We made the acquaintance of Lenore Albert on the Covid-19 Long-Haulers Discussion Group on Facebook. She’s an artificial intelligence expert and a Covid patient who collected data and made this data visualization set and several others on her website.
It’s fascinating for what it shows about what symptoms are most common at what time over the population of 900 patients surveyed.
Here’s what she wrote about it.
By Lenore Albert
MinedMedia is an AI expert non-profit organization focusing resources and partnerships on
policy analysis for the common good.
An early defining project focuses on the Long-Haul impact of Covid 19. Surveying 900 patients
and applying our artificial intelligence analysis, consistent symptom patterns are revealed over
time; these results are presented on the MinedMedia site for health policy analysis,
physician/patient education and as guidance for further research on long-term Covid impact.
The research produced five statistical classes or sets of common symptoms, each with their
own patterns of variability. For example: in the symptom class where involuntary muscle
twitching is 16% common, the surveys indicate 15% also experience body chills and 19%
experience nausea, however this grouping has substantially less incidence of fast heart rate
(tachycardia) over 100 beats per minute over five minutes. These indices of different classes of
symptoms are in the first menu: “Long Covid Symptom Survey Results” – double click on each
symptom to see an inset chart on the associated applicability to each statistical class.
On the MinedMedia’s Long Haulers page, a second icon “Symptoms Over Time” contains an
animated map showing the migration of symptoms. It shows, for example, that Brain Fog is the
least significant symptom at the start of Covid, but by the third week it’s the second-most
common, reported in about 1/3rd of patients. By Month #5, Brain Fog has returned to initial
levels and by month #7 it’s half of the initial level. The intermix of symptoms over time is one
of the most compelling results. https://covid.mastermined.tech/gallery/racebar/
Medical awareness of patterns
Medical awareness of patterns of Long Haul Covid symptoms in patients has a profound effect
on patient treatment.
Longer-term effects of Covid-19 are not well understood. Consequently, surveys indicate
doctors have often focused on “mood-enhancing” prescriptions, including sedatives, with
significant levels of Doctor denial of patient symptom validity. Yet all researchers indicate so
little is understood about the effects of this virus that respect for patient experience is a wellspring of research guidance.
Among the 900 survey participants, there are numerous anecdotal reports of patients referring
to the MindedMedia site during doctor discussions, resulting in a substantial change in the
attitude of analysis of patient symptoms.
On the site, the “Long Hauler” data sets indicating the frequency and severity of symptoms are
presented at various time intervals which helps patients “know what to expect” based on the
experiences of a substantial set of 900 other patients.
Using data for theories and additional research
This effort also informs doctors, researchers and policy analysts.
Several medical researchers and research institutions are making use of MinedMedia data and
reports to help qualify theories and to inform additional research.
Further partnership and collaboration is encouraged. It is believed this on-going research
warrants expansion in scope and in depth of analysis.
The artificial intelligence engine used in MinedMedia is built on Python, an industry standard
machine learning framework. This implementation of AI has deployed credentials in other
market sectors including the media industry and in the legal profession. Its contribution to this
non-profit initiative includes analysis of the vast numbers of variables of patient reports, along
with pattern-recognition of symptom sets which would require substantially more time and
resources if analyzed without AI resources.
MinedMedia Covid-19 surveys and analysis come initially from involvement in the Covid-19 Long
Haulers Discussion Group of approximately 10,000 participants on Facebook. Approximately
10% agreed to take part in multiple surveys tracking symptoms over time. The results show
distinct patterns of symptom spikes, abatement, renewal, long-haul endurance and
remediation.
The analysis of the Covid symptom incidence indicates two distinct spikes of symptoms, in
March and July, derived from MinedMedia’s surveys. The early December surveys of Covid
patients didn’t indicate any spike from the return of children to school in September; a followup survey in January will include symptom set analysis from Thanksgiving travel.
All patient data is strictly confidential and there is no individualized information distribution of
any type.
It is believed substantial additional insights from this ongoing set of surveys are likely.
Participation is welcome by all interested individuals, medical research groups, media
organizations and health policy analysts.
