• Abusing kids causes more than long-term mental damage

    Someone sent me this study – “Does Childhood Misfortune Increase Cancer Risk in Adulthood?”:

    Objective: To address the inconsistent findings on whether childhood misfortune increases adult cancer occurrence. Methods: This study uses longitudinal data from the National Survey of Midlife Development in the United States (MIDUS) that first sampled 3,032 respondents aged 25 to 74 during 1995-1996. A series of logistic regressions were estimated separately for men and women to test whether the effect of childhood misfortune on adult cancer was largely cumulative or specific to the type or profile of misfortune.

    There were three research questions that the authors wished to answer:

    1. Does childhood misfortune increase cancer risk in adulthood?
    2. Is the effect of childhood misfortune on adult cancer risk largely cumulative or specific to the type or profile of misfortune?
    3. What pathways link childhood misfortune to cancer occurrence during adulthood?

    This is how childhood misfortune was measured:

    (1) family receipt of welfare or Aid to Dependent Children assistance for a period of 6 months or longer; (2) self-report of being financially worse off than other families; (3) less than a high school education for father (or mother if father was absent); (4) lack of male in household; (5) parental divorce; (6) parental death; (7-10) physical abuse by father, mother, sibling or other; (11-14) emotional abuse by father, mother, sibling, or other; and (15-16) self-report of poor mental or physical health at age 16.

    Without getting too much into the weeds, the authors wanted to see if any of these factors were related to an increased risk of getting cancer. They also wanted to see if the effects of any of these “misfortunes” were cumulative, or additive over time. Even the idea of this depresses me. It’s not bad enough that you have “misfortune”, such as emotional or physical abuse. The hypothesis of this study was that these bad things also had the extra effect of making you more likely to get cancer.

    It turns out some do. It also turns out there are some complex relationships between gender and how these effects play out.

    Men who were physically abused by their fathers were more likely to get cancer. Women who were physically abused by their mothers were also more likely to get cancer. But if there was frequent mental and physical abuse, then it didn’t matter what the sex of the perpetrator was – it led to an increased risk for cancer.

    There are, of course, limitations to this study. Any study of this nature is subject to what we call recall bias. People with cancer may remember things differently than people without cancer. There’s also the possibility that potential participants with the most misfortune were jailed or even killed, making them inaccessible for the study.

    I just don’t care, though. It’s not like we need any further reason to try and prevent children from being abused. But this study points to the fact that childhood misfortunes, especially ones that occur over and over, can have not only a mental cost, but a physical one as well. There seems to be some link between stress and long-term damage to their person that can lead to cancer. Abusing children not only hurts their souls; it hurts their bodies, too.


    • Caution should be applied when correlating phenomena. Just because the misfortune occurred earlier than the cancer doesn’t mean there is a causal relationship.

      Perhaps those who are prone to misfortune are so because they are predisposed to taking risks or exercising poor judgement with themselves and their children. No doubt parents pass these traits onto their children via both nuture and nature.

      • You’re technically correct here, as we’ve mentioned many, many times on the blog. You need an RCT to prove causation.

        But let’s own that there will never be one for this. We cannot randomize kids to abusive situations. We can’t even do a prospective cohort study, since the second we know there’s abuse, we’d have to intervene. So studies like this, especially ones that try to get at repeated, cumulative effects, may be the best we get. Let’s not let the perfect become the enemy of the good.

        Moreover, unlike some other studies, where we might want to weight the benefits versus harms (like eating meat or drinking alcohol), there’s no plus side at all to abuse. It may even be that you’re right, and it’s all behavioral. So let’s intervene before it’s too late.

        • The RCT one could do (and, perhaps has been done) would be interventions to educate parents about the dangers of abuse and about how to avoid it. Then one could look years later at the children of parents who (randomly) received that intervention versus a control group that (randomly) did not. There would still be an issue of abuse that was observed during the active phase of the RCT. That would have to be stopped. But, chances are unobserved abuse would unfortunately occur in the intervening years between the intervention/control phase and the outcome observation. The (not directly tested) hypothesis would be that kids in the control group are more likely to be abused. The tested one is that they suffer health effects, like cancer, later in life.

