for me, is not a design research because prospective it refers to the directionality of the study, indicating that there was follow, especially this directionality is true for cohort studies, while the cross-sectional measured simultaneously exposure and outcome, there is no follow up..

Is different than for a cohort survey of a cross measurement.

I don't think it is a legit description. Prospective studies usually refer to prospective cohort studies where a group of similar individuals are followed up over time. Cross-sectional studies measure parameters at a single point of time. Repeated cross-sectional study is a related term where a cross-sectional study is repeated at regular or irregular intervals, but the sample population may not be the same for different time points.

I dont think it is right. for corss-sectional study, you usually used retrospective data in the past to support your hypothesis, however, prospective research mostly focuses on collecting data in future.

Yes. The term prospective means you begin to colect and generate information forward. If you colect the information or measures the variables only one time (simultaneously dependent and idependent variables), you  have a cross sectional design. If you measures the variables two or more times you have longitudinal prospective design.

Prospective cross-sectional is contradictory. Study design is always relative to the outcome variable. Eg if the exposure/predictor variable was measured before disease, then the study is a prospective study of the association between exposure variable X and outcome variable Y. Cross-sectional means that the association between variables is estimated at a single time point. Without seeing a more detailed example, prospective cross-sectional sounds non-sensical. 

People just use terms even if they are not needed. I just saw a paper that mentions.." We used an observational cross-sectional correlational survey design"....not sure what was the need for all that. Keep it simple, either you have an observational study or some form of experimental study.

In cross sectional studies you can review records, it mean that the information was collected for people that were not thinking in  future research, that information can be incomplete, biased. Or the researcher can collected the information that they want to use. Some times researcher can use both sources of information from de records and from interviewing to the people but any thing you do is not prospective.

I agree with Rohit P Ojha

Prospective means 'look forward' and usually refers to collecting data over time after a baseline observation (e.g. 'five years of prospective followup in a cohort study').  'Cross-sectional design' usually means all variables were collected at the same time.  So usually, calling a study a 'prospective cross-sectional study' would be nonsensical.  One possible exception is that someone has done a cross-sectional analysis using the baseline of a prospective cohort study, but if that's what they did it would be clearer if they called it a 'cross-sectional analysis' (rather than 'x-sectional design').  Another less charitable possibility is that some authors use methodological terms that they know connote rigour (e.g. 'prospective') even though they don't really understand what the terms mean, or hope others don't.

The design of a study (cross-sectional, cohort...) is distinct of data collection chronology (retrospective, prospective).  "Prospective cross- sectional" or "Retrospective cohort study" could be an appropriate term which inform both about design an chronology.

Thank you all of you who gave answer to my question. Your different responses helped me to deal with this problem.

Prospective cross-sectional study is wrong. Some may add unnecessary terminology to the study designs. Prospective  and cross-sectional both are different study designs. Prospective study done with the follow-up (Collection of data) study and cross-sectional study without follow-up. cross-sectional study is like a survey. The collection of data at one point of time. Their is no need of follow up in  cross-sectional. So, two studies in one study; how it possible. Hence, Prospective cross-sectional study is a wrong study to my knowledge. 

My understanding to the terms  of retrospective and prospective  is that they are related to time. If the researcher collects data on events which have already occurred in the past, it is a retrospective study design. If the researcher starts collecting data as they occur from the moment the study starts on wards (in the future) it is a prospective study design. Whether these are descriptive (no comparisons) or cohort (comparative , controls are used) is another issue. In cross-sectional studies the essential element in design is the fact that interview, examination, observation, testing...etc is done only once and thus the results describe the point prevalence of the condition or characteristics understudy. A cross-sectional study may take some time (days, weeks or even months ) to complete but you need not to call it prospective because there no element of follow up.  

Mind you, in a cross-sectional study, a retrospective component may be introduced. Suppose one wants to assess the extent of prevalence of smoking among youth. The question will be do you smoke or not? The results will give point prevalence rate at the time of the study. you are not interested in the time when youth started smoking.

2. However if you are interested in the timing of starting smoking, the question could be modified to become as follows: do you smoke or not? for those who answered yes you can add further questions about the date of starting smoking and other aspects. Thus if the interview covers a defined group of youth (School students for example) you can now calculate both the point prevalence and the incidence rates

among these youths

I hope the point is now clear in the light of myanswers and the answers of other colleagues          

Thank you for your anwser .

 Suppose, we are conducting a study now (2017) with previously collected data say the year 2002-2005 (exposure and outcome already occurred) and we want to know the prevalence of certain disease (crossectional). In this case, we can say retrospective crossectional study because we are working with already collected data (retrospective) and looking at the prevalence of disease in the data now (crossectional study).

