Micro-moment-based Interventions for a Personalized Support of

Healthy and Sustainable Ageing at Work: Development and

Application of a Context-sensitive Recommendation Framework

Georgios Athanassiou

, Maria Pateraki

2,3

and Iraklis Varlamis

2,4

Department of Ergonomics, IfADo, Leibniz Research Centre for Working Environment and Human Factors,

Ardeystr. 67, 44139 Dortmund, Germany

Institute of Computer Science, Foundation of Research and Technology, 100 Nikolaou Plastira str.,

71003 Heraklion, Greece

School of Rural, Surveying and Geoinformatics Engineering, National Technical University of Athens,

15780, Athens, Greece

Department of Informatics and Telematics, Harokopio University of Athens, Omirou str. 9, 17778, Athens, Greece

Keywords: Recommendation Systems, Micro-moments, Occupational Safety and Health, Well-being, Work Ability,

Sustainable Work, Recommendation-based Interventions.

Abstract: The paper outlines the sustAGE system, a smart solution that builds upon strategic technology trends, such as

Internet-of-Things, machine learning and recommender systems, to support sustainable work environments

and increase wellness at work and well-being with a focus on the ageing workforce. Acknowledging the

interrelation of the work and private arrays for healthy ageing, the developed solution utilizes a

recommendation-based approach providing personalized warnings and preventive recommendations

regarding occupational risks, as well as personalized cognitive and physical training activities for the off-

work context with the overall goal of maintaining Work Ability and enabling sustainable work. The piloting

of the proposed solution in two critical industrial domains provides promising results towards the use of

personalized recommendation-based interventions for the working context and beyond for improving

workers’ occupational safety and health, performance and general well-being.

1 INTRODUCTION

As the overall population in the EU is getting older,

the proportion of relatively older employees in all

occupational sections is also increasing. The

combination of this fact and the decision of most EU

member states to increase the retirement age limits to

at least 65 to 67 introduces specific challenges in

terms of enabling factors for sustainable work (Belin

et al., 2016).

Ensuring sustainable work is a multi-faceted goal

and includes the common and combined consideration

of work-related and general contributing factors. In

terms of Occupational Safety and Health (OSH) and

in accordance with the European legal framework, the

notion of sustainable work considers good working

conditions and the control of all risks for physical and

mental health and across the entire occupational

lifespan (Belin et al., 2016). In addition, the needs of

vulnerable occupational groups, and the (age-)

appropriate design of working conditions and

respective interventions in order to meet these needs,

is also concomitant to the definition of sustainable

work (Belin et al., 2016). Consequently, a specific

focus should be given to specific risks and challenges

for OSH for the occupational group of older workers

(50+).

Another concept that points out the mutual

influence of work and general health and well-being

for sustainable work and sustainable ageing is Work

Ability (WA).

WA defines the balanced interplay

between occupational demands and the (subjectively

perceived) ability of individuals to effectively cope

with these demands (Ilmarinen & Ilmarinen, 2015).

Its development and maintenance depend on several

core, individual as well as contextual, work-related

factors that are tightly interlinked:

1) health and

functional capabilities; 2) competence/expertise; 3)

values, attitudes and motivation towards work; and 4)

job/task characteristics. Ilmarinen & Ilmarinen

(2015) expand the boundaries of the working context

Athanassiou, G., Pateraki, M. and Varlamis, I.

Micro-moment-based Interventions for a Personalized Support of Healthy and Sustainable Ageing at Work: Development and Application of a Context-sensitive Recommendation Framework.

DOI: 10.5220/0010723000003063

In Proceedings of the 13th International Joint Conference on Computational Intelligence (IJCCI 2021), pages 409-419

ISBN: 978-989-758-534-0; ISSN: 2184-3236

409

by including private aspects, such as social

environment as well as overall physical and mental

well-being that demonstrate a mutual

interdependence with work-related factors for WA.

