Assessing Registration and Screening Technologies for

Efficient Mass Vaccination and Public Health Monitoring

Eva K. Lee

1,2,3 a

and Kevin Yifan Liu

1,2

The Data and Analytics Innovation Institute, Atlanta GA 30309, U.S.A.

Georgia Institute of Technology, Atlanta GA 30322, U.S.A.

Accuhealth Technologies, Atlanta GA 30310, U.S.A.

Keywords: Point-of-Dispensing, Vaccination Clinic, Public Health Monitoring, Adverse Effect, Registration, Screening,

Barcode Scanner, QR Code, Immunization Information System, Vaccination Efficiency, Systems Simulation,

Performance Optimization, Resource Allocation, Nonlinear Mixed Integer Program.

Abstract: Vaccine data collection during mass vaccination campaigns is a difficult task due to the lack of a unified

system; yet, accurate and timely documentation is essential for monitoring efficacy and adverse effects. In

this study, we evaluate five electronic registration and screening technologies to test for how quickly

immunizations could be delivered and recorded given the different physical and cyber requirements of the

different technologies. Using time−motion studies and service data analysis from influenza vaccination

campaigns, we demonstrate operations and tracking efficiency with throughput improvements of 16% to 45%.

Based on these findings, we propose a prototypical unified system for dispensing, monitoring, and assessment

that is interoperable with existing immunization and electronic medical record systems. This paper highlights

the potential of electronic technologies to significantly enhance processes in vaccine administration and data

management. With the resource-constrained public health setting, the design emphasizes on minimally-

enhanced technology requirements to achieve seamless data and process management and improved

operations efficiencies. The system is flexible, scalable, and adaptable for different types of medical

countermeasures.

1 INTRODUCTION

During a pandemic or other vaccine-preventable

disease outbreak, it is essential to vaccinate as many

people as quickly as possible. Delaying vaccination

can lead to widespread illness and loss of lives and

serious burdens to healthcare systems (https://www.

cdc.gov/globalimmunization/fast-facts/index.html;

Lee, E. K., et al., 2015; Miller, M. A., et al., 2008;

Tisoncik, J. R., et al. 2009; Wein, L. M., et al., 2003).

There has been active research in advancing

operations and logistics to maximize vaccination and

mass dispensing throughput under time and resource

constraints (Cot, C., et al., 2021; Danzon, P. M., et al.

2005; Ferreira, L. S., et al., 2022; Hupert, N., et al.,

2002; Lee, E. K., et al., 2006a; Lee, E. K., et al.,

2006b; Prieto Curiel, R., et al., 2021; Wagner, C. E.,

et al., 2022; Washington, M. L., 2009). To establish

real-time, locale-specific and scalable capabilities for

public health decision-making, working with the

https://orcid.org/0000-0003-0415-4640

Centers for Disease Control and Prevention (CDC),

Lee et al. developed RealOpt©, a large-scale

information decision support system that seamlessly

integrates a disease modeling engine, a simulation

module and optimization technology into a unified

all-purpose response system. RealOpt allows real-

time analysis to maximize operations efficiency,

optimize staffing and resource allocation, analyze

layout design, and mitigate disease spread (Lee, E. K.,

et al., 2010; Lee, E. K., et al., 2013; Lee, E. K., et al.,

2017; Lee, E. K., et al., 2021; Kwong, J. C., et al.,

2010).

The recent COVID-19 pandemic caused by the

SARS-Cov-2 virus underscores the importance of the

rapid development of effective medical

countermeasures. Through Operation Warp Speed, a

partnership between the Departments of Health and

Human Services and Defense that aimed to help

accelerate vaccine development, the first COVID-19

vaccine was distributed to the public by December

156

Lee, E. K., Liu and K. Y.

Assessing Registration and Screening Technologies for Efﬁcient Mass Vaccination and Public Health Monitoring.

DOI: 10.5220/0013571100003967

In Proceedings of the 14th International Conference on Data Science, Technology and Applications (DATA 2025), pages 156-170

ISBN: 978-989-758-758-0; ISSN: 2184-285X

2020. Still, the United States reported over 111

million cases, resulting in over 1.2 million deaths.

Worldwide, over 7 million died from the disease.

Despite the vaccine development success, the

initial rollout of COVID-19 vaccines in the U.S. was

chaotic. There were supply shortages, distribution

issues, confusion over eligibility and a serious lack of

documentation. Many states struggled to set up

efficient appointment systems. Each state had its own

vaccine sign-up website, often run by the state health

department or local county health offices. Some of

these systems crashed due to high demand. Hospital

and healthcare provider systems used their own

patient portals for scheduling vaccine appointments,

and major pharmacy chains set up scheduling systems

too. With no centralized booking system, people had

to check different sites — state portals, pharmacy

chains, hospitals, and mass vaccination sites —

hoping to find an available slot. Many people signed

up on multiple websites, leading to wasted

appointment slots. The high traffic overwhelmed and

crashed scheduling systems, causing long wait times,

and raised access issues for seniors and underserved

communities. At the vaccination sites, clients were

given a paper card with their name, the vaccination

date and type, and a short note regarding where to

report an adverse effect.

Currently, there is no unified or integrated system

for effective vaccine data collection. Unless steps are

taken, inefficiencies such as these could occur during

a future emergency. Information regarding uptake is

critical for monitoring adverse effects and vaccine

efficacy. This is especially important when a newly

developed vaccine is being dispensed or when

multiple doses of vaccine are needed per person.

These data can provide valuable information during

an emergency and serve multiple purposes for public

health planning and research (Kwong, J. C., et al.

2010).

In this paper, we describe and analyze five types

of electronic technologies used for registration and

screenings in vaccination clinics. We contrast their

functionalities, usability and operations performance

based on time-motion studies and service data

collected during actual influenza vaccination

campaigns (“Campbell mass dispensing flu clinic”

CDC Report, n.d.; “Platte mass dispensing flu clinic”

CDC Report, n.d.; “Denver public health closed POD

for flu vaccination” CDC Report, n.d.; “Philadelphia

department of public health Closed POD for flu

Vaccination” CDC Report, n.d.; “CDC closed POD

for flu vaccination” CDC Report, n.d.) (Section 3.1).

We evaluate their dispensing performance under an

optimal dispensing clinic design (Section 3.2). Our

analysis shows that these electronic technologies can

improve operations and tracking efficiency with 16%

to 45% increase in overall throughput. Based on our

assessment and analyses, we propose a unified

prototypical registration and screening system with

integrated information flow for vaccination that can

be used for dispensing, monitoring and assessment

(Section 3.3). The system is interoperable to the local

Immunization Information system and electronic

medical record systems. The design is flexible and

adaptable for different types of medical

countermeasures and can be used by a broad spectrum

of regional public health departments.

2 METHODS AND DESIGN

This paper presents the first study using time-motion

techniques to scrutinize point-of-dispensing

operations by analyzing five different (inexpensive

and practical) electronic technologies for data input

and collection. RealOpt, an informatics-decision-

support enterprise system used by over 14,000 public

health emergency response users across 50 states, is

used to simulate and optimize the dispensing

operations to contrast the performance of each

technology. Through hundreds of time-motion

studies, it has proven that RealOpt is able to predict

the performance of the actual system throughput well

(within 95% to 105%). Hence, the analysis herein

should offer useful foundations and insights into the

potential improvement that can be achieved by the

electronic data collection technologies. Figure 1

briefly layouts the schematic method and design of

our study.

Figure 1: A schematic layout of the method and design.

Assessing Registration and Screening Technologies for Efﬁcient Mass Vaccination and Public Health Monitoring

157

2.1 Technologies for Registration and

Screening

A variety of methods have been used for client

registration and screening (Nanji, K. C., et al., 2009;

Pereira, J. A., et al., 2012). Based on a catalogue of

available electronic technologies developed by CDC

(Oak Ridge Associated University, 2016) and several

hundreds of time-motion studies on mass vaccination,

five representative technologies were investigated,

including barcode, mobile and magnetic stripe

scanners and web-based software modules. They

were chosen by public health leaders as candidates

and were supported by comparison results (done

independently) to be practical and cost-effective in

actual settings.

