IN SEARCH OF ANALYTIC INSIGHT – COVID19

Over the last several weeks while hunkered down in imposed quarantine most of us anxiously watch the numbers involving the COVID19 pandemic. By and large the general population looks at the change in the number of people infected and the death rate. From this superficial basis emotions occur, but are they valid concerns?

The numbers we see include;

·        New Cases

·        People Under Investigation (PUI)

·        People Under Monitoring (PUM)

·        Recoveries

·        Deaths

From these raw numbers we arrive at personal conclusions relative to whether we are winning the war or loosing the battle. There are concerns as to accuracy and authenticity. Some of these impressions are the result of routine reporting conditions, whereas others are claims (yet to be proven). These include,

·        Timelyness of data reporting

·        Failure to report (specifically deaths)

·        Failure to adopt a consistency as to whether a PUI is a person who has been subjected to testing, and awaiting results, or whether its simply a person with unconfirmed symptoms. According to various governmental clinical agencies a PUI is a person who fills out a form reporting suspect symptoms.

·        Absence of test run but failed as a result of test kit quality or unsatisfactory clinical administration, and

·        Of late there has been revisions in terminology and classifications which may appear to be more definitive definitions but poses challenges in the reformation of data groups previously.

It is my contention that there are voids that inhibit the ability to contrive a speculation relative to trends which will ultimately jeopardize predictive abilities.

·        Absence of tested awaiting results numbers – initial onset cases,

·        Number of inconclusive and faulty tests,

·        Lack of unified statistics globally (although there are World Health Organization(WHO) guidelines each country has used what they had available in order to the spirit of result collection), as a result infectious movements and resolving actions are difficult to use. Further the authenticity of national data is suspect at present and as a result a determination must be made whether to include, exclude and what course of political actions must be taken as it relates to domestic concerns.

·        Aside from the aforementioned jeopardy items we cannot overlook that as an ancillary component to the clinical data regarding COVID19 there are also demographic and personal statistics that would be used to further qualify the status and clinical delivery for the viral condition.

Piece-Part Combinations

Don’t be fooled by charts and graphics that depict a known set of data elements. We can speculate as to what they mean but a more definitive understanding becomes necessary in order to enact public policy and clinical care.

One observation is that analytic objectives are vague or non-existent which leads us to speculate what they represent. It’s not just for the curiosity of the general public but also for use by government entities. We can objective separate viral activities into at least three(3) stages;

               Stage 1 - Intake

·        Number of suspected cases reported (formerly PUI)

·        Number of suspected cases confirmed (by test)

Stage 2 – Treatment

·        Number of cases in quarantine (formerly PUM)

·        Number of cases under clinical care (also PUM but severity qualified)

Stage 3 – Outcomes

·        Number of deaths from quarantine

·        Number of deaths under clinical care

·        Number of negative cases /recovery condition

·        Number of re-occurrences (excluded from intake numbers to avoid duplication)

From this set of data we are then able to drive interpretive numbers. For example, the ratio of reported vs. confirmed cases who be indicative of a reduction or growth in the epidemic. Without this comparison all we see is changes in reported numbers which may not reflect an increase in the virus (only suspect conditions). Another examples would be to compare the mortality rates between quarantine and clinical care, and the number of recoveries by type of care condition. It is essential that we advance from analytic conjecture on the fly to a definitive pragmatic paradigm. This involves precise definitions, strengthened data care integrity and a definitive set of baseline analytic derivative that reflect an appropriate conclusionary basis.

The Overlay

Data and analytics does not lay totally in a void that is absent of external variables. It is touched by statistical profiles about the populations from which this point specific data is derived. As mentioned previously there are demographic and population group profiles that give added insight that permit making policy decisions. For example, isolation or clearing of quarantines areas, protection of at risk populations (either to exposure or for urgent right care options) or treatment plan options. Again these elements must be considered at the same time as the definition and qualification of all viral analytic data. We must be alert to hinderances in which nations have the inability to comply, and this leads to the question of whether a global reporting and analytic solution is necessary.

Big Picture

What we have seen, even with sparse analytics is what starts out small can grow to cataclysmic proportions. It then becomes a matter of geopolitical consideration with a universal system is need and ultimately how it can be properly cared for.  It is said that ‘information is power’ and as trusting as we all might be we have also seen on many occasions that mankind doesn’t always hold a caring view toward others. Even during these times of uncertainty we still see the ugly elements of finger pointing, wild speculation, hatefilled intentions and counterproductive acts. Even those that should be in a neutral position have been raised up in suspicion of pandering and biased policy opinions. Is there a simple answer? I would hope so but this will involve intense discussions between governments and the analytic community to acts as a servant to its benefactors (those who would benefit). It will also mean that aside from just analytics the community must become stewards of security and the authorized use and release of information. Finally the community of analytic professionals must be visionaries and not set in terms of absolutely static thinking. As time advances our obligation is to address discoveries, realities and contentions to the work that is being performed. It’s more than a table or a graph, speculation based on a single set of numbers or a broad conclusion that the data is fully representative of the population group.

Reaching deep also means a profound understanding of statistical extrapolation to achieve a level of value correlation to an overall population. It’s more than just a percentage representation but must embrace actual numbers viewed as a simple when compared against a total population size. This will then statistically permit you to form a level of confidence, a tolerable error rate and precision.

Next Steps

In the present context the analytic journey has already commenced. The need to act is at an urgent state. Best of minds must be compelled to address the short comings, the rectification of gaps (definitions, minimum collective set…) and to stay engaged for the long term. This will involve professional and personal commitment and do so objectively without personal gain.