A Proposed Framework for Integrating Digital Triage with 3D Human

Model for Intuitive Health Visualization and Monitoring

Md Jobayer Hossain Chowdhury

, Mohamed Mehfoud Bouh

, Abdullah Al Noman

Nadia Binte Rahman Peeya

, Shah Manan Vinod

, Syed Usama Hussain Shah Bukhari

Prajat Paul

, Forhad Hossain

and Ashir Ahmed

Faculty of Information Science and Electrical Engineering, Kyushu University, Fukuoka, Japan

Keywords:

Digital Triage, Human Digital Twin, Healthcare Informatics, Electronic Health Records, Health Visualization,

Patient Monitoring, Artiﬁcial Intelligence.

Abstract:

This paper presents a novel integration of digital triage protocols with three-dimensional human digital twin

models to enhance patient assessment and clinical decision-making in healthcare. We investigate how Elec-

tronic Health Record (EHR) data can be transformed into intuitive, anatomically-relevant visualizations that

map health parameters to speciﬁc body regions using color-coded indicators. Building upon the B-logic

framework from Portable Health Clinic systems, our approach creates personalized 3D patient models that

dynamically represent health status through targeted visual cues—from BMI and vital signs to biomarkers and

lifestyle factors. The system architecture incorporates anthropometric data and facial recognition to generate

individualized avatars, while large language models provide contextual healthcare suggestions based on de-

tected risk factors. This integration addresses limitations in current EHR-based triage systems, particularly

regarding alert effectiveness and protocol compliance. While the system shows potential for enhanced visual-

ization, practical implementation may face challenges in data availability, privacy, and clinical validation. The

proposed visualization methodology offers healthcare providers and patients an intuitive interface for health

monitoring, potentially improving engagement, comprehension, and clinical workﬂow in both emergency and

routine healthcare settings.

1 INTRODUCTION

Over the past decade, the adoption of Electronic

Health Record (EHR) systems has transformed mod-

ern healthcare delivery. As of 2021, 96% of non-

federal acute care hospitals in the United States had

implemented certiﬁed EHR systems, compared to

only 7.6% in 2008 (Ofﬁce of the National Coordina-

tor for Health Information Technology (ONC), 2021;

https://orcid.org/0009-0008-5311-9191

https://orcid.org/0000-0002-7716-7007

https://orcid.org/0009-0002-5360-8667

https://orcid.org/0009-0008-8657-1651

https://orcid.org/0009-0004-2398-9795

https://orcid.org/0009-0003-2755-9568

https://orcid.org/0009-0002-2243-6078

https://orcid.org/0000-0002-3593-0860

https://orcid.org/0000-0002-8125-471X

Jiang et al., 2023). A global survey by the Organi-

sation for Economic Co-operation and Development

(OECD) across 27 countries revealed that only 15

had achieved nationally uniﬁed EHR systems, under-

scoring persistent challenges related to interoperabil-

ity and fragmentation (Slawomirski et al., 2023).

While developed nations are progressively inte-

grating EHR and Electronic Medical Record (EMR)

systems, adoption rates in developing countries re-

main low, ranging from 5% to 30%, primarily due

to limited infrastructure, ﬁnancial constraints, and the

lack of standardized health data frameworks (Dere-

cho et al., 2024). One key advancement enabled by

EHR systems is digital triage—automated protocols

that prioritize patient care based on clinical urgency.

However, traditional EHR-based alert systems often

suffer from poor protocol compliance, alert fatigue,

and suboptimal visualization interfaces. For instance,

a randomized controlled study showed no signiﬁcant

Chowdhury, M. J. H., Bouh, M. M., Al Noman, A., Peeya, N. B. R., Vinod, S. M., Bukhari, S. U. H. S., Paul, P., Hossain, F. and Ahmed, A.

A Proposed Framework for Integrating Digital Triage with 3D Human Model for Intuitive Health Visualization and Monitoring.

DOI: 10.5220/0013567200003970

In Proceedings of the 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2025), pages 329-336

ISBN: 978-989-758-759-7; ISSN: 2184-2841

329

Table 1: Triage System Categories in Healthcare.

