The recent H1N1 outbreak has been fertile ground for discussion on this site and elsewhere. Borrowing a term that Stephen once brought onto this blog, there is a veritable “smorgasbord” of objects, concepts, strategies and technologies to examine and reflect upon as the world comes to grips with a bona fide pandemic event (I’ll save for Carlo to verify this claim). One such first-order concept which I would like to introduce today is “situation awareness.” The term has bubbled up a great deal lately in discussions around H1N1, especially within the various operational structures and organizations, such as HHS and CDC, taking the lead on the response side of things.
At its core, situation awareness, or SA, basically describes knowing the right information, at the right time, in the right way and in the right amount, in order to make the right decisions to improve or protect health. Although it is ostensibly applicable across all domains of public health, as a concept it is usually associated – or in any event, it eventually ties itself back to – public health emergency preparedness and response. To give an example from H1N1, response has been very much guided by demands for information that go beyond what we might call “traditional” public health surveillance and epidemiological investigation. Certain disease detection and surveillance technologies have been employed in tracking the disease, but the sources of information are disparate and varied, requiring a great deal of filtering and integration to paint a meaningful picture. From aggregate disease reporting to case level data; media and internet search term tracking to hospital bed and patient tracking; virologic surveillance to border surveillance, the array of data being produced across heterogeneous sites that constitute an increasingly global health security apparatus is staggering. Contrast these information needs with the 1976 Swine Flu event, and one begins to see a stark shift in terms of the kinds of techniques brought to bear in developing useful information for response.
Many of us blogging on VSS are familiar with various public health capabilities in the context of preparedness and response. Community mitigation strategies, countermeasure delivery, mass prophylaxis and vaccination – these are just some of the capabilities to which public health is held accountable in carrying out its preparedness and response functions. More to the point, knowledge about the state of readiness and performance of these various capabilities is also very much at the core of situation awareness. Think of knowledge about capabilities as something like the second half of a feedback loop in the production of situation awareness, the first half of which is information gleaned from signals (like flu surveillance, or internet search term use, or BioSense, etc.) in the external environment. In theory, marrying these two produces a picture not just of a health threat in near real-time, it also produces a picture of what can be done about it – what resources can be brought to bear – in like time, suggesting what likely outcomes will be. This, in turn, can suggest future states, with future interventions and future outcomes, etc etc.
For students of public health, this is a far cry from the kind of rationality that has governed public health surveillance historically: namely, a focus on the aggregation of a variety of population-level data to determine risk factors (and interventions) associated with health outcomes. I would like to suggest that public health’s traditional techniques – surveillance, epidemiological investigations, laboratory testing, etc.- are, at least at certain sites and in certain contexts, being marshaled and modified in service to the emergent need (rationality?) for situation awareness. My core argument, which I’d like to test out in this forum, is that situation awareness embodies the demand for a new kind of knowledge, one premised less on the need for ever greater quantities of data, per se, and more on the need for timely, actionable information. Again turning to H1N1, this may account for the seemingly odd development that public health officials are no longer concerned, per se, with the accuracy of aggregate case counts. Generating that information is extremely resource-intensive and, to the point I am trying to make, not all that useful for public health decision-making. At the level of actual practice, this number, the all-important “numerator” in calculating disease incidence, and “denominator” in calculating case fatality ratios – and the very core of traditional epidemiological techniques – has been backgrounded in favor of additional types of information: characteristics in severe cases, transmissibility in specific settings, such as schools and hospitals, utility of community mitigation strategies such as school closures (VERY controversial, for about a thousand reasons), and viral susceptibility to treatment and prophylaxis.
* * *
How has situation awareness emerged? What problems does it seek to address? What are its techniques? Its boundaries? What configuration of practices and knowledge arose to form it? What demands does it place on individuals? On other assemblages and apparatuses? These are complex questions worthy, I think, of further pursuit. Lyle, of course, has done a stellar job of getting at many of these issues in his discussion of BioSense and syndromics, and I should very much like to see continued discussion of, for example, the employment (failure?) of real-time disease detection technologies in the context of H1N1. A broader investigation tracing back the concept may prove fruitful. A cursory look shows situation awareness to have a very extensive history in military applications, notably around operator performance of a variety of technologies, including aircraft and other combat vehicles. (Good) pilots have, in fact, been held as exemplars in the field, having been very heavily scrutinized for how they are able to make sense of, and inject order into, a vast amount of very disparate and very quickly changing information in a high-stress environment – and achieve desired outcomes on top of that. Tracing back contemporary public health situation awareness across its various lineages in aviation, psychology, operations research and the military, could prove a very informative project.