At the point-of-dispensing sites, through time-

motion study, we recorded the completion of each

task by observing and documenting staff members

performing their duties and interacting with clients.

Using RealOpt (Lee, E. K., et al., 2010; Lee, E. K., et

al., 2013; Lee, E. K., et al., 2017; Lee, E. K., et al.,

2021; Kwong, J. C., et al., 2010), a CDC-

recommended vaccination clinic is designed and used

to analyze the vaccination throughput performance

and contrast the strengths and limitations of each

technology. Table 1 summarizes the five data

collection technologies and its usage information.

Table 1: Summary of the five studied sites.

Gillette,

omin

Platte City,

Missouri

Denver,

Colorado

Philadelphia,

Atlanta, Georgia

Vaccination period

7am–7

m 10am–2

m 12

m–4

m 8am–1

m 8:30am–3

POD location

Public health

facilit

Middle school Public health

facilit

Police

uildin

State facility

Dispensing module

Walk-in open

POD for all

residents

Drive-through

and walk-in

open POD for

adult residents

(18 years or

older

)

Closed POD

for employees

with valid

badges

Closed POD

for first

responders

and their

family

member

Closed POD for

site employee

Data collection technology

Barcode

scanner and

Wyoming's

Immunization

Registry

(

)

Dispense

Assist

Handheld

Automated

Notification

for Drugs and

Immunizations

(

HANDI

)

IDWedge &

AutoFill

Countermeasure

and Response

Administration

(CRA)

We briefly describe each technology below.

2.1.1 Barcode Scanner Linked to an

Immunization Registry

Barcode scanner linked to an Immunization Registry

was used at the Wyoming event. A Honeywell 4600G

Barcode scanner was used to scan the linear and 2D

barcodes on the Wyoming Drivers’ licenses. The

scanned information includes name, address, date of

birth, and gender. This information is populated onto

the Wyoming's Immunization Registry (WyIR). If the

scanning is successful, the system searches for

existing information linked to the driver’s license. For

those without valid licenses (new licenses or children)

or if the scanning failed, the demographic information

is input manually. Vaccine information is also entered

manually into the registry.

The entire setup is rather straightforward.

Scanning itself does not require an internet

connection. The scanners can be connected to the

computer via USB ports. A software program is used

to sync the barcode scanner to WyIR, allowing the

scanned data to be populated onto WyIR in real-time.

WyIR is an Immunization Information System

(IIS) and within it there is a password-protected mass

vaccination module, among other modules. Lot

numbers and other vaccination data (vaccinators,

manufacturers, clinic dates, etc.) can be input into

WyIR prior to the POD operations. Internet

connection is required to use the WyIR.

There are two levels of access within WyIR: entry

access and full access. Vaccine administrators login

via entry access to view data, conduct queries on

clients, and perform data entry. Full access allows

inventory management, and more complex tasks.

Vaccine can also be ordered from WyIR.

The barcode scanner-WyIR setup allows direct

import of client demographic information, reduces

data entry of client information after the dispensing

event, hence saving time and reducing potential

errors. The state health department provides technical

assistance for WyIR through email or telephone.

2.1.2 Dispense Assist

Dispense Assist is an online tool developed by the

Johnson County Department of Health and Environ-

ment in Kansas (https://www.kshcc.com/ dispense-

assist.html). It is used by multiple county health

departments across the nation. Dispense Assist collects

both client information (name, address, phone number,

date of birth, and gender) and vaccine information (the

vaccine type (CVX code), manufacturer (MVX code),

expiration date, injection site, route of administration,

provider and type and publication date).

Clients fill out the registration form either through

a web browser or a mobile app. The questions cover

demographic information and medical screenings.

The medical screening determines which medication

or treatment the client is eligible for. One family

member can fill out forms for an entire family. The

address input in the first form can be saved as a

DATA 2025 - 14th International Conference on Data Science, Technology and Applications

158

default address and pre-populated to subsequent

forms. Multiple vouchers can be saved on one device.

Dispense Assist interfaces with barcode

scanners (Motorola Symbol) and printers. The

system generates QR codes that can be scanned

directly from the mobile app. Vouchers can be

printed or saved to a mobile device. Alternately, it

can be sent to an email account set up by the clinic

and printed when the client arrives at the POD.

Scanning the QR code adds an entry to an Excel

spreadsheet. This spreadsheet can later be exported

to electronic medical records (EMRs) after being

reformatted to fit the EMR's specifications.

Real time usage requires an Internet connection, a

barcode scanner that can read QR codes ($80) and a

USB connection between the scanner and a laptop.

Once the QR code is scanned, the corresponding

coded data will pop up on the laptop screen for POD

staff to review. A server that can handle thousands of

simultaneous hits is required. Dispense Assist can

also be downloaded and used off-line. Staff can

record vaccination information on the form and scan

them into the database later. The system has Spanish

translation capability through a web-link.

2.1.3 Handheld Automated Notification for

Drugs and Immunizations (HANDI)

Handheld Automated Notification for Drugs and

Immunizations (HANDI) is a mobile device

application developed by Denver Public Health to

improve data capture and efficiently administer

immunizations for closed POD operations. The

system consists of two parts: a backend database

system and a front-end handheld device (iPod Touch

and scanner). It uses a three-step process to collect

data: (a) client pre-registration; (b) medical

contraindication; and (c) vaccination information.

Clients register online through a website setup via

HANDI. Once registered, the client’s employee ID is

scanned using a scanner attached to the mobile device.

The scanner reads the magnetic stripe on the ID. If the

identification card is not readable or available, the

information will be manually entered into the device.

Otherwise, registered information of the client will be

loaded for verification. Any additional information can

be updated at this point. Once scanned into the system,

it helps to monitor contraindications and track

immunizations, and treatment plans administered

during mass intervention events.

At the POD vaccination station, the nurses scan

the information to identify the vaccinator, lot number,

dosage and site name before administering the

vaccine. The information on the mobile device is sent

to the server after the POD is closed. The server

bundles the data from each station and the complete

data set is securely transferred to a designated

database or registry (e.g., a state immunization

registry). The application server stores the

information in a SQL-server database and can export

the data to other designated formats. For example,

HANDI can convert the data to HL7 and send its

records to any system that accepts HL7.

2.1.4 IDWedge & AutoFill IDWEdge

IDWedge & AutoFill IDWedge (Tokenworks)

supports scanning of drivers’ licenses and military

IDs from all U.S. states and all Canadian provinces.

Used in conjunction with IDWedge, AutoFill is a

Windows application that can be configured to

automatically populate forms with information from

scanned IDs. Prior to the clinic, a simple user defined

formula is generated to specify the field order (first,

last, etc.) and the keystrokes (tab, arrow up, enter)

which are sent to a database.

At the dispensing site, staff swipe drivers’

licenses or state identification cards using IDWedge

to collect client identification information. The data

prefills into the customized Microsoft Access

database in real time. The licenses scanned and their

vaccination records are automatically updated. For

children who do not have valid drivers’ licenses, their

parents’ licenses are scanned, and the children’s

personal information is input manually.

AutoFill is configured to automatically populate

the Access form with the client data. The Access form

also contains checkboxes for screening questions

(manually entered by the data collectors) and

vaccination information. The manufacturer, lot

number, and date of administration are copied for all

clients, while the injection site and provider are

manually entered for each vaccination. After the

event, the Access databases from each computer are

downloaded, merged, and manually cleaned up,

removing duplicates and populating missing fields.

The final merged database is uploaded onto the IIS.

2.1.5 The Countermeasure and Response

Administration (CRA)

The Countermeasure and Response Administration

(CRA) is a web-based application developed by CDC

contractors. Prior to the event, staff customize the

form manually and select various fields to record. The

only client ID collected is the employee ID number.

