Respectful Insolence

Being involved in clinical research makes me aware of the ethical quandaries that can arise. Fortunately for me, for the most part my studies are straightforward and don’t provoke much in the way of angst over whether what I am doing is ethical or whether I’m approaching a line I shouldn’t approach or crossing a line I shouldn’t cross. However, there’s lots of research that flirts with the unethical and sometimes even crosses the line. Institutional review boards are there to oversee the ethics of clinical trials and protect the human subjects who participate in them, but they don’t always function as they should. Moreover, even clinical trials that are not at all unethical can be compromised if the investigators cross the line in other ways.

Ways such as using their own children for research subjects.

There was a fascinating article in the New York Times over the weekend that didn’t come to my attention until yesterday and that I’m surprised not to have seen mentioned more in the blogosphere. It was entitled Test Subjects Who Call the Scientist Mom or Dad, and it disturbed the hell out of me.

Even before his son was born, Pawan Sinha saw unique potential.

At a birthing class, Dr. Sinha, a neuroscience professor at the Massachusetts Institute of Technology, stunned everyone, including his wife, by saying he was excited about the baby’s birth “because I really want to study him and do experiments with him.” He did, too, strapping a camera on baby Darius’s head, recording what he looked at.

Dr. Sinha is among a new crop of scientists using their children as research subjects.

Other researchers have studied their own children in the past, but sophisticated technology allows modern-day scientists to collect new and more detailed data. The scientists also say that studying their children allows for more in-depth research and that the children make reliable participants in an era of scarce research financing.

“You need subjects, and they’re hard to get,” said Deborah Linebarger, a developmental psychologist who directs the Children’s Media Lab at the University of Pennsylvania, who has involved her four children in her studies of the effect of media on children.

Am I off base in thinking this thing gives me a serious case of the heebie-jeebies?

Worse, in many cases I would argue that not only is this bad parenting and unethical, but it’s bad science as well. It’s drifting towards a line that none of these scientists would acknowledge that they are drifting towards, but the parallels will be disturbing.

As the article points out, it’s not as though this is a new phenomenon, although it appears as though it may be more prevalent than in the past and becoming more so. For example, Jonas Salk injected his own children with the polio vaccine. Of course, back in those days, it may have been a bit more understandable. Polio was a scourge; it was feared. Public swimming pools were shut down in the summer because of polio outbreaks, and panics regularly occurred as recently as 50 or 60 years ago. A parent’s desire to protect his children from this disease may have overridden Salk’s caution, and back then there were no institutional review boards, at least not in anything near their current form or level of rigor, to supervise and judge the ethical conduct of clinical trials. No doubt Salk was also interested in proving to the world that his vaccine was also safe, and what better symbolic demonstration of its safety than for him to give it to his own children?

So what about experiments that do not provide what is potentially such a huge benefit, experiments that are being done strictly for knowledge? For example, what should I make of the case of the scientist who subjected his children to MRIs, justifying it thusly:

Dr. Toga said some nonscientists have said: “Why would a parent subject their kid to the dangers of M.R.I.? You should be ashamed of yourself.”

His response: “All I’m doing is taking a picture. Nobody loves my kids more than me. Would I ever do something that would endanger them?”

I smell the rancid stench of rationalization here.

Although very, very safe, MRIs are not without risk. No medical procedure is. Also, they are more than just “taking a picture.” For one thing, many types of MRI require the injection of intravenous contrast agents, agents that can cause adverse reactions, sometimes even life-threatening. True, the chance of such a reaction is very, very small, on the order of 1 in 100,000 or even lower, but why on earth would a parent subject his child to even that risk for no benefit to the child? Also, consider that an the administration of contrast requires that an IV be placed, or, at the very least, that an intravenous injection of the agent be given. Children in general do not like having IVs placed; they are painful. Adults can deal with such minor pains as an IV or injection quite well, but children often cannot. Then there is the issue of having to lie very, very still inside of a large metal tube with machinery that clanks. I’ve known adults who completely freak out with claustrophobic sensations when inside the MRI machine. It takes them an enormous effort of will in order to keep still. I’ve even had the occasional patient for whom I’ve had to prescribe a sedative or anti-anxiety drug for them to make it through the procedure. Imagine placing a child in such a machine. Indeed, for really young children, general anesthesia may be required to do imaging studies such as CT or MRI scans, because they simply cannot stay still long enough.

All of this, however, is relatively peripheral to the real reason why using one’s children as a research subject in one’s own research protocol is unethical, and it’s an incredibly obvious one. Specifically, it’s a huge and unreconcilable conflict of interest:

The role of the parent is to protect the child,” said Robert M. Nelson, director of the Center for Research Integrity at Children’s Hospital of Philadelphia. “Once that parent becomes an investigator, it sets up an immediate potential conflict of interest. And it potentially takes the parent-child relationship and distorts it in ways that are unpredictable.”

Indeed. Some of these investigator-parents even sign their children’s consent slips for participation in the study, which is truly appalling from an ethical standpoint. How any scientist can justify that simply boggles the mind. It doesn’t matter if the study is benign and totally safe. It doesn’t even matter if both parents are in agreement. If one of the parents of a child being considered for a study is an investigator in the study, there is an unacceptable conflict of interest. The parent can’t be an objective advocate for his or her child because emotional and professional investment in the clinical research precludes it, and the researcher testing his own child cannot be an objective collector of data. Such parents may delude themselves into thinking that they are their child’s staunchest advocates, but they cannot be if there is a professional advantage to enrolling their own children in their study. They may also delude themselves into thinking that they can be objective when taking measurements based on their children, but they cannot and their bias runs a huge risk of tainting whatever observations they make. Yes, they may, as described in the article, justify to themselves enrolling their own children in their research protocols by arguing that the study must be perfectly safe, so safe that they see no problem in enrolling their own child into it. But that is not enough. They may also minimize risks and maximize their estimate of potential benefits–so much so that they might even go as far as this:

Some researchers sign required parental consent forms, and some have spouses sign.

“I sign my own permission slips,” said Gedeon Deak, whose three sons have participated in his cognitive science studies at University of California, San Diego. He tells review boards his subjects are a “sample of convenience,” not randomly selected, but he has seen no need to specify that his sons are among the participants.

“If they’re your kids and you want to ask them questions, you can,” he said. “If you want to put your kid into a drug trial, that’s different.”

No, it’s not. Really, it isn’t, except perhaps in degree. Worse, it’s arguably bad science, particularly in behavioral studies, because the pre-existing parent-child relationship can color how the children behave as research subjects and thus affect the actual results obtained in the study:

Dr. Toga said research had allowed his children to “come to work with Dad,” and get brain-scan printouts for show-and-tell. “They were so determined to please their father that they would lie still,” he said.

And how might that desire to please the parents affect research in other ways? Think about that for a moment.

Also think about this. Remember Dr. Jon Poling and his daughter Hannah? Of course you do. Hannah Poling is a child with a mitochondrial disorder who developed symptoms suggestive of regressive autism that her parents (and the antivaccine movement) blamed on vaccines. Less than a year ago, she became the poster girl for the “reinventing” of the antivaccine movement, which tried to use her case as “evidence” that vaccines cause autism in children with “genetic susceptibility.” Unfortunately, Hannah Poling was also the poster child for why parents should not be investigators who enroll their children in clinical studies that they are involved with. This is how I put it last year:

Dr. Poling was a co-author on a case report in 2006 about a girl with a mitochondrial disorder who underwent developmental regression after being vaccinated, a case report that has been widely cited as “evidence” for the claim that vaccines can cause autism in the presence of mitochondrial disease (which, conveniently enough, we’re told is much, much more common than previously thought). He disclosed neither that he was a plaintiff in the Autism Omnibus nor that the subject of the case report for which he was co-author was in fact his daughter, both serious breaches of reporting potential COIs, as far as I’m concerned.

What really shocked me about this article is that parents could not only enroll their children in clinical studies for which they were the principal investigator, but they could even do it without informing the institutional review board charged with protecting human subjects. The only word I can think of for this is appalling. If it is possible for a scientist to ethically enroll his own child in a clinical study for which he is the principal investigator (and I do not believe that it is, except perhaps under a few very limited situations), at the very minimum I would argue that full disclosure to the committee charged with overseeing the study and guarding the welfare of its human subjects would be a bare minimum requirement that would have to be met before researchers could enroll their own children in their own studies. Apparently it’s not.

