Epi in the News

Hi everybody...
Another interesting read... The Radium girls disaster is well known and deeply saddening. The court case that followed was epic! And the movie is amazing as well!

The question is regarding designing a case-control study to investigate an occupational risk factor... I think some of the decision to be taken would be:
1) Define the population to be studied or from whom the cases and controls need to be selected, and also to define what constitutes a case- either based on clinical evaluation or lab findings

2) Another decision to decide the controls, and what variables to be used to match with the cases- and would have to be independent to the exposure. I think its important to match the occupation as well... since the aim is to assess the possibility of occupational exposure.

3) The next decision would be to decide the exposure to be evaluated. Some studies look at all the exposures possible according to the person's occupation- there are matrixes that are usually used that determine all the possible exposures related to every occupation. Or else, there is concern regarding a particular exposure- then it would be important to determine the nature of the exposure... if to include all forms of exposure or to only consider a severe exposure as exposure. A decision would need to be taken regarding how to define exposure- based on the amount or frequency or duration of exposure. Since it's a new disease, maybe it would be better to look at all possible occupational harmful exposures.
Maybe its also relevant to exclude controls who have any other occupational-related diseases.

4) The next decision would be to decide how to collect information about the exposure. Recall bias is a problem, but if matrixes like I have explained can be used it might reduce the bias. Then it becomes important to determine the extent of the exposure- may be using employment records or timesheets to note the duration and extent of exposure rather than just using the memory of the subjects.

What do you all think?

Fathima

Thank you Fathima and Hany.

I think Fathima's idea of a job exposure matrix is excellent. I think that the best evidence that radium is associated with a particular outcome would be to show a graded response - ie that people with the greatest exposures to radium were most likely to end up with the outcome.

I agree with Hany that having several outcomes makes everything more complex. Ideally each outcome could be examined separately, but if numbers were insufficient then a composite outcome could be used.

As Fathima and Hany have said with a case control study the cases have to have the outcome and the controls have to not have it. I would have thought that several different cases and controls could be used. Firstly women in the same role in the factory (other dial painters), secondly women in other roles (cleaners, administrators), thirdly women in other factories. They should be matched by age. Each case and control could be given a score for how much radium exposure she had, as per Fathima's exposure matrix. I guess dial painters who licked their brushes had most exposure, dial painters who didn't lick their brushes less exposure, women in other roles in the dial painting factory, like cleaners, might have had a little exposure, and women in other factories might have none. Length of time working in the factory could also be used in the exposure matrix to add to the sophistication of the grading.

What do others think?

Judith