WA remains generally at a high level between the

ages of 20 and 65, however a cross-domain, cross-

gender decline as well as an increase of individual

differences has been observed after 45 years

(Ilmarinen, 2006; Ilmarinen & Ilmarinen, 2015). This

is also in accordance with the inter-individual

variance in the way changes (i.e. decline) in physical

and cognitive abilities ensue with age. Furthermore,

it seems that the availability and maintenance of an

adequate level of personal resources, such as general

well-being and health status, can moderate the effects

of ageing on perceived WA. This suggests the

necessity of personalized approaches that consider

inter-individual variance in order to maintain and

strengthen the respective work-related resources that

affect WA for older (i.e. 50+) employees (Ilmarinen

& Ilmarinen, 2015).

Recommender systems provide individuals with

targeted, cue-based guidance for an appropriate

alignment of goals and goal-promoting behaviours

(Sardianos et al., 2020). In the context of targeted

interventions for human factors/ergonomics and OSH

practice, recommender systems can support

interventions for OSH and well-being of employees

by monitoring behaviour and providing targeted and

systematic recommendation-based guidance for

appropriate responses.

With sustAGE we propose a holistic approach for

recommendation-based interventions for OSH and

well-being in order to promote sustainable work and

Work Ability of older employees. This in turn is

expected to reduce the likelihood of older workers to

leave work by considering their health state in

relation to working conditions and demands.

The recommendations are the tool for gradually

shaping better user profiles, for notifying or alerting

users with respect to OSH risks and recommend

appropriate counter actions, and for motivating them

to engage in activities that promote physical and

psychosocial well-being. In order to strengthen the

effectiveness of these recommendation-based

interventions, sustAGE introduces a contextual

aspect for the delivery of recommendations, which

relates mostly to the appropriate timing of delivery,

but implicitly considers the general conditions within

which a recommendation is issued. This is

implemented through the concept of Micro-moments

(MiMos). MiMos are cues of special interest to the

daily routine of each worker. sustAGE uses them to

trigger MiMo-relevant recommendations.

The recommendation-based interventions

incorporated in sustAGE are based on a respective

conceptual framework and a focused consideration of

age-specific risks and challenges for OSH and well-

being and are being currently piloted in two industrial

domains: manufacturing industry and maritime

logistics/port operations.

In the following Section 2, we shortly outline the

conceptual framework laying at the core of sustAGE

and describe the specification of recommendation-

based OSH and well-being interventions for the

purpose of meeting the needs of two pilot

occupational environments. Section 3 discusses the

practical aspect of the actual implementation of the

framework in two working setups, as well as for the

off-work context. Section 4, illustrates some first

results concerning the acceptance of the proposed

intervention from the workers that participated in the

pilots, the impact that it had on them and the

perceived subjective workload when using sustAGE.

Finally, Section 5 summarizes our findings and

provides an outlook on the next steps of this work.

2 SETTING THE STAGE: AGE,

OSH AND WELL-BEING, AND

DOMAIN-SPECIFIC NEEDS

FOR INTERVENTIONS

The design of sustAGE is relying on the development

of a generic framework for recommendation-based

interventions that outlines the most salient risks and

challenges for OSH and well-being of older

employees regardless of the specific occupation

domain. It constitutes the first step for deriving age-

appropriate interventions for the support and

improvement of OSH and general well-being of older

employees in the working context and beyond.

Further on, and beyond the consideration of

general definition of age-related criticality and

probability of the occurrence of OSH risks, planned

interventions need to take into consideration the

specific characteristics and demands of tasks and

occupations in different industrial domains (Belin et

al., 2016). Hence, the proposed framework serves

also as supporting base-line for the domain-specific

definition and prioritization of interventions.

This section provides a short outline of the

conceptual framework for the sustAGE interventions

and provides insights for the process of substantiation

of interventions for specific occupational contexts of

implementation with distinguishable characteristics

and needs.

SmartWork 2021 - 2nd International Workshop on Smart, Personalized and Age-Friendly Working Environments

410

2.1 A Framework for

Recommendation-based

Interventions for OSH and

Well-being

The objective of the proposed framework is to

provide guidance for the implementation of a holistic

solution towards healthy aging, by jointly improving

OSH and the well-being of older employees.

postulates the well-documented correspondence

between work-related factors, general well-being and

Work ability. It provides the basis for a definition of

the most prevalent risks and challenges for OSH and

well-being for older employees that will be founded

on scientific evidence and promotes the idea of the

tight interdependence of the work and private arrays

for sustainable work and WA.