As others have said there is an important distinction between 'research design' and 'direction of measurement of exposures and outcomes'.  For the exposure 'typical attendance in class' and the outcome 'average grades':

Designs:

Cross-sectional: both exposure and outcome measured concurrently (e.g. number of classes attended last term, and average grades last term).  One sample, one set of measurements relating to the same time point or narrow time window.

Prospective cohort: Cohorts defined on exposure now at baseline (e.g. sample across the range of class attendance from poor to perfect; create attendance cohorts by stratifying from poor to perfect).  

Retrospective cohort: Cohorts defined on exposure in the past (e.g. create poor-perfect attendance cohorts using student attendance in official school records from 5 years ago).

Direction of measurement:

Prospective measurement of outcome - follow students forward from today to measure academic performance

Retrospective measurement of outcome - find past academic performance in school records

Confusion comes because it is possible to have retrospective measures in prospective designs, and vice versa:

Retro measures, prospective design: at baseline of a prospective cohort study (with cohorts defined across the range of baseline poor-perfect attendance as above), ask students about past study habits, previous grades, schools attended, parents' education - anything that has already occurred.

Prospective measure, retrospective design: start measuring students' performance at the beginning of the retrospective study, even though the cohorts (poor-perfect attenders) were defined using data from five years ago.  In this way, a retrospective design can have five years of retrospective outcomes and five years of prospective outcomes - ten years of follow-up in a study that takes five years to complete.

Because cross-sectional designs have only one set of concurrent measures, neither the design nor the measures can be described as 'prospective'.  However, a cross-sectional design can have retrospective measurement because any question about something that occurred in the past is by definition a retrospective measure.

It is possible that authors use the term 'prospective cross-sectional' to refer to prospective recruitment: we will start enrolling students today and recruit for one month.   But the elements of the cross-sectional design are not affected whether we enrol over a day or a month because there is only one set of measures taken, and so this term is misleading if it is used to imply true prospective measurement involving active followup.

I think the long list of comments made above is sufficient to verify that in general a term like "Cross-sectional prospective study" is not acceptable in epidemiological designs

I have seen this sort of design in some journals. I wonder why and how these papers were accepted for publication in the first place. With due respect, and from an epidemiological point of view, the so called "retrospective cross-section design" is absolutely wrong. It is an academic and epidemiological suicide.

very interesting information

I agree with Mark R Speechley: Retrospective, Cross-Sectional Design is possible (eg. Review of Hospital Records at one point) but not Prospective Cross-Sectional Design due to absence of a follow-up period which is a very important characteristic of Prospective Studies. Survey studies done at one time is just a plain Cross-Sectional Design, not Prospective.

Hello, I have the same question. i am designing a research proposal where i would like to collect data from medical record, exposure and outcome already occurred (data of 2016 to 2017). May call it retrospective, cross sectional study design?

Any collection of data on previously occurred investigation or intervention/outcome would be classed as retrospective study

Once again I believe there is widespread confusion between research designs (retrospective cohort, cross-sectional, prospective cohort) and the direction of measurement of exposures and outcomes (retrospective: 'look back' at things that have already happened; concurrent: things that exist in the present; prospective: 'look forward' at things that will happen in the future).

If you are using administrative data to identify exposure cohorts in the past (e.g. people who had surgery vs. medicine 10 years ago) I would call it a retrospective cohort design because your exposure groups are defined as having occurred in the past. Any outcomes you measure from the charts that occurred after the exposure of interest but before the study start are measured retrospectively. If you intend to also analyze outcomes that occur after the study start, they would be prospectively measured outcomes (in a retrospective cohort design).

Another source of confusion is between cross sectional design and cross-sectional analysis. If you correlate BMI at time T with depressive symptoms at time T, it is a cross-sectional analysis, whether the data come from baseline questionnaires in a prospective cohort design, from medical charts in a retrospective cohort design, or from a cross-sectional survey.

No wonder people find it difficult sorting out these terms, but I have found it helpful to clarify design, measurement or analysis, and work from the above first principles.

I think prospective refers to the direction of measurement and not time. In that respect I am not sure whether it can be a Prospective cross sectional since cross sectional refers to one time measurement

The comments above are sufficient to show that that in general terms like "Cross-sectional prospective/retrospective studies" are not acceptable in epidemiological designs.

this is interesting , recently i faced the same problem . We interviewed post GBS patient in one setting and try to evaluate their treatment and outcome and complication during their management. Only survived patient we interviewed and not any record file for crosschecked . So there is potential for recall bias . However what type the study will be . Is not it cross sectional retrospective data!