Triage Category Implementation Con-

text

Representative Systems Key Characteristics

Emergency Depart-

ment Triage

Emergency Depart-

ments, Pre-Hospital

ESI(Emergency Severity

Index), CTAS(Canadian

Triage and Acuity Scale),

MTS(Manchester Triage Sys-

tem), ATS(Australasian Triage

Scale)

Traditional, rule-based methods that

help prioritize who gets care ﬁrst.

AI Driven EHR Integration, Tele-

health, AI-Supported

Platforms

HopScore, SERT(System

for Emergency Risk Triage),

TriageGO

Uses AI and machine learning to predict

risk and suggest actions.

Disaster & Mass

Casualty

Emergency Situations,

Pandemic Response

START(Simple Triage and

Rapid Treatment), Jump-

START, Triage Sieve, SAVE

Helps in chaotic events by categorizing

patients.

Specialized Pediatric, Mental

Health

JumpSTART, MHTS ( Mental

Health Triage Scale)

Tailored for speciﬁc populations like

children or people with mental health

needs.

improvement in triage compliance through passive

EHR alerts, highlighting the need for more intuitive

and dynamic triage strategies (Holmes et al., 2015).

This paper proposes a framework integrating dig-

ital triage protocols with three-dimensional human

digital twin (HDT) models. By leveraging anthro-

pometric data, facial recognition, and color-coded vi-

sualizations, the system aims to transform structured

EHR data into an anatomically meaningful 3D repre-

sentation. This approach enables healthcare providers

and patients to monitor and understand health risks

more intuitively.

Research Question: How to integrate and visual-

ize medical triage with a 3D human model?

To address this question, the paper introduces

a framework that links B-Logic-based triage from

Portable Health Clinic (PHC) systems with AI-

generated digital avatars. The system also incorpo-

rates large language models (LLMs) to provide per-

sonalized health suggestions based on mapped risk

factors. The proposed methodology targets both

emergency and routine care, particularly in low-

resource settings, offering a potential pathway toward

more accessible and personalized healthcare monitor-

ing.

2 AVAILABLE TRIAGE SYSTEM

AND ITS LIMITATION

In an EHR-integrated triage workﬂow, clinicians

(often triage nurses) enter a patient’s initial in-

formation—vital signs, symptoms, and chief com-

plaint—directly into a module of the EHR. This dig-

ital form captures key data points and often enforces

required ﬁelds to ensure completeness (Aronsky et al.,

2008).

The integration of triage systems within Elec-

tronic Health Record (EHR) platforms represents a

signiﬁcant advancement in healthcare informatics and

clinical decision support. Multiple triage frameworks

now operate within these digital environments and

can be categorized into four primary classiﬁcations

based on implementation context and clinical focus.

Emergency Department Triage systems, such

as the widely adopted Emergency Severity Index

(ESI) (Aronsky et al., 2008; Liu et al., 2022) and

Canadian Triage and Acuity Scale (CTAS) (Ofﬁce

of the National Coordinator for Health Informa-

tion Technology (ONC), 2021), Australasian Triage

Scale(ATS) (Ebrahimi et al., 2015) , Manchester

Triage System (MTS) (Azeredo et al., 2015) offer

structured protocols used in emergency and urgent

care settings to prioritize treatment based on patient

acuity and resource requirements.

Mass Casualty Incident (MCI) Triage protocols

are speciﬁcally designed for disaster scenarios and

large-scale emergencies. Systems such as Simple

Triage and Rapid Treatment (START), JumpSTART

(for pediatric patients), and the Triage Sieve catego-

rize patients using a color-coded classiﬁcation (im-

mediate/red, delayed/yellow, minor/green, and expec-

tant/black) to optimize resource allocation (Bazyar

et al., 2019; Wang et al., 2022).

Technology-Assisted Triage includes informatics-

driven systems like HopScore (Levin et al., 2018),

the SERT (System for Emergency Risk Triage), and

TriageGO (Johns Hopkins Medicine, 2022), which

utilize artiﬁcial intelligence and EHR integration to

predict clinical risk and suggest appropriate care path-

ways in real time.

Specialty Population Triage frameworks are de-

signed for speciﬁc groups such as pediatric or men-

SIMULTECH 2025 - 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

330

Table 2: B-logic triage system.

