Key findings from public health

Healthy "people" aren't localized rocks, but are normally well-interconnected bidirectionally into the social fabric around them.

Social connectivity is the most robust predictor of internal, "physiological", "biomedical" outcomes, such as morbidity, mortality, survival rate of surgery, resistance to infection, level of depression, outcome of diabetes, obesity, "mental" health, you name it.

Prevention is a thousand times more cost effective than repair. ( A lesson from software engineering and many other fields as well.)

The caring human loving touch of another individual is very important to human health and healing. Infants who aren't touched do poorly or simply die.

All interesting social phenomena (such as relationships, jobs, teams, family, stress, love, sex, the economy, depression) involve intimately bidirectional feedback loops.

But, classical statistical measures and attitudes, based on prediction of yields of crops, assume critically that causality is defined in one direction only, and that all phenomena of interest can be "isolated" from context and one part of it varied by the experimenter while other parts of it are "held constant." None of that applies to "complex adaptive systems", including social systems, which are inextricably interconnected, context-dependent, interdependent, and riddled with bidirectional feedback loops. Since the tools and expertise breakdown when applied to these areas, rather than admit that the tools and expertise are inadequate, the problem space is instead defined as "non-scientific" or "soft-science" and demeaned as unimportant or "non-scientific."

Possibly due to such schizophenia, the US "healthcare" system behaves as if none of the above solid empirical facts were known. There is no focus on social connectivity, less than 2% of the budget is spent on prevention, and machines and processes have replaced people at the bedside. People are treated like machines, and diseases are treated as if they were independent of each other and the rest of peoples lives. "People" are reduced to "patients". "Caregivers" are too busy to stay and chat for a while with "patients" and are increasing renamed "providers" which is ironic, since mostly they consume resources, particularly money, while being forced by "the system" to be too busy to stick around and observe the actual outcomes of their "treatments" on the people they serve. It's a lose-lose scenario, disliked by the patients, disliked by the caregivers, and apparently continues to exist because it's loved by the insurance companies. The whole thing needs to be rethought based on the above new facts of life.

Perhaps, not surprisingly then, the outcomes of the US Healthcare system are terrible, compared to peer countries. Infant mortality is something like 19th in the world. Costs are huge but a recent study showed that the BEST quartile of US citizens (the rich) have health outcomes worse than the WORST quartile of British citizens in the UK. (ref ?). Depression, obesity, diabetes are widespread and rampant epidemics in the US.

But, efforts to build healthcare interventions that are designed around social connectivity and whole persons are demeaned and ridiculed as being "non-scientific", or avoided because the feedback loops make computing "p-values" problematic for academic researchers, for whom such mathematical bases for certainty are held with a sort of blind obsession despite the fact that the assumptions of the theory (General Linear Model) don't fit the problem they're trying to address.

The result is that the most effective interventions are known, and involve teams of people assisting individual humans to modify or control their behavior and life style, but the advocates of these interventions are academically shunned and have to present their work in embarrassment in back rooms. The Office of Behavioral and Social Science Research (OBSSR) within NIH is treated like an awkward in-law.

Probably the single best book that summarizes interventions in health care that actually work is Health Program Planning : An Educational and Ecological Approach by Lawrence W. Green and Marshall W. Kreuter, now in it's fourth edition. (c) 2005 McGraw Hill, initial version written in 1961. It was around that year that non-communicable diseases began to replace communicable diseases as the leading causes of death, disability, and impaired quality of life, but the older, biomedical model had a very tightly held death-grip on the "health care industry."

On page 3 of that book the authors note:

Ecological approaches have proven difficult to evaluate because the units of analysis do not lend themselves to rand assignment, experimental control, and manipulating characteristic of preferred scientific approaches to establishing causation. Although the linear isolatable cause-effect model of scientific problem solving remains the point of departure for the training of health professionals, practitioners find ... they cannot ignore the contextual reality that health status is unquestionably influenced by an immensely complex ecological system. ...