In addition, it considers the importance of

individual differences in terms of ageing, OSH and

general well-being for the experienced level of Work

Ability and propagates the necessity of implementing

personalized interventions in order to accommodate

the individual needs and characteristics.

The framework points out the need to detect and

to effectively manage existing risks and to prevent

future risks through targeted multi-level

interventions.

It differentiates between three types of

recommendation-based interventions that all

contribute to the goal of sustainable work: those that

target workplace behaviour modification; those that

target structural (i.e. working conditions conditions)

modification; and interventions that target

improvement of general well-being. Further on, it

acknowledges the interrelations and the benefits of a

joint implementation of all intervention types for the

purpose of sustainable work. Figure 1 illustrates the

sustAGE conceptual framework.

The framework provides a rather generic model

for risks and challenges of OSH and well-being that

can accommodate the implementation of targeted

interventions. The design of such interventions will

depend on the characteristics of different

occupational environments, and on the technical

specifications of the sensory environment that

supports user monitoring and the respective

triggering of recommendation.

Finally, it emphasizes user-generated feedback

with respect to the valid assessment of user status as

a necessary precondition for the effectiveness and

acceptance of recommendation-based interventions

by the users as well as for the adjustment of the

triggering indicators and the assessment of

improvements of users’ overall health and well-being.

Figure 1: the sustAGE framework for OSH, well-being and

sustainable work.

2.2 The sustAGE Use Cases for the

Working Context

sustAGE examines two highly demanding working

environments with different requirements that at the

same time are crucial for Europe’s economy:

assembly line work in the manufacturing

(automotive) industry; and maritime logistics

(port/cargo operations).

Based on a thorough context-specific review of

the two industrial domains of assembly line work and

port operations, specific OSH risks for the use-case

scenarios are outlined which can be utilized for

targeted recommendation-based interventions within

sustAGE.

The domain-related specification of OSH risks

was based both on reviews of available literature as

well as on the collection of respective empirical data

from two industrial partners cooperating in the

implementation and evaluation of sustAGE. The

purpose of the data collection consisted in obtaining

information regarding OSH risks, existing practices

and potential solutions as well as cognitive,

psychosocial and physical aspects of work and well-

being directly from the end-users in the two

occupational sites and thus tailor interventions to the

specific needs in an optimal manner. The data

collection took place between September and October

2019 in two respective sites.

An overall of N = 77 individuals participated in

the data collection in both sites (manufacturing = 30;

port operations = 47). 68 thereof were male and 9

were female. The age mean of the participants was

M=53,4 (SD=4,9, Min=41; Max=66). The

participation was voluntary. The test battery included

10 questionnaires and five psychometric cognitive

tests available in the two languages of the sample

(Italian and Greek). In addition, a short interview was

conducted in order to elicit workers’ perception of

age-related aspects of work, and existing practices

with respect to managing OSH risks, work

Micro-moment-based Interventions for a Personalized Support of Healthy and Sustainable Ageing at Work: Development and Application

of a Context-sensitive Recommendation Framework

411

organization and workload, work adaptations in the

face of unanticipated occurrences.

2.2.1 Specification of Interventions for OSH

in the Manufacturing Use Case

OSH risks for older workers in the manufacturing

industry depend on task characteristics and overall

working conditions. In partially automated hybrid

human/machine mixed-model assembly lines, despite

the technological improvements physical abilities and

cognitive skills are still required in order to carry out

more complex production stages not fully automated

such as final assembly operations.

Workplaces and tasks in the manufacturing sector

are characterized by high repetitiveness of short-

cycled tasks and activities, shift work, dependence on

the pace given by the line, awkward body postures,

occasionally lifting heavy objects, and application of

force that is associated with strain to specific body

parts and muscle groups involved in the

assembling/manufacturing task (mostly neck and

upper extremities) (Belin et al., 2016; Boenzi et al.,

2015; Guerreiro et al., 2017; Otto & Battaïa, 2017).

