PU580 Unit 2 Exposure Assessment Comparison Example

Four Ways to Estimate the Same Exposure, Ranked by What Each One Actually Measures: A Unit 2 Exposure Assessment Comparison for Organophosphate Pesticides in Pregnancy in a Composite Farming Valley

[Student Name]

Purdue University Global

PU580: Environmental Epidemiology

Unit 2 Exposure Assessment Comparison

[Instructor Name]

[Date]

The farming valley and its study are composite; the four assessment methods, their error structures and the research cited are real.

The Question Each Method Is Serving

A cohort of 480 pregnant women in a valley where orchards and row crops are sprayed between March and September. The question is whether higher exposure during pregnancy is associated with a measure of infant neurodevelopment. Every method below is an attempt at one variable: exposure during a specific window of weeks.

Method One: Distance From Treated Fields

1. What it measures: proximity, combined with pesticide use reports to give an area estimate.

2. When: any period, including retrospectively, which is its main advantage.

3. Cost: low. Requires addresses and a use database, no contact with participants.

4. Error structure: non-differential and large. It assigns the same exposure to a woman who works in the orchard and a neighbor who never leaves an air-conditioned house, which biases toward no association in most analyses.

Verdict: the right tool for generating a hypothesis across thousands of births, the wrong tool for estimating a dose.

Method Two: Questionnaire

1. What it measures: reported activities, residence, occupational tasks, household pesticide use and hygiene practices.

2. When: any period, with recall quality falling across time.

3. Cost: moderate, mostly interviewer time.

4. Error structure: potentially differential, since a mother of a child with a developmental concern may recall exposures differently from a mother whose child is well, which is the bias that makes retrospective questionnaires weak for this question.

Verdict: necessary for the pathways no measurement captures, such as whether shoes are removed at the door, and insufficient alone.

Method Three: Urinary Metabolites

1. What it measures: internal dose, integrating all routes, including diet.

2. When: the past few days only. Organophosphate metabolites clear quickly, so one spot sample represents a narrow window and not a trimester.

3. Cost: high per sample, and the cost structure pushes studies toward too few samples.

4. Error structure: low systematic error and high within-person variability. A single sample misclassifies a woman's average exposure badly, and the fix is several samples per trimester rather than a better assay (Barr & Needham, 2002).

Verdict: the only method that measures dose, and only for the days before collection. Studies in farmworker families have shown that both the metabolites and the home environment vary within the same families over a season (Bradman et al., 2007). A biomarker is not automatically the best exposure measure. It is the best measure of a few days, which is only the best measure if the few days are the ones that matter.

Method Four: Silicone Wristbands

1. What it measures: integrated airborne and dermal-contact exposure to the parent compounds over the wear period, typically a week.

2. When: prospectively, during the window the study chooses.

3. Cost: moderate, with laboratory analysis the main cost and good compliance in practice.

4. Error structure: measures the environment around the person rather than the internal dose, so diet contributions are missed and metabolism is not captured.

Verdict: the best compromise for a defined window, and the method with the least published validation against health endpoints.

How They Perform Against the Study's Actual Question

The question names a developmental window, so timing dominates. Distance can cover it but cannot distinguish within it. A questionnaire can place activities in it but depends on recall. Urine measures days, so covering a trimester needs repeated sampling, which is a budget decision rather than a methods decision. Wristbands cover a week prospectively and can be repeated. The recommendation is a tiered design: distance for the full cohort, wristbands in two one-week periods per trimester for a random subsample of 150, two urine samples per trimester in that same subsample, and a questionnaire for everybody. The tiered structure exists because the methods measure different things and because validating the cheap method against the expensive one in a subsample is what allows the full cohort to be used at all. Combining methods this way, rather than choosing one, is where exposure assessment in environmental epidemiology has been heading for two decades (Nieuwenhuijsen et al., 2006).

What Each Method Would Cost for This Cohort

Putting prices beside the methods changes which design is realistic, so the comparison carries them. Distance assignment for all 480 women costs about 6,000 dollars in staff time once the use database is obtained. The questionnaire costs roughly 40 dollars per participant in interviewer time, or 19,000 dollars. Urine at two samples per trimester for 150 women is 900 samples at about 180 dollars each, which is 162,000 dollars and the single largest line in the study. Wristbands at two weeks per trimester for the same 150 women come to 900 bands at roughly 95 dollars analyzed, or 85,000 dollars. The tiered design therefore costs about 272,000 dollars in exposure assessment, against 500,000 dollars for urine alone across the whole cohort at the same sampling density, which nobody would fund. The arithmetic is the argument: the tiered structure exists because the best measurement cannot be afforded for everybody, and a study that pretends otherwise ends up with one spot sample per woman and a null result it cannot interpret.

What the Comparison Cannot Fix

Three limits survive any combination. Exposure before conception and in the first weeks, before enrollment, is unmeasured by every method here except distance. Mixtures: these women are exposed to several compounds and the analysis will report one class, which misattributes any joint effect. And the measurement error in the cheap method, once quantified in the subsample, can be corrected statistically only under assumptions about its independence that the valley's working patterns probably violate.

References

Barr, D. B., & Needham, L. L. (2002). Analytical methods for biological monitoring of exposure to pesticides: A review. Journal of Chromatography B, 778(1), 5-29. https://doi.org/10.1016/S1570-0232(02)00035-1

Bradman, A., Whitaker, D., Quiros, L., Castorina, R., Claus Henn, B., Nishioka, M., Morgan, J., Barr, D. B., Harnly, M., Brisbin, J. A., Sheldon, L. S., McKone, T. E., & Eskenazi, B. (2007). Pesticides and their metabolites in the homes and urine of farmworker children living in the Salinas Valley, California. Journal of Exposure Science and Environmental Epidemiology, 17(4), 331-349. https://doi.org/10.1038/sj.jes.7500507

Nieuwenhuijsen, M., Paustenbach, D., & Duarte-Davidson, R. (2006). New developments in exposure assessment: The impact on the practice of health risk assessment and epidemiological studies. Environment International, 32(8), 996-1009. https://doi.org/10.1016/j.envint.2006.06.015