If any of these scientists who have enrolled their own children in their own clinical trials happen to see this post, he or she will not be happy at what I am about to write next. Being as involved as I am in the movement to combat the antivaccine movement, I am reminded by parents such as those described in my article of another group of parents willing to turn their children into guinea pigs. I’m talking, of course, about parents who fervently believe that vaccines or the thimerosal that used to be in vaccines can cause autism. In their search for a cure, these parents will “study” their own children and subject them to woo after woo. Remember what Jenny McCarthy said about her son Evan: “He is my science.” Or consider Kent Heckenlively, who has subjected his daughter to quackery after quackery. Or what about the “studies” of Mark and David Geier using Lupron to treat autistic children? At least one of the parents on the actual dubious IRB that was supposedly charged with overseeing the study actually had a child whom she had enrolled on the study. If researchers can enroll their own children on their own study, why would it be wrong for parents with children on a study to be on the IRB overseeing that study?

Think I’m being harsh? Consider this: At least Jenny McCarthy and her ilk fervently believe that what they are doing is the only hope they have to “cure” their children of autism. They’re wrong–totally deluded, even–so much so that it’s not unreasonable to speculate that if they didn’t believe so fervently that biomedical quackery could cure their children of autism they wouldn’t subject them to it. If you put yourselves into their world view, you can see how they might think that the benefits of “curing” autism would far outweigh the risk involved.

What’s the excuse of investigators like the ones desscribed in this NYT article?

Comments

  1. #1 Petri
    January 22, 2009

    Yikes! From a science perspective I can’t see how testing your children makes any more sense than testing on yourself. Researchers who use themselves as subjects are generally viewed as quacks, why not those that use family and friends. The COI is insurmountable, so that not only do your risk your children, you risk them for no benefit, because your research will always have that dark shadow of conflict of interest obfuscating any real result. What a waste.

  2. #2 ehoffman
    January 22, 2009

    There’s an interesting article in this month’s issue of WIRED on a similar topic – a man whose daughter has a genetic disorder that’s never been seen before. The father happens to have an MD and was heavily involved in consulting for genetics firms, so he’s got a lot of connections, and has basically turned his life’s work into finding the cause of her symptoms. He’s even sequenced her DNA.
    I think this is the link.
    I can see how that would make for bad science… but I also feel for the father’s desperation. Is it more acceptable because he’s not a clinical PI and it’s not really a study? Or is this just as bad as the examples above? I’m not trying to be snarky, just honestly asking for your opinion.

  3. #3 skepticpedi
    January 22, 2009

    It’s beena while since my undergrad psychology classes, but upon reading this postI immediately recalled learning that B.F. Skinner had locked his daughter in an experimental “skinner box”. I remembered that she was severely damaged psychologically and even committed suicide. Of course I did about 1 minute of online research and discovered that this was an urban legend and that young baby Skinner grew up in perfect mental health. I’ve also recently had the opportunity to unlearn a large group of residents regarding a similar urban legend regarding Jaime Lee Curtis that all had been taught by college and medical school professors. I’m now left wondering how much of what we are taught by supposed experts is accurate and even more content with my skeptical approach to life.

  4. #4 Mario
    January 22, 2009

    Doesn’t this sort of thing go back to B.F. Skinner?

  5. #5 catgirl
    January 22, 2009

    If scientists use their own children in a study, that sample is obviously biased. The children are much more likely to do what their parents are expecting of them, even in cases that seem relatively objective.

    This is especially true for Deborah Linebarger, who is trying to use her children to study the effects of media. Any conclusion she gets from this ‘study’ is not useful because it could be caused either by the media, or the expectation of the children’s parent.

  6. #6 Mu
    January 22, 2009

    I can see a difference between “working on a cure for my child” and “getting scientific data for paper 259″. As a parent, I’m sure I’d do pretty much anything if I have a seriously ill child, and having a MD shouldn’t preclude other parents from doing so. Exploiting this for financial or scientific personal gain a la Poling is despicable, but not any worse than the other cases cited before (it’s not reported that Poling did any “experimental treatments”).
    Is there any authority supervising those IRBs btw? Reading some of the reports on the Lupron “study” makes me wonder if we’d accepted Mengele’s research if he’d just past the Auschwitz SS IRB beforehand.

  7. #7 Art
    January 22, 2009

    Certainly experimentation of one’s own children can be bad long before we get into ‘shock-the-monkey’ operant conditioning, implants and live vivisection.

    On the other hand, every child is an experiment of sorts. A semi random blending of DNA into a unique form which will go forth and produce results and make his/her mark on the world independent of any guiding PI or restraining experimental protocols.

  8. #8 vlad
    January 22, 2009

    B.F. Skinner daughter is one of the most respected people in the behaviorism field. In that field getting a good eval. from her elevates your status greatly.

    “For one thing, they require the injection of intravenous contrast agents, agents that can cause adverse reactions, sometimes even life-threatening.” No, not always. They are not required for some brain scans, wife’s had several. Also you can not compare MRI and CT scans. CT scans use a great deal of radiation, much more than a standard x-ray. MRI uses magnetic fields.

    Science wise I will agree that the validity of that subject may become questionable but not always. The safety COI is dependent on the study. IE: If you have a child with ADD/ADHD or ASD and your trying to do fMRI scans with regards to brain function (not requiring contrast agent BOLD) as long as the names are pulled I don’t see the issue.

    I don’t agree that this is a matter of degree issue with regards to drug trial. A contrast free MRI has no known risks, drug trials are there to determine if there is risk.

  9. #9 vlad
    January 22, 2009

    Skinner daughter is Julie Skinner Vargas.

  10. #10 BA
    January 22, 2009

    Skinner had two daughters, Julie Vargas, a leader in the field of behavior analysis as mentioned by vlad, and particularly accomplished in the study of verbal behavior. The daughter who used the baby tender (humidity and temperature controlled device that allowed for avoiding blankets and providing a clean safe place to sleep and such) as a crib (she was by no means raised in a box) is Deborah Skinner. A well adjusted artist who very much loved her father.

    I also know Dr. Sinha, a rather intelligent and personable neuroscientist at MIT. Pawan is a vision researcher and there is certainly a distinction between very closely following your child’s development (like Sinha, Skinner, Piaget) and performing intrusive experiments. The latter is very questionable, if not outright unethical, and the former informed observation. Both of my children have been in numerous experiments, mostly as developmentally matched peers to compare development in children with developmental disabilities. I do not feel there is anything wrong with their participation.

  11. #11 Tracy W
    January 22, 2009

    “True, the chance of such a reaction is very, very small, on the order of 1 in 100,000 or even lower, but why on earth would a parent subject his child to even that risk for no benefit to the child?”

    Parents subject their children to similar risks on a regular basis. If you ever take your child out onto a road, be that as a pedestrian, on a bike, or in the safety seat in the car, you are exposing them to the risk of being hurt or killed in a car crash. Now how many parents have taken their kids out onto a road for reasons other than the kid’s own immediate benefit (say dropping the kid off at daycare, or off to visit a sick relative or something).

    As for experimenting on your kids, there’s no clearly-established best practice for raising children – and a fair number of things the “experts” have fervently claimed in the past have turned out to be dead wrong. So every parent raising a child is experimenting.

    I agree with your concerns about the likelihood of bias in observations, making such experiments unreliable (of course, observations of your own children might be a way of generating some hypotheses for further testing). And there are parents who can go off the rails and hurt their children in the name of science. But if there’s something inherently unethical about exposing your children to even a very minor risk for no immediate benefit, then nearly every parent is unethical.

  12. #12 Heather
    January 22, 2009

    No, there is nothing inherently unethical about engaging in everyday activities with your children, but something judged to be unethical by current standards for test subjects in general does not become ethical because it is your own child.

  13. #13 Orac
    January 22, 2009

    What Heather said.

    I child out onto a road, be that as a pedestrian, on a bike, or in the safety seat in the car, you are exposing them to the risk of being hurt or killed in a car crash. Now how many parents have taken their kids out onto a road for reasons other than the kid’s own immediate benefit (say dropping the kid off at daycare, or off to visit a sick relative or something).

    The difference is that these are activities of daily life. It’s damned near impossible to exist in this society without being a pedestrian, riding in a care, etc. And, as you point out, there is often an immediate benefit to these things. Or, in the case of sports (which I’ve written about before), there’s the potential for better health, learning cooperation and teamwork, etc.

