Data prep

COVID-19 breakthrough cases among Harford County residents in Maryland

(Page updated at 2022-02-24 20:11:48)
Data from:
* COVID Link breakthrough cases data by MDH, as of 02.24.2022 * COVID Link VOC/VOI data by MDH, as of 02.24.2022
* The number of total new COVID-19 cases by MDH, as of 2022-02-25
* Vaccination data from Immunet, as of 12.03.2021
* COVID19 hospitalization data by MDH, as of 02.19.22

See Annex for detailed note about the data.

1. What is the level of breakthough cases?

1.1. Trend of breakthrough cases - absolute number

A total of 151412 COVID-19 breakthrough cases have been identified in Maryland. In Harford county, 5986 cases have been reported. Below figure shows the daily number of breakthrough cases over time.

(NOTE: the 7-day average is the mean during a week leading to the date.)

1.2. Trend of breakthrough cases - percentage

To date, among people who have been fully vaccinated in the county (n=154890), 3.86% have had a breakthrough case - assuming no multiple breakthrough cases per person (and assuming the number of vaccinated people has remained stable since early December 2021).

Following figure shows cumulative percent of fully vaccinated people who have breakthrough cases - but only through December 16, 2021 (due to changes in the Immunet data that are accessible since early December).

We initially expected the cumulative percent to:
- fluctuate initially, depending on the pace of numerator increase (number of breakthrough cases - see above figure) vs. pace of denominator increase (number of people fully vaccinated), and then
- level-off, based on expected vaccine effectiveness, and then possible
- increase due to decreasing immunity and/or variants.

(Note: percent = 100* ‘# cum cases on day X’ / ‘# cum fully vaccinated people on day X-14’)

1.3. Level of new cases

Note:
* Non breakthrough cases = total cases - breakthrough cases. This assumes date in total cases is based specimens collection date, which was used in breakthrough case data. This may not be correct, but the implication is minor as long as there is only minimal report delay.
* Fully vaccinated population on day X = population who received second/final dose 14 days ago (X-14).
* Not fully vaccinated population = total population - fully vaccinated population * See detailed numerators and denominators below.

1.4. Can we estimate vaccine effectiveness - to prevent cases?

Below figure shows expected percent of breakthrough cases based on vaccine efficacy and vaccine coverage, according to Chen and Orenstein (1996). But, this method requires routine (unbiased) surveillance system…

2. What are characteristics of breakthough cases?

2.1. Timing of breakthough cases since the second/final dose

In Harford county, 50% of breakthrough cases occurred within NA days after receiving the second dose. 75% of breakthrough cases occurred within NA days.

In Maryland, similarly, 50% of breakthrough cases occurred within NA days after receiving the second dose. 75% of breakthrough cases occurred within NA days.

But, by age group, the distribution is little different. See below.

2.1.A. Survival analysis

No ID to merge immunet data and breakthrough data. need to check all COVID link reports variables.

2.2. Other characteristics among breakthrough cases by period: prior to vs. during the last 60 days

COVID-19 breakthrough cases by group: prior to vs. during the last 60 days
Group Group.label in the last 60 days: Harford county, Number in the last 60 days: Harford county, % in the last 60 days: MD, Number in the last 60 days: MD, % Prior to the last 60 days: Harford county, Number Prior to the last 60 days: Harford county, % Prior to the last 60 days: MD, Number Prior to the last 60 days: MD, %
By Age
11 and Under 268 6 6644 6 9 0 515 1
12-19 275 7 7566 7 141 8 4570 10
20-29 459 11 14698 13 235 13 7496 17
30-39 723 17 18982 17 369 20 8398 19
40-49 685 16 17543 16 289 16 7245 17
50-59 648 16 17506 16 302 17 6875 16
60-69 588 14 14852 13 254 14 5044 12
70+ 534 13 13367 12 207 11 3559 8
By death
No 4132 98.9 110074 99 1793 99.3 43500 99.5
Yes 48 1.1 1084 1 13 0.7 202 0.5
By hospitalization
No 3854 92.2 103031 92.7 1747 96.7 42612 97.5
Yes 326 7.8 8127 7.3 59 3.3 1090 2.5
By Having.Symptoms
missing 3259 78 84714 76 792 44 21228 49
No 118 3 4184 4 425 24 7679 18
Yes 803 19 22260 20 589 33 14795 34
By Variant.Lineage
Not tested 4090 97.85 109312 98.34 1598 88.48 39548 90.49
VOC: Delta 11 0.01 117 6.48 2477 5.67
Wild Type 25 0.6 543 0.49 20 1.11 528 1.21
NA 65 1.56 1292 1.16 47 2.6 719 1.65
By Congregate.Housing
0 4112 98 108164 97 1778 98 43212 99
1 68 2 2994 3 28 2 490 1