Dale—thanks for starting us off with this post. I was particularly struck by your point that the types of information that are being pursued in the context of the H1N1 outbreak differ significantly from those pursued in traditional public health work. Instead of calculating the numerator and denominator of cases in order to estimate case fatality rates, information gathering has focused on understanding “characteristics in severe cases, transmissibility in specific settings, such as schools and hospitals, utility of community mitigation strategies such as school closures, and viral susceptibility to treatment and prophylaxis.” I think the concrete shifts at stake here are fascinating: how is the population as object of epidemiology reconstituted and given new forms, no longer treated as an homogenous probabilistic object?
I first encountered the term situation awareness (or situational awareness) at the 2006 Syndromic Surveillance Conference. A woman from CDC described how BioSense was being shifted from an “early warning” to an “situational awareness” system. In fact, real-time health surveillance in general has increasingly abandoned the goal of early warning or event detection in favor of some kind of ‘situation awareness’ in the course of an event. Where are cases increasing most rapidly? Which hospitals are in danger of being overwhelmed? How are public health announcements and interventions impacting the epidemic spread?
The current H1N1 outbreak is a good example. The recognition of an ‘event’ had much more to do with diagnostic and particularly laboratory surveillance than any ‘real-time’ surveillance system. However, New York City heavily used syndromic surveillance during the course of the outbreak there. One of the key insights it provided was that ‘worried ill’ (those with ailments which were probably not H1N1, and were certainly not serious) were flooding the emergency room. This led them to the health departments to make pubic announcements discouraging visiting the ER except in case of severe symptoms. What is interesting about this is that syndromic surveillance in this case was not revealing anything about the actual epidemic of disease but about the behavior that accompanied the outbreak. However, for public health today I would suggest that such behavioral effects are as significant as biological effects in the management of an epidemic event.
One note of hesitation: I wonder if the term ‘situation awareness’ actually is covering up the concrete shifts or lines of mutation going on here. It seems a little like a buzzword the way it is used in some preparedness circles, and while I think significant changes are at stake, the relation between this term and the concepts and practices we are following is I think worth keeping in question.
Thanks Dale for this interesting post. My question is a simple one: What exactly is new about situation awareness in your view? Has the problem that situation awareness seeks to address not always been at the core of public health? Or is it not the problem that is new, but the “solution” that it provides? But again, the fact that public health experts would stop counting cases at some point is not really surprising. (In fact, physicians and local public health officials have often accused federal public health agencies (like CDC) of being obsessed with case counting …). The fact that they would be interested in specific developments at specific sites is obvious. The fact that information needs to be confirmed, filtered and integrated… again, isn’t that what they have been doing all the time? Have new technologies emerged (in the broadest sense, including technologies of the self) that are faciliting such work? I think Lyle’s question is a good one: where exactly are the mutations occurring? If I understand Dale correctly, this is still an open issue, as his questions at the end of his post indicate.
As to syndromic surveillance and H1N1, I wonder how good an example it really is: If this is the main contribution of syndromic surveillance to the mission of public health, it is a rather modest one. What physician working in an ER could miss the fact that they are suddenly overrun by worried ill cases? And what epidemiologist working at the NYC Department of Health would be surprised to get a phone call from that physician? In other words: an event like this is entirely predictable and hard to miss once it occurs. It seems as if this technology is still searching for a unique problem to which it can provide a unique solution (and thus legitimize itself).
Two very good responses, thank you both. To Lyle’s points, I would note that stressors on the healthcare system, such as burden on EDs and hospitals more broadly, are very much in line with how HHS thinks about situation awareness. Note that I say “HHS.” CDC activities are also oriented to situation awareness to some extent, although such activities tend to fall heavily within the “lane” of public health. The distinction between the two (public health and the healthcare system) is worth noting both because of the distinct apparatuses that operate in each (with expected contestations and resistance at certain points of intersection), but also because the two are increasingly combined in presenting a relatively new “information picture” for senior decision-makers.
In any event, I think Lyle’s point regarding health behaviors and their effects in the context of emergent events is well-taken and, I might suggest perhaps in response to Carlo’s point, distinctly “new” as a type of salient information that public health is increasingly regarding as relevant to its mission. This is not to say there are not tensions here. I spoke with a senior epidemiologist recently about the social effects / impact of school closures on communities (e.g., strain on parents’ & work schedules, consequences for student education if prolonged closure, etc.), and his reply was basically to dismiss the notion that public health should either collect data about those things, or take into consideration those sorts of effects when making decisions. This view may be held more by epidemiologists than senior leadership, who are more sensitive (and vulnerable) to the political implications of mitigation measures (and their “social” effects) of this type.
