Cross Sectional Studies

Cross Sectional Studies Purposes

• descriptive without comparisons or evaluation of associations between an independent and dependent variable

  • summary of characteristics of the study sample

  • may or may not have null hypothesis

• can also examine associations between variables

• often conducted from survey data

Cross Sectional Studies

• identification of exposure group and assessment of outcome occurs at the same point in time

  • cannot conclude which came first: exposure or outcome

• also referred to as “prevalence studies”

  • examines the prevalence of exposures, risk factors, or disease

Prevalence

• frequency that something is present in a sample at a given point in time

• includes new cases and existing cases

• number of number with the condition of interest at the time of assessment divided by the number of people studied

Incidence

• new cases occurring over time

Cross Sectional Study Sampling

• identify the target population

• select individuals that represent that target population

• random or convenience sampling

Random/Probability Sampling

sample more likely to represent the target population

Convenience sampling

• sample is “Easy” for investigators to obtain

• less likely to represent the target population

Data collection

• can be prospective or retrospective

• in-person/telephone interviews

• mailings and internet surveys

Prospective data collection

• not currently in existence

• collected from interviews, mailings, internet

Retrospective data collection

• previously colleced

• medical records

• insurance databases

Pros/Cons of in-person or telephone interviews

• may allow clarification of responses

• requires transportation to interview site or a telephone

Pros/cons of mailings and internet surveys

• less time-consuming and less costly

• depend on reading or educational level of the respondent

• requires computer and internet access or a permanent address

Cross sectional study advantages

• inexpensive and fast

• no lost to follow up

• can study more than one exposure and more than one outcome

• can determine prevalence of exposure and outcome

Cross Sectional Study Limitations

• prevalence only tells us how many have the outcome at that point in time

• may not be appropriate for conditions of short duration

• no info on how long it takes to develop the outcome

• no info on incidence (development of new cases over time)

• cannot draw conclusions about the temporal relationship between exposure and outcome (cannot infer causality)

• response/participation bias

• self-reporting

• subject selection

Response/participation bias

• people who agree to participate differ from those who do not

• non participants tend to be sicker, less well educated, and less socially-stable, and less reliable

  • sample may not be representative of population of interest

• authors should state what the participation rate is and some explanation as to potential differences between participants/nonparticipants

  • assess response rate to survey; 60% is considered a good benchmark