    But if there’s something inherently unethical about exposing your children to even a very minor risk for no immediate benefit, then nearly every parent is unethical.

    I’m sorry, but this doesn’t even make sense; you yourself said there often is an immediate benefit to activities such as going places in a car, bicycling, etc.

  14. #14 Despard
    January 22, 2009

    Behavioural psychologist here. I’m going to brave the possibility of being subjected to some not-so-respectful insolence and disagree with you, Orac. I think that in some cases it is perfectly ethical to use one’s own children in studies. I’ll cover whether it is bad science in a followup comment.

    To my mind it all depends on the kind of study. Say it’s a questionnaire study about kids’ food preferences run by a PhD student on a sample of children that happens to contain one of the PI’s own children. Nothing unethical about that – the kid only has to answer questions that have already been approved by an IRB for use on children.

    How about a behavioural psychophysics study on how kids perform reaching movements to various objects. This time you have a trial structure that involves the kid making repeated reaches while their movements are recorded. Again, as long as the study itself passes the ethics board, no ethical problem.

    Where I think it gets tricky is in two distinct areas. The first is when the study itself may cause some physical or psychological harm to the child. I think that in these instances parents should recuse their own children from taking part as there is obviously a conflict of interest (their interest in the science vs. the kid’s interest in not being harmed). It’s hard enough to get these kinds of studies past ethics boards in the first place.

    The second is where the presence of the parent in the study may cause psychological harm to the child, irrespective of the study itself. For example, the need of the child to please the parent might cause them to feel worthless and unloved if they don’t do the task ‘properly’, or whatever. This can be avoided in the study design, however.

    One case I heard of where I think that using their own child was unethical: a researcher (I forget the details, sorry) wanted to know whether children could understand pictures if they grew up in an environment where they had never been exposed to any images. This meant censoring TV, covering cereal boxes, etc. Really hard work. At the end of the study he let his now 3-year-old see pictures for the first time, and the kid had no problem in recognising them. I think that was unethical because the researcher didn’t know the outcome from the beginning (although obviously he had a pretty strong hypothesis) and it could have resulted in real psychological harm to the child if he was wrong.

    In this I agree with Heather’s comment above:

    No, there is nothing inherently unethical about engaging in everyday activities with your children, but something judged to be unethical by current standards for test subjects in general does not become ethical because it is your own child.

    So to sum up, I really do think it depends on the kind of study you’re doing. Some are ethical, some are not.

  15. #15 Despard
    January 22, 2009

    Now to tackle the ‘bad science’ part of the argument.

    Being aware of bias, as we know, is an integral part of being a scientist. The main argument here is that introducing an extra source of bias into an experiment in the form of one’s own child is a bad thing. I generally agree with this.

    But then, again, you have to consider the particular experiments that are being performed. Let’s go back to our PhD student or postdoc running the questionnaire study with the PI’s kid (see above). Now, it’s entirely possible that the kid has been primed by the PI as to how to answer the questions, but I consider it highly unlikely for two main reasons. First, a good scientist would be aware of his or her own potential to bias the study and avoid even a mention of it with the child previously. Second, it would be extremely difficult to prime someone on how to answer a questionnaire study, even unconsciously.

    The biggest source of potential bias in this case, assuming that the parent/PI is a food researcher, is that the kids may have been taught from an early age about ‘good’ and ‘bad’ foods, to the extent that that biases their responses. Maybe they want to answer in a certain way to please their parents. But even if that’s the case, so long as the cohort does not only consist of the PI’s kids and is actually part of a random sample, it doesn’t really affect the results at all. And if you’re worried, you can always do a statistical test to work out whether the PI’s kids responses varied hugely from the mean of the responses as a whole, in whatever direction.

    Second example: the movement study. Again I find it very hard here to work out how kids could be biased to move in a certain way, thus affecting the results, unless the parents literally trained them to do it beforehand! In this case I don’t think that including your own child, as long as it’s as part of a good random sample, is a problem at all.

    If the person is running the study on his/her own child, however, I’d be a bit more concerned. Even if that were the case, the use of appropriate controls and an extra check to make sure that the child wasn’t performing significantly differently to the other children would reassure me.

    Just a note by the way on Petri’s comment:

    Researchers who use themselves as subjects are generally viewed as quacks, why not those that use family and friends.

    I disagree with your assertion that researchers who use themselves as subjects are generally viewed as quacks. For example, many vision researchers experiment on themselves and often publish papers with only themselves and their co-authors as subjects. This way of working is an established and accepted norm in the field. (I realise that this sounds a bit like ‘hundreds of vision researchers can’t be wrong, can they?’ but the point is that they are not viewed as quacks, certainly not in the psychological community.)

    So what would constitute bad science then? Well, the sort of thing that always does: shaky methodology, unfalsifiable hypotheses, and relying on your preconceptions to drive your interpretation rather than what the data tell you. Those are the mistakes that the Jenny McCarthy et al. crowd make. Good scientists should not make those kind of mistakes – and should be doubly careful to avoid them when using their own children as subjects.

  16. #16 Mike
    January 22, 2009

    As was said earlier, every child is an experiment from an observational point of view. Two children, 20 months apart in age: One grows up disliking books, the other devours them. What were the differences that might have contributed to that? The reader has very limited musical talent but the non-reader is a virtuoso violinist. Another layer of questions arises.

    The small number of data points makes any theories or conclusions almost impossibly worthless. If that is the case, what would be the motivation to create harmful conditions for some or all of one’s children?

  17. #17 daedalus2u
    January 22, 2009

    The recent winner of the Nobel prize for discovering that Helicobacter pylori caused ulcers experimented on himself by drinking a culture of the bacteria. That is not an experiment that an IRB would be likely to approve because it does pose non-negligible risk. It also is not likely to be consistent with the Declaration of Helsinki. It would be completely wrong for him to have given the bacteria to drink to anyone else without IRB approval.

    People can subject themselves to things which can not be done to them experimentally, for example cosmetic surgery, sky diving, drug abuse, kidney donation, partial liver donation.

    If you are going to do an “experiment” on anyone but yourself, you have to follow the Declaration of Helsinki. In the US, this means you must go through an appropriate IRB. Even if the experimental subject is your dependent, you must go through an IRB. To not do so makes the experiment unethical and it should be rejected as such by any reputable journal.

    Going through an IRB is to prevent exploitation of the subjects. Children can be exploited by their parents too. If a parent does use their child as an experimental subject without going through an IRB, then the child will have the right to sue for being exploited when that child comes of age. The child can sue the institution that the parent worked at, not just the parent.

    There is a difference between the ethics of treatment vs. the ethics of experimentation. Regarding Jonas Salk giving his children polio vaccine early; that is treatment not necessarily experimentation. If he subsequently exposed his children to polio, that would be experimentation and would be highly unethical (and I am quite sure he didn’t do it).

    Regarding using your own children as controls; if you do, then they are not blinded. In a behavioral study, the PI might perceive their own child’s behavior as “more normal” than it actually is, simply due to familiarity. In terms of instrumental studies, i.e. blood chemistry for example, again, it isn’t a blinded or “random” sample. If the same PI uses the same controls for many experiments, that isn’t good science.

  18. #18 The Perky Skeptic
    January 22, 2009

    The NYT article made my skin crawl, too. I suspect I’m just too close to the issue to think rationally about it. That said, I cannot help but think that any experiment involving one’s own child is the very embodiment of selection bias, innit? :)

  19. #19 Orac
    January 22, 2009

    Where I think it gets tricky is in two distinct areas. The first is when the study itself may cause some physical or psychological harm to the child. I think that in these instances parents should recuse their own children from taking part as there is obviously a conflict of interest (their interest in the science vs. the kid’s interest in not being harmed). It’s hard enough to get these kinds of studies past ethics boards in the first place.

    But this irredeemable conflict of interest is what makes the act of a parent enrolling his child on a clinical study he is doing inherently of dubious ethics at best. One of the major principles of human subjects research, aside from beneficence, is autonomy. The major documents defining the ethical conduct of human subjects research, the Helsinki Declaration and the Belmont Report, both emphasize that potential subjects must be free to accept or refuse offers to participate as subjects in research without fear of reprisal or that they will not get standard of care treatment. Moreover, both documents identify children as vulnerable population for whom research protections must be even more stringent, because they cannot decide for themselves.

    A parent who is also the researcher doing a study seriously endangers that protection for his child because of the inherent conflict of interest. It doesn’t matter if the human subjects study is completely safe; it’s still of questionable ethics.