2.3. Characteristics comparison: breatkthrough vs. non-breakthrough cases

Work in progress using Redcap/Prepmod data

2.4. Hospitaization: breakthrough vs. regular cases

NEDSS is unique in redcap data.
NEDSS is NOT unique in hospitalization data…

Work in progress. YJ to double check ID variables in COVIDlink reports…

4. Among Harford county residents who were hospitalized, vaccination status at the time of specimen collection - weekly trends since March 2021 and monthly trends since the beginning of the pandemic

Below figures are weekly trends for Harford county, based on MDH’s “Weekly COVID Hospitalizations Line List”, received on 02.19.22.

4.1. Broken down only by vaccine status

Note: These are absolute number of hospitalization cases. The rate of hospitalization is lower among full-vaccinated people than among non-fully vaccinated people, even when the absolute numbers appear to be similar.

4.2. Broken down by vaccine status AND age group

Note: These are absolute number of hospitalization cases. The rate of hospitalization is lower among full-vaccinated people than among non-fully vaccinated people, even when the absolute numbers appear to be similar.

4.3. Broken down only by detailed age group

###Annex Data source 1. Breakthrough cases: Reports, Cases w Vax Breakthrough (>=14 days) 4 and Cases w Vax Breakthrough (>=14 days) 3, from COVID Link (http://mdcovid.lightning.force.com/).

Breakthrough cases are identified, based on the following - according to Alex Goode at MDH:
“1. COVID-19 case is matched to the CRISP master patient index based on Firstname, Lastname, DOB, Gender, Address - pulls in something called EID - Enterprise Identifier
2. Immunet data is separately matched to the CRISP master patient index on Firstname, Lastname, DOB, Gender, Address
3. Immunet Data is added to the contact object where contact.EID = immunet.EID
4. We provide the Immunet Names and DOB to ensure the match is accurate, since there is a very small chance of mismatching, although this match is usually > 99% reliable and catches variations in name and address entry, and accommodates simple transpositions of dates of birth, etc in one of the sources - still allowing the records to match.”

Data source 2. New COVID-19 cases: from MDH (https://opendata.arcgis.com/datasets/0573e90adab5434f97b082590c503bc1_0.geojson)

Data source 3. Number of people fully vaccinated: Immunet data from MDH.

Data source 5. COVID-19 hospitalization data; updated data from MDH received on Fridays.

Data source 5. VOC/VOI cases: A “report”, Cases with VoC, from COVID Link (http://mdcovid.lightning.force.com/).

Also, see below for further information about testing for variants (source: Email communication with Monique Duwell, MDH, Chief, Center for Infectious Disease Surveillance and Outbreak Response; July 29, 2021)

"1. Can you tell us roughly what percent of COVID-19 cases are tested for variants in Maryland? Currently, and, if different, in the past as well?
Based on the sequencing results reported to our surveillance program, about 15% of COVID cases are being sequenced. This is likely an under-estimate of the true % of sequencing being performed statewide.

  1. Is the testing done randomly or more focused among certain patients (e.g., by age, by vaccination status, by severity/admission status)?
    We focus on sequencing patients who are hospitalized, deceased, fully vaccinated, live in high-risk settings, and who have traveled. However, we are also obtaining sequencing results on many patients who do not fall into these categories, since some labs sequence all patients, regardless of risk factors.

3.Do you know the general turnaround time for the variant test result: from specimen collection to data entry in the COVID link?
It varies greatly, largely due to the variability in amount of time it takes for labs to complete sequencing, and the amount of time it takes for them to upload the data to GISAID or report to us directly. I would say that about 3-4 weeks is the average turnaround time."