Information including the vaccine type (CVX code),

manufacturer (MVX code), dose number, expiration

date, dose volume and unit, injection site, lot number,

Assessing Registration and Screening Technologies for Efﬁcient Mass Vaccination and Public Health Monitoring

159

date of administration, and provider can be recorded.

In addition, the VIS type and publication date are

collected with the latter given to the client.

At the vaccination event reported herein, four

fields were selected: employee ID number, date of

administration, vaccine lot number, and injection site.

The date of administration is pre-populated with the

event date and the vaccine lot number is pre-

populated with the lot number of the last vaccine

scanned. During the event, when a new batch of

vaccine is used, the staff scans the vaccine. The

vaccinator only records manually two fields, the

employee ID number and the injection site.

CRA can work as a stand-alone system to support

jurisdictional operations if an Internet connection is

not available. Aggregate data and reports can be

produced for individual clinics and client populations.

CRA can also create recall reminders for future doses

that clients may need. These reminders can be sent to

both providers and clients. After the event, all

recorded information is exported to a single file and

sent to the employer’s electronic medical record.

2.2 Analyzing Operational

Performance via a Common

Vaccination Clinic Layout

To gauge the operational performance, we derive

experiments to analyze the importance of each

technological component with respect to POD

operations and the overall performance. Figure 2

shows a medical dispensing clinic process flow

recommended by CDC for mass medical

countermeasure dispensing and commonly used by

public health sites.

Figure 2: A medical dispensing clinic used for mass

vaccination.

2.3 Simulation-Optimization

Computational Platform and

Multi-Objective Resource

Allocation

2.3.1 The Simulation-Optimization

Computational Platform

RealOpt is designed for responding to emergencies,

including biological, chemical, radiological, nuclear

incidents and natural disasters. Figure 3 shows an

overview of RealOpt-POD for biological defense and

the analytic methodologies that are embedded in it. In

the frontend, users interact with the system through

the multi-panel, cognitive-analytics interface. In the

backend, the simulation-optimization module

consists of the large-scale simulator and the rapid

optimizer. The solution engine works by running

optimization and simulation in an interlacing manner.

The system has been used for mass vaccination (and

diagnostic testing) events including seasonal flu,

Ebola, Hepatitis, and COVID-19. It has been used for

fire, flood and hurricane responses (for setting up

shelters, food/medical/supply distribution sites, and

command-control logistics etc.), Haiti earthquake

emergency response relief, Japan Fukushima

radiological response (rapid screening and

decontamination, food/shelter, supplies

distributions), hurricane Sandy response, optimizing

Ebola treatment center operations, and predicting and

containing Zika disease spread. The system has also

been used by numerous hospitals/clinics for

optimizing hospital workflow and operations

efficiency (Lee, E. K., et al., 2015). The system

allows management of multiple resource types

simultaneously, for example, labor, equipment, beds,

financial investment, medical supplies, and

countermeasures (Lee, E. K., et al., 2010; Lee, E. K.,

et al., 2013; Lee, E. K., et al., 2017; Lee, E. K., et al.,

2021; Lee, E. K., Li, Z. L., Liu, Y. K., & Leduc, J.,

2021; de Mesquita PJB, et al., 2021).

Figure 3: The architectural design of RealOpt-POD.

2.3.2 Nonlinear Mixed-Integer Program for

Optimizing Resource Allocation and

System Performance

Within RealOpt, system optimization can be

performed to ensure the best operations and system

performance (throughput, wait-time, queue length,

DATA 2025 - 14th International Conference on Data Science, Technology and Applications

160

utilization, etc). The resource allocation is modeled

via a nonlinear mixed integer program (NMIP).

Resources can include labor, equipment, computer,

etc. Constraints in the model include: (a) maximum

limits on wait time and queue length (which is

dictated by the capacity of the waiting room in the

facility and also the desire to maintain sufficient

distancing to minimize potential infection); (b) range

of utilization desired at each station; (c) assignability

and availability for each resource group, and resource

types at each station (i.e., the skill set and the numbers

available); and (d) maximum limit on the cycle time

of the individual. Mathematically, the model

parameters are defined as follows:

 R: the set of resource groups.

 𝐓



: the set of resource types in resource group

r, 𝑟∈𝐑.

 S: the set of services in the process flow.

 𝐒



⊆𝐒: the set of services in which resource

type i in resource group r can be assigned. This

models the assignability of the resource (e.g.,

based on skills of workers).

 𝑘



: the cost of assigning a resource of type i in

resource group r to station j. 𝑟∈𝐑, 𝑖∈𝐓



, 𝑗∈

𝐒



 𝑚



and 𝑚



: the maximum and minimum

number of resources of type i in resource group

r that may be assigned to station j. 𝑟∈𝐑, 𝑖∈

𝐓



, 𝑗∈𝐒



 𝑛



: the number of available resources of type i

in resource group r. 𝑟∈𝐑, 𝑖∈𝐓



 𝑤



, 𝑞



, and 𝑢



: the average wait time, average

queue length, and average utilization rate,

respectively, at station j. 𝑗∈𝐒.

 c: the average cycle time (i.e., the length of time

a customer spends in the system).

 𝜃: the average throughput (number of customers

served in a specified period).

The decision variables for this problem are 𝑥



∈

𝐙



: the number of resources of type i in resource

group r assigned to station j. 𝑟∈𝐑, 𝑖∈𝐓



, 𝑗∈𝐒



We can represent the cost at each station j as

𝑔





∑

𝑘



𝑥



(

,

)

∈𝛀



, 𝑤



, 𝑞



, 𝑢





, 𝑗∈𝐒, where 𝛀





(

𝑖,𝑟

𝑟∈𝐑, 𝑖∈𝐓



, 𝑗∈𝐒





. The total system cost

depends on the cost at each station, and on system

performance variables, such as cycle time and

throughput. Thus, we can represent the total cost as

𝑓

∑

𝑔

∈𝐒

, 𝑐, 𝜃. Here, 𝑔



and 𝑓 are functions that

are not necessarily expressible in closed form. We can

formulate a general representation of the multiple

resources allocation problem as

Min

𝑧=𝑓



𝑔



∈𝐒

, 𝑐, 𝜃



s.t.

𝑚



≤𝑥



≤𝑚



∀ 𝑟∈𝐑, 𝑖∈𝐓



𝑗

∈𝐒



(1)

𝑥



∈𝐒



≤𝑛



∀ 𝑟∈𝐑, 𝑖∈𝐓



(2)

𝑤

(

𝑥

)



≤𝑤



𝑞

(

𝑥

)



≤𝑞



𝑢



≤𝑢

(

𝑥

)



≤𝑢



∀

𝑗

∈𝐒 (3)

𝜃

(

𝑥

)

≥𝜃



𝑐

(

𝑥

)

≤𝑐



(4)

𝑥



∈𝐙



∀𝑟∈𝐑, 𝑖∈𝐓



𝑗

∈𝐒



(5)

Constraint sets (1) to (5) form a NMIP problem

for cost minimization under the constraints of

multiple resources allocation and stochastic system

performance. Constraint sets (1) and (2) describe the

resource availability for each service or decision.

Constraint sets (3) and (4) ensure that services satisfy

safety guidelines to minimize potential cross

infection, e.g., individuals are not waiting for too

long, the queues are of reasonable size, workers are

not overwork, individuals do not spend excessive

amount of time inside the dispensing facility, etc.

We note that Constraint sets (3)-(4) are system

parameters in the simulation, and performance

variables in the optimization. Since some of the

functions in the objective and constraints are not

necessarily expressible in closed form, the system

optimizes the overall outcome performance using the

cycle time as a surrogate measurement and the effect

on different processes and the global return. The

problem is proven intractable by commercial systems.