Cross-sectional prospective study" is not possible because both cross-sectional & prospective study are of different type study design; one is prevalence study & other one is incidence.

Review of record based data is also a cross sectional study design because of one point study & most of the studies compare outcomes only .

What are some better descriptions of cross-sectional studies using data from records, or collected in the past for study purposes? Cross-sectional study using retrospectively collected would not be correct (data can't be collected retrospectively) so maybe cross-sectional using previously collected data or medical record data. And include whether analytical or descriptive.

From the STROBE Elaboration article:

We recommend that authors refrain from simply calling a study ‘prospective’ or ‘retrospective’ because these terms are ill defined.29 One usage sees cohort and prospective as synonymous and reserves the word retrospective for case- control studies.30 A second usage distinguishes prospective and retrospective cohort studies according to the timing of data collection relative to when the idea for the study was developed.31 A third usage distinguishes prospective and retrospective case-control studies depending on whether the data about the exposure of interest existed when cases were selected.32 Some advise against using these terms,33 or adopting the alternatives ‘concurrent’ and ‘historical’ for describing cohort studies.34 In STROBE, we do not use the words prospective and retrospective, nor alternatives such as concurrent and historical. We recommend that, whenever authors use these words, they define what they mean. Most importantly, we recommend that authors describe exactly how and when data collection took place.

http://www.equator-network.org/reporting-guidelines/strobe/

Thank you Mr John for the clarification!

Thank You all for your explanation. I faced difficulty to differentiate these terms but now I got it clear.

I agree with Jayanti Semwal

I agree with DR. Mbong Eta

It's not possible at all. Cross-sectional study has no chance of either prospective or retrograde.

When it comes to classifying the study designs, one very crucial thing to remember is that, especially, analytic studies, the temporal relationship (i.e. looking to see whether the cause precedes the effect). in contrast of prospective studies such as cohort and RCT, cross sectional studies cannot determine the temporality. You may collect new "primary" data, but still you are not sure whether the study participants have the exposure and outcome of interest, and at a single point in time you can assess both of them, while prospective studies must be sure at base line the study population are free from the outcome and then you need time to follow them to assess the outcome.

It all depends on how one defines the terms.

For me, a prospective study is one in which nobody has taken the measurements. The researcher is in a position to specify the settings under which the data should be collected. In contrast, a retrospective study is one in which somebody has already taken the measurements. The researcher cannot alter how the readings were made. Retrospective or prospective is a classification based on time.

For me, a cross-sectional study is one in which the parameter of interest is recorded once per subject. A cohort study is one in the parameter of interest is measured after an intervention, and after a defined follow-up period. The difference between the two in the fact that the sample was at a given a position at one time, and another position after follow-up. The classification is based on the fact of the follow-up.

The two classifications are not mutually exclusive. A study may be retrospective OR prospective (not both), as well as cross-sectional OR cohort (but not both).

For example:

First: I look at hospital records between 2000 and 2010 to measure how many persons, undergoing ultrasound, had gall bladder stones (retrospective). It turns out that 10% of all abdominal ultrasounds (=800 patients) showed gall bladder stones (retrospective, cross-sectional). Of these patients, four-fifths were female (still retrospective, cross-sectional).

Second: Among the patients WITHOUT gallstones, I select 800 at random (Group 2). It turns out that 50% were female (now the study is retrospective, cross-sectional, and case-control).

Third: I look at the records from 2010-2020, looking to see if these gallstone patients were readmitted with pain. It turns out that of 800 persons, 80 were readmitted with pain in the period 2010-2020 (now the study is retrospective, cohort. It is no longer cross-sectional).

Fourth: I look at the records from 2010-2020, looking to see how many of the 800 Group B patients were readmitted with pain. It turns out that out of 800, 6 were readmitted with pain in the period 2010-2020 (now the study is retrospective, cohort, and case-control).

Now substitute 2020-2021 where 2000-2010 appeared, and 2021-2022 where 2010-2020 appeared. Automatically, "retrospective" becomes "prospective".

Suneet Sood, 12-12-2020

The short answer is NO

A prospective study design implies follow-up visits with multiple tests. A cross-sectional study design implies one visit or single test. A study cannot be prospective and cross-sectional at the same time.

Unfortunately, these terms are frequently misused in the literature, despite good intention. A study cannot be both prospective and cross-sectional. It is either one or the other. The former involves longitudinal collection of data going forward, while the latter involves collection of data of a snapshot in time.