Continuous exposure of older workers to

suboptimal working conditions can lead to

discomfort, pain and health impairments due to work-

related musculoskeletal disorders (MSD) and

susceptibility for respective injuries resulting in

health-related drop outs from work and drawbacks in

productivity especially as workers’ age advances

(Belin et al., 2016; Guerreiro et al., 2017)

Another important aspect of OSH that is

associated with MSD and working conditions is task-

related fatigue. Older workers in manufacturing have

reported that fatigue mainly manifests itself in strain

in lower and upper extremities and the neck.

In addition, the acquired data suggest that older

workers may demonstrate lower levels of WA

(measured with the Work Ability Index). Figure 2

illustrates the overall WA score for the sample in the

manufacturing context.

Figure 2: Work Ability overall scores-manufacturing

industry.

This result can be associated with the reported low

autonomy and control over the process and the

amount of work when compared to older workers in

port operations that may imply an imbalance in the

perceived task demands and available resources (see

Figure 3).

Figure 3: Relative frequencies – Influence of workers on

work volume Port/manufacturing (COPSOQ Item B.3-02:

“Can you influence the amount of work assigned to you?”

1=always;5=never).

OSH risks in the manufacturing context can be

mitigated by providing workers with more control

and decisions over the working process (e.g. break

management) and with tasks that vary in cognitive

demands; introducing job redesign measures for

optimizing workload such as job/work station

rotation and task switching; and monitoring awkward

postures and physical and mental workload and

fatigue in real-time and providing support

(information and opportunities for micro-breaks with

alternative activities) for effective mitigation.

The respective areas for interventions that target

ageing manufacturing workers will have to focus on

existing risks for OSH and well-being for the overall

workforce and prioritize aspects that correspond

highly with advanced working age in terms of work

demands, available compensation capabilities, and

potential consequences for safety and health.

Recommendation-based interventions should

focus on the preventive management of specific OSH

risks and hazards related to the repetitive execution of

the same task and the associated mental and physical

workload and fatigue; to the accumulated strain

through awkward body postures during assembly

work; the introduction of individual micro-breaks for

a short-term mitigating recovery in the face of fatigue

effects; and the potential utilization of the short

intervals between each assembly task cycles for short

physical (stretching) exercises.

Hence, respective recommendations should

trigger and enable the optimization of body part

stressing and mitigation of physical fatigue (that may

SmartWork 2021 - 2nd International Workshop on Smart, Personalized and Age-Friendly Working Environments

412

lead to muscle strain injuries and MSD on the long

run), as well as the optimization of the mental

demands of consecutive and repetitive micro-tasks of

medium and high complexity that may lead to mental

fatigue and/or monotony effects. Apart from the

preventive avoidance of injuries, discomfort and

suboptimal physical and mental workload,

recommendation-based interventions must also

address the monitoring of environmental conditions

in order to provide a proper working environment.

2.2.2 Specification of OSH Interventions for

the Maritime Logistics Use Case

The port and cargo handling sector has seen the

introduction of technological aids and sophisticated

supporting solutions in the last decades, signalling a

shift from labour-intensive towards capital-intensive

operations (Turnbull, 2011). Despite the

technological developments, port operations remain

challenging for OSH, as port workers are still exposed

to occupational hazards and discomfort factors as

well as face the risk of severe occupational accidents

(Antão et al., 2016). In addition, containerisation is

characterized by a rather high degree of repetitiveness

and reduced mental demands that may result in

decreased attention during work and also overall

long-term decline of cognitive abilities due to lack of

challenging work-conditions and the respective

work-related resources.

The respective areas for interventions that target

older port workers will have to focus on existing risks

for OSH and well-being for the overall workforce and

prioritize aspects that correspond highly with

advanced working age in terms of work demands,

available compensation capabilities, and potential

consequences for safety and health.

Recommendation-based interventions that focus

on OSH and the well-being of older port workers

should focus on the preventive management of

specific OSH risks and hazards of the working set of

cargo handling. In addition, different occupational

groups are involved in port operations that face

different risks and challenges for OSH due to the

specific nature of the respective tasks, special

consideration should be taken with respect to the

group-related needs for implementing context-

appropriate recommendations. The current

recommendation system focuses on the two most

significant occupational groups for port operations

involving containerized cargo, crane operators (COs)

and dockworkers (DWs), as these groups have been

found to experience the most OSH challenges during

cargo handling at the port (Walters et al., 2020).