RealOpt is designed to overcome such computational

bottlenecks by interweaving rapid system simulation

and optimization (Lee, E. K., et al., 2010; Lee, E. K.,

et al., 2013; Lee, E. K., et al., 2017; Lee, E. K., et al.,

2021).

Given a clinical process configuration with

various service distributions for each process and

decision point and associated performance metrics,

RealOpt simulates the entire clinical process to

acquire the cycle time and the system performance

under the configuration. The output simulation

statistics are then input into the stochastic NMIP

optimization model where performance and resources

are optimized. The resulting optimization output is

entered back into the simulation to obtain the next

system performance metrics. The simulation-

optimization iterates until no further improvement is

achieved.

Assessing Registration and Screening Technologies for Efﬁcient Mass Vaccination and Public Health Monitoring

161

3 RESULTS

3.1 Time-Motion Study and Data

Collection for Modeling

3.1.1 System Functionalities

Based on system functionalities and time-motion

studies, we summarize key features and user

experiences for each registration and data collection

technology in Table 2.

Table 2: Comparison of the five registration technologies.

Scanner+

Dispense

Assist

HANDI

IDWedge

& AutoFill

CRA

Pre-re

istration online

Yes - via

website or

mobile app

Yes - via

internal

website

No No

Onsite re

istration

Yes - via

website or

mobile a

Yes - via

internal

website

No No

ut mechanism: how client data are read into the s

stem

Barcode

scanner

Barcode

scanner

Magnetic

stripe

scanne

Magnetic

stripe

scanne

Manual

input

Identification used as input for client data

Driver’s

license

System

generated

QR code

Employee

Driver’s

license

Employee

Pediatric

Manual

input

QR code

Not

supporte

Manual

input

Not

supporte

Data collected - Client data

Name

Yes Yes Yes Yes Yes

Address

Yes Yes Yes Yes

Birthda

Yes Yes Yes Yes

Insurance status

Yes

ID t

Yes

Gende

Yes Yes Yes

Phone numbe

Yes

Vaccine data

Vaccination date

Yes –

manual

input

Yes -

manual

input

Yes -

scanned

Yes -

manual

input

Yes -

manual

input

Vaccine dose numbe

Yes -

scanne

Vaccine dose volume and unit

Yes -

scanne

Vaccine ex

iration date

Yes -

manual

Yes -

manual

Yes -

scanned

Vaccine data

Vaccine in

ection site

Yes -

manual

Yes -

manual

Yes -

scanned

Yes -

manual

Yes -

manual

Vaccine lot numbe

Yes -

manual

Yes -

manual

Yes -

scanned

Yes -

manual

Yes -

manual

Vaccine manufacturer

(

MVX code

)

Yes -

manual

input

Yes -

manual

input

Yes -

scanned

Yes -

manual

input

Yes -

manual

input

Vaccine

rovide

Yes -

manual

Yes -

manual

Yes -

scanned

Yes -

manual

Yes -

manual

Vaccine route of administration

Yes -

manual

Yes -

scanned

Vaccine type (CVX code)

Yes -

manual

Yes -

manual

Yes -

scanned

Yes -

manual

VIS date

iven to client

Yes -

manual

Yes -

manual

VIS t

e &

ublication date

Yes -

manual

anization of data durin

/after events

Database t

Interfaced

directly

with IIS

Generates a

generic

Excel form

Data stored

in local

SQL-server

database

Data stored

in local

Access

database

Data stored

in the local

database

Imports data electronically to database (merge automatically to the

database)

Yes -

Interfaced

directly

with IIS

No -

generate

Excel files

that needs

manual

operation

Yes - The

server

bundles the

data from

each station

and the

complete

data set is

securely

transferred

to a

designated

database or

istr

Yes - data

stored in

Access

database

Yes - data

stored in

local

database

Transfer data to other formats (EMRs, etc.)

No - data

directly

reported to

IIS

No - data

stored in

local Excel

files and

needs

manual

operations

to transfer

to other

formats

Yes - can

transmit the

data to

other

designated

formats,

convert to

HL7 and

send its

records to

any system

that accepts

HL7

No - data

stored in

local

Access

database

and needs

manual

operations

to transfer

to other

formats

Yes - data

exportable

to other

databases

including

the

electronic

medical

records and

CDC’s

occupationa

l health

database

DATA 2025 - 14th International Conference on Data Science, Technology and Applications

162

Table 2: Comparison of the five registration technologies

(cont).

Organization of data during/after events

uire internet connection durin

events

Yes – for

connecting

to IIS

during

operations

Yes - for

onsite

registration

(5/4G or

LTE

connections

for mobile

phones will

also work)

Yes - for

onsite

registration;

but

essentially

only

Ethernet is

needed as it

is an

internal tool

for closed-

POD

No – unless

connecting

to IIS is

necessary

No – does

not need

connect to

any

database in

real time;

standalone

deployment

available;

registration

and

vaccination

are

performed

by the same

nurse

Scalabilit

for lar

er events

eed of data in

Fast with

scanner,

slow with

manual

Fast with

scanner,

slow with

manual

Fast

Fast with

scanner,

slow with

manual

Slow

Need for reliable servers

Yes Yes Yes No No

Observations durin

actual events

Trainin

Just-in-time

training;

workers

performed

well

Just-in-time

training;

workers

performed

well

Team has

5-year of

experience

Just-in-time

training;

workers

performed

well

Difficult for

vaccinators

to use,

more prone

to entry

errors

Pediatric

children

without

driver’s

license take

longer to

rocess

Via QR

code, same

as adults

Not

supported

children

without

drivers’

license take

longer to

rocess

Not

supported

Scanne

Failed to

scan some

drivers’

licenses

No scanner

available

on site. Pre-

filled forms

had to be

emailed

and printed

on site

Occasionall

y failed to

scan some

drivers’

licenses.

Registration

Automatic

Required

clients to

online,

many

forgot

Only

limited

employees

showed up;

unclear if

system is

scalable for

lar

e events

Automatic

mechanism

to input

client

information

automatical

Health information

Not

connected

to EMR;

Health

information

needs to be

manually

filled out

clients.

Not

collected

Not

collected

Not

collected

Not

collected

3.1.2 Service Time Distributions

Observations were recorded for all clients entering

the vaccination site. Table 3 below shows the staffing

and total throughput (number of observations

collected) at each site.

Table 3: POD hours, staffing and throughput of the five

sites.

Gillette,

Wyoming

Platte City

Missouri

Denver,

Colorado

Philadelphia,

Atlanta,

Georgia

Vaccination period

7am

–

7pm 10am

–

2pm 12pm

–

8am

–

1pm 8:30am

–

3pm

POD type

Open POD Open POD Closed POD Closed POD Closed POD

Total throughput

650 adults

50 children

200 275 1,400 310

Staffing

8 staff, 4

volunteers

10 staff, 4

volunteers

4 11 10

Nurses

14 7 2 14 5

During time-motion studies, we observe several

factors that influence the service time.

 Technology familiarity: For Scanner+WyIR,

Dispense Assist, and IDWedge systems, POD

workers received just-in-time training on the use

of the technology. For HANDI, the staff have used

it for multiple years and are quite proficient in its

usage. For CRA, workers log in a few days prior

to the vaccination event to learn of its usage and

to pre-set the dispensing information.

 Incomplete technology: Scanning devices were

not available for the QR codes generated from

Dispense Assist to fully take advantage of its

capability. If equipped with proper scanning

devices to scan the generated QR codes (as the

technology is designed for), the service times will

be significantly shortened.

 Human factor: Staff served at varying rates

depending on how busy the clinic was. Staff

tended to operate faster when the queue was long,

and slower when the clinic was not busy.

To address some of these factors, we perform

three steps: (a) Within each observed time, we

subtract from it the time spent on leisure chat between

the worker and the client (which we also collected).

(b) We remove outliers that are greater than the 90

percentiles. (c) We use only the service times

reflecting staff performance when the demand is high.

