Withdrawing from a study and taking your data with you (created 2010-05-15).

This page is moving to a new website.

Someone asked me what the phrase "you can withdraw from the study at any time" really means. Can a research subject withdraw and take their data with them (that is, ask that their data be expunged from the database)? What if they raise the objection after the data analysis is done, because they don't like the results of the study. Can they ask for their data to be expunged then? What if they raise the objection after the data is published?

It helps to distinguish reasons why patients withdraw. Some withdraw because they just don't want to be bothered with another survey, another clinic visit, another blood draw, or whatever. Others don't like the intervention because the pills make their skin turn green. It is probably safe to assume that these people don't mind letting you use the data you've collected so far, as long as you stop any further intervention. There is indeed value to the researchers for keeping the data accumulated so far. For example, if you only have baseline data on subjects who withdraw, that still allows you to compare the demographic profile of withdrawals to the rest of the sample to see if there is a systematic difference that could produce serious biases.

Other subjects might want any existing data about them expunged from the database. It might be due to a sense of loss of privacy, such as in a genetic study, or it may be that they have developed a sense of moral outrage about the research itself. A newly developed concern about privacy or a sudden sense of moral outrage is unusual. If a subject has these concerns, they will probably feel that way during the consent process and just refuse to join. These sudden revelations during the research may be a sign that the informed consent process was flawed.

How can we tell if a subject is withdrawing out of a sense of inconvenience rather than because of deeper concerns that might lead them to want their data expunged? It is rare, I suspect, for a researcher to ask a patient who withdraws if they want their data expunged, unless the IRB requires the researcher to ask such a question. There are studies where consent was impossible during the initiation of data collection but is asked as soon as practical. In these studies, it is vital to expunge the data if the subject refuses consent.

Also, many patients withdraw from the study by just ceasing all contact with the researcher, such as stopping to keep their appointments, failing to respond to surveys, as opposed to actively asking to withdraw. So you may never get a chance to even ask.

The whole research process does not make it easy to find out if a subject wants their data expunged. If a subject makes such an objection known in spite of the obstacles, I believe you should respect it. There may be some exceptions in a medical study where information about dangerous side effects of a drug might be lost, but in most other settings you should respect the subjects wishes. Failure to do so is just asking for trouble. What happens if you refuse to expunge their data and they go public with their complaint. Do you want to hold onto their data so badly that you would be willing to weather all the bad publicity that you'd get?

Can subjects ask to have their data expunged after they've seen the results of the research? Sad to say, such an opportunity is not possible in most research settings. Most subjects are never aware of the results of the research, and if they become aware, it is only after the data is published. Researchers should share at the earliest practical time the results of the research (as well as the individual data results for a particular patient if those results have clinical value), but it just isn't done.

If a subject objects after seeing the publication, their objection is not going to have any practical impact. They are trying to shut the barn door after the horses have gotten out. In theory, I suppose, a "correction" to the article could be published, if a large number of subjects make a late objection, but no journal is going to allow this to accommodate one or two late objections. Be honest about this, but do let them take the data out of any second or third publication down the road. Again, if subjects develop a sense of moral offense, that may be an indication of the failure of the original consent process.

So in short, the research process itself makes it difficult for a research volunteer to get their data expunged. If such a request occurs in spite of these obstacles, this may be a sign that the informed consent process was flawed. Why are they finding out just now, for example, that your research might lead to greater use of stem cells?