    The reason is simple. Because children cannot decide for themselves, their parents must decide for them, the assumption being that a parent has the best interests of the child as his or her primary consideration. When the parent plays two roles, it is no longer a given that the parent’s conflict of interest won’t color a decision to give permission for the child to participate. Moreover, because the human subjects protections for children must be greater and more stringent than they are for adults, the argument that an adult (as in the H. pylori example) can choose to risk himself doesn’t fly, any more than it flies when parents choose faith healing over medicine for their child.

    The bottom line is that this is a very clear cut ethical problem, and it’s a very high bar for it to be overcome.

  20. #20 rose
    January 22, 2009

    While I agree that there are ethical and scientific concerns involved in allowing one’s child to participate in research project where the parent is the PI, I can’t accept that there is any level of similarity between the risks associated with participation in a clinical trial and the risks associated with performing some memory tasks on a computer (or any other type of cognitive test) or a non-contrast fMRI scan. If I wasn’t prepared to have my own child participate in an fMRI scan or a cognitive test, then as a scientist I wouldn’t feel right asking other parents to do the same for my studies.

  21. #21 Orac
    January 22, 2009

    The level of risk or safety is almost irrelevant. Autonomy is viewed as one of the ethical pillars upon which any human subjects protection must rest, and parents’ conflict of interest seriously undermines both autonomy and the higher level of protection accorded to children in research. It truly doesn’t matter if it’s a drug trial or a test of doing memory tasks on a computer. The COI undermines subject autonomy.

  22. #22 perceval
    January 22, 2009

    There’s a big difference between observational studies that are clearly designed and described as such and formal experiments, and there’s another difference between simple behavioural or cognitive experiments and more invasive studies. If the parent could in any way consciously or subconsciously influence the behaviour of their child in an experiment, they shouldn’t encourage their children to take part. I also agree wrt MRIs.

    However, in other fields and for other purposes, I think studying your child is justified. Much of what we know about child language development we owe to mothers and fathers who diligently recorded and noted down their children’s utterances. To a much lesser extent, I also do this with my own kids; I’ve been blogging their language development and have been on the lookout for evidence of interesting phonological developments. I’ve also been known to covertly get my daughter to say a certain word to test a hypothesis about the linguistic rules she’s developing. Researchers are very conscious of observer bias, and of the potential to seriously annoy your offspring. Collecting a good database of child language data is tricky unless you have the unique access of a parent and many of the observationally generated hypotheses have in the mean time been tested experimentally. If parents publish about their own kids, this is always, ALWAYS fully disclosed and acknowledged.

    (Full disclosure: My minor was in linguistics, and I’ve taught language development, although I’m not a child language expert.)

  23. #23 G Felis
    January 22, 2009

    Firstly, I feel compelled to reference a recent XKCD that is hilarious and wrong, and highly relevant to this post.

    Secondly, and more substantively: The excessively rigid and hyper-legalistic nature of Orac’s most recent response troubles me, even though he is surely right – and the reason for this self-contradictory response rests in distinction missing from the discussion so far. Mommy or Daddy saying “Let’s play a computer game!” which also happens to be a memory task experiment is clearly no more unethical than “Let’s play a computer game!” when it’s an educational game that focuses on memory tasks (but which is not part of an experiment)- in an ordinary sense of the ethical in terms of right and wrong, harmful and harmless. However, from the standpoint of professional ethics and the ethical reasoning which underpins scientific experimentation – IRBs, Helsinki, etc. – it is obviously and incredibly problematic for an experimenter to have parental authority over whether or not a subject child participates in an experiment: As Orac says so clearly, this undermines the protection of autonomy at the heart of experimental ethics – which is a matter of insitutional practice and professional conflict of interest, not simply a matter of immediate and specific case-by-case ethical considerations.

  24. #24 desiree
    January 22, 2009

    i agree with orac here. sure, it’s not always unethical, but it’s just more likely to be. and it’s not like there’s any shortage of kids out there, so no one really has to use their own.

    i think the burden of proof, so to speak, is on the researchers. why is it necessary to use your own kid?

  25. #25 vlad
    January 22, 2009

    “The COI undermines subject autonomy.” Children do not have autonomy, by law. So this would be the same issue if a parent were to enroll their child in an experiment, no matter how safe when the parent is not the PI. Using this argument children should not be used in any experimentation ever. How does one develop therapy for children without involving them in testings? ABA or any other behavioral or cognitive therapies have to be tested on children eventually. I maybe now or later but the first time a therapy is tried on a child (who can not by law consent) your are experimenting on someone who can not consent.

  26. #26 IBY
    January 22, 2009

    The earliest incidence of child experiment I can think of is John Watson’s highly unethical experiment. Though the childe wasn’t his, it was another woman’s.

  27. #27 Orac
    January 22, 2009

    “The COI undermines subject autonomy.” Children do not have autonomy, by law. So this would be the same issue if a parent were to enroll their child in an experiment, no matter how safe when the parent is not the PI. Using this argument children should not be used in any experimentation ever.

    No, it doesn’t. You’re too hung up on the word “autonomy” in the sense used for an adult. Think of it this way. In essence, the parent is the guardian of a child’s autonomy vis a vis whether or not to enter into a research protocol. The conflict of interest involved in being both parent and PI is such that it is very, very difficult for a parent to exercise autonomy of consent on behalf of the child with regard to deciding whether to enter that child on a study. Remember, the Belmont Report and Helsinki Declaration demand greater protections for vulnerable populations such as children than they do for competent adults.

    Even more so than adults, when it comes to children, any aspect of a trial that might compromise ethics must be minimized or, if possible, eliminated. A conflict of interest as glaring as that of a parent being both parent of a research subject and PI of the protocol that the research subject is taking part in is easily eliminated by not enrolling the child on the protocol. Because the COI is so easily eliminated, to justify not eliminating it requires an even more compelling benefit that can come from including the child. Admittedly, this can cut both ways, because for some forms of research it can be viewed as a benefit for the child to participate, making the argument that a child should not be denied the opportunity to participate just because his parent is the PI, but these examples are few and far between.

  28. #28 Sam
    January 22, 2009

    Vlad,

    Sure any child does not have autonomy, and so can not decide for themselves whether to participate in an experiment. However, the parent does have autonomy, which, presumably is directed in the best interests of the child. Thus if they notice an unacceptable amount of distress in their child they can cancel the experiment, without any tangible consequence to themselves.

    For a parent who is also the experimenter, there is the conflict of interest in that they face the consequence of losing potential data. Now, I would argue that most scientists would put their child’s well-being before the collection of data, but the conflict of interest can skew perception, so why take the risk?

    I think Orac’s latest argument is the most sensible, and should have been the main thrust of his original post.

    Talking about the slight risk of MRI was a red herring. By the way, as a parent I would by happy to give an MRI to a mature child that requested one. The learning benefit far outweighs the risk IMHO. I just would not use it for scientific data.

  29. #29 BA
    January 22, 2009

    daedalus says:
    “Regarding using your own children as controls; if you do, then they are not blinded. In a behavioral study, the PI might perceive their own child’s behavior as “more normal” than it actually is, simply due to familiarity. In terms of instrumental studies, i.e. blood chemistry for example, again, it isn’t a blinded or “random” sample. If the same PI uses the same controls for many experiments, that isn’t good science.”

    Not knowing which child produces which data set is a good starting place to guard against bias but one does not necessarily need or want a random sampling for certain types of observational research. Sometimes we are looking for typical samples or attempting to discover the nature of the phenomenon at hand. For psycholgists working with developmental issues, we want to see what kids do in specific situations. The assumption of normally distributed responding is a slipshod notion with psychological contructs. True, large unbiased samples are critically important in large scale medical or psychological treatment research but bringing components of interevention or research that seeks to discover laws of nature are more exploratory and do not necessarily begin with test vs control groups.

    That said, a good observational study blinds raters and the coding in many of these studies remains blinded. Observational studies, particularly regarding development, are very important to inform good educational practices. Looking at the behavior of children in the context of play with a peer or some such scenario allows us to gauge typical responding. We may have somewhat good information on the developmental milestones children meet and when but for what it is that a typical child does in certain situation is critical and not readily available. Take joint attention as an example. We know what it is and when it develops but not how it is typically displayed (at least we don’t know this very well). We know it is deficient in individuals with autism but do we teach them to do it all the time or should our expectations take into account what we see with typical kids? And different kids will respond differently. Once we make some observations to better understand the phenomena we are dealing with we can investigate on a larger and larger scale.

