Paul Erwin, Associate Editor of the American Journal of Public Health, recently wrote about the establishment of a Sentinel Practitioner Surveillance System for Policy Change Impact, or what might be called “sentinel policy surveillance.” The network of twelve diverse health officers will be trying to identify and share instances of harmful impact from Trump administration policies.
Erwin is suitably circumspect about what such a network can do. It is, he writes, no replacement of research, and, indeed, may be reporting perceived or feared effects as often as real ones. I found the idea intriguing to ruminate on, though. What follows are some scattered thoughts about the concept. I hope readers will add theirs. Mostly I am interested in how the practice fits with general policy surveillance and public health law research.
I like the recognition that legal acts can have health consequences. The whole idea of legal epidemiology, including policy surveillance, is to identify those effects and understand the mechanisms of effect as rapidly as possible to inform policy refinement and change. But part of the case for scientific public health law research is rigor. We recognize that the causal processes of legal effect are as complicated as any other modes of behavioral and environmental impact in public health. It is therefore important to be guided by theory in defining phenomena to observe and measures for accurate observation. In other words, to monitor early effects of any particular policy change, we first ought to have a theory and causal model in mind to define the kinds of effects we would expect to see.
The second thing we need first is a baseline. Under the circumstances, observers will be primed to see effects consistent with the political concerns that motivated the sentinel surveillance in the first place. For the surveillance observations to be useful, we need some documentation of the prior state of affairs from which the observed events differ. I’m not sure how one does that in the policy setting, but I suppose it can be figured out given more time and space than allotted to a blog post.
We need enough sites. Observing similar effects of a new policy at several surveillance sites would somewhat strengthen the inference that the effects were caused by the policy change.
All these things point to another necessary element: surveillance of the primary policies at issue. The details of the policy should be observed as part of the theorizing and modeling process. The adoption and effective dates are important to designing the surveillance, establishing the baseline and making assessments of causation.
Finally, sentinel policy surveillance should be embedded in a longer process of legal epi evaluation. The data collected in the sentinel surveillance system has immediate value as an early indicator of impact, but its utility grows as observations accumulate and the opportunity for rigorous research arises.
How do these “criteria” play out in Erwen’s first surveillance report? The policy chosen for the test was Trump’s executive order on immigration, issued on January 27, 2017. This is an interesting choice, because on its face the order pertains to admission of aliens, not to domestic policy related to immigrants already admitted or people already in the country illegally. Nonetheless, the network was asked, “Have you identified
public health practice consequences of the Executive Order on Immigration? If so, what are they? If not, have concerns about potential impact prompted you/your agency to action?” Nine sites responded; three reported “immediate budget impacts.”
Thinking about my criteria helps me see why I was a bit uncomfortable reading the Erwen report. Defining the content of the policy is important, because there is nothing in the order about state or local health services and no legal reason why the executive order would lead to state budget cuts or freezes. One can hypothesize political/normative mechanisms, but if the theory is that the order or the Trump administration has enabled anti-migrant acts and attitudes, that should be explicit and some case for how a border control executive order led to state health agency budget decisions should be offered to explain the suggestion that these cuts were a result of the order. The baseline matters, too. State budgets fluctuate over time. The fact that three budget cuts happened after the executive order is a great instance of bias bait: so tasty to bite on that hook, especially if you don’t like Trump or his orders.
I think sentinel policy surveillance is a great idea. I applaud Erwin and his network for kicking it off. Having run it through the brain once, I have the sense the idea needs work.