To address those systems in our planning, we must first be able to see them ...
By definition, ecological sub-systems do not operate in isolation from one another ... [but] interact with one another to influence health. [We need] a kind of ecological map or "web" or "systems model" enabling us to visualize the network of relationships that need to be taken into account as we plan our intervention strategy tailored to the unique circumstances of the target population and the place where they live and work.

The primary tools up to this task are described by John Sterman in his tome Business Dynamics, 999 pages in length. The simpler techniques of mapping on a white-board is known as Causal-Loop Diagramming or CLD. These qualitative webs can be assigned some semi-quantitative values, such as directionality and general magnitude (large, small, strong, weak) and then simulated using tools such as Vensim (tm).

That, however, is a lot of work. "Systems thinking" didn't show up in the MPH curriculum until 2006, and is absent, by that name, in most courses, even at leading universities. Only MIT and Worcester Polytechnic Institute seem to have embraced these tools, although the Ross School of Business at the University of Michigan is starting to build a systems thinking program after the auto industry started demanding it.

Note that the pressure for innovation here is from business, and the academics are lagging behind, sometimes kicking and screaming, in stage 2 of Schopenhauer's three stages:

All truth passes through three stages. First, it is ridiculed. Second, it is violently opposed. Third, it is accepted as being self-evident.
Arthur Schopenhauer

So, this pretty much summarizes the state of affairs today. Johns Hopkins Bloomberg School of Public Health has started a new department of Health Behavior along the lines of the new theory, but most health and public health people are famously non-quantitative, and so they are attempting to think through such problems mentally, unassisted by available tools used in other industries for over 50 years now in systems dynamics.

And, the biomedical establishment has a strong lock on most thinking and peer-review journals, and alternates denial and violent opposition to the "new paradigm" which it perceives as a throwback to mystical soft thinking instead of a more general version of the scientific method that can embrace feedback loops and complex adaptive systems without distortion of the tools or violation of the assumptions behind the models and statistics.

Even at Hopkins in the department of Epidemiology, the ratio of new thinkers to old-paradigm thinkers is essentially 3 to 70, and this new paradigm is ridiculed, rejected, opposed, despised, by most old-school thinkers who wish the answer to health had stayed down the microscope, under control, where they had strong muscles and good intuition - instead of showing up increasingly outside the window of the lab, in the social fabric of society, in all the places the scientists grew up despising and where their tools and muscles and intuition all fail.

So, where does that leave us humans?

Apparently, we can't expect either academics or health care workers to take the lead in fixing this terrible mess, and business is going to have to get down to business and do something about it.

(This is not without precedent - the center of innovation in the USA has increasingly moved out of universities and into businesses, despite the very strong marketing campaign with the opposite message. Witness the pulling-teeth it's taken to get systems thinking into the Ross Business School curriculum.)

Business today is much more cybernetic on a real-time basis than academia, and utilizes "good enough" models which, with cybernetic feedback control, get the job done and produce the desired outcomes - - while driving academics crazy because the underlying models are "so bad."
The National Institutes of Health is still heavily dominated as well by biomedically oriented researchers of the old school, who resist the new paradigm.

So, with a few exceptions, industry money may be the only way to advance health care in serious ways, and address the findings at the top of this post sometime this century when we're still alive to care about it.

We have, as in so many of M.C. Escher's paintings, (see this link:
http://en.wikipedia.org/wiki/Image:Escher_Waterfall.jpg
created a world that is locally-sensible and globally nonsense, but few people working locally are motivated to address the global wrongness, and no Masters or PhD student or young researcher would be encouraged to tackle a "large" problem, and so it sits there, unaddressed by academia and a thorn in the side of everyone: patients, doctors, nurses, payers, industry.
Like Escher's paintings, one is hard pressed to see or point to exactly "where" the wrongness is, and yet, standing back, it's clearly wrong.

That's where things are today.


[ M.C. Escher website: http://www.mcescher.com/ ]