Two more points in reply. First, again to Carlo’s question about what is new here. This is a very good question and, in fact, I wrestled with it in an early draft of my initial post. I think it is fair to say that much of what has and is transpiring with respect to H1N1 is familiar. At its core, the response revolves around traditional public health capabilities like laboratory services, surveillance and epi investigations, not to mention interventions like mass vaccination, which falls directly under the purview of public health. My sense, though, is that situation awareness and its attendant “inputs” – like the relatively recently formulated notion and set of practices that constitute “biosurveillance” – move substantially past these traditional orientations and activities. I’ll mention just two aspects that differentiate it from traditional public health.
Briefly, I think situation awareness demands the kind of aggregation, filtering and integration of information that is light years different than prior configurations of public health information needs or data collection and analytic practices. What technologies are new? What relationships are forged across sectors (and apparatuses)? Several come to mind: global disease surveillance systems that work to forge new relationships with historically volatile political jurisdictions and shape the behavior of life scientists, such as now occurs in the middle east and the former Soviet Union; ties to private industries and retailers that sell over-the-counter medicines and equipment for early warning purposes; an increasing trend to self-report, blog and inquire about health-related issues, some of which is monitored and analyzed utilizing quite complex algorithms to detect search and reporting trends and the like. This list goes on. Mind you, I’m not taking the position that this is the inevitable march of progress; I’m taking the position that these are incontrovertibly different in their configuration and effects than traditional “core” public health techniques and rationalities.
The second aspect to SA to mention here is early warning and detection. Lyle noted in his response that these aims have fallen somewhat by the wayside; I might qualify this statement by suggesting that this is still (very much) an aim of syndromics and BioSense – and therefore SA – especially at the federal level. The recently published National Biosurveillance Strategy for Human Health makes very clear that early warning and detection are (still) central to thinking around situation awareness. There seems, however, to be a general sense that achievement of this aim was elusive in the H1N1 case and, to Lyle’s point again, he is exactly right in saying that syndromic technologies provided something more akin to supplemental information for decision-makers than anything approaching an early sign of trouble. On the flip side, I suspect that some days were shaved off of the date of detection of the outbreak by virtue of very specific technologies (successfully used to detect H1N1 cases #1 and 2), which were deployed to enhance surveillance coverage and the timeliness of lab reporting in very specific contexts – in other words, to enhance situation awareness.
Is SA a buzzword, the concept du jour in a field that perennially reinvents itself? Maybe. But we should be cautious in dismissing it if it is: the transience of a term is not a correlate of its novelty in practice, and should not imply that it has no effects. There are real strategies being developed, incorporating real technologies, integrating disparate sources and types of real information, which are collectively reformulating and reconfiguring institutional relations between domains and shaping expectations of behavior among the lay public and across myriad occupational groups in ways that are unfamiliar in the history of even recent public health. But this is not a monolithic movement; it is advancing at some sites of practice, and remains virtually unheard of in others. Where, how and under what conditions the concept is being implemented is, to me, a question of primary importance (I’m constantly reminded that a case of measles will activate an entire health department’s public health outbreak management machinery; multiple cases of norovirus in a nursing home may not raise an eyebrow).
In any event, there is a lot to mull over here, but also a lot more that can be said. I would welcome a further round of exchange if any points we have raised merit additional consideration.
The point about a ‘buzzword’ I think, is that it inaccurately describes a diversity of novel practices and concepts as a “monolithic movement”. You mention a number of specific changes taking place in diverse health-related fields: from a propensity to blog about health and monitor those blogs to collaboration with OTC pharmaceutical retailers or Soviet weapons scientists. I know that at the federal government level there is a continuing effort to “integrate” all of these diverse data sources into some kind of meaningful information, which I have heard sometimes described as a “common biological operating picture”. It will be very worthwhile to follow the development of these systems as they are certainly moving the range of data sources and surveillance populations far beyond the norms and forms of human disease surveillance. But I think its early to say whether this actually is a coherent new apparatus—“situation awareness”—or on the contrary is only a problematization of an existing apparatus, with new techniques and discourses spilling out of the boundaries of public health.
We all face these two problems daily. Situational awareness and information pollution. How do you know what you need to know to work and live better. And at the same time not experience a cognitive overload of either too much or untrustworthy knowledge? This is especially important in the case of health care.
The key is to gather the right information ( but not too much), being able to objectively analyze it, and make projections based on the analysis. It means being able to do something useful with the information.