    Does the chemist not first start with producing a single reaction and then replicate directly and systematically that reaction to discover its limits? There is not just one kind of research that is scientific. Though I’m certainly in favor of not subjecting any persons to undue risk, I am also not opposed to studying the behavior of those I know.

  30. #30 Orac
    January 22, 2009

    I think Orac’s latest argument is the most sensible, and should have been the main thrust of his original post.

    Yeah, you’re probably right; in retrospect, I now agree that I should have emphasized that aspect more.

    Still, I’m rather surprised at the vehemence of some of the commenters defending parents who enroll their own children in research studies for which they are the PIs. I would have thought that the conflict of interest embodied in such behavior. Perhaps it’s the difference between a physician and clinical researcher’s viewpoint and that of basic scientists who don’t think their human subject research has any potential for harm, being pure research rather than the testing of interventions. Even so, human subject research is human subject research, and autonomy is paramount. In retrospect, perhaps I should have used the word frequently used in the Belmont Report, “voluntariness,” and likely that would have provoked less of a response. It’s an ugly mess of a word, though.

  31. #31 vlad
    January 22, 2009

    “Because the COI is so easily eliminated, to justify not eliminating it requires an even more compelling benefit that can come from including the child.” Here I agree whole heartedly, just don’t think it should be frowned upon to this extent (comparing putting a camera on your babies head to Mengle). Some of the studies may not be scientifically rigorous as we would like or have much of a point, a classification that can be used on more than a few IRB approved research studies. A few profs from my undergrad career come to mind.

  32. #32 bsci
    January 22, 2009

    Going back to Petri’s first comment, ” Researchers who use themselves as subjects are generally viewed as quacks.” This is false. For a good (though opinionated) history of self-experimentation, read “Who Goes First” by Lawrence Altman
    http://www.amazon.com/Who-Goes-First-Self-Experimentation-Medicine/dp/0520212819/ref=sr_1_1?ie=UTF8&s=books&qid=1232661649&sr=1-1

    As for experimentation on children, let’s flip the question. You are doing a longitudinal study of brain development on children. It has passed IRB approval and requires a structural MRI data every year from early childhood to adulthood. There are no MRI contrast agents being used. The data is static and there is no way for the parent/child relationship to bias data collection (except perhaps the child is more or less motivated to keep still). This example is extremely similar to some of the studies mentioned in the article.

    Assuming there is a slight, unknown risk to MRI, is it ethical NOT to consider using ones own children while recruiting others children? There might be legitimate reasons why a researcher’s child shouldn’t participate in a specific study, but assuming a study design can’t be biased by the parent/child relationship and there are no child-specific contraindications, why is it ok to allow others to take risks with your children that you are not willing to take yourself?

  33. #33 Karl Withakay
    January 22, 2009

    “Mommy or Daddy saying “Let’s play a computer game!” which also happens to be a memory task experiment is clearly no more unethical than “Let’s play a computer game!” when it’s an educational game that focuses on memory tasks (but which is not part of an experiment)”

    Not so clear to me. What if the child says, “I don’t want to play a computer game, your games suck!”? What is the parent/PI’s motivation for pushing the child to play anyway. What if the memory games causes stress for the child because they don’t do well? The parent/PI has a conflict of interest in regards to whether or not to push the child to complete the game in the experiment, whereas with the game, the parent’s only concern the the benefit and welfare of the child.

    I don’t know whether to groan or laugh every time I see someone post, “Researchers are very conscious of observer bias”, implying that the bias will not get in the way of obtaining good, meaningful data. The whole point point of double blinded RCT’s is that either researchers aren’t as aware of their biases as the think, or that they aren’t able to fully overcome those biases in spite of their awareness.

    The conflict of interest for parents exists outside the experimental world, of course. Do you want your child to star in the play/ be a doctor/play QB for their benefit, or because you want the pride & prestige of being the parent of the star/doctor/QB?

  34. #34 Karl Withakay
    January 22, 2009

    bsci,
    “Assuming there is a slight, unknown risk to MRI, is it ethical NOT to consider using ones own children while recruiting others children?”

    The parent/PI is not in an objective, unbiased position to make that judgment, that’s the point.

    I think your argument actually works against you. That is exactly the kind of decision a parent/PI might make, but it isn’t made in the best interests of their child, it’s made in a sense of honor and fairness to all the other subjects. You have just put the rest of the world’s interests above those of your child.

  35. #35 Laura
    January 22, 2009

    “…The small number of data points makes any theories or conclusions almost impossibly worthless. If that is the case, what would be the motivation to create harmful conditions for some or all of one’s children?”

    Outside the research sector bride prices are a motivation some parents have to create harmful conditions for their female children. Could there sometimes be a financial motivation within research too?

    “i think the burden of proof, so to speak, is on the researchers. why is it necessary to use your own kid?”

    Because you can’t spank other people’s kids for disobeying you the way you have the right to spank your kids for disobeying you?

    “‘Mommy or Daddy saying ‘Let’s play a computer game!’ which also happens to be a memory task experiment is clearly no more unethical than ‘Let’s play a computer game!’ when it’s an educational game that focuses on memory tasks (but which is not part of an experiment)’

    “Not so clear to me. What if the child says, ‘I don’t want to play a computer game, your games suck!’?”

    Then you can spank the child or make the child sit in the corner for a time-out or take away the child’s computer privileges for a week or whatever, but only if you’re one of the child’s parent/guardians yourself

  36. #36 bsci
    January 22, 2009

    Karl,

    What if the non-researcher parent consents and children of researchers also get some extra confirmation through the IRB to make sure that there’s no improper pressure?

    The decision is made out of a sense of honor and fairness, but why is that a problem? If a researcher thinks that a study is reasonably safe, why should they have any problem with including their own children in the sample? If researcher thinks a study isn’t safe enough for their own children, why should they be willing to allow other people’s children to assume that risk? This doesn’t mean researchers’ children must be included in all studies, but if you can’t honestly say you’d be willing to include your children in study, why is that study being done?

  37. #37 Karl Withakay
    January 22, 2009

    “He tells review boards his subjects are a “sample of convenience,” not randomly selected…”

    Am I the only one who thinks an IRB should at least raise their eyebrows at this phrase?

    “Sample of convenience,” reeks with at least the possibility of a vulnerable population such as orphans, the homeless, military personnel, persons in mental institutions, etc. (as in Identical Strangers, Tuskegee airmen, MK ultra, etc)

    “…but he has seen no need to specify that his sons are among the participants.”

    If it’s not an issue, why not mention it? (He’s obviously aware that some people think it matters.) Is he afraid the IRB might not agree with him?

  38. #38 Joshua Zelinsky
    January 22, 2009

    Doing research on oneself isn’t necessarily a sign of crankdom. In some cases people have done so because it would be clearly unethical to subject someone else to the experiment in question. The most obvious such example is Werner Forssmann who later got a noble prize for the work he did which started by him shoving a catheter into his own beating heart.

  39. #39 Karl Withakay
    January 22, 2009

    bsci
    “Then you can spank the child or make the child sit in the corner for a time-out or take away the child’s computer privileges for a week or whatever, but only if you’re one of the child’s parent/guardians yourself”

    I really think you’re making my argument for me.
    Is your motivation to discipline the child, or get your research done? Is your decision to spank in the best interest of the child or your study?

  40. #40 Despard
    January 22, 2009

    Just to make it clear, I’m not advocating PIs using their own children in clinical trials at all. I’m noting that in many behavioural studies, there is really no ethical problem with using your own kid.

    This comment:

    Not so clear to me. What if the child says, “I don’t want to play a computer game, your games suck!”? What is the parent/PI’s motivation for pushing the child to play anyway. What if the memory games causes stress for the child because they don’t do well? The parent/PI has a conflict of interest in regards to whether or not to push the child to complete the game in the experiment, whereas with the game, the parent’s only concern the the benefit and welfare of the child.

    is fair enough. I guess it’s up to the parent to decide whether they want to push the child to perform the experiment. That, to my mind, is the ethical question. If the kid is happy to do it and the experiment itself has passed ethical review, I just don’t see the big deal.

    I will think further on the conflict of interest problem.

  41. #41 Karl Withakay
    January 22, 2009

    Whoops, meant to address that last one to Laura, this one’s for bsci.