Parameter Lower Warning Green Yellow Orange Red Upper Warning

Height (cm) <100.0 >200.0

Weight (kg) <25 >100.0

BMI <25 ≥25 & <30 ≥30 & ≤35 >35

Waist (cm) Male <40.0 <90.0 ≥90.0 NA NA >120.0

Waist (cm) Female <40.0 <80.0 ≥80.0 NA NA >110.0

Hip (cm) <40.0 >120.0

Waist Hip Ratio Male <0.90 ≥0.90 NA NA -

Waist Hip Ratio Female <0.85 ≥0.85 NA NA -

Temperature (C) <33.0 ≥37.0 & <37.5 ≥37.5 NA NA >39.0

HBsAg negative positive

Smoking positive

Urine Sugar - +- Others

Urine Protin - +- Others

Urinary Urobilinogen +- Others

Oxygenation of Blood (%) >100 ≥96 ≥93 & ≤96 ≥90 & <93 <90 <92

Blood Pressure Systolic <70 <130 ≥130 & <140 ≥140 & <180 ≥180 >220

Blood Pressure Diastolic <50 ≤85 ≥85 & <90 ≥90 & <110 ≥110 >140

Blood Sugar (mmol/dl) PBS <3.0 <7.78 ≥7.78 & <11.11 ≥11.11 & <16.67 ≥16.67 >30.0

Blood Sugar (mmol/dl) FBS <3.0 <5.56 ≥5.56 & <7.0 ≥7.0 & <11.11 ≥11.11 >20.0

Blood Hemoglobin (g/dl) >18.0 ≥12.0 ≥10.0 & <12.0 ≥8.0 & <10.0 <8.0 <6.0

Pulse Rate (beats/min) <50 ≥60 & <100 ≥50 & <60 <50 OR ≥120 >130

Arrhythmia Normal Others

Blood Cholesterol (mg/dl) <120.0 ≤200.0 >200.0 & ≤225.0 ≥225.0 & <240.0 ≥240.0 >300.0

Blood Uric Acid (mg/dl) Male <2.5 ≥3.5 & ≤7.0 >7.0 & ≤8.0 ≥8.0 >12.0

Blood Uric Acid (mg/dl) Female <2.5 ≥2.4 & ≤6.0 >6.0 & ≤7.0 ≥7.0 >12.0

tal health patients. These include systems like Jump-

START (Bazyar et al., 2019; Wang et al., 2022)

and the Mental Health Triage Scale (MHTS) (Broad-

bent et al., 2007), which tailor triage protocols to the

unique needs of their respective populations.

The triage systems perform effectively within

their domains, offering structured support for clini-

cal decision-making. However, they face limitations

in patient engagement due to health literacy gaps.

Many patients struggle to interpret medical data and

triage outcomes presented in standard EHR inter-

faces. Moreover, the lack of intuitive, body-mapped

visualizations reduces clarity and makes it harder for

users to understand their health status. These issues

highlight the need for more accessible and patient-

friendly triage solutions.

3 PORTABLE HEALTH CLINIC

AND B-LOGIC

The Portable Health Clinic (PHC), developed by

Kyushu University and Grameen Communications,

delivers telehealth services to underserved rural ar-

eas using a portable briefcase with diagnostic tools.

Health data is sent to a remote call center, where

doctors review EHRs and provide consultations via

telemedicine. A color-coded triage system (green

to red) guides patient prioritization. Integrated

with the PHC, the B-Logic framework uses prede-

ﬁned medical parameters to classify patients by risk

level, enabling efﬁcient diagnosis and resource allo-

cation (Ahmed et al., 2013). table 2 shows the B-logic

triage system.

4 CONCEPT OF 3D

VISUALIZATION OF PHC

TRIAGE SYSTEM

The PHC system collects patient data and stores it in

a database, automatically assigning each patient to a

color-coded triage category. The new proposed sys-

tem will generate a ’patient digital twin’ and use these

data to provide customized suggestions and visual-

izations, allowing individuals to view and understand

their health status through a personalized digital rep-

resentation. In ﬁg. 1, it shows the system architecture

of the proposed system.