Figure 4 shows the adjusted service time with the

associated best-fit distribution (Chi-Square

Goodness-of-Fit Test) for registration and scanning

(data collection) for the five technologies and the

current manual process. The boxplots in Figure 5

Assessing Registration and Screening Technologies for Efﬁcient Mass Vaccination and Public Health Monitoring

163

contrasts these service times for each technology. It

shows that scanner-type technology shares

comparable median time while computer input

requires a longer time. The figure depicts clear time

variance among users or types of clients.

Data collection time for

barcode scanner with WyIR

𝑙𝑜𝑔𝑛𝑜𝑟𝑚𝑎𝑙(21.84,9.45)

Data collection time for

Dispense Assist

𝑙𝑜𝑔𝑛𝑜𝑟𝑚𝑎𝑙(16.74,8.67)

Data Collection time for

HANDI

𝑙𝑜𝑔𝑛𝑜𝑟𝑚𝑎𝑙(16.57,4.99)

Data collection time for

IDWedge

𝑙𝑜𝑔𝑛𝑜𝑟𝑚𝑎𝑙(20.01,9.68)

Data collection time for

CRA

𝑙𝑜𝑔𝑛𝑜𝑟𝑚𝑎𝑙(28.10,13.64)

Data collection time for

manual input

𝑙𝑜𝑔𝑛𝑜𝑟𝑚𝑎𝑙(40.23,16.39)

Figure 4: Service time distributions with the associated

best-fit distribution for each technology after adjustment:

using only service times reflecting performance of staff

when demand is high. This excludes time spent on leisure

chats and outliers.

We note that across the five sites, there is marginal

difference in the service times for vaccination,

medical evaluation, and post-vaccination observation

respectively.

During the actual events, 90% of the clients in

Gillette, Wyoming event had valid Drivers’ licenses,

60% of the clients pre-registered at the Platte City

Missouri site, 86% of the clients pre-registered for

HANDI in Denver and 87% of the clients in

Philadelphia had valid Drivers’ licenses. We note that

at the Wyoming site, the form-filling process by the

clients themselves does not affect the performance of

the technology nor the POD process.

Figure 5: Boxplots contrasting the adjusted service times

for each technology.

3.2 Operations and Performance

Efficiency

Table 4 summarizes the associated service time

distributions and decision probabilities that are

collected from the time-motion studies. Using the

vaccine clinic layout in Figure 2 and these service

distributions, we simulate and perform systems,

operations and resource allocation optimization using

RealOpt for each of the technologies to obtain the best

throughout under the same number of staff (30) and

necessary equipment resource constraints. We report

some comparison findings below.

Table 4: Service time distributions and probabilities at

decision blocks obtained via time-motion studies of actual

influenza vaccination events.

Service Station Service Time Distribution

(all units in seconds)

Registration & Scanning Specific to each technology

as shown in Fi

ure 4

Best fit distribution obtained using combined data from

all five sites

Vaccination

𝑙𝑜𝑔𝑛𝑜𝑟𝑚𝑎𝑙

(

32.20,10.59

)

Medical Evaluation

𝑙𝑜𝑔𝑛𝑜𝑟𝑚𝑎𝑙

(

61.03,21.54

)

(Post-vaccination)

Observation

𝑢𝑛𝑖𝑓𝑜𝑟𝑚 (51.03,78.43)

Scanning Device versus Web Input Versus

Manual Approach. Figure 6 compares closed-POD

performance using magnetic strip scanners (HANDI,

IDWedge) versus input via website (CRA) versus the

manual approach. Magnetic strip scanners (HANDI

and IDWedge) gather the demographic information

of clients from their employee IDs or drivers’

licenses, while CRA requires workers to manually

input data onto the computer. When the pre-

registration percentage (for HANDI) or the

percentage of valid drivers’ licenses (for IDWedge)

are low, the resulting performance is similar to CRA.

During the actual vaccination events, 86% of the

Time (sec)

ens

0 10203040

0.00 0.02 0.04

Time (sec)

ens

0 5 10 15 20 25 30

0.00 0.02 0.04 0.06

Time (sec)

ens

10 15 20 25

0.00 0.02 0.04 0.06 0.08

Time (sec)

Density

0 5 10 15 20 25 30

0.00 0.02 0.04 0.06

Time (sec)

Density

10 15 20 25 30 35 40

0.00 0.02 0.04 0.06

Time (sec)

Density

10 20 30 40 50 60 70

0.00 0.02

WyIR DispenseAssist HANDI IDWedge CRA Manual

Technology

Data Collection Time (seconds)

DATA 2025 - 14th International Conference on Data Science, Technology and Applications

164

clients pre-registered, and 87% had valid drivers’

licenses. Thus, both HANDI and IDWedge produce

higher throughput than when using CRA by 14%.

However, we note that CRA results show that even

with a simple web-based system, the gain in

throughput can be significant (15% increase) when

compared to the traditional manual process. This

confirms that automatic scanning is essential for

efficient closed-POD operations and should be

incorporated into any data collection technology

Figure 6: Maximum throughput achieved when using

magnetic strip scanners, web-based input, and manual

registration respectively, under varying pre-registered

percentage or percentage with valid drivers’ licenses.

QR-Code versus Drivers’ ID. What is more efficient

for registration, QR codes or drivers’ licenses?

Scanner+WyIR and IDWedge use drivers’ licenses to

populate client demographic data automatically onto

the computer, while Dispense Assist generates a QR

code for each client (family) during online/mobile

registration for automatic population. Such

registration can be done either pre-event, or with on-

site kiosks or a mobile app. Figure 7 contrasts the

maximum throughput achieved by scanners

(Scanner+WyIR and IDWedge) versus QR-based

(Dispense Assist). When pre-registration and valid

drivers’ licenses percentages are both below 15%,

there is a marginal difference in the performance

among these technologies. However, Dispense Assist

shows a clear advantage as pre-registration picks up.

During the actual vaccination events, 60% of clients

pre-registered, while over 85% of clients have valid

drivers’ licenses. Uniformly, Dispense Assist

processes 20% more clients under the same resource

and time constraints. The IDWedge site performs

better than the Scanner+WyIR site. This may reflect

the fact that the former was a closed-POD operation,

where staff have used the same technology for five

years already; hence the POD operations ran more

efficiently. The clients were very comfor with the

environment. In general, QR code technology is very

competitive in an open-POD environment for

automatic registration. ID scanners are equally

competitive in closed-POD operations when most

clients have valid drivers’

licenses.

Figure 7: Maximum throughput achieved using QR codes

versus ID scanners, under varying pre-registered or

percentage with valid drivers’ licenses. Note that WyIR

uses onsite form-filling for medical screening while

IDWedge staff conducts verbal medical screening.

Verbal Medical Screening versus Form-Filling

Before Scanning. Both Scanner+WyIR and

IDWedge scan the drivers’ license to gather

demographic data automatically. In addition, clients

in Wyoming were asked to fill out a preliminary

medical screening form (by themselves) prior to

scanning, while Philadelphia IDWedge site

conducted onsite medical screening verbally during

scanning. Figure 7 (Blue versus Green) shows that the

throughputs are similar when the percentage of valid

drivers’ licenses is below 35%. Beyond 35%,

IDWedge is more competitive with slightly higher

throughput (10.5%) than Scanner+WyIR. When

every client has an ID, the difference in throughput is

only 1.9% (203 clients). During the actual events,

similar percentage of clients (90% vs 86%) had valid

IDs. The results show that verbal medical screening

(especially used in conjunction with the online EMR

or personal health records) is more efficient. The

time spent by nurses to review each paper form is not

negligible and may take up to a couple of minutes. In

addition, extra physical space is needed to

accommodate clients filling medical screening forms

onsite, which can potentially create congestion.