    “The decision is made out of a sense of honor and fairness, but why is that a problem? If a researcher thinks that a study is reasonably safe, why should they have any problem with including their own children in the sample?”

    Because they have a conflict of interest. There’s a world of difference between being willing to include their children in the sample and actually doing it. If a PI is asked if they are including their own children, the response should be something like, “I’d be willing to include my son, but it’s a conflict of interest to be both the investigator, and the one performing the experiment, because I might be willing to do things to my own child that I was reluctant to to do someone else’s, such as spank them. If I treat the subjects differently, I skew the data.”

  42. #42 Karl Withakay
    January 22, 2009

    “That, to my mind, is the ethical question. If the kid is happy to do it and the experiment itself has passed ethical review, I just don’t see the big deal.”

    Who judges if the child is actually happy to do it, or if they are inwardly miserable but just trying to make mommy or daddy happy? The best way to avoid the conflict of interest is to eliminate its possibility by just not going there.

    The big deal, I guess is that the world doesn’t have big red lines painted on the ground letting us now when we’re about to cross one. It doesn’t go from black to white, from play a game to electrodes in the brain; there’s infinite shades of gray in the middle.

  43. #43 Mu
    January 22, 2009

    There might also be a big legal pitfall. If your kid is the one in ten thousand who develops meningitis on your new vaccine, the hospital insurance carrier or the vaccine injury program will be seriously unhappy about that consent form signed by the PI parent.

  44. #44 bsci
    January 22, 2009

    Karl,

    Who judges if the child is actually happy to do it, or if they are inwardly miserable but just trying to make mommy or daddy happy? The best way to avoid the conflict of interest is to eliminate its possibility by just not going there.

    This true for ALL child consents? Why can any child participate in a study? If a parent likes reading about science and is excited to have their child participate in a science study, there is also pressure on the child. There might a difference from the research/parent pressure, but the pressure is still there and still real. If the child misbehaves during the study, they might be punished. Do all studies of children have followups to see how parent treat misbehaving children afterwards? What about payment? The money is usually nominal to minimize this type of behavior, but if a misbehaving child means that the family won’t be getting money from repeat visits that’s also pressure.

    There are also studies where there is no possibility of data skewing. For MRI structural scans, a person either stays still and the data is usable or they move too much and the data is garbage. I feel rationalization goes both ways. It’s easy to say that you’d be willing to include your child if there’s no way that your child will actually be included. Someone who’s actually considering including their child might think a bit harder about safety when designing a protocol.

  45. #45 Sam
    January 22, 2009

    bsci,

    My understanding is that the conflict of interest problem we are discussing regards other interests that could be expected to skew the parent’s ability to look out for their own child’s interests. I think it is pretty clear to most of us how money, or obtaining convenient data, could weigh against the parents responsibility to the child. Few of us would see a devoted Science reader as facing the same problem–but sure it is not a black and white thing. I guess there could be a few parents that would actually disregard a child’s distress because they really want to advance science by the smallest of increments. I have never heard of one, however.

    For other posters regarding skewing data:
    I don’t think the problem about parent PIs is at all about potentially skewing data, which is a separate concern regarding experimental design.

  46. #46 Karl Withakay
    January 22, 2009

    I mentioned skewing data as one of the potential pitfalls; I feel strongly that that pitfall is at best, a secondary concern to the COI regarding the child’s welfare.

    It could also be said that if a researcher is willing to include their own child in a study, that could influence another parent to defer to the researcher’s judgment rather than their own in regards to whether to allow their own child to participate. “If it’s OK for his kid, I won’t worry about my kid.” If the investigator’s judgment is skewed by his desire to do the study, his COI has not only compromised his judgment, but the judgment the other parent(s) as well.

    bsci, surely you can admit there’s likely to be more perceived pressure on the child if they know it’s mommy or daddy’s study they are participating in?

    The decision to participate in the study is not a decision the child can make. The decision therefore falls on their guardian as the protector & custodian of their welfare. The combo PI/parent has dual interests, the welfare of their child, and their study. That is an inherent conflict.

    In regard to the concept that the researcher should not do anything to anyone else’s child they wouldn’t do to their own, I ask the following: is OK to experiment on a child without their parents’ consent as long as the PI performs the same experiments on their own child? Obviously not, the parent must be allowed to make that decision for the child, not the investigator. The decision must be made by someone who has only the child’s interests at heart.

  47. #47 G Felis
    January 22, 2009

    Karl Withakay: What seems to be unclear to you is not my example of a harmless experiment, but what you added to my example without any real warrant. As far as I can tell, you’re suggesting that it’s impossible on the face of it for any experiment to be harmless because a parent who is also an investigator might be such a jerk that the desire to have an experimental subject would cause them to punish their child if the child proved uncooperative. For my part, I don’t think it’s unreasonable to assume, just for the sake of argument, that at least *some* scientist parents might actually love their children and not be such raging assholes that they would punish their children at the first sign of resistance to participating in an experiment. In which case, an experiment in which a child engages in a fun educational activity which is exactly like an activity they might otherwise engage in (except that data is collected on their performance) would indeed be harmless. I also didn’t make a universal claim that no experiment like my sketchy example could never ever be harmful, so generating counter-examples advances no logical point.

    In any event, the substance of my argument – which you seem to have quite missed in your zeal to quibble – was that the question of whether any given experiment is potentially harmful is ultimately quite separate from the fundamental violation of the subject autonomy principles at the heart of research ethics. It is the matter of subject autonomy which generates the very real conflict of interest problem Orac is talking about – making all this focus whether an experiment is harmless or not a red herring.

  48. #48 DRK
    January 22, 2009

    “Dr. Sinha’s wife, Pam, although a scientist herself, did not rest easy.

    She was “quite opposed to this idea of experimentation” on their son, Dr. Sinha said, so “it had to be done surreptitiously, whenever she would go out or when I would take him out in his little BabyBjorn.”

    It is still, he said, “a sore topic between us.””

    Dr. Sinha was strapping a camera on his baby’s head to see what the child looked at. Non-invasive, yes, but how often is a child the subject of experimentation in spite of the stated objection of one of its parents? This seems wrong to me. (It seems, too, as if the data would be quite affected by only recording when the mother was not around, given that babies spend a lot of time looking at their mothers).

  49. #49 Karl Withakay
    January 22, 2009

    G Felis,
    You are in error if that is actually your understanding of what I am saying, otherwise thanks for the straw man.

    I don’t believe I was the one who brought up punishment at all; I believe Laura brought that up as an option. Punishment actually never occurred to me until Laura brought it up.

    I agree totally with what Orac has said above; I can’t think of anything he has said in this post that I disagree with..

    My point is basically that coming up with examples in which the experiment would be harmless is the red herring, partly because one can always come up with counter examples. That would seem to either require two different protocols, one for no potential of harm whatsoever where the PI can include their own child, and one for any potential of harm (and someone to decide which is which), or one protocol for all situations that does not allow the investigator to include their child in the study, or at a minimum includes full disclosure of the inclusion of the investigator’s own child in the study and extra rigorous review/supervision by an IRB or other third party.

  50. #50 bsci
    January 22, 2009

    Sam and Karl,
    There are real cases where the pressure on non-PI parents is even more than PI parents. Let’s say a child has ADHD or autism. If they sign up for a study they get to be observed and diagnosed by some of the world’s experts for free. At the end of the study, they will receive the best available treatment for their child. All they need to do is sign up for a bunch of tests or drugs that have some level of risk. That’s much more pressure than anything in the PI/child relationship.

    As for pressure, is there any more pressure if it’s mommy or daddy’s study or a friend of mommy and daddy or a co-worker of mommy and daddy or a big important person who mommy and daddy say can help me. Not that this pressure can’t create conflicts of interest, but to exclude all potential conflicts of interest using blanket rules would stop virtually all child research. This needs to be handled on a case-by-case basis and if people seem to be participating in a study for the wrong reasons, they need to be excluded.

    I understand your example of a PI/parent letting science cloud their own child’s safety and allowing that to bias other’s decisions. First, I’m not sure how often this happens compared to the opposite situation where a PI might not think if she’d be willing to include his own children. That said, there’s an easy solution to this. It is a HIPPA violation to publicly release the names of study participants. That includes one’s own children. There is no situation where a PI should tell a potential study participant whether one’s own’s children are involved. That would both be a biasing of the consent process and a HIPPA violation. If anything that’s an issue with some of the NY Times examples.

