cross sectional studies can be used in analytic causation inference if the development of the condition does not cause loss to observation or change of exposure category  eg analysis of chronic conditions in a workforce in relation to exposures which do not change with time or which are affected by the condition. See the attached paper on COLD in coal mining

I speak only from my own experience I may be wrong but I also see it: Can be used for one part of the study, only where the section is made, or for the entire study must be used analytical analysis or few times classify cross-sectional studies and combined with analytical method !!

The above example is very good but I think with us in medicine can go only a combination of these two studies?

Dear Habib,

Hello, after a long time, I became happy to see your message, and thank you for the interesting question.

In my opinion, Cross-sectional study is a specific manner for data collecting. It means, if data collecting restricted to a specific point in time, it is cross-sectional study and the obtained data are cross-sectional data.

Now, after data gathering, we can use the data as information about naturally occurring e.g. health status, so, it will be a descriptive study.

But, we can use the data as information about exposure and outcome, in order to test our hypothesis, so, it will be an analytical study.

Finally, cross-sectional study is a specific data collecting method, which the collected data can be used for description or analytical decision.

Of course, in a cross-sectional study, our initial design shows the final use of the data, for example, collecting data about exposure status and outcome, shows the obtained data will use for the analytical purpose, but, collecting data about naturally occurring information, shows the obtained data will use for a simple description.

Best

Mehdi

Dear Dr.Emami,

I forgot to say a very brief definition of cross sectional study: this study as I see it, it is a study, that section only a moment studies were done and as such can not give a complete overview of the entire study, but only one of its lower part, therefore I was I could use as a comprehensive analytical study!

Regards from Winter cold Sarajevo

Jasenko Karamehic

Hi Dr. Emami,

I, too, have wondered this. Here is my argument, which may look a lot like yours.

To demonstrate causal inference a la Bradford Hill, there are several features that could be demonstrated in a cross-sectional study, and they are: strength of association, dose-response association, consistency, specficity, and coherence. "Biologic plausibility" doesn't count in my mind; you can always explain that away.

The key piece of evidence we are missing from a cross-sectional study is temporality, which you get in a cohort study (analytic), and you only theoretically get from a case-control study since you often are not able to document the exposure happened prior to the outcome (and this is an analytic design). So the answer to "why is a cross-sectional design NOT analytic?" cannot be "you cannot show temporality" because case-control studies can't do that either usually, and are subject to recall bias.

So, in summary, I don't know why they call cross-sectional "descriptive" and not "analytic". I do cross-sectional studies, and if the exposure and outcome are related, I enumerate the evidence for causal inference (as I described above). I put a link that is a helpful guide.

I think we can just ignore the label and present the evidence of causality to get around this. Certainly, why do a study relating exposure and outcome if you are not looking for cause? I agree with you.

Good luck with your cross-sectional study!

-Monika

http://www.drabruzzi.com/hills_criteria_of_causation.htm

Cross-sectional studies can also be classified as analytical studies. 

If we only describe the disease distribution in terms of place, person and time, it is observational study.

When we also determine any type of association by means of cross table, it will become analytical study. 

The name cross-section belongs to the fact that both exposure and outcome are measured simultaneously. So only disadvantage is about the temporal relationship. 

Indranil, do you have any peer-reviewed reference you can share with us that we can use when writing up a paper? I like your perspective, and actually agree with you, but I do not think it is a dominant perspective, and I cannot find any references that say what you way. Thanks!

Such type of question mostly asked during academic discussion. Thanks for putting this question. I think when we analyze the cross sectional study, then we get about the who, where and when, and the participants those facing any health problem or other participants are not. Then we can go further for the query of “why”. So then we can have hypothesis “Why” this Why. Then this can be accepted or rejected.

A cross sectional study can be both descriptive as well as analytical.

When you are testing a hypothesis and have a diseased and non-diseased group or an exposed and non exposed group and you are testing difference in the 2 groups by reporting prevalence ratios it is analytical.

The cross-sectional study could be analytic, but you can not prove causation because the exposure and outcome are assessed simultaneously. The first respondent (James Leigh) mentioned 'the causation' in occupational settings. May be in this situation you can prove that exposure precedes the outcome. You can show the association between exposure and outcome by using the Chi square test.

There are broadly two types of studies:

1. Descriptive Studies whose goal as the name implies is to describe or characterize a population or group

2. Analytic Studies whose goal is to critically analyze or test hypotheses regarding Association or Causality of "exposure" factors.

Cross sectional studies can be done for either purpose but not both at the same time. Cross sectional study done for description may be used to generate hypotheses but an independent study must be done to prove them. Some researchers conduct descrptive stdies and split them randomly into to groups and usee one set to generate hypothesess and the other group to validate them. This approach however is qustionable and may lead to fallacious conclusions.

Yes, they are analytic, and they can indicate associations.

After carefully thinking of read colleague, I accept that it can be and analytical, so far I think it may be just study section without analytical components!

Thanks

You can develop a cross-sectional study only for descrIptive purposes (obtaining prevalences). However, most of times these studies also allow to evaluate associations estimating prevalence ratios and performing hypotheses tests. In that case, you can consider them as analytic studies.

yes cross sectional study may be considered as analytical study. it depends how you are analyzing the data. of course there are certain limitation. usually we use cross sectional study for describing time, place and person distribution. However it may be a first step to find out the possible association between independent and dependent variable to further decide for designing a long term study.

Dear Monika Wahi, I missed ur query. Just now noticed ur response.

References:

1. Park K. Park's Textbook of Preventive and Social Medicine. Banarsidas Bhanot. Jabalpur. 24th edition. 2017.

2. Roy RN, Saha I. Editor. Mahajan & Gupta. Textbook of Preventive and Social Medicine. Jaypee Brothers Medical Publishers. New Delhi. 4th Edition. 2012.