Why we should think critically when reporting on childhood adversity

[This is the first of two posts. Find part 2 here — Eds.]

What do juvenile crime, high medical costs, and short life spans all have in common?

If you believe recent health reporting on these and dozens of other topics, all of these can be traced back to something traumatic that happened to people when they were young.

Adverse childhood experiences (ACEs) is a field of study that dates back to a seminal study published in May 1998. But it is now very much in vogue as a way to explain a wide range of social ills. One sign of its policy currency was the California governor’s appointment this year of an oft-cited ACEs researcher and pediatrician as the first surgeon general of California.

I wrote recently about checking our biases when doing an investigative story. When it comes to ACEs, I think we all could do a better job checking our biases. The zeal with which we turn to ACEs as a root cause of so many things borders on the religious.

Why are so many people so fervently in the thrall of the ACEs hypothesis? Because it just seems to make intuitive sense.

How could neglect by a child’s parents not have a powerful downstream effect?

How could witnessing a crime – or repeated crimes – as a child not do serious physiological damage?

How could the stress of being bullied at school not create a toxic inner world that alters the body down  to the cellular level?

The answer is: We simply don’t know enough yet to be able to say that all of those downstream effects are really tied to ACEs.

Here is the first problem: It is very hard to establish that A caused B and then caused C when you are looking backwards over a trail of health records. That is known as a retrospective study, and such studies come with limitations. The original ACEs study, an effort by the Centers for Disease Control and Prevention and Kaiser Permanente, analyzed health records retrospectively to look for patterns.

Here is the second problem: This foundational study – while admirable for its ambition – took a snapshot comprising just seven months of survey data. It examined health data in Kaiser patients who took a survey between August 1995 and March 1996. A second wave of surveys were sent in 1997. Later studies took additional snapshots of patients in generally the same population but not necessarily the same people. For the one about life expectancies being shorter for people with higher ACEs scores, researchers looked at the same CDC-Kaiser surveys that were completed between 1995 and 1997 – a total of 17,337 adults surveyed – and then tried to use death records to figure out whether those same people had died between then and 2006. It found that 1,539 people had died.

So, think about that. Out of 17,337 adults surveyed, 1,539 adults had died 20 years later. Then, the researchers found an even smaller subset of those deceased adults who had scored six or more out of 10 on the ACEs questionnaire. The researchers wrote:

“People with six or more ACEs died nearly 20 years earlier on average than those without ACEs … .”

Concluding that the experiences (divorce, for example) led to life-altering diseases and shortened life spans based on so little data — a total of 20 deaths — is problematic to say the least, as the authors acknowledge in the study.

Here’s the third problem: That original ACEs study was based on 9,500 members of the Kaiser Permanente HMO in Southern California. Even if we assume it is large enough to draw some good conclusions from the retrospective review of the data, subsequent studies have been far smaller in scale. Most are small studies with limited scopes that are difficult to repeat. A recent study on college students and ACEs included 239 people. A study on ACEs and chronic pain included 141 people. A study on ACEs and lupus included 269 people. It is difficult, to say the least, for a small study looking at the self-reported incidences of traumatic events to establish that those events – assuming they all happened the way people say they remember them – led to the downstream impacts observed by the researchers.

The underlying problem with much of the ACEs research is that it relies on correlations between childhood adversity and later-in-life health and behavioral problems. But we all know correlation is not causation. One of the CDC and Kaiser studies says, for example, “our findings were based on a retrospective survey which made it difficult to distinguish the causal impact of ACEs on premature mortality.”

So many things can happen over a person’s life that isolating adverse childhood experiences from the millions of influences and experiences that followed is difficult.

So, why not do a study that tests whether ACEs really shorten life spans by two decades? The same study notes:

In theory, prospective studies related to child maltreatment would avoid these potential biases. In practice, given the social and legal implications, this type of studies are nonetheless difficult to conduct.

Scientists can’t corral a bunch of kids, neglect half, nurture the other half, and see what happens.

That doesn’t mean ACEs as a concept doesn’t have validity and isn’t worthy of further research. But too much of a focus on ACEs might actually distract us from tractable solutions, too. By blaming huge swaths of societal ills on things that happened to us decades ago – as opposed to things that we are doing to each other or doing to ourselves in the hear and now – it removes agency. It allows us to create a level of abstraction between the problem and the source, avoiding the really tough questions that topics like juvenile crime, health care costs, and chronic disease demand. If, as other recent health stories suggest, half of all kids suffer from ACEs and ACEs actually transmit their damaging power across generations, then what can even the best interventions do to reverse this course?