  51. #51 bsci
    January 22, 2009

    Karl,
    That would seem to either require two different protocols, one for no potential of harm whatsoever where the PI can include their own child, and one for any potential of harm (and someone to decide which is which), or one protocol for all situations that does not allow the investigator to include their child in the study, or at a minimum includes full disclosure of the inclusion of the investigator’s own child in the study and extra rigorous review/supervision by an IRB or other third party.

    You have more than two situations here. There are many scenarios and many of them actually happen. There needs to be strong oversight and a uniform review system, but you can’t make blanket rules on this topic.

  52. #52 Karl Withakay
    January 22, 2009

    bsci
    Are you sure that would be a HIPPA violation? I can release my medical information to anyone I want to. I can consent to anyone else releasing my medical information under whatever conditions I choose to allow as well. (Presidential candidates do both every election cycle.)

    Presumably a minor’s guardian has these same rights.

    Do I not have the right to allow an investigator to release my name as a study participant if I so choose? Certainly I have the right to announce I am study participant. (Assuming I didn’t sign an NDA agreeing not to)

    If the PI is also the guardian of the minor, do they loose their HIPPA rights to release the minor’s medical information or allow it to be released by a third party?

    If I can announce that I am in a study, I can also announce that my child is in a study. Does that change if I am an investigator in the study?

    As to your other point, I’ll defer to a researcher such as Orac on whether I’m being too absolute and rigid in my position.

    I’ll just say, that while it does not specifically address this subject, the Declaration of Helsinki kind of seems like a blanket guideline, if not rule, to me, specifically to take the judgment call out of certain situations.

    I’ll also defer to Orac as to whether you can have a uniform review system without at least some blanket rules on some topics or not.

    *** I’ll just note here that while I’m leaving that last line in, I acknowledge that you did not advocate no blanket rules, just no blanket rule on this topic.

  53. #53 JuliaL
    January 22, 2009

    Excellent post.

    In the parent-child relationship, the parent’s focus needs to be on doing what is best for the child. Even though a young child may not be able to name the problem, there is always a danger that he/she will feel used by the parent as a convenient object rather than related to as a deeply-loved person. Not healthy parenting, no matter how benign the experiment itself may be.

    Though I had a close and loving relationship of more than sixty years with my now deceased father, I can still remember one point in our relationship when I felt unsafe, unprotected, somehow just not right: When I was five, he decided to see how long a word he could get me to memorize the spelling for. I promptly succeeded in the first word he chose: “antidisestablishmentarianism.” It was an easy and apparently benign experiment.

    But as an adult, I realize what that unpleasant feeling was. For just a little while, I felt pushed away, the real me ignored, used like an object. No doubt some children used in experiments by their parents don’t feel the break in the normal personal relationship. But others may, no matter how innocent the experiment seems.

  54. #54 Laura
    January 22, 2009

    “‘Then you can spank the child or make the child sit in the corner for a time-out or take away the child’s computer privileges for a week or whatever, but only if you’re one of the child’s parent/guardians yourself’

    “I really think you’re making my argument for me.”

    Thank you. ;) I’m glad you got that I was describing an argument, not agreeing with it. :)

    “Is your motivation to discipline the child, or get your research done? Is your decision to spank in the best interest of the child or your study?”

    More good questions.

  55. #55 bsci
    January 22, 2009

    Karl,

    Perhaps I’m wrong on the HIPPA violation. At the very least, the consenting process includes a list of risks. In my mind, telling the guardian of a potential participant that the risks are negligible enough that you let your own child participate, would be a fundamental alteration of that section of the consent form that would need to be formally processed through an IRB for approval (and would probably not get approved).

    I do agree that there are situations where blanket statements make sense, but whether a PI’s child is allowed in a study is not one of them. The declaration of Helsinki understands this and most of it is focused on goals and guidelines rather than specific things that are or are not permitted. There are certain studies where death is sometimes an acceptable risk and others were a minor rash is unacceptable. Both can fit within the declaration.

  56. #56 Broken Link
    January 22, 2009

    Orac,

    You aren’t the only one to find the Poling case appalling. The journal editor did too, and went so far as to publish a letter, indexed in pubmed, to that effect:

    http://jcn.sagepub.com/cgi/reprint/23/9/1090.pdf

    It takes a lot to get a journal editor to speak so frankly.

  57. #57 DLC
    January 22, 2009

    Okay, so let’s say Dad’s a developmental psychologist.
    Let’s say he wants Junior to take a skills test.
    Let’s further say that Dr Dad thinks Junior should score 90 of 100 on the test. What happens when Junior only scores 70?

    What happens to Junior when he sees Dad’s disappointment on the test ? What happens when Dad tests him again, hoping for a better result ?
    Tell me none of you have ever disappointed your parents, and seen that look on their face and felt the hurt from it.

    ::Shrug::: not scientific ethics perhaps, but questions every would-be parent-researcher needs to ask themselves before including their child in the study.

  58. #58 Tracy W
    January 23, 2009

    The difference is that these are activities of daily life. It’s damned near impossible to exist in this society without being a pedestrian, riding in a care, etc.

    Indeed, but it would be possible I think for most parents to stay home with their kids a lot more than they do. We used to be loaded into the back seat of the car to go off to meet various relatives and friends at the airport, who could have taken public transport to my parents’ house. I don’t think that makes my parents bad parents.

    Or, in the case of sports (which I’ve written about before), there’s the potential for better health, learning cooperation and teamwork, etc.

    And in the case of medical research there’s the potential for better health, learning cooperation, etc, as the research would hopefully contribute to the expansion of knowledge. Plus some experiments are interesting to participate in in their own right. I was quite interested to find out my blood type as a kid.

    There are serious problems in that a parent who is a researcher may well be conflicted on giving proper consent, and the use of your own children may make the research results unreliable, I agree. You are quite convincing on that. But the “small remote risk for no particular benefit” argument I don’t think is valid, or if it is, it captures an awful lot of ordinary parental behaviour.

  59. #59 moneduloides
    January 23, 2009

    Honestly, in most cases I would say using your children in your studies is unethical, but not all. For instance, @ my undergrad there was a lot of biomechanics work being done in one of the anthro labs that involved walking on treadmills and force plates. One of the graduate students often had her son in the pilot studies (he was anywhere from 1-5 during this time). All he ever did was run across a platform with little ‘sensor balls’ attached to a few specific points on his body.

    I’m not sure I can really see any harm in this, as it poses absolutely no threat to the child whatsoever. That being said, the cases you described are certainly ethically questionable. This isn’t black and white; it never is.

  60. #60 Rogue Medic
    January 23, 2009

    A few things from the original article:

    “I sign my own permission slips,” said Gedeon Deak, whose three sons have participated in his cognitive science studies at University of California, San Diego. He tells review boards his subjects are a “sample of convenience,” not randomly selected, but he has seen no need to specify that his sons are among the participants.

    Why has he seen a need to hide this information from the IRB?

    How can we expect this researcher to be objective, when he feels the need to withhold data from the IRB?

    If he won’t provide this information, what else falls into his individually determined category of information that he has seen no need to specify?

    Dr. Linebarger changed some of her procedures after her son Alec, at about age 5, answered a question by saying “his parents didn’t listen to him and that sometimes he felt lonely,” she recalled. “I was just floored,” she said. “I sort of assumed that we have this wonderful son-mother relationship. I decided I needed to be more careful. I was worried I would be biasing anything I did.”

    Well, we can’t have a researcher interacting with the test subject outside of the confines of the study. Or is it that we can’t have a child interact with a parent unless the study calls for it?

    When your child/research subject says, “his parents didn’t listen to him and that sometimes he felt lonely,” it is reasonable to avoid interaction that the child/research subject might perceive as professionally distant. I am jumping to the conclusion that the research has something to do with the child/research subject’s statement. Instead of taking the child/research subject out of the research, she creates even more distance between her child/research subject and her.

    I was worried I would be biasing anything I did.”

    Somebody just does not seem to get it. Her priority seems to be the research. Is it any surprise the child/research subject seems to feel more like a research subject than a child?

    Then for the child on The Truman Show, the video taping has these precautions

    Dr. Roy has students who catalog recordings sign confidentiality agreements, and each handles only video or audio, in 15-minute, randomly ordered snippets. They are asked to report anything “potentially embarrassing,” he said, but usually “when there’s something juicy or controversial the recording is off, and if it isn’t, good luck finding it.” Plus, every room has an “oops button” to erase regretted utterances instantly.

