COVID19: KPIs for Dedomenocracy

My term dedomenocracy refers to an imaginary government by data and urgency as opposed to present-day democratically elected representatives inspired by enlightenment confidence in the enduring products of science.

For a variety of reasons and particularly for practicality, modern governments rely on infrequent passing of new laws that consequently stay in effect indefinitely.   The laws may be inspired by up to date science at the time, but once passed, the laws remain fixed even as new data undermines the original justifications.   It is very difficult to pass a new law and removing an old badly thought-out law requires passing a new law.    As a result, the law becomes Law and we’re comfortable with that because we believe science is as close as we can get to getting direction directly from God himself, so the law is presumed permanent.

In contrast, my particular concept of dedomenocracy has no permanent laws and makes no pretense of believing that the rules have any enduring justifications.   Instead, rules are imposed autocratically but from algorithms and data outside of any human influence in the rule itself.   This concept has a democratic role in collecting and verifying data and in agreeing on the algorithm and the weights of various consideration.   Also, democratic process determines when there is an urgency for a new rule.   A majority expression of a sense of urgency requiring a new rule triggers the algorithm to process all available data (historical and recent, but with more weight on recent data) and evaluates the democratically expressed weights of different considerations to come up with some ruling that everyone must follow.   The ruling will always expire within a short period of time that may vary based on circumstances but in no circumstance will last longer than a couple years.   Also, the ruling is completely free of any precedence or even of rationality.

For example, when the current COVID19 condition first appeared it took a long time before the population would have reached a majority consensus that there was a reason to be concerned.   If the governments hadn’t over-reacted, it is possible we never would have reached the necessary majority to trigger a ruling at all.   Obviously, we would have been paying attention and adapting accordingly, but without government telling us to panic, it will take a lot longer for people to panic.     Assuming that the conditions did rise to the point to trigger a ruling from the algorithm, the rulings could have been similar “stay at home” orders we currently see.

Alternatively, the rulings could have been something completely different and maybe unrelated to virus at all.    For example, the ruling could have considered the importance of the current election cycle and made a ruling to motivate people to get more politically involved, ironically reducing social distancing instead of increasing it.   The algorithm would certainly consider the current medical data about the spread and severity of the disease, but it would also consider everything else.

As mentioned, the population would use democratic processes to select the weighting of various measures and various objectives, but this is done in advance of any need to run the algorithm.   When the algorithm is run, it uses the already agreed upon algorithms.   I assume that absent any specific problem to solve, the popular choice of algorithm weightings would be to optimize future outcomes at the expense of short term losses.   As a result, I can imagine the algorithm to accept the short term losses for the disease to have a more robust future for those who remain.

While this sounds very different from the current democratic processes, I think it is closer to the intents of a democratic process.   When we elect people to occupy public office, we really hope that the person will make wise decisions when emergencies occur.  Unfortunately, because we elect human individuals they are free to make decisions that may neglect our original intentions.    My proposal essentially elects a goal for the algorithm that will make future rules.    When there is a need for a ruling, the algorithm is bound to consider that goal.

A goal of optimizing future benefits for the survivors could very well accept that there will be short term losses.    I do expect we would subject candidate algorithms to  simulations to be sure it makes reasonable decisions.   This again is similar to how we require candidates to debate each other, but the difference is that the elected algorithm will reliably act consistently with our expectations.

Especially during my lifetime, there is little consistency of what a candidate campaigns on and what he does when in office.   I prefer to have a government by a democratically elected algorithm than a democratically elected human.

When considering a situation like the current epidemic, the algorithm will consider the remaining demographics a couple years in the future.   Clearly we can expect more losses from this disease and we know the losses will be highest among a sub-population likely experience loss of life anyway in the next couple years.   Thus there are two possible futures where both involve a surviving generation that has already mourned the losses of the prior years.  One future is where that surviving generation has resentful memories of abuses and deprivations during the period, and the other instead has the sense of guilt that they may have been able to do more to help.    The choice of the algorithm is a choose between these two (or other) choices that a combination of good and bad.   The choice is which population will have a brighter future.   I suspect the latter group that felt guilt at not sacrificing more: they are not resentful of the system and they are motivated to make future sacrifices for the benefit of others, at least to some degree.

The first group that resents the system for forcing them to sacrifice for what appears (to them) to be nothing  because there was still a large loss of life from their perspective.    They have no objective way of knowing how many more lives would have been lost if they hadn’t made those earlier sacrifices.   They will not be as cooperative with their peers going forward.

That reminds of another principle of dedomenocracy.   The objective is to minimize the number of rules in force at any time with the goal of having no rules at all.   Meeting this objective involves encouraging a population that cooperates and acts with consideration of others.   This is not a utopia vision because clearly people will be people and do inconsiderate and selfish things.   In a dedomenocracy it is sufficient for the population to be not get to the point where they trigger the urgency threshold that forces a new autocratic and authoritarian ruling.    People will certainly push the envelop, but enough will be cautious to not push things that far because there really is no way to know what kind of ruling the algorithm would give, except that is it likely to be very harsh and authoritarian so that the objectives can be met in the shortest time possible.