7000

7500

8000

8500

9000

9500

10000

10500

11000

11500

12000

0% 20% 40% 60% 80% 100%

Maximum throughput

Pre-registration % or % with valid drivers' licenses

Closed POD maximum throughput: scanner

(HANDI, IDWedge) vs web-base (CRA) vs manual

input

HANDI: pre-event registration %

IDWedge: % with valid drivers' license

CRA: web-based entry

Manual input

8000

8500

9000

9500

10000

10500

11000

11500

12000

0% 20% 40% 60% 80% 100%

Pre-registration % or % with valid drivers' licenses

Maximum throughput: scanner (WyIR, IDWedge)

vs QR code (Dispense Assist)

WyIR: % with valid drivers' license

Dispense Assist: pre-event registration %

IDWedge: % with valid drivers' license

Assessing Registration and Screening Technologies for Efﬁcient Mass Vaccination and Public Health Monitoring

165

During a pandemic, this may increase infection

transmission within the dispensing facility; hence

proper crowd control such as the use of multiple

rooms or open space may be needed to facilitate

distancing among clients while completing the form.

Pre-Registration Impact. Dispense Assist uses

online pre-event registration and medical screening,

whereas IDWedge performs the entire process within

the POD (during scanning). We compare the

performance of these two technologies to gauge their

efficiency (Figure 7 Red versus Green). When less

than 15% of the clients pre-register or have valid ID,

performance of both technologies are similar. When

the pre-registration rate or percentage of valid

licenses range from 15% and 80%, Dispense Assist

performs better than IDWedge, while the maximum

difference in throughput is achieved at 49% with 10%

throughput difference. At a 60% pre-registration rate

for Dispense Assist and 87% valid ID for IDWedge,

there is only a marginal difference in the throughput.

We caution that Dispense Assist was used during an

open POD event (serving a more diverse population).

Hence its performance is remarkable when compared

to the closed-POD IDWedge results. Public

campaigns to encourage pre-registration during

mass dispensing events can help with the overall

efficiency and throughput

Open POD versus Closed POD. Dispense Assist is

used for general open POD events. The QR codes

generated via an online/mobile pre-registration is

scanned on-site. HANDI is used for closed POD

events to scan employees’ IDs. Both technologies

require pre-registration. Employee ID is an efficient

method for data input since every employee has one,

and with the same and standard data fields. Figure 8

show the maximum throughput under different pre-

registration rates. The two technologies have similar

trends as the pre-registration rate increases. At a low

pre-registration rate, HANDI has slight advantages in

the throughput. The results show that when

operating an open POD, QR codes are a good

alternative for employee IDs. QR code technology is

flexible in gathering information. Besides

demographics, and premedical screening questions,

additional information such as contact phone

numbers, primary physician, etc. can also be encoded.

This information can become crucial for health

monitoring and follow-up, or second dose reminder.

Although such information may also be recorded in

some employee IDs/records, they may not be

uniformed across different organizations.

Figure 8: Maximum throughput achieved for QR-based

technology (Dispense Assist) versus employee ID scanner

closed POD (HANDI) under varying pre-registered

percentage.

Figure 9a: Maximum throughput under an optimal POD

setup for a 6-hour shift using the observed time-motion

study data for each technology.

Figure 9b: Percentage throughput increases with respect to

manual registration is calculated as (x-manual)/manual *

100%.

Optimal Throughput For a 6-Hour Shift. Figure 9a

and 9b compare the maximum throughput of the five

technologies under the medical POD setup (Figure 2)

operating for a 6-hour shift. Each throughput is

obtained using the observed (real) time-motion study

data with the associated percentage of clients with

valid IDs and pre-registered. We also contrast the

results against the ideal scenario where all clients

have valid IDs and pre-registered (orange bar). In

contrast to the manual approach (Figure 9b), the CRA

8000

9000

10000

11000

12000

0% 20% 40% 60% 80% 100%

Pre-registration %

Maximum throughput achieved: open POD vs

closed POD

Dispense Assist: open POD QR-based

HANDI: closed POD employee ID scan

5000

10000

15000

WyIR: 90% vs 100%

valid drivers' license

DispenseAssist:

60% vs 100% pre-

registered

HANDI: 86% vs

100% pre-

registered

IDWedge: 87% vs

100% with valid

drivers' license

CRA

Maximum throughput

Maximum throughput under optimal POD

setup for a 6-hour shift

10%

20%

30%

40%

50%

WyIR: 90% vs

100% valid drivers'

license

DispenseAssist:

60% vs 100% pre-

registered

HANDI: 86% vs

100% pre-

registered

IDWedge: 87% vs

100% with valid

drivers' license

CRA

% increase over manual

input

Percentage of throughput increase w.r.t.

manual registration

DATA 2025 - 14th International Conference on Data Science, Technology and Applications

166

web-based approach increases the throughput by

16%. Scanning drivers’ ID or using QR codes (even

with just 60% of client pre-registration) can increase

the throughput by 30%. If more people pre-register,

the gain could be as high as 45%.

3.3 A Prototypical Electronic System

for Data Collection During Mass

Dispensing

Based on our findings, we design a computerized

system for data input and collection during dispensing

and post-event monitoring and assessment. Figure 10

shows the design structure and architecture of the

system. The system includes four key modules:

The digital registration module facilitates online

or on-site mobile registration. A QR code in

generated which stores the client information (e.g.,

name, address, zip code, immunization history etc.) in

standard format. QR codes are widely used in daily

activities including information gathering,

identification and mobile payment, air transportation

security (e.g., boarding passes), attendee ID at

conferences, and retails and sales. It is both reliable

and scalable.

The scanning module with a scanner, the client

information will pop up rapidly allowing staff to

review and cross reference to prepare for vaccination.

Scanning the barcode from vaccines saves time and

reduces errors. The scanning module streamlines the

POD operations, improves efficiencies and staff

utilization.

The database and interoperable encrypted

transfer module establishes a functional platform for

data analytics and interoperable data transfer. The

data collected is organized into various formats to

support key functions. SQL-like data queries and

management functionalities are employed to support

data analysis. HL7 and EMR-standardized format are

used to push data to immunization registries and local

EMRs. This is important for IIS data transfer — IIS

is critical for overall state level planning and

management beyond the individual POD site. The

system design is modulized with simple and

extensible architecture. This allows for flexible

expansion of system functions, ease of maintenance,

and connectivity to an EMR for pre-event cross-

reference on medical / vaccination history and post-

event monitoring and assessment of the efficacy of

the medical countermeasures.

The client-communication interface module is

critical for monitoring adverse effects and

determining vaccine efficacy. This is especially

important when newly developed vaccines are being

dispensed. Public health efficacy for the vaccines can

only be obtained when clients report their response /

health effect after vaccination. The communication

interface module is essential also for sending out

updates on vaccines, or reminders (to both providers

and clients), when multiple doses of vaccine are

needed per person. These features were sorely missed

(and are still missing) during COVID-19 vaccination

(Apartsin, K. A., et al., 2021; Chekol Abebe, E., et al.,

2022; Riad, A., et al., 2021; Yadav, T., et al., 2023).

This system captures the best among the five

technologies, it is practical, implementable, and

scalable.

Figure 10: The design architecture of an electronic system

for data collection during mass vaccination.

3.4 Contribution

Documenting clients, screening, and vaccinations

administered is of particular importance during mass

vaccination, since information regarding uptake is

critical for monitoring adverse effects and vaccine

efficacy. This is especially crucial when newly

developed vaccines are being dispensed, or when

multiple doses of vaccine are needed per person.

Hence, it is important that vaccine providers

document vaccinations administrated accurately and

quickly and upload this information onto the local

Immunization Information System. These data can

provide valuable information during an emergency

and serve multiple purposes for public health

planning and research. Currently, there is no uniform

or integrated system for effective vaccine data

collection. In this paper, we describe and analyze five

types of electronic technologies for registration and

screening in vaccination clinics. We contrast their

functionalities, usability, and operations performance

based on time-motion studies and service data

collected during actual influenza vaccination

campaigns. We evaluate their dispensing

performance under an optimal dispensing clinic

Assessing Registration and Screening Technologies for Efﬁcient Mass Vaccination and Public Health Monitoring

167

design. Our analysis shows that these electronic

technologies can improve operations and tracking

efficiency with 16% to 45% increase in overall

throughput. Based on our assessment and analyses,

we design a prototypical registration and screening

system with integrated information flow for

vaccination that can be used for dispensing,

monitoring, and assessment. The system connects to

the local Immunization Information system and

electronic medical record systems. The design is

flexible and adap and can be used by a broad spectrum

of regional public health departments.