          • The problem is that abuse is so hard to measure, and so relatively rare, that the size of the study would be so huge as to be practically (and financially) impossible.

            You’d also need some sort of education that works. That alone has not been able to be well studied.

        • Did they investigate as to whether the abused kids eventually were more at risk for other unhealthy habits, such as smoking, drinking, or drug use?

    • At first glance, this seems like more of a matter of imperfect data collection and correlation not implying causation. There is nothing that I can think of that seems like really biologically plausible mechanism for child abuse to cause cancer. The physical beatings themselves are unlikely to be the case; otherwise, we would see very high incidence of cancer among boxers and football players. Perhaps something about the mind gets maladjusted as a result of the abuse, which makes it more difficult for the body to kill cancer in its developing stages? Seems unlikely to me.

      • 1) I can totally believe that living under constant stress could affect hormonal or steroid levels in some way to cause long term damage to your body.

        2) It’s totally possible that there’s a third factor here in the mechanism we can’t measure. Abuse -> XXXX -> cancer. I don’t care. Stop the abuse.

    • I’ve been following the lively discussion between Austin and Aaron on twitter and with all due to respect to both of you (and you guys really are awesome), I think the discussion is off the mark in the following way:

      This study is interesting NOT because it tells us abuse is bad. We knew that already. If abuse did not cause (or was associated with) cancer, it would still be horrendous.

      This study is interesting because it tells us something about a potential mechanism by which cancer arises. That’s interesting and important.

      We’re not doing RCTs of abuse versus no abuse to see if its related to cancer. Agree it won’t happen and shouldn’t.

      An RCT that examines whether an intervention is effective at reducing abuse is great…and if it reduces cancer rates 40 years later, fine. But, I agree that that the reason to do that RCT is to see if it reduces abuse, not to see if abuse is related to cancer.

      Fundamentally — this is an interesting study because of the mechanism of cancer issue. It probably does have serious confounders and we’re stuck trying to decipher whether the association is real or just hopelessly confounded. More work here that better handles confounding would be helpful.

      • Totally agree. But I’m also human, and my post (and reaction) were speaking to my visceral reaction.

        It’s also possible that there is some mechanism that isn’t biologic. I didn’t want to overreach. This study doesn’t give us the mechanism, it just suggests it’s there. I’d love for more work figuring that out.

    • Chronic stress (and cortisol elevation) seems like a plausible mechanism for increased cancer risk, since chronic stress is linked to immunosuppression.(1) There could also be an increase in certain risky behaviors associated with abuse that the study did not adequately adjust for (I don’t know what adjustments they made, since I only saw the abstract). The gender linkage is interesting, but I have no plausible explanation for why that would be the case.

      Regardless, there’s no doubt that child abuse is harmful, and serious efforts should be made to prevent it from happening.

      1. Glaser R, Kiecolt-Glaser JK. Stress-induced immune dysfunction: implications for health. Nat Rev Immunol. 2005;5:243–251.

    • I am not surprised by the findings, although am deeply concerned. Abuse in childhood has also been shown to predict the development of type 2 diabetes in adult women. This relationship is not completely explained by the increase in BMI seen in women who have suffered abuse. Type 2 diabetes is also associated with increased cancer risk, but the underlying mechanisms are unclear. It is speculated that the chronic stress reaction (or altered state of mental health associated with the abuse) leads to an altered state of physical health conducive to the development of these conditions. Upregulated glucocorticoid / stress steroid production, insulin resistance, and chronic systemic inflammatory responses have been hypothesized as potential mechanisms for associations between diabetes and cancer, and may also be important factors in stress-related physical health problems.