This is why I concede that a dedomenocracy could have first imposed a short, one month, order for everyone to stay home except for essential activities.   This was harsh.   The difference is that the dedomenocracy intended the rule to educate the population of the new risk that needs to be taken seriously.    The democracy intended this to be the permanent solution to stop the spread until either the virus disappears or there is a vaccine.

The dedomenocracy is run by an algorithm that tunes society to operate through new circumstances with the objective of finding something that can be sustained with the assumption that the new circumstances will never go away.   It would be free to choose more practical options that supports a steady and even high rate of new cases and new deaths.    Based on what we know right now, we may never eliminate this virus nor find some RNA vaccine.   A dedomenocracy would instead focus on reaching a steady state of new cases matching the closed cases and on provisioning the medical infrastructure to keep the active case load below the available capacity.

There are multiple practical things we can do in short term to adapt to a continuous stream of new cases:

  1. Improve testing, disinfectant, quarantine, and general cultural practices to reduce the spread of the disease to a manageable level (instead of striving for zero).
  2. Improve treatment at each stage of progression to avoid requiring escalation to more extensive care
  3. Improve treatment to shorten the treatment period at each stage to shortest time possible, thus having patients spend less time in treatment
  4. Permit medical staff to practice and learn new methods with a steady stream of new patients as long as those patients will continue to need help.   (Note that this steady case load permits keeping staff current in their skills in contrast to the current strategy that may end up with flare ups spaced out years apart so that each new time requires new people to retrain and this results in the early excess casualties resulting from staff being out of practice.)

In contrast to the current policies that have the goal of spending whatever it takes to bring new cases to zero and new deaths to zero, this operational approach accepts that the virus is now a fact of nature that must be accommodated with a mix of adequate capacity and better treatments.

It is from this perspective that I take issue with the current reporting of the COVID19 situation.   The current reporting is presenting numbers that are biased toward the current objectives of reaching zero new cases and zero new deaths, something I doubt will ever happen, even with a vaccine.   As a result of this bias, we get daily reports of absolute numbers of new cases, new hospitalizations, new deaths, new recoveries.   We also get the cumulative numbers broken down by country, province, or city and this is presented as some kind of score for who is doing worse than others.   The ones with the biggest numbers are doing the worst.

From an operational perspective that accepts this virus as a fact of nature, these metrics are completely meaningless.    We will learn to accept that there will be new deaths from this disease just like we accept that we will always have deaths due to cancers, strokes, heart attacks, organ failures, etc.   The absolute numbers and accumulated numbers are definitely tragic and regrettable, but they are also mostly unavoidable although we continue to strive to reduce them little by little.

The important key performance indicators we need now are daily reports of current utilization and excess capacity.   Based on current numbers it appears reasonable to assume that we can reach constant average number of daily new patients.   What we need to do is focus on making sure that the excess capacity is there to handle the expected variation around a hopefully steady mean value.   If there is a growing trend, then we need to be sure to plan on expanding capacity to match that growth.

A different strategy would apply if we have to accept exponential growth, but for now I’m assuming a manageable growth or even a steady rate of new cases.

Metrics of interest for an operational approach include:

  • for each hospital or clinic,
    • the current capacity of beds and intensive care units (including ventilators) available for treating this infection in particular
    • the current utilization of the available capacity terms of percentage of capacity (not absolute numbers).
    • The current throughput of the different stages, how long a patient occupies each level of treatment, and how long before the cases are closed
    • The current rate of success for each stage of treatment where success means not having to escalate to the next stage (the last stage being death).
  • For each region where patients may be reasonably moved around
    • similar capacity, utilization, and success data but in aggregate terms
    • additional information about how far patients are from their homes or how many facility transfer they needed
    • location information to localize hotspots where patients are coming from
    • trending and prediction to recommend future allocations to meet future needs, metric is projected increases in funding needs
  • for each higher level of administration: state, country, continent, etc
    • similar aggregates of capacity utilization and success rates
    • similar planning processes with predicted increases in funding needs

The operational perspective accepts that the infection and its consequences is now a fact of nature that we need to plan around just like we currently accept that we will occasionally experience earthquakes, hurricanes, locust plagues, and crop blights.   Each episode is tragic and will have casualty counts, but the measures of merit are in terms of our capacity to respond to each outbreak without collapsing society.

The absolute and accumulated numbers are only important for political gains for electing parties and politicians.   A government by data and urgency has no such politician elections so these numbers are not meaningful as long as the survival rate is sustainable.  From an operational perspective, what matters is maintaining sufficient capacity to handle the new cases.

While the operational perspective would welcome some solution that makes the problem disappear such as a vaccine, its primary focus is on the more practical and promising approaches to find ways to improve treatments and reduce treatment times and this will eventually lead to a reduction of capacity that needs to be maintained.

It is feasible that we can adjust to live our lives close to the normalcy of just a few months ago while gaining comfort in accepting the persistent risk of getting this disease, just as we accept other risks in life.   That comfort will come from confidence that we have adequate capacity to handle the new patients, and that we are improving treatments so those patients’ conditions do not worsen and their recoveries are fast.

There will be deaths.   The deaths are tragic, but we can live with those consequences as long as we can maintain the capacity to deal with new cases.    Implicit in maintaining the capacity is optimizing the health of the entire economy where everything is deemed essential.   A dedomenocracy (of my imagination, of course) is better suited to react to this kind of new reality than is our current democracies.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s