    This starts out reading like a top secret document. Then it non sequiturs into the suggestion that embarrassing stuff would not be likely to happen on camera, and if it did, you couldn’t find it. Finally it mentions their ability to delete anything that might not have looked the way they want it to. That just screams objectivity at me.

  61. #61 Laura
    January 23, 2009

    “I don’t believe I was the one who brought up punishment at all; I believe Laura brought that up as an option. Punishment actually never occurred to me until Laura brought it up.”

    Bingo. “Ways such as using their own children for research subjects” could be accurately rephrased as “Ways such as using people whom they’re allowed to spank without safewords for research subjects.”

  62. #62 Irene Delse
    January 23, 2009

    Am I the only one to think Laura is a troll? The subject is already ethically thorny enough, but she keeps thinking up outrageous comparisons, apparently with no other reason than pushing people’s buttons. Typical trollish behaviour.

  63. #63 wfjag
    January 23, 2009

    Dear Orac:

    I hope you’ll revisit this subject again. Yes, it’s true that people like Skinner and Piaget used their own children for psych experiments. That was, however, long ago and ethical standards have been tightened since then. The current standard are, among other things, to prevent getting on an ethical slippery slope.

    The Poling case raises not only the issue of whether parents can validly (and ethically) give consent for their children as experimental subjects, it also raises the issue of financial interests as a conflict of interest. When Dr. Poling decided to experiment on his child, the litigation before the Vaccine Court was on-going. Among other things, there have been questions raised about how “healthy” Hannah Poling was before the vaccination. She may not have been the healthy, active, social, happy 19 month-old, as represented later, and there is concern that Dr. and Mrs. Poling’s “recollections” changed after it was realized that if her condition could be linked to receiving vaccines, there was a source of recovery.

    Lot’s of interesting questions and many of the commentators have provided very insightful comments and valuable links.

  64. #64 Julie Stahlhut
    January 23, 2009

    Actually, there’s another big problem with experimenting on your own kids, even if you could reduce the physical and psychological risk to the child to zero. Experiment on one child, and your N = 1. But even if you have a brood the size of the Duggar family and experiment on all of them, your experiments are pseudoreplicated because of their shared environment and genetic similarity. Your N is really still 1.

    Social science studies often have small sample sizes simply because of the difficulty of getting suitable subjects, but they still require true replication. So, I would consider any experimental results under these circumstances to be about as compelling as experimental data obtained from one fruit fly.

  65. #65 Karl Withakay
    January 23, 2009

    Julie Stahlhut,
    Good point about the lack of diversity in the sample in such a scenario; at best you’d have a pilot study showing further research on a larger, more diverse sample group might be warranted. Your data could really only be used to draw conclusions about your child/children, and to hypothesize on children in general.

  66. #66 Laura
    January 23, 2009

    “Am I the only one to think Laura is a troll? The subject is already ethically thorny enough, but she keeps thinking up outrageous comparisons, apparently with no other reason than pushing people’s buttons. Typical trollish behaviour.”

    What was so outrageous about the comparisons? My parents both spanked me (though on the arm and face, not the hands), and they had the law on their side (even if it didn’t solve my myopia or make me respect them in the long run). I’ve seen bride prices and spanking mentioned (and even defended in the names of culture and parents’ rights) so often in a range of other forums, so they just seemed relevant examples to a question of whether or not parenthood makes a person guaranteed to put his or her child’s interests first during a conflict of interest and whether or not a researcher having custody of one of his or her subjects could make a difference.

  67. #67 Laura
    January 23, 2009

    “Actually, there’s another big problem with experimenting on your own kids, even if you could reduce the physical and psychological risk to the child to zero. Experiment on one child, and your N = 1. But even if you have a brood the size of the Duggar family and experiment on all of them, your experiments are pseudoreplicated because of their shared environment and genetic similarity. Your N is really still 1.”

    …or at least their shared environment, even if you have a family more genetically diverse than Josephine Baker’s.

  68. #68 albatross
    January 24, 2009

    ISTM that there’s an interesting conflict here. Suppose Orac and I are both considering signing up our children for some medical experiment, in which Orac has an interest of some kind.

    a. On the one hand, I have less conflict of interest than Orac, so I can make the risk/reward decision without the added problem of trying to discount my desire to get more data for my study or something.

    b. On the other hand, Orac will be able to independently evaluate the risks to his kids in ways I’m simply not equipped to do. Where I’ll be entirely depending on the IRB to have vetted the experiments to make sure they’re not unduly risky or stressful, he’ll probably have the personal experience and knowledge needed to evaluate it himself, in addition to whatever the IRB decided.

    Most of the people saying they think it’s okay to use their kids in these studies are working from (b)–from their own knowledge of the field, they’re very confident that their kids are not being put into danger or unreasonable amounts of stress. Now, there’s clearly a problem with conflict of interest here–maybe they can’t be neutral about this decision, because they really want the study to succeed. (More fundamentally, if they thought this kind of study was unethical, they probably wouldn’t have proposed it.) But in a lot of the cases above, the people responding are saying something like “look, this is obviously not a problem for my kids, who are expected to answer a questionaire administered by one of my coworkers.” And it’s honestly hard, in that case, to see where the harm will come in. (There might be biases introduced into your research, but assuming questions
    that made it past the IRB in the first place, it sure seems hard to imagine how letting your kids answer those questions will harm them.)

    I have no idea how to resolve this. If I were doing usability studies, I might very well get my wife and kids to try using the thing I was studying, regardless of whether I could include that in the data.

  69. #69 DebinOz
    January 24, 2009

    As an epidemiologist, I know how hard it is to recruit subjects for studies. If I hear of a study being run by a reputable bunch of folks, I will enroll my children (and me) to further the interest of science and medicine. My daughter was in a chicken pox vaccine study in 1994, when she was a baby. My three kids have been in many a cognitive type of study, and my child who is blind and has Asperger’s has been studied over and over, and has been the subject of a master’s thesis. His neurological quirks are so interesting that I am grateful that someone is learning from him.

    As an investigator myself, I have only used my children to test questionnaires on (I have specialised in nutrition-based diseases, and design various dietary intake questionnaires). I guess you could also say that I have experimented on my children using different child-raising techniques.

    One of my professors at UCB used to walk around with a little colony of fleas strapped to his leg, because he was an expert in infectious disease epidemiology.

    I truly am not sure that if I was in the field where I could effectively ‘invade’ my children’s bodies in any way, which way I would go. But then again, I am not insane nor a megalomaniac, so hopefully my choices would be fine.

  70. #70 Emily
    January 24, 2009

    It seems to me that using one’s own children for behavioral science things would lead to bad science. However, not being a behavioral scientist, I am not fluent with the ins and outs of the involved protocols. But I am interested to get people’s take on progeria. I was at the American Society for Cell Biology meeting in San Francisco last year, and I watched a very interesting series of talks on the discovery of a gene responsible for progeria, the premature aging disease. The effort was spearheaded by an MD/PhD mother whose son was born with progeria. Within I think five years, they discovered the responsible gene and are in clinical trials for a treatment. They did not discuss whether her son was one of the members of the trial group. Currently, there are no treatments for progeria and it is invariably fatal by the teenage years. It is also a very rare disease- their n=29, and they had to scour the globe to find even that many affected children.

    She spoke but didn’t say whether or not her son was in the trial group. What a heartbreaking position she is in- she isn’t the MD responsible for running the trial but she is still very involved in the science. If she has chosen to enroll her son, is it better since she isn’t the trial PI? And how do you reconcile the need to *good* science (double-blind, placebo-controlled) vs desparately trying to extend the life of your son?

    Just thought it was an interesting case from a lot of areas and was interested to get everyone’s take.

  71. #71 Anna K.
    January 25, 2009

    It’s considered stupid and unethical for doctors to treat their own family members because it’s hard for them to be objective. For the same reasons, I would doubt the data of any researcher who used a close friend or family member as a research subject. That is just ridiculous. Plus poster above pointed out the problems with a non-randomized trial, so I won’t go over that ground again.

    Admittedly I didn’t take the time to read the article, but where the heck was the IRB?? Sheesh.

  72. #72 Bhetti B
    January 25, 2009

    A book that seems directly related to this is The Alchemist’s Daughter, if you have the time for a bit of fiction.

The site is undergoing maintenance presently. Commenting has been disabled. Please check back later!