4.1 Digital Twin with Real-Time Triage

To make the digital twin with a real-time triage sys-

tem, we need to break it down into two parts.

A Proposed Framework for Integrating Digital Triage with 3D Human Model for Intuitive Health Visualization and Monitoring

331

Figure 1: Proposed framework.

4.1.1 Silhouette Estimation from

Anthropometric Data

Almost every EHR system collects the anthropomet-

ric data of a patient, such as height, weight, etc. Also,

PHC has the patient’s anthropometric data. We will

use the data to generate a speciﬁc silhouette of the pa-

tient. In ﬁg. 2, it shows how the anthropometric data

will be the input of an AI model that will make the

silhouette of the patient.

Figure 2: Anthropometric data to silhouette structure.

4.1.2 3D Face Construction from Image Data

EHR systems have patient images in their database.

As ﬁg. 3 suggests, the AI model will create a 3D face

based on the image, and the face will merge with the

human silhouette to make a clone of the individual

patient.

Figure 3: User image to 3D face structure.

4.2 Mapping Health Parameters to

Body Parts

The system will map each key health parameter to

a speciﬁc body region on the 3D model, indicating

where the effect of that metric is most visible or rele-

vant. Here’s a breakdown of the parameters and how

to visualize them on the body

4.2.1 Height, Weight, BMI (Body Mass Index)

These relate to the overall body. A common approach

is to reﬂect BMI by the overall silhouette. For in-

stance, the entire ﬁgure could be outlined or ﬁlled

with a color representing whether the BMI is normal

or high. A green full-body glow for normal BMI, vs.

orange or red if BMI is in overweight/obese range,

immediately signals the category.

4.2.2 Waist, Hip Measurements, and Waist-Hip

Ratio

To emphasize the abdominal and hip region, a col-

ored band can be drawn around the waist or hips of

the model. For instance, a ring or outline at the waist

level may be displayed in green to indicate a healthy

circumference, while red can denote measurements

beyond the risk threshold. Since the waist-to-hip ra-

tio serves as a single risk indicator, the entire mid-

section, including the stomach and hip area, can be

color-coded to reﬂect risk levels. A high ratio, indica-

tive of central obesity, may be represented by a red-

colored region. This visual mapping enables users to

perceive an expanding red belly when waist size be-

comes a concern. Additionally, a subtle translucent

“slice” or disc around the waist can be incorporated to

display the numeric value of the ratio. A straightfor-

ward approach involves highlighting the torso, partic-

ularly the abdominal area, with severity-based color

coding for waist and hip metrics.

SIMULTECH 2025 - 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

332

Table 3: Health Parameters Mapping.

Health Parameter Mapped Body Re-

gion

Visualization Approach

Height, Weight, BMI Overall body Color-coded whole body (green/yellow/red for BMI), nu-

meric values for height/weight

Waist, Hip Measure-

ments

Abdomen and hip re-

gion

Colored band around waist/hips, midsection highlighted

based on risk

Body Temperature Forehead, head,

whole body

Heatmap overlay (blue to red gradient), forehead glow

Blood Pressure Arms, chest highlighted arms for BP, heart icon on chest for pulse

Blood Oxygenation

(SpO

)

Fingertips,

lungs/chest

Glowing ﬁngertip (green/yellow/red), lung overlay in

color

Blood Sugar (Glu-

cose)

Fingertip, hands,

veins

Hand highlight, blood droplet icon, color-coded veins

Hemoglobin (Hb) Circulatory system Blood vessels colored red (normal) or blue (anemic), pale

skin tone

Cholesterol Heart/chest Heart turns color for cholesterol risk, artery clog icons

Uric Acid Joints (feet, knees) Foot/knee highlights for uric acid buildup

Smoking Lungs, mouth Lungs overlaid with smoky texture or colored (yel-

low/red)

Urine Parameters Kidneys, bladder Kidneys and bladder highlighted, color-coded for risk

4.2.3 Body Temperature

Body temperature is commonly measured at the fore-

head or ear, but fever affects the entire body. An ef-

fective method for visual representation is a heatmap

overlay that spans the entire body. This can be im-

plemented using a gradient shader, where cooler tem-

peratures are represented in blue and elevated tem-

peratures in red. Given that the normal human body

temperature is approximately 37°C, a simpliﬁed ap-

proach could involve using green to denote normal

temperature ranges and red to indicate fever. To en-

hance clarity, speciﬁc regions, such as the forehead,

can be emphasized using a thermometer icon or a red

glow, aligning with the common practice of forehead-

based temperature checks. Alternatively, the model’s

facial or forehead region can dynamically change

color—retaining a normal skin tone when within the

healthy range and turning ﬂushed red when fever is

detected. If a full-body heatmap is employed, care

should be taken to ensure it does not interfere with

other visual overlays. In such cases, a subtle overall

tint adjustment—such as a slight red hue when fever

is present—can effectively signal an elevated temper-

ature. Given that body temperature is represented as a

single numerical value, a minimalistic approach, such

as a small colored indicator (e.g., a red dot on the fore-

head), may also sufﬁce while maintaining an intuitive

and informative visualization.

4.2.4 Blood Pressure and Pulse Rate

Blood pressure and pulse rate are critical indicators

of circulatory health, typically measured on the arm

and closely associated with cardiovascular function.

To effectively visualize these vitals, a model can

highlight speciﬁc anatomical regions where measure-

ments commonly occur. For instance, the upper arm

or wrist—locations used for blood pressure monitor-

ing—can be color-coded to indicate status, with green

representing normal levels and red signaling hyper-

tension. Additionally, an icon of a heart or artery

can be placed on the arm to reinforce the associa-

tion with circulatory health. For pulse rate visual-

ization, an intuitive approach is to use a heart sym-

bol on the chest that dynamically animates to mimic a

beating heart. This icon could change color to reﬂect

pulse rate abnormalities—red for tachycardia (ele-

vated heart rate) and blue for bradycardia (low heart

rate). Some avatar-based monitoring systems already

implement similar features, where heart icons adjust

color based on real-time vital signs. Since blood

pressure and pulse rate are interrelated, a dual rep-

resentation could enhance clarity: the heart symbol

on the chest can reﬂect pulse rate through animation

and color changes, while the arm region can indicate

blood pressure status. In cases of severe hypertension,

an extended visualization—such as highlighting the

blood vessel network in red—can effectively convey

cardiovascular strain. Conversely, hypotension or a

dangerously low pulse could be depicted using a blue

A Proposed Framework for Integrating Digital Triage with 3D Human Model for Intuitive Health Visualization and Monitoring

333

tint or a slow pulsating animation. For user interac-

tion, a clickable interface where selecting the arm or

heart symbol provides precise numerical readings of

blood pressure and pulse rate would enhance usabil-

ity. However, even without interaction, a color-coded

system ensures immediate recognition of circulatory

health status at a glance

4.2.5 Blood Oxygenation (SpO

)

Blood oxygen saturation (SpO

) is a vital parame-

ter typically measured at the ﬁngertip using a pulse

oximeter or inferred from lung function. In a digital

twin, it can be visualized by highlighting the ﬁnger-

tip or lungs. A color-coded glow—green (normal),

yellow (moderate), red (low)—can indicate oxygen

levels, with pulsing effects enhancing visibility. For

anatomical context, the lungs may be tinted red to

signal respiratory distress. Combining ﬁngertip and

lung highlights offers intuitive feedback. Interactive

elements, like clicking for exact values, can further

improve user engagement and health monitoring clar-

ity.

4.2.6 Blood Sugar (Glucose)

Blood glucose levels are typically monitored through

ﬁnger-prick tests or continuous glucose monitors

(CGMs) placed on the arm. While glucose

metabolism affects the entire body, an effective vi-

sual representation should focus on intuitive indica-

tors, such as the hands (where blood tests occur) or

a blood droplet symbol to signify sugar levels. A

simple and clear method is to highlight the ﬁnger-

tip, where traditional glucose tests are performed. A

small droplet icon can be placed on the ﬁngertip,

changing color to reﬂect blood sugar status: Green for

normal glucose levels Yellow for moderate elevation

Red for high blood sugar (hyperglycemia) For a more

anatomical approach, a vein or artery overlay could

be used to signify blood sugar levels, though this is

a more abstract representation. If the model includes

visible veins, they could subtly change color based on

glucose concentration. However, to maintain clarity

and usability, a color-coded highlight on the hands is

a more direct and intuitive approach. User interac-

tion can be enhanced by allowing the glowing hand to

be clickable, displaying real-time blood sugar read-

ings. In some medical visualization systems, high

blood sugar is represented across multiple organs due

to its long-term effects on areas such as the kidneys,

eyes, and nerves, but this level of detail may be un-

necessary for general use. If an anatomical focus is

preferred, the pancreas (responsible for insulin pro-

duction) could be highlighted, though most lay users

may not immediately recognize its location. Alter-

natively, a small glucose meter icon placed near the

model could provide additional clarity. However, fol-

lowing the established visual scheme, a color-coded

hand region remains the most intuitive and immedi-

ately recognizable indicator of blood sugar status.

4.2.7 Hemoglobin

Hemoglobin (Hb) plays a crucial role in the blood’s

oxygen-carrying capacity, directly inﬂuencing cir-

culation and tissue oxygenation. A decrease in

hemoglobin levels, indicative of anemia, may be rep-

resented through visual changes in the circulatory sys-

tem. For instance, if a model includes arteries and

veins, normal hemoglobin levels could be depicted

with bright red vessels, while anemia might be illus-

trated using a dull blue or gray hue. In the absence

of detailed vascular representation, an alternative ap-

proach could involve using a blood drop icon over

the torso or adjusting the overall skin tone—rosy for

normal hemoglobin and pale or bluish for low lev-

els. Given hemoglobin’s impact on energy and oxy-

genation, visual cues such as highlighting the chest

(symbolizing the heart and circulation) or the arm

veins (where blood is commonly drawn for testing)

may enhance interpretability. Maintaining consis-

tency with oxygen-related indicators, such as linking

hemoglobin visualization to the chest or arterial path-

ways, can further reinforce its physiological signiﬁ-

cance.

4.2.8 Cholesterol

Cholesterol and uric acid are distinct physiologi-

cal markers, each associated with speciﬁc body sys-

tems. Cholesterol is primarily linked to cardiovascu-

lar health, while uric acid is connected to joint func-

tion, particularly in conditions like gout. To visu-

alize cholesterol levels, the heart or arterial system

can serve as a focal point, with color-coded indi-

cators—such as a red or orange hue—to signify el-

evated cholesterol and potential cardiovascular risk.

Additional elements, such as an artery-clogging icon,

could further reinforce this association. For uric acid,

visualization can be centered on the joints, with a fo-

cus on areas most commonly affected by gout, such

as the big toe, knees, or hands. A practical approach

would be to highlight the foot or toe joint when uric

acid levels are high, as gout frequently manifests in

these areas ﬁrst. Alternatively, a generic joint icon,

such as a knee, could be used to represent broader

joint-related risks. If highlighting multiple joints

becomes complex, selecting a single representative

joint—such as the knee or foot—provides clarity

SIMULTECH 2025 - 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

334

while maintaining effectiveness. A user-interactive

model could allow access to speciﬁc details by click-

ing on the heart for cholesterol-related data and the

foot or knee for uric acid levels, ensuring intuitive en-

gagement with the health metrics.

4.2.9 Smoking

Smoking signiﬁcantly affects lung health and in-

creases disease risk. In the 3D model, this can be

visualized by highlighting the lungs—healthy lungs

appear normal, while smoker’s lungs are tinted gray

or black. A color gradient (green to red) can indicate

smoking intensity. Since it’s a lifestyle factor, a sim-

ple visual toggle or overlay can signal smoking status,

enhancing user awareness of its health impact.

4.2.10 Urine-Related Parameters

Urine-related parameters, such as urine sugar, ke-

tones, and kidney function markers, are closely tied

to the renal system, including the kidneys and blad-

der. A clear visualization of these metrics can be

achieved by mapping them to the anatomical loca-

tions of these organs. The kidneys, positioned in the

lower back, could be highlighted from either the rear

view or subtly shown from the front as faint outlines

on the sides. The bladder, located in the lower ab-

domen, can also serve as a visual indicator for urine-

related issues. Color-coded cues can effectively com-

municate abnormal ﬁndings. For instance, healthy

kidneys and bladder could be depicted in green, while

abnormal readings—such as proteinuria or elevated

creatinine—could prompt a shift to orange or red.

A more simpliﬁed approach might use a single uri-

nary tract icon, such as a kidney symbol on the ab-

domen, to consolidate renal health indicators. How-

ever, given the distinct roles of the kidneys and blad-

der, representing both individually enhances clarity.

If the model allows rotation, users could view the kid-

neys from the back, reinforcing anatomical accuracy.

By linking color changes to speciﬁc urine test abnor-

malities, this approach provides an intuitive and direct

way to visualize renal function concerns.

4.3 Healthcare Suggestion

The system will incorporate a triage-based approach

using a large language model (LLM) to provide per-

sonalized healthcare suggestions and motivate pa-

tients toward better health management. By analyzing

key health parameters, the model will assess risk lev-

els and generate tailored recommendations. For ex-

ample, if high cholesterol or elevated blood pressure

is detected, the system may suggest lifestyle changes

such as dietary modiﬁcations, increased physical ac-

tivity, or medical consultation. Beyond medical ad-

vice, the system will also focus on patient motivation.

Instead of merely presenting risk factors, it will use

positive reinforcement and actionable steps to encour-

age behavior change. If a patient shows early signs

of dehydration or kidney strain, the system might

prompt hydration reminders and explain the beneﬁts

of maintaining optimal ﬂuid balance. Similarly, for

smokers, it could offer quitting strategies, highlight

immediate health beneﬁts, and suggest resources for

smoking cessation. The integration of an LLM al-

lows for a dynamic and engaging interaction, where

responses are not only medically relevant but also em-

pathetic and motivating. By adapting to patient needs

and health trends, the system can enhance patient en-

gagement, encourage proactive healthcare decisions,

and ultimately contribute to improved long-term well-

being.

5 DISCUSSION

This study presents a concept of integration digi-

tal triage system with human digital twin models to

improve healthcare visualization and patient moni-

toring. By mapping clinical parameters to speciﬁc

body regions on a 3D model, the system offers an

intuitive interface that enhances both patient under-

standing and clinician decision-making. It holds par-

ticular promise in underserved areas, supporting re-

mote assessments and low-resource healthcare deliv-

ery through the Portable Health Clinic model.

Despite its potential, the system has notable lim-

itations. Accurate avatar generation depends on reli-

able anthropometric and facial data, which may not

always be available, especially in rural or under-

resourced settings. The use of facial recognition also

raises ethical and privacy concerns, highlighting the

need for strict data protection and informed consent.

Additionally, reliance on AI for triage decisions

introduces risks, including algorithmic bias and mis-

interpretation of diverse clinical presentations. These

risks are especially relevant in global health contexts

where population data may be underrepresented in

training datasets.

Future work should include clinical validation,

user studies, and the integration of AI methods to en-

sure fairness and trust. Addressing these challenges is

critical for safe, effective, and ethical implementation

across diverse healthcare settings.

A Proposed Framework for Integrating Digital Triage with 3D Human Model for Intuitive Health Visualization and Monitoring

335

6 CONCLUSION

This paper has presented a conceptual framework for

integrating digital triage protocols with 3D human

digital twin models to enhance healthcare visualiza-

tion, patient monitoring, and decision-making. The

proposed system leverages anthropometric data and

facial recognition to create personalized 3D models

that visually represent health parameters in anatom-

ically relevant locations. By implementing a color-

coded visualization scheme based on the B-logic

triage framework, the system enables intuitive inter-

pretation of complex health data. The incorpora-

tion of LLM-based healthcare suggestions further en-

hances the system’s utility by providing personalized

recommendations and motivational prompts based on

detected risk factors. This combination of visual rep-

resentation and actionable guidance represents a sig-

niﬁcant step toward more patient-centered healthcare

monitoring. The technology has particular promise

for remote healthcare delivery in underserved com-

munities, building upon the Portable Health Clinic

model. While technical challenges remain in imple-

mentation and integration with existing EHR systems,

the approach offers a promising path to improve pa-

tient engagement, enhance clinical decision-making,

and ultimately advance healthcare delivery through

more intuitive and accessible health information vi-

sualization.

REFERENCES

Ahmed, A., Inoue, S., Kai, E., Nakashima, N., and Nohara,

Y. (2013). Portable health clinic: A pervasive way to

serve the unreached community for preventive health-

care. In Distributed, Ambient, and Pervasive Inter-

actions: First International Conference, DAPI 2013,

Held as Part of HCI International 2013, Las Vegas,

NV, USA, July 21-26, 2013. Proceedings 1, pages 265–

274. Springer.

Aronsky, D., Jones, I., Raines, B., Hemphill, R., Mayberry,

S. R., Luther, M. A., and Slusser, T. (2008). An inte-

grated computerized triage system in the emergency

department. In AMIA Annual Symposium Proceed-

ings, volume 2008, page 16.

Azeredo, T. R. M., Guedes, H. M., de Almeida, R. A. R.,

Chianca, T. C. M., and Martins, J. C. A. (2015). Ef-

ﬁcacy of the manchester triage system: a systematic

review. International emergency nursing, 23(2):47–

52.

Bazyar, J., Farrokhi, M., and Khankeh, H. (2019). Triage

systems in mass casualty incidents and disasters: a re-

view study with a worldwide approach. Open access

Macedonian journal of medical sciences, 7(3):482.

Broadbent, M., Moxham, L., and Dwyer, T. (2007). The

development and use of mental health triage scales in

australia. International journal of mental health nurs-

ing, 16(6):413–421.

Derecho, K. C., Caﬁno, R., Aquino-Caﬁno, S. L., Isla Jr,

A., Esencia, J. A., Lactuan, N. J., Maranda, J. A. G.,

and Velasco, L. C. P. (2024). Technology adoption of

electronic medical records in developing economies:

A systematic review on physicians’ perspective. Dig-

ital Health, 10:20552076231224605. Version A.

Ebrahimi, M., Heydari, A., Mazlom, R., and Mirhaghi,

A. (2015). The reliability of the australasian triage

scale: a meta-analysis. World journal of emergency

medicine, 6(2):94.

Holmes, J. F., Freilich, J., Taylor, S. L., and Buettner, D.

(2015). Electronic alerts for triage protocol com-

pliance among emergency department triage nurses:

a randomized controlled trial. Nursing research,

64(3):226–230.

Jiang, J., Qi, K., Bai, G., and Schulman, K. (2023). Pre-

pandemic assessment: a decade of progress in elec-

tronic health record adoption among us hospitals.

Health Affairs Scholar, 1(5):qxad056.

Johns Hopkins Medicine (2022). Tool developed to assist

with triage in the emergency department. Accessed:

March 10, 2025. https://www.hopkinsmedicine.org/

news/articles/2022/11/tool-developed-to-assist-with-

triage-in-the-emergency-department.

Levin, S., Dugas, A., Gurses, A., Kirsch, T., Kelen, G.,

Hinson, J., et al. (2018). Hopscore: An electronic

outcomes-based emergency triage system. Agency for

Healthcare Research and Quality.

Liu, N., Xie, F., Siddiqui, F. J., Ho, A. F. W., Chakraborty,

B., Nadarajan, G. D., Tan, K. B. K., Ong, M. E. H.,

et al. (2022). Leveraging large-scale electronic health

records and interpretable machine learning for clinical

decision making at the emergency department: pro-

tocol for system development and validation. JMIR

research protocols, 11(3):e34201.

Ofﬁce of the National Coordinator for Health Information

Technology (ONC) (2021). Adoption of electronic

health records by hospital service type, 2019-2021.

Accessed: March 10, 2025.

Slawomirski, L., Lindner, L., de Bienassis, K., Haywood,

P., Hashiguchi, T. C. O., Steentjes, M., and Oderkirk,

J. (2023). Progress on implementing and using elec-

tronic health record systems: developments in oecd

countries as of 2021.

Wang, J., Lu, W., Hu, J., Xi, W., Xu, J., Wang, Z., and

Zhang, Y. (2022). The usage of triage systems in mass

casualty incident of developed countries. Open Jour-

nal of Emergency Medicine, 10(2):124–137.

SIMULTECH 2025 - 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications

336