4 CONCLUSIONS

This paper addresses the critical issue of efficient

vaccine data collection during mass vaccination

campaigns. Recognizing the importance of accurate,

timely documentation for monitoring adverse effects

and efficacy, we evaluate five electronic registration

and screening technologies to test for how quickly

immunizations could be delivered and recorded given

the different physical and cyber requirements of the

different technologies. Through time−motion studies

and service data analysis from influenza vaccination

campaigns, we demonstrate and throughput

improvements of 16% to 45%. Based on these

findings, we propose a prototypical integrated system

designed for dispensing, monitoring, and assessment,

emphasizing its flexibility and interoperability with

existing immunization and medical record systems.

During a pandemic or other vaccine-preven disease

outbreak health emergency, it is essential to vaccinate

as many people and as quickly as possible.

Documenting vaccines administered during the event

is particularly important, and vaccination administra-

tion information (i.e., patient and vaccine information)

will need to be reported to the jurisdiction’s

Immunization Information System quickly. Advanced

technologies play an irreplaceable role in healthcare

and vaccine dispensing and can be used to accelerate

the processing rate of vaccination clinics.

Few studies have investigated the time staff spend

registering clients and collecting data in a mass

vaccination or other dispensing event. This paper

presents the first study using time-motion techniques

to scrutinize point-of-dispensing operations by

analyzing five different (inexpensive and practical)

electronic technologies for data input and collection.

There are clear advantages of electronic input

technologies compared to traditional manual input

approaches. They automate or eliminate labor-

intensive tasks such as searching and typing, and

reduce manual-entry errors. These technologies can

improve POD throughput, operations efficiency, and

quantity and accuracy of collected data.

The five technologies analyzed herein offer

different input mechanisms. Scanner+WyIR and

IDWedge both scan Drivers’ drivers’ licenses to

populate demographic information, while the

vaccination information is input by staff members.

Both systems can be linked in real-time to a

jurisdiction’s immunization information system. This

assists during health screening of clients the

determination of the type of vaccination/medication

that should be used. CRA, on the other hand, allows

scanning of the vaccination information but requires

manual input of clients’ demographic information.

The two technologies that benefit from pre-

registration (HANDI and Dispense Assist) also utilize

scanning devices to facilitate input of information.

Moreover, Dispense Assist is the most flexible by

generating a QR code using data input by clients.

During observation, most on-site registrations are

done by manual form filling, even though electronic

form-filling options are available (the mobile app for

Dispense Assist, and the on-site web-based

registration for HANDI). This creates additional

workloads for both clients and staff: clients have to

spend extra time in the POD filling out the forms, and

staff need to spend extra time and pay attention to

reading the handwritings when inputting these hand-

filled forms to a computer system. Although

implemented not very efficiently at the influenza

vaccination clinic, Dispense Assist does provide a

very promising solution to overcome this issue. The

gain in throughput is remarkable.

A mobile app or website for registering clients

and disseminating knowledge regarding the type of

medical countermeasures to be dispensed is practical

and useful to the public. These sites can be accessed

easily via web browsers on clients’ mobile devices

and generate barcodes or QR codes that can later be

used for scanning. This setup can significantly reduce

the workload for POD staff and the time clients spend

inside PODs. It also improves accuracy in the data

collected.

RealOpt is a live evolving informatics decision-

support enterprise system and has over 14,000 public

health emergency response users across 50 states.

Among the hundreds of time-motion studies

conducted for real dispensing events, the results from

RealOpt simulation analysis predict the performance

of the actual system well (within 95% to 105%).

Hence, the analysis present herein provides useful

foundations and insights into the potential

DATA 2025 - 14th International Conference on Data Science, Technology and Applications

168

improvement that can be achieved by the electronic

data collection technologies.

Our analysis shows that electronic technologies for

registering and screening can improve 16% to 45% in

overall throughput. Moreover, it is not necessary for

a health department to seek the most advanced, most

cutting-edge electronic data collection technologies.

Our findings show that simple technlogy increment

(e.g. a scanner) can already provide improved

operations efficiency and data accuracy. Health

departments should carefully analyze their demand and

environment and choose the technology that best suits

their needs. Simple-to-use is key for just-in-time easy

training. Automation reduces input errors. These are all

key elements to consider during selection. Health

departments should incorporate these devices into their

routine vaccination process. This ensures staff

confidence in using the devices and that proper device

maintenance is performed.

We propose a prototypical integrated system

designed for dispensing, monitoring, and assessment,

emphasizing its flexibility and interoperability with

existing immunization and medical record systems.

The prototypical electronic system for data collection

during mass dispensing (a) transfers most of the time-

consuming tasks to clients before their arrivals to

improve efficiency; (b) enables fast and computerized

onsite registration through the online / mobile

registration tool; (c) improves data / process flow and

efficiency and reduces errors by automating the

vaccine data collection using scanning devices; (d)

connects clients to vaccine reminders, and the vaccine

effect registry where clients can report their response

and also receive vaccine update information; (e) is

easy to use and maintain; and (f) is flexibility and

scalability. In our design, we propose a synthesis of

the various techniques that would prove more efficient

than the least efficient methods while requiring a

minimally-enhanced technology and methods. This is

critical for the resource strapped public health

organizations. The system can easily accommodate

new devices to reflect the rapid advances in electronic

devices, systems, and software.

At the time of this writing, our recommendation

has been adopted by CRA. CRA has incorporated the

scanner for registration and screening, affording 45%

throughput improvement over the manual input.

This work highlights the potential of electronic

technologies to significantly enhance vaccine

administration and data management. The design is

flexible and adaptable for different types of medical

countermeasures and dispensing purposes. For

example, it can be used for prophylactic medical

countermeasure dispensing against biological attack,

or decontamination and health registration for

radiological incidents. The online registration can be

adapted based on the nature and purpose of the event.

ACKNOWLEDGEMENTS

This material is based upon work supported by the

Centers for Disease Control and Prevention. The

views and conclusions contained in this document are

those of the authors and should not be interpreted as

necessarily representing the official policies, either

expressed or implied, of the Centers for Disease

Control and Prevention.

The authors thank the students Lavannya Atri,

Aissatou B Bah, Yu Cao, Joseph Malecki,

Danielle

Mathis,

Kyumin Lee, Zhuonan Li, Parthiv Naresh,

Hunter Stroud, Pavan Thaker, and Xin Wei from

Georgia Institute of Technology for assisting in time-

motion study and data collection. We give special

thanks to Amy Parker Fiebelkorn and Dr. Samuel

Graitcer from CDC National Center for

Immunizations and Respiratory Diseases for their

invaluable guidance, insights, comments, and edits on

the paper. We thank the anonymous reviewers for

their insightful comments.

REFERENCES

Apartsin, K. A., Sverdlova, N. A., Kireeva, V. V., Riad, A.,

& Klugar, M. (2021). International COVID-19 Vaccines

Safety Tracking Study (CoVaST-RU): Participation of

the Russian Federation. Profilakticheskaya Meditsina,

24(12).

Campbell mass dispensing flu clinic, Gillette, Wyoming.

CDC report: Centers for Disease Control and Prevention.

CDC closed POD for flu vaccination, Atlanta, Georgia. CDC

report: Centers for Disease Control and Prevention.

Chekol Abebe, E., Mengie Ayele, T., Tilahun Muche, Z.,

Behaile T/Mariam, A., Dagnaw Baye, N., Mekonnen

Agidew, M., & Asmamaw Dejenie, T. (2022).

