th percentile and 513 microgram/liter at the 90th percentile.
More than 80 % of the pregnancies have been exposed to arsenic concentrations of more than 10 micrograms/liter. Studying 'different levels of exposure (where possible) enables investigators to assess trends or dose-response relationships between exposures and outcomes–e.g., the higher the exposure, the greater the rate of the health outcome. The presence of trends or dose-response relationships lends credibility to the hypothesis of causality between exposure and outcome.' So, yes to this question, a strength for the Matlab study.
1)Are the groups recruited from the same population with uniform eligibility criteria?
Mostly Yes. According to the study, pregnancies were recruited from the Matlab study area from two service centers. One is the International Center for Diarrhoeal Disease Research that has been running a health and demographic surveillance system since 1966. It also runs a Maternal, Child Health, and Family Planning program in Matlab: this is the first service center. The other is a government service center which provides government subsidized health care to residents in Matlab.
Although this study recruits from two clinics in the area, the second service center is a general health care center. So, the study population has been recruited from the same population with mostly uniform eligibility criteria (the second service center may include pregnant women among its patients), leaving some room for weakness or bias in the study. To emphasize strictly uniform eligibility criteria, the study area would utilize pregnancy subjects from the same clinical population: Maternal, Child Health, and Family Planning Centers.
2)Exposure assessed prior to outcome measurement.
Mostly Yes. In order to determine whether the exposure causes the outcome, the exposure has to come before the outcome. Many cohort studies start with studying samples of exposed and non-exposed subjects and then follow them forward in time to assess risks from the exposure.
This cohort study relies on pregnancy data in the area from 1991-2000. Although it is stated that workers collected information on life-time drinking water sources for all inhabitants above 4 years of age by going from house to house, information about these water sources was validated using census records from 1974, 1982, and 1996. So, prior exposure to arsenic from records in 1974 and 1982 may substantiate the fact of prior exposure; however, the year 1996 falls within the pregnancy study period, introducing a weakness or bias into the study.
For more information on how to evaluate further strengths and weaknesses of this cohort study, please refer to the links. Thank you!