I agree with Clement; a study cannot be cross-sectional (snapshot) and prospective. Prospective studies are longitudinal in nature (more than one round of data collection; baseline and subsequent rounds)

Consider a situation where I wish to know the prevalence of smokers among patients hospitalized for colon cancer.

FIRST study: I look at the hospital case files for the past 10 years, and find that 30% of the patients were smokers. This is retrospective data, not prospective (classified by conditions of data collection). This is cross-sectional, not cohort (classified according to follow-up).

Unfortunately the resident doctors who collected the data did not categorize the patients into 0-5 sticks/day, 5-10 sticks/day, and >10 sticks/day. This is data that I want, but the data was not recorded per my needs. I cannot change the conditions under which the data was recorded.

SECOND study: So I interview the next 100 patients, over the next 1 year, this time collecting the information according to my needs. This is prospective data, not retrospective (classified by conditions of data collection). This is still cross-sectional, not cohort (classified according to follow-up).

In your second study, is data collected only once from the 100 patients?

I still think there is confusion between cross-sectional design and cross-sectional analysis. You can perform a cross-sectional analysis in a prospective design but you cannot perform a prospective analysis in a cross-sectional design. This confusion may have led someone to classify a study design incorrectly as 'prospective cross-sectional'.

1. If we correlate current tobacco smoking and shortness of breath in a cross-sectional study it is a cross-sectional analysis because both variables are measured concurrently. Suppose we find that non-smokers are more likely to have shortness of breath than smokers. Our inference that non-smoking causes shortness of breath would be false if our findings were due to smokers with shortness of breath quitting smoking before the study, and being recorded as non-smokers.

2. If we correlate current tobacco smoking and shortness of breath in the baseline of a prospective cohort study it is still a cross-sectional analysis because both variables are still measured concurrently. We can still produce the same (probably) biased inference as in 1. The fact that this is a prospective cohort design is irrelevant because we have analyzed data from one timepoint (i.e. cross-sectionally.)

3. We can minimize biased inferences in a longitudinal design with prospective measures of outcome (i.e. a prospective cohort study; prospective = 'look forward' for outcomes in exposure groups defined at baseline.) We identify smokers and non-smokers without shortness of breath at baseline ('inception cohort') and follow them forward over time to see which group has a higher probability of developing shortness of breath.

This is why we can see cross-sectional analyses in both cross-sectional and prospective designs but we can never see prospective analyses in cross-sectional designs.

I agree with Mark

Cross-sectional is like taking a snapshot while prospective involves following up over a period of time.

Cross-sectional is like taking a snapshot while prospective involves following up over a period of time.

Cross-sectional is like taking a snapshot while prospective involves following up over a period of time.

The confusion arises from the need to qualify the timing of a cross-sectional study. A cross sectional review of patients in the last 10 years for example compared to a cross-sectional review of patients visiting a facility over the next 10 years will lead to the use of the terms retrospective and prospective even when they have nothing to do with the design. The ordering of the words in my opinion should do it. Obviously you cannot have a cross-sectional prospective/retrospective study

I agree with you Mark R Speechley , what happens if you collect data prospectively (injury over time) and you take a snapshot of a physiological parameter (power). I assume this is what you mean a cross-sectional analysis in a prospective design. What design would the title include?

Hi NIco

Suppose the snapshot is taken at Followup #1 in a prospective cohort study. The association between power at Followup 1 and number of injuries at Followup 1 is a cross-sectional analysis because both variables are measured concurrently. I think many people would agree that 'cross-sectional analysis' should be in the title of the paper. It would be misleading to have 'prospective' or 'longitudinal' in the title because even though the injury data were collected prospectively they are not being analyzed prospectively (e.g. by comparing the incidence rate of injuries measured over followups between those with low/high power at baseline.). It's also relevant what the hypothesis is (dependent and independent variable). Do we think power affects injuries? (more muscle power increases opportunity for tissue damage? or perhaps more power increases ability to avoid injuries?) Or the other way around (those with injuries have less muscle power due to de-conditioning after the injuries, or perhaps they have more power from rehab following the injuries?). These different interpretations highlight the limitation of cross-sectional analyses with respect to temporality if DV and IV are measured concurrently.

Hi Mark,

Thanks for this. This is was my conclusion. A cross sectional analysis would be appropriate for the title. Thanks again for the comment

If I enroll patients at baseline, and collect follow-up data once after 12 months of enrollment to see the outcome, how do I call such study? Is prospective about timing or design? Would it be appropriate to call such study a cohort study knowing that we are only assessing the frequency after 12 months while the patient is still exposed to the drug of concern?