Evaluation and comparison of post-vaccination adverse

effects among Janssen and Oxford-AstraZeneca

vaccinated adult individuals in Debre Tabor Town: A

cross- sectional survey in Northwest Ethiopia. Human

Vaccines and Immunotherapeutics, 18(6).

Cot, C., Cacciapaglia, G., Islind, A. S., Óskarsdóttir, M., &

Sannino, F. (2021). Impact of US vaccination strategy on

COVID-19 wave dynamics. Scientific Reports, 11(1).

Danzon, P. M., Pereira, N. S., & Tejwani, S. S. (2005).

Vaccine Supply: A Cross-National Perspective. Health

Affairs, 24(3).

de Mesquita PJB, Roush R, Sarkar J. County-Level Planning

for Efficient Distribution of Emergency Medical

Countermeasures with RealOpt Software. Health Secur.

Assessing Registration and Screening Technologies for Efﬁcient Mass Vaccination and Public Health Monitoring

169

2021 Sep-Oct;19(5):532-540. doi: 10.1089/hs.2021.00

26. Epub 2021 Oct 4. PMID: 34609918.

Denver public health closed POD for flu vaccination,

Denver, Colorado. CDC report: Centers for Disease

Control and Prevention.

Fast Facts on Global Immunization, Centers for Disease

Control and Prevention. https://www.cdc.gov/global-

immunization/fast-facts/index.html

Ferreira, L. S., de Almeida, G. B., Borges, M. E., Simon, L.

M., Poloni, S., Bagattini, Â. M., da Rosa, M. Q. M., Diniz

Filho, J. A. F., Kuchenbecker, R. de S., Camey, S. A.,

Kraenkel, R. A., Coutinho, R. M., & Toscano, C. M.

(2022). Modelling optimal vaccination strategies against

COVID-19 in a context of Gamma variant predominance

in Brazil. Vaccine, 40(46).

Hupert, N., Mushlin, A. I., & Callahan, M. A. (2002).

Modeling the public health response to bioterrorism:

Using discrete event simulation to design antibiotic

distribution centers. Medical Decision Making, 22(5

SUPPL.).

Kwong, J. C., Foisy, J., Quan, S., Heidebrecht, C., Kolbe, F.,

Bettinger, J. A., Buckeridge, D. L., Chambers, L. W.,

Crowcroft, N. S., Dhalla, I. A., Sikora, C. A., Willison,

D. J., Pereira, J. A., Boulianne, N., Brien, S., Deeks, S.,

Elliott, R., Finkelstein, M., Guay, M., … Quach, S.

(2010). Why collect individual-level vaccination data?

CMAJ. Canadian Medical Association Journal, 182(3).

Lee, E. K., Atallah, H. Y., Wright, M. D., Post, E. T.,

Thomas, C., Wu, D. T., & Haley, L. L. (2015).

Transforming hospital emergency department workflow

and patient care. Interfaces, 45(1). https://doi.org/10.12

87/inte.2014.0788.

Lee, E. K., Chen, C. H., Pietz, F., & Benecke, B. (2010).

Disease propagation analysis and mitigation strategies

for effective mass dispensing. AMIA ... Annual

Symposium Proceedings / AMIA Symposium. AMIA

Symposium, 2010.

Lee, E. K., Li, Z. L., Liu, Y. K., & Leduc, J. (2021). Strategies

for vaccine prioritization and mass dispensing. Vaccines,

9(5).

Lee, E. K., Liu, Y., Yuan, F., & Pietz, F. H. (2021). Strategies

for Disease Containment: A Biological-Behavioral-

Intervention Computational Informatics Framework.

AMIA Annu Symp Proc. 2021.

Lee, E. K., Maheshwary, S., Mason, J., & Glisson, W.

(2006a). Decision support system for mass dispensing of

medications for infectious disease outbreaks and

bioterrorist attacks. Annals of Operations Research,

148(1).

Lee, E. K., Maheshwary, S., Mason, J., & Glisson, W.

(2006b). Large-scale dispensing for emergency response

to bioterrorism and infectious-disease outbreak.

Interfaces, 36(6).

Lee, E. K., Pietz, F. H., Chen, C. H., & Liu, Y. (2017). An

interactive web-based decision support system for mass

dispensing, emergency preparedness, and

biosurveillance. ACM International Conference

Proceeding Series, Part F128634.

Lee, E. K., Pietz, F., Benecke, B., Mason, J., & Burel, G.

(2013). Advancing public health and medical

preparedness with operations research. Interfaces, 43(1).

Lee, E. K., Yuan, F., Pietz, F. H., Benecke, B. A., & Burel,

G. (2015). Vaccine prioritization for effective pandemic

response. Interfaces, 45(5).

Miller, M. A., Viboud, C., Olson, D. R., Grais, R. F., Rabaa,

M. A., & Simonsen, L. (2008). Prioritization of influenza

pandemic vaccination to minimize years of life lost.

Journal of Infectious Diseases, 198(3).

Nanji, K. C., Cina, J., Patel, N., Churchill, W., Gandhi, T. K.,

& Poon, E. G. (2009). Overcoming Barriers to the

Implementation of a Pharmacy Bar Code Scanning

System for Medication Dispensing: A Case Study.

Journal of the American Medical Informatics

Association, 16(5).

Oak Ridge Associated University. Catalog of electronic

technologies used for data collection at vaccination

clinics. Washington, DC: National Association of

County & City Health Officials; 2016.

Pereira, J. A., Quach, S., Hamid, J. S., Heidebrecht, C. L.,

Quan, S. D., Nassif, J., Diniz, A. J., Van Exan, R.,

Malawski, J., Gentry, A., Finkelstein, M., Guay, M.,

Buckeridge, D. L., Bettinger, J. A., Kalailieff, D., &

Kwong, J. C. (2012). Exploring the feasibility of

integrating barcode scanning technology into vaccine

inventory recording in seasonal influenza vaccination

clinics. Vaccine, 30(4).

Philadelphia department of public health Closed POD for flu

vaccination, Philadelphia, Pennsylvania. CDC report:

Centers for Disease Control and Prevention.

Platte mass dispensing flu clinic, Platte City, Missouri. CDC

report: Centers for Disease Control and Prevention.

Prieto Curiel, R., & González Ramírez, H. (2021).

Vaccination strategies against COVID-19 and the

diffusion of anti-vaccination views. Scientific Reports,

11(1).

Riad, A., Schünemann, H., Attia, S., Peričić, T. P., Žuljević,

M. F., Jürisson, M., Kalda, R., Lang, K., … Klugar, M.

(2021). Covid-19 vaccines safety tracking (Covast):

Protocol of a multi-center prospective cohort study for

active surveillance of covid-19 vaccines’ side effects.

International Journal of Environmental Research and

Public Health, 18(15).

Tisoncik, J. R., Belisle, S. E., Diamond, D. L., Korth, M. J.,

& Katze, M. G. (2009). Is systems biology the key to

preventing the next pandemic? In Future Virology (Vol.

4, Issue 6).

Wagner, C. E., Saad-Roy, C. M., & Grenfell, B. T. (2022).

Modelling vaccination strategies for COVID-19. In

Nature Reviews Immunology (Vol. 22, Issue 3).

Washington, M. L. (2009). Evaluating the capability and cost

of a mass influenza and pneumococcal vaccination clinic

via computer simulation. Medical Decision Making,

29(4).

Wein, L. M., Craft, D. L., & Kaplan, E. H. (2003).

Emergency response to an anthrax attack. Proceedings of

the National Academy of Sciences of the United States

of America, 100(7).

Yadav, T., Kumar, S., Mishra, G., & Saxena, S. K. (2023).

Tracking the COVID-19 vaccines: The global landscape.

In Human Vaccines and Immunotherapeutics (Vol. 19,

Issue 1).

DATA 2025 - 14th International Conference on Data Science, Technology and Applications

170