i am planning to collect blood samples from patients with a particular disease after their first visit in my department with the aim to compare the blood levels of some biomarkers with the general population. But this disease is rare, so i need around a year to collect a significant number of patients as well as data. Blood will be collected only during the first visit. As cross-sectional study by definition involves data collection at a specific point of time, what type is my observational study ?
Hello Than Stef
I guess it is case-control study
How You will do samplig from general population?
In case-control studies (CCS) participants are selected based on the outcome status.
In a cross-sectional study (CSS), participants are selected based on the inclusion and exclusion criteria for the study.
CCS answers the question "What happened", You know the outcome before the study.
CSS - "What is happening just now" - You measure markers (factors) and outcomes at the same time during the study.
Please, point out the hypothesis of Your study?
Design are built for casual inference.
P.S. Cross-sectional study time frame may be some years (rare situation).
May be useful :
Article Methodology Series Module 3: Cross-sectional Studies
Article Methodology Series Module 2: Case-control Studies
Thank you for the suppl links!
My null hypothesis is that patients with alopecia areata with a lifetime incidence of 2% and a prevalence of 0.1 percent (so in my country at a specific point of time i expect around 10000 people) do not have increased autoantibodies when compared to the general population. But i cannot go around the streets and ask for people to come or ask them if they know someone with hair loss, I wait for them to come to me for treatment. And i do not know the result already. So i guess this is a cross-sectional study with a long recruitment period. Another question is that if i wait for a year to collect the samples of 100-150 patients i can claculate the annual incidence of these patients for my clinic but how can I extraoplate the data for the country? i think it is not feasible isnt it?
The time at which you collect the data is only important to the degree that you expect timing to make a difference. From what you say, I see no reason why time by itself is important; what may be important is the length of time since the onset of the disease along with the rate of progression of the disease; the person's age might be a factor, and for some skin conditions there might be seasonal variations.
1 hypothesis about the level of antibodies
typical case-control design. The disease is known (case), and control is those who are not under disease.
2 "annual incidence of these patients for my clinic"
this is a cohort study
you observe all clinic visitors for a year and calculate the rate of visits for the alopecia
3" can I extraoplate the data for the country"
If your clinic serves a region that is homogeneous within the country - then the second question "annual incidence of these patients for my clinic" becomes "annual incidence of these patients for my region", if the region is a representative sample for the country - go ahead.
Incidence = (the number of new cases during a specific period) divided by ( total midperiod population at risk)
alopecia incidence = new cases of alopecia during one year/number of citizens in the region the same year
###
two different scientific questions - two different studies and design. It's OK :)
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA