Fomite Transmission in Airports Based on Real Human Touch Behaviors

Abstract

The public areas of airports are often bustling, raising the risk of infectious diseases spreading through fomites. We recorded 21.3 h of video at three airports, focusing on nine common areas (e.g., boarding and check-in areas) where people touch surfaces. We analyzed 25,925 touches to create a model for how microbes spread from surfaces to humans through touch. The airport mask-wearing rate is high (96.1% in non-restaurant areas), but it is lower (22%) in restaurants. Passengers touch their mucous membranes more often (10.3 times/hour) in restaurants compared to other areas (1.6 times/hour on average). Wearing a mask can significantly reduce the risk of obtaining a virus through direct contact with hands and mucous membranes. If everyone in non-restaurant areas wore masks, the viral intake fraction could be reduced by up to 97.4% compared to not wearing masks. People touch public surfaces the most in self-service check-in areas, at a rate of 473.5 times per hour. Disinfecting public surfaces or hands twice per hour could reduce the viral intake fraction in each area by 27.7% or 15.4%, respectively. The findings of this study provide valuable data support and a scientific foundation for implementing interventions aimed at mitigating fomite transmission within airport settings.

1. Introduction

Fomites comprise both porous and non-porous surfaces or objects that are susceptible to contamination by pathogenic microorganisms, serving as potential vehicles for transmission [1]. A wealth of research underscores the pivotal role of fomites in propagating respiratory and gastrointestinal diseases [2,3,4,5,6].

The transfer of viruses from contaminated surfaces to clean hands, followed by the potential cross-contamination of seven additional pristine surfaces by virus-laden hands, can contribute to the widespread contamination of indoor environments [7]. For instance, a study conducted in aircraft cabins revealed that viruses can contaminate nearly all accessible surfaces within a short span of 6 h [8]. In office environments, if a door handle is contaminated, it could result in over 300 surfaces being infected in just a mere two hours [9]. Touching mucous membranes plays a pivotal role in determining the exposure risk through fomite transmission. When both contaminated hands and fomites are present, viruses can easily enter mucous membranes.

Airports serve as crucial hubs connecting different cities and countries, characterized by crowded environments and extensive interpersonal interactions. Viruses can persist on surfaces within airport environments, potentially exposing passengers to infection during their stay. Assessing the infection risk in various areas of the airport becomes paramount when dealing with significant outbreaks, as it allows for the development of effective preventive and control measures tailored to different risk levels. However, the comprehensive cleaning of all environmental surfaces can impose a heavy burden on airport cleaning personnel. Therefore, focusing on “critical” environmental surfaces at airports can help reduce operational costs [10].

When assessing the risks of indoor infectious disease transmission through fomites, mathematical [11,12] and conceptual models [13] are often employed to forecast exposure risks. However, current research presents two primary limitations. Firstly, these studies frequently overlook the impact of human touch behaviors, leading to inaccuracies in simulating fomite transmission. Secondly, investigations tend to focus solely on hand hygiene and surface disinfection, with minimal consideration for interventions that address human behaviors to mitigate parasite transmission. Hence, there is an imminent need to integrate surface transmission models that encompass real human behaviors, offering a more precise comprehension of individual exposure risks and the effectiveness of intervention measures.

This study involved gathering real touch data from airports, establishing a surface touch database and constructing a Markov chain-based fomite transmission model. We conducted quantitative analyses of passengers’ infection risks in different airport areas and assessed the effectiveness of interventions such as mask-wearing, handwashing, and surface disinfection. This model is applicable to all viruses transmitted through surfaces [14].

2. Methods

2.1. Data Collection

2.1.1. Video Recording and Analysis

Six participants were employed to collect videos, capturing passenger surface touching behavior in nine designated areas (manual check-in area, self-service check-in area, escalator, restaurant, charging area, shopping area, waiting area, boarding area, and baggage claim area) across two airports in the north and one airport in the south. The 9 areas are represented in Figure 1. By handholding or employing tripods to support mobile devices such as smartphones and a Gopro, we ensured that these devices were positioned at the same ground level as the experimenter’s eyes. We selected areas with a higher concentration of passengers for recording, such as waiting areas near boarding areas where flights were scheduled to depart. When capturing passenger touch behavior, we typically used side angles to minimize obstructions that might obscure the passengers’ hand movements. We recorded a total of approximately 21 h of video footage. A team of six trained video analysts systematically analyzed each second, recording every visible touch within the designated areas. A 64-bit Pot Player multimedia player (https://potplayer.daum.net/ accessed on 23 September 2021) was utilized to view the videos, allowing for pausing and the analysis of every second. During the analysis of personnel behavior data, we assigned a numerical ID to each passenger to distinguish them, with our analysis focusing primarily on passenger touch behavior data. After completing the analysis of passenger touch behavior data in various airport areas, we took measures to delete all videos to ensure that the video content remained confidential. Additionally, video analysts signed confidentiality agreements regarding the content of their video analyses. It is noteworthy that our research has received approval from the Ethics Committee of Beijing University of Technology (No: CJXB03). The resulting data were then accurately recorded in a table (samples shown in Table S1). To ensure quality control, the author randomly verified 0.1% of the data, achieving an accuracy rate of 98.9%. The detailed records include time, individual’s ID (each analyzed passenger has an independent numerical ID), gender (male or female), the specific surface touched by their hand (whether left or right), the owner of the surface (categorized into four groups: surfaces belonging to the passenger who touched it, people accompanying the passenger, strangers, or public surfaces), and the mask status of the passenger (whether they were wearing a mask or not). The personal information of passengers, including their names, ages, and contact details such as phone numbers and emails, was not collected. As the type of mask (e.g., N95 respirator, surgical mask) does not impact exposure risk via fomite transmission, we chose not to collect this information. Instead, we recorded whether passengers were wearing masks or not, allowing us to determine the mask-wearing rate, defined as the proportion of time passengers wore masks during their stay.

Through video analysis, we classified a total of 108 surfaces into 5 primary categories across 9 areas. The 5 primary categories include mucosa (mouth, nose, eyes), hands, body (encompassing the trunk, arms, legs, as well as the face and neck), private surfaces (such as mobile phones and backpacks), and public surfaces (such as check-in counters and restaurant tables). Table S2 offers detailed information concerning fomite transmission, including the virus transfer rate between surfaces and hands [15], as well as surface area. The number of primary surfaces in each area is recorded in Table S3.

The surface touch was defined as any contact lasting for one second or longer between the hand and a solid surface [16]. The average surface contact frequency was calculated as the ratio of the surface contact times and the total time all passengers spent in that area. Personal touch frequency refers to the ratio of a passenger’s surface touch times to the duration of their stay in the area.

2.1.2. Online Survey

From 31 March 2022 to 21 April 2022, we administered a survey on the personal behaviors of passengers on China’s largest online survey platform (https://www.wjx.cn accessed on 31 March 2022). The survey consisted of 11 questions related to this study, including the total duration of time spent in airports, the length of time spent in each area of the airport (with 0 indicating no stay), the number of people accompanying the respondent on their trip, and the method of check-in (manual or self-service). If any of the following conditions were met, the questionnaire was considered invalid.

One respondent could only submit the questionnaire once. As per the stated requirements, a total of 1536 valid questionnaires were collected online. The comprehensive questionnaire is provided in Part I in Supplementary Materials.

2.1.3. Flight Information

We collected a total of 383 non-stop flight details, including flight departure time, arrival time, and aircraft model for a certain workday in April 2022 at a northern airport in China, via one of the leading online ticketing platforms in China, Ctrip Travel (https://flights.ctrip.com/online/channel/domestic accessed on 12 April 2022). Assuming that all cabins were fully occupied and using the known aircraft models, we estimated that there were a total of 67,050 passengers at the airport on that day. While surface touch behaviors were collected at three airports, we focused on analyzing the exposure risk and effectiveness of non-pharmaceutical interventions (NPIs) solely at the airport for which we had flight information.

Based on the personnel flight information, we were able to determine the departure time of the passenger’s flight. From the questionnaire, we could determine the duration of time a passenger stayed at the airport, which eventually enabled us to ascertain the time of their arrival at the airport. For instance, if a passenger’s flight departed at 10:00 and they spent 2.5 h at the airport, it could be inferred that their arrival time at the airport was 7:30.

2.2. Fomite Transmission Model

In this study, we only considered the exposure risk through fomite transmission route. In the initial parameter setting, the following assumptions were made:

(1) After conducting a sensitivity analysis of the initial number of infected passengers among airport travelers (Figure S1), we assumed that the proportion of initially infected passengers was 1/1000. This suggests that out of the 67,050 passengers, 67 individuals were infected with the virus. These infected passengers were identified based on an interval system, where all passengers were assigned an ID number from 1 to 67,050 according to the order in which they entered the airport. The index patients were selected based on their ID numbers, which followed a pattern of 1, 1001, 2001, and so on up to 67,001.

(2) At the start of the simulation, 1 unit of virus was initially placed on each of the index patients’ left and right hands. Upon each contact with their mucosa, 1 unit of virus was transferred and evenly spread throughout their hands.

(3) Not all areas of an object can be touched by hands (for instance, the back of a table is seldom touched). Therefore, this study defines the valid contact area as the area that was frequently touched [16]. It is assumed that all viruses on the valid contact area were uniformly distributed.

(4) Once susceptible individuals touch their mucosa, the viruses on their hands enter it at a certain transfer rate. However, the viruses on the mucosa do not transfer onto their hands.

(5) The transfer rate of viruses between hands and surfaces, which refer to the proportion of pathogens transmitted from hands/surfaces to surfaces/hands after hand–surface contact [15], can be influenced by various factors, including pressure and humidity, among others [17]. Since we had insufficient data, we assumed a constant transfer rate between hands and a specific surface in this simulation (Table S2).

(6) The disinfection frequency refers to the number of disinfection procedures conducted per hour [18]. In this study, we set the surface disinfection frequencies at 0.2, 0.5, 1, and 2 times per hour, with a disinfection efficiency of 99.99% [19,20,21]. Surface virus load was cleared to zero at the designated time.

(7) When a person wearing a mask touches their own nose and mouth, it is considered as touching the outer surface of the mask.

The fomite transmission of the virus can be expressed by the following equation [14]:

$$\left\{\begin{array}{l}{V}_s(t+1)=V_s(t)+{\displaystyle \sum _{i_s(t)}^{n_s(t)}}\left[A_{hs}\left(\frac{V_h(t)}{A_h}{r}_{hs}-\frac{V_s(t)}{A_s}{r}_{sh}\right)\right]-\frac{I_{ac-s_i}}{60}{V}_s(t)\\ V_h(t+1)=V_h(t)+A_{hs}\left(\frac{V_s(t)}{A_s}{r}_{sh}-\frac{V_h(t)}{A_h}{r}_{hs}\right)-\frac{I_{ac-h}}{60}{V}_h(t)\end{array}\right.$$

(1)

In this equation, the variables $h$ and $s$ represent the hand and the surface; $V_h\left(t\right)$ and $V_s\left(t\right)$ are the viral loads on the hand and surface at time $t$, while $i_s\left(t\right)$ represents the specific surface that was touched by the hand at that time; $n_s\left(t\right)$ indicates the total number of surfaces that were touched by hands at time $t$; $A_{hs}\left(t\right)$ represents the valid contact area between the hand and surface; $A_s$ and $A_h$ are the total areas of the surface and hand, respectively; $r_{sh}$ and $r_{hs}$ are the virus transfer rates from the surface/hand to hand/surface, respectively; $I_{ac-s}$ and $I_{ac-h}$ represent the inactivation rates (min⁻¹) of the virus on the surface and hands, respectively.

To simulate the behavior of passengers touching surfaces in a specific area of an airport, we utilized the following discrete-time Markov chain to model the sequence of surface touches made by airport passengers:

$$p_{i,j}=P(X_{n+1}=j|X_n=i)$$

(2)

where $X_n$ represents the surface touched at time $n$, $i$ and $j$ refer to the respective surfaces, and $P$ represents the conditional probability. For instance, in the case where $i$ represents a phone and $j$ represents a backpack, $p_{i,j}$ stands for the probability of a person touching the backpack after touching the phone.

The flow chart of the fomite transmission model is shown in Figure 2.

Viral load refers to the quantity of active viruses present in the mucosa. In the case of a passenger who has been exposed to a virus, their personal exposed virus load refers to the amount of the virus that has been absorbed by their mucous membranes. To evaluate the risk of exposure, we utilized a metric known as the viral intake fraction, which is calculated as the ratio of the viral load on the mucosa of each susceptible person to the total amount of virus generated by the index patients. The personal exposed viral intake fraction is defined as the ratio of personal exposed virus load to the total amount of virus produced by the index patients. Meanwhile, the virus fraction is defined as the ratio of the viral load on each surface to the total amount of virus produced by the index patients.

3. Results

3.1. Video and Questionnaire Results

A total of 21.3 h of video footage was recorded, and over 240,000 pieces of data on human touch behaviors were collected (samples are shown in Table S1), concerning 1760 passengers in nine different areas (

Table 1. Surface touch data volume.

Airport’s Area	Video Duration (s)	Number of Recorded Passengers	Data Volume (Person·Second)	Number of Touches
Manual check-in area	6349	76	22,915	3140
Self-service check-in area	12,973	111	9848	2390
Escalator	12,452	1047	29,212	4258
Restaurant	10,549	111	59,977	7474
Charging area	5573	29	20,948	1820
Shopping area	8294	9	10,060	416
Waiting area	15,843	192	68,449	3786
Boarding area	1107	139	11,025	2036
Baggage claim area	3600	46	9233	605

). A total of 25,925 touch events (an example of which is shown in Table S4) were recorded across nine areas.

The mask-wearing rate was 96.1% (95% CI 87.9–99.9%) in all areas except for restaurants, where it was only 22.0% due to meal consumption (

Table 2. Personnel wearing masks and staying status.

Airport’s Area	Mask Wearing Rate a (%)	Probability of Stay b (%)	Duration for Stay (min)
Manual check-in area	98.5	46.9	22.9
Self-service check-in area	95.6	53.1	16.1
Escalator	98.9	97.6	0.4
Restaurant	22.0	76.8	22.1
Charging area	90.6	76.8	22.2
Shopping area	98.2	77.3	20.8
Waiting area	87.3	99.5	28.5
Boarding area	100.0	100.0	1.3
Baggage claim area	99.7	96.6	12.1

Notes: ^a Mask wearing rate is the ratio of time of passengers wearing masks and the duration of stay. ^b The probability of stay is derived from the questionnaire, representing the proportion of passengers staying in a specific area among the valid survey responses.

). The probability of passenger stay in various airport areas was 80.5% (95% CI 48.1–99.9%), and the mean duration of stay was 18.0 min (95% CI 0.25–60 min). The boarding area had the highest passenger stay probability, which was 100%, as passengers checked in and boarded there. Conversely, the manual check-in area had the lowest probability, which was 46.9%, due to the availability of self-service check-in. Finally, the waiting area, manual check-in area, and charging area had the longest average stay durations of 28.5 min, 22.9 min, and 22.2 min, respectively.

3.2. Surface Touching Behavior

The highest mucosal touch frequency (10.3 times/hour) was observed among passengers in restaurants, followed by charging areas (4.6 times/hour) and waiting areas (1.9 times/hour) (

Table 3. Passenger touch behavior.

Airport’s Area	Average Surface Touch Frequency a (Times/Hour)									Personnel Touch Time Ratio b (%)
	Mucosa	Hand	Body		Private Surface	Public Surface	All Surface	Male	Female
			Main Body	Face Neck
Manual check-in area	1.1	13.7	49.3	11.3	283.7	55.8	414.9	423	401.2	90.3
Self-service check-in area	0.5	17.0	22.5	12.1	789.9	473.5	1315.5	1299.2	1335.0	90.9
Escalator	1.1	18.6	44.7	16.3	311.0	68.8	460.4	449.5	477.9	95.8
Restaurant	10.3	11.6	29.1	18.6	251.6	33.1	354.4	379.9	314.6	94.3
Charging area	4.6	11.0	42.5	11.0	163.1	18.4	250.6	245.6	257.4	91.9
Shopping area	2.2	4.5	44.7	87.2	283.8	156.4	578.8	562.4	594.2	86.8
Waiting area	1.9	9.7	32.8	15.0	115.6	9.6	184.6	255.1	137	99.3
Boarding area	1.3	71.5	39.2	10.4	489.1	4.2	615.8	711.5	552.1	97.3
Baggage claim area	0.0	10.9	13.3	7.8	154.0	4.3	190.3	142.2	308.1	98.8

Notes: ^a The average surface touch frequency is calculated as the ratio of the number of touches on the surface and the stay time of all passengers. ^b Personnel touch time ratio is the ratio of the time that people in the area touch the surface and the total time that all people stay.

). Self-service check-in areas recorded the highest touch frequency on public surfaces (473.5 times/hour) due to passengers repeatedly touching the screen of the machine to retrieve their flight tickets (Table 3). Passengers in self-service check-in areas also exhibited the highest personal touch frequency (1315.5 times/hour), followed by those in boarding areas (615.8 times/hour), as these areas require a quick ticket pickup or check-in process (Table 3). Additionally, Figure S2 portrays the probability of duration per touch for passengers in different airport areas. Waiting areas, baggage claim areas, and boarding areas have the longest average duration per touch (36.2 s, 95% CI 1–274.0 s; 30.3 s, 95% CI 1–182.5 s; and 23.2 s, 95% CI 1–77.0 s, respectively) (Figure S2). Passengers waiting for their flights typically use their cell phones to pass the time, leading to longer durations per touch. Approximately 40.8% of touches last for no more than 3 s, with less than 0.5% lasting longer than 5 min (Figure S2). Gender differences were observed in surface touching behavior in the baggage claim area, with female passengers touching surfaces 165.9 times more frequently than male passengers per hour (Table 3). The waiting, baggage claim, and boarding areas had the highest personal touch rates, with rates of 99.3%, 98.8%, and 97.3%, respectively (Table 3).

The Markov touch probability matrix of the entire airport is presented in Figure 3, which comprises 9 areas and a total of 108 surfaces (matrices in different areas can be viewed in Figure S3). It is evident from the matrix that passengers have a substantial likelihood of touching the same surface consecutively, as indicated by the light blue diagonal grids. Furthermore, passengers tend to touch their mobile phone surface (surface 11) after contacting a specific surface.

In various areas of the airport, passengers touched their body surfaces at an average rate of 112.4 times per hour, with an average of 59.8% of passengers engaging in such contact (Figure 4a,b). On personal surfaces, the surface of mobile phones emerged as the most frequently touched surface in the majority of areas. It boasted an average touch frequency of 173.1 times per hour, and approximately 67.4% of passengers chose to interact with their mobile phone surfaces. When it comes to public surfaces, the machine screens in the self-service check-in area were touched at a rate of 371.7 times per hour (95% CI 73.9–974.7 times per hour), and this surface attracted the engagement of 82.0% of passengers. Similarly, the touch frequency of escalator handrails was measured at 214.6 times per hour (95% CI 76.0–600.0 times per hour), with 36.9% of passengers opting to interact with this surface.

3.3. Exposure Risk

As shown in

Table 4. Viral intake fraction, viral load, and proportion of exposed passengers in each area.

Airport Region	Personal Exposed Viral Intake Fraction a (Mean (95%CI))	Personal Exposed Viral Load b (Mean (95%CI))	Proportion of Exposed Passengers c (%)
Manual check-in area	6.0 × 10−10 (4.6 × 10−32–1.3 × 10−9)	8.4 × 10−8 (6.4 × 10−30–1.8 × 10−7)	12.1
Self-service check-in area	3.3 × 10−9 (2.5 × 10−22–3.1 × 10−9)	1.2 × 10−8 (3.4 × 10−20–4.2 × 10−7)	5.1
Escalator	5.3 × 10−8 (4.7 × 10−14–7.9 × 10−9)	7.1 × 10−6 (6.3 × 10−12–1.1 × 10−6)	<0.1
Restaurant	2.3 × 10−9 (5.8 × 10−18–2.0 × 10−8)	1.0 × 10−6 (2.5 × 10−15–9.0 × 10−6)	95.0
Charging area	2.3 × 10−9 (4.8 × 10−13–1.1 × 10−8)	4.3 × 10−7 (8.9 × 10−11–2.0 × 10−6)	11.4
Shopping area	2.6 × 10−10 (9.4 × 10−22–1.4 × 10−9)	3.9 × 10−8 (1.4 × 10−19–2.1 × 10−7)	22.9
Waiting area	1.1 × 10−10 (1.7 × 10−40–3.4 × 10−11)	1.9 × 10−8 (2.9 × 10−38–5.8 × 10−9)	25.3
Boarding area	1.8 × 10−9 (5.9 × 10−26–4.5 × 10−9)	2.5 × 10−7 (8.0 × 10−24–6.1 × 10−7)	<0.1
Baggage claim area	0.0	0.0	0.0

Notes: ^a Personal viral intake fraction is the ratio of the virus exposure to the total amount of virus produced by the infected. The mean (95% CI) of the personal exposed viral intake fraction only considered those exposed passengers. ^b Personal viral load is the amount of activated virus on the mucosa, and the mean (95% CI) of the personal exposed viral load only considered those exposed passengers. ^c The proportion of exposed passengers is the ratio of the number of exposed passengers (personal viral load > 0) to the total number of passengers that were in the area.

, among the 9233 data recorded in the baggage claim areas, no instances were observed of individuals touching their mucous membranes. Consequently, the viral intake fraction for all passengers in this area was considered to be 0. On the other hand, it was found that in escalators, exposure risks (viral load) for passengers who were exposed were greater than in other areas, but only 0.01% of passengers were exposed. The percentage of exposed passengers was highest in restaurants (95%), followed by waiting areas (25.3%) and shopping areas (22.9%). The distribution of the personal virus intake fraction by location is presented in Figure S4.

Taking norovirus as an example, during the peak infection period, the number of virus particles in each gram of feces can exceed 10¹¹. It is estimated that an infection can occur with as few as 10 to 100 norovirus particles [22,23]. Assuming a patient has 0.1 g of virus particles on their hands after using the restroom [24], and susceptible individuals only need to be exposed to 100 virus particles to become infected, the infection threshold for the virus intake fraction is approximately 10⁻⁸ (Figure 5). This means that in the restaurant area, there might be approximately 3.8% of passengers infected, and in the charging area, there might be approximately 0.3% of infections.

Table 5. Virus fraction on human body, private surface, and public surface in different areas ^a.

Airport’s Area	Manual Check-In Area		Self-Service Check-In Area		Escalator		Restaurant		Charging Area		Shopping Area		Waiting Area		Boarding Area		Baggage Claim Area
Hand (×10−8)	Hand	5.9	Hand	27.0	Hand	14.3	Hand	3.9	Hand	3.5	Hand	3.5	Hand	0.8	Hand	0.5	Hand	2.5
Body (×10−8)	Main body	11.2	Main body	16.1	Main body	4.8	Main body	4.0	Main body	10.2	Main body	3.8	Main body	4.9	Main body	5.6	Main body	0.6
	Face neck	0.9	Face neck	5.0	Face neck	0.4	Face neck	1.4	Face neck	0.7	Face neck	6.3	Face neck	2.2	Face neck	0.4	Face neck	0.3
Private surface (×10−8)	Mask	9.3	Mask	22.6	Phone	1.1	Phone	3.1	Phone	6.2	Phone	2.2	Phone	1.2	Mask	1.3	Phone	1.0
	Baggage	10.6	Bag	35.2	Mask	0.4	Mask	7.8	Mask	9.6	Mask	10.6	Mask	2.3	Bag	4.9	Baggage	4.3
	Bag	15.7	ID card	20.2	Baggage	1.3	Bag	3.3	Bag	3.9	Bag	25.9	Bag	4.4	Reticule	2.7	Bag	13.8
	ID card	4.9	Reticule	18.8	Bag	2.6	Tissue	15.5	Tissue	3.4	Reticule	81.4	Reticule	10.7	Hat	0.9	Coat	5.7
	Boarding pass	33.7	Boarding pass	134.5	Boarding pass	1.6	Boarding pass	1.5	Book	1.7	Boarding pass	1.8	Food Packing bag	1.1	Boarding pass	16.6	Reticule	1.8
Public surface (×10−7)	Luggage tray	17.4	Machine screen	3.3	Escalator handrail	530.7	Chair	10.0	Table board	233.5	Commodity	0.7	Trolley	10.0	Trolley	4.8	Trolley	102.8
	Check-in counter	111.5	Machine	0.4	Escalator windshield	5.5	Table	126.0	Chair	4.4	Settlement counter	102.0	Chair	5.2	Check-in counter	15.8	—

Notes: ^a The top 5 and 2 surfaces with the highest virus fraction were selected for private and public surfaces, respectively.

presents the average virus fraction on hands, bodies, private surfaces, and public surfaces across various areas. Interestingly, public surfaces in certain areas (excluding the self-service check-in area) were found to have a higher virus fraction than that on human bodies and private surfaces. In escalators, the virus fraction on hands was greater than that of human bodies and private surfaces, potentially because commuters rarely came into contact with other surfaces after being exposed to viruses on public surfaces, such as escalator handrails during their brief rides.

3.4. Interventions

It is worth noting that, except for the restaurant, if all passengers wore masks in other areas, the virus intake fraction could be reduced by 97.4% (Figure 6). In restaurants where passengers cannot wear masks during their meals, when 20% of passengers wore masks in restaurants, the virus intake fraction decreased by 32.0% in comparison to no mask-wearing. The virus exposure rate for different mask-wearing rates in nine areas is presented in Table S5.

Figure 7 illustrates how disinfection frequency influences individual virus intake fraction reduction. Targeting the most frequently touched surfaces in each area, the optimal disinfection efficacy is achieved with a frequency of twice per hour. Under this regimen, there was a reduction in passenger intake fraction by 5.8% (95%CI 0.0–18.1%). By disinfecting the surfaces with the highest viral scores on both body and personal surfaces, a disinfection frequency of two times per hour led to a decrease in passenger intake fraction by 1.7% (95%CI 0.0–6.9%). However, adopting the practice of replacing masks every half hour resulted in a minimal reduction of only 0.9% (SD, 0.1%) in passenger intake fraction. Conducting disinfection on public surfaces or performing handwashing twice per hour led to reductions in the virus intake fraction of 27.7% (95%CI 3.1–65.2%) and 15.4% (95%CI 0.9–38.5%), respectively, in each area. Based on the simulation results, we found that in areas such as escalators, restaurants, charging areas, shopping areas, and boarding areas, the disinfection of public surfaces proved more effective than handwashing when the disinfection frequency was lower than twice per hour. Conversely, in the self-service check-in areas, handwashing surpassed public surface disinfection in effectiveness.

4. Discussion

This study focused on airport settings and examined human touching behaviors, personal preferences in airports, and mask wearing in nine different airport areas. By establishing a surface transmission model that incorporates these factors, the study analyzed the exposure risk and effectiveness of interventions in each area.

Mucous membrane touch behavior directly determined the risk of surface infection, but it was influenced by factors such as occupation and indoor environment. In our study, passengers exhibited a mucous membrane touch frequency of 10.3 times per hour in airport restaurants, whereas in non-dining areas of the airport, the mucous membrane touch frequency was 1.6 times per hour. When compared to other studies on mucous membrane contact frequencies, such as those conducted in classroom environments with students, they were observed to have a mucous membrane contact rate of approximately 10.1 times per hour, with 36.3%, 31.1%, 26.7%, and 6.0% of contacts involving the mouth, nose, eyes, and combinations of these areas, respectively [25], and in office settings, research indicated that graduate students had an average mucous membrane contact rate of 35.0 times per hour (eyes: 2.8 times/hour; nose: 14.9 times/hour; mouth: 17.1 times/hour) [16]. Furthermore, customers and staff dining in restaurants had an average mucous membrane touch frequency of 40.0 times and 7.0 times per hour, respectively [14]. Therefore, our study found that passengers at the airport demonstrated relatively lower mucous membrane contact frequencies. This observation could be related to passengers wearing masks. Studies during the COVID-19 pandemic showed that wearing masks can reduce the frequency of touching mucous membranes. For example, Chen et al. compared tourism marketing videos on “Youtube” before and during the pandemic and found that wearing masks reduced the frequency of people touching their mucous membranes [26]. Shiraly et al. observed the number of times people in public places touched their faces and mucous membranes and concluded that non-mask wearers were more likely to touch mucous membrane areas than mask wearers [27]. This indicates that wearing masks can not only effectively block airborne and large droplet transmission [28] but can also significantly reduce the frequency of mucous membrane contact.

In hospitals, 20% to 40% of healthcare-associated infections involve cross-contamination through the hands of medical staff, who can become contaminated from direct contact with patients or indirect contact with environmental surfaces [29]. High-touch surfaces in medical settings, such as doorknobs, medical equipment, etc., are potential sources of infection [30,31]. During the COVID-19 pandemic, da Silva et al. sampled high-touch surfaces in public transport areas, including counters, seats, etc., and found a positivity rate of 55.9% [32]. By observing the touch behaviors of passengers in different areas of the airport, we identified several high-touch surfaces, including counters for manual check-in and seats in waiting areas. The significance of these high-touch surfaces in the transmission of pathogens cannot be overlooked, as they can become reservoirs for viruses and bacteria. Therefore, the thorough cleaning and disinfection of high-touch surfaces, along with the promotion of personal hygiene awareness, are crucial measures to reduce the risk of infection.

Maintaining hand hygiene and regularly cleaning environmental surfaces are closely associated with reducing viral infections. In a community that highly emphasizes reinforcing hand hygiene, the incidence of gastrointestinal and respiratory diseases significantly decreased by 31% and 21%, respectively [33]. Moreover, research on the impact of environmental cleaning on the risk of virus transmission through fomites indicated that a single cleaning activity can reduce the average risk of viral infection by 6.98% to 17.06% [34]. In our simulation of disinfecting public surfaces in different areas of the airport, we conducted frequent disinfection of all public surfaces, a practice that may prove challenging to implement in real-world public transport settings. Currently, researchers are focusing on antimicrobial surfaces to reduce the risk of pathogen transmission. The application of antimicrobial surfaces can rapidly deactivate certain bacteria or viruses, thereby lowering the risk of passenger exposure to contaminants in public areas [35,36,37,38].

However, this study has several limitations. Firstly, our primary focus was on the transmission of the virus through hand-to-surface contact, and we did not consider the deposition of the virus from exhaled air or speech-generated droplets onto surfaces. This scenario could lead to a reduction in the viral load on surfaces, thereby lowering the risk of passenger-to-passenger transmission via contaminated surfaces. Secondly, the collection of passenger touch behavior data took place during the COVID-19 pandemic. Due to strict national regulations during this period, passengers were mandated to wear masks in public spaces [39]. Wearing masks reduced the likelihood of passengers coming into contact with mucous membranes, consequently lowering the risk of infection through fomite transmission. Finally, due to the lack of absolute data on the quantity of viruses transferred from mucous membranes to hands, our study relied on virus intake fractions, providing relative estimates associated with exposure risk rather than specific infection risk calculations.

In the future, we can employ Computational Fluid Dynamics (CFD) simulations to more precisely estimate the virus deposition on surfaces resulting from passengers’ talking and breathing. By integrating these data with our contamination spread model, we can accurately determine the viral load on public surfaces and assess the risk of infection to individuals. Additionally, we can conduct experiments by applying a known quantity of MS2 or ϕX174 bacteriophages to simulate passenger touch behaviors in different airport areas [40]. Subsequently, we can measure the bacteriophage levels on surfaces in various airport areas to validate our simulations of surface viral loads. This approach will provide a more accurate foundation for developing preventive and control strategies for contamination spread within airports.

5. Conclusions

This article conducted an analysis of passengers’ touch behaviors in different areas of the airport, utilizing real human touch behaviors as input parameters to establish a fomite transmission model. The objective was to assess the relative infection risks across various airport areas and the effectiveness of intervention measures. Passengers exhibited the highest frequency of mucous membrane contact in the restaurant area, with an average of 10.3 occurrences per hour. Additionally, passengers most frequently touched public surfaces in self-service check-in areas, shops, and escalators, with frequencies of 473.5, 156.4, and 68.8 occurrences per hour, respectively. Using norovirus as an example, high-risk areas were identified, notably the restaurant and charging areas, with infection rates of 3.8% and 0.3% among passengers in these respective areas. Requiring 100% mask-wearing in non-restaurant areas can reduce passengers’ virus intake fraction through surface transmission by 97.4%. Furthermore, disinfecting public surfaces or hands twice an hour can reduce the virus intake fraction by 27.7% or 15.4% in each respective area. This research can assist airports in developing precise cleaning and disinfection plans for frequently touched public surfaces, particularly in high-risk regions. This will contribute to reducing the risk of pathogen transmission and ensuring a safer and more hygienic airport environment.