Interventions addressing DWs should target the

optimization of heavy physical workload and fatigue

(that may lead to muscle strain injuries and MSD).

The acquired data for the port environment confirms

this issue as dock workers have reported that fatigue

corresponds with back aches, as well as

musculoskeletal discomfort and pain in upper and

lower extremities.

In addition, as port operations take place in the

open, workers are exposed to environmental conditions

such as high or low temperatures. The effects of

environmental conditions (especially heat stress)

should also be targeted as heat stress is associated with

cardiovascular impairments which can be crucial for

older workers. Although other environmental

conditions (such as noise and vibration) seem to be

more prevalent for the manufacturing context, these

should also be monitored and, if feasible, addressed in

an appropriate manner. Figure 4 illustrates the

workers’ appraisal of the relevance of respective

hazards for their working context.

Figure 4: Relative frequencies of stated OSH hazards in

port operations and manufacturing.

Safety awareness during operations must be

supported in order to prevent potential accidents, and

the monitoring of environmental conditions is

important for avoiding workers’ exposure to extreme

conditions (M. Cezar-Vaz et al., 2014; M. R. Cezar-

Vaz et al., 2016; Walters et al., 2020).

Respectively, interventions addressing COs

should target effects of environmental conditions,

work-related physical and mental strain effects

associated with demands for sustained attention,

responsibility, and awareness for the safety of people

on the ground, musculoskeletal discomfort through

continuous and constrained arm/hand postures and

repetitive head and neck movements and prolonged

stationary work can create preconditions for MSD

and chronic fatigue as well as psycho-social negative

responses such as anxiety through the increased

responsibility for the safety of others (Yakub & Sidik,

2014).

Micro-moment-based Interventions for a Personalized Support of Healthy and Sustainable Ageing at Work: Development and Application

of a Context-sensitive Recommendation Framework

413

2.3 Specification of Recommendations

for General Well-being

In addition to the recommendations within the

workplace, the system incorporates gamified

cognitive training measure supports the maintaining

cognitive skills that are related to work performance

and general health and well-being and have been

found to decline with age (e.g. memory, executive

functions, processing speed) (Anguera et al., 2013;

Ballesteros et al., 2014; Green & Bavelier, 2003;

Strobach & Karbach, 2021). At the same time,

maintaining physical fitness also corresponds with

mental fitness and stress reduction for older

individuals and this has been also associated with

overall healthy ageing (Colcombe & Kramer, 2003;

Gajewski & Falkenstein, 2015; Smith et al., 2010)

sustAGE unobtrusively and under the premise of

users’ consent utilizes information beyond the

working context through monitoring users’ patterns

of overall physical activity. Based on this

information, the system proposes (physical and

cognitive) training activities over the day that

promote wellness and support maintenance of

cognitive skills and an adequate physical fitness, and

also inform users over performance improvement

over time in order to motivate the sustainable

habituation of an overall healthy lifestyle. In addition,

off work interventions address sleep quality as this

issue correlates with work-related aspects such as

chronic fatigue, performance detriments at work and

absenteeism.

Respective recommendation-based interventions

consist in reminders for practice physical and

cognitive skills on a regular albeit dynamic manner

(i.e. taking advantage of any idle time that users may

have while at home).

In order to improve physical fitness, the system

recommends off-work physical activities several

times a week. In order to increase the

recommendation acceptance, the system considers

the location of the users two hours after they have

finished work, and if they are at home, they are

recommended to have 1 hour of physical activity.

Cognitive Games (CGs) provide a joyful

intervention to compensate for age-related declines in

cognitive functions. Six different CGs are used to

provide users with the ability to exercise their

cognitive skills. The system recommends the users to

participate in a cognitive game challenge every

second day or three times a week, but the users are

also able to train with the available games any time

they wish. Similar to the physical activity

recommendation, the users are recommended to play

CG in the afternoon when they are at home. The

acceptance of the provided recommendation is

automatically verified by the CG app back end which

logs the hours that each user has played, his/her

scores to the games and the difficulty level that is

currently attributed to the given user.

With respect to sleep quality, the number of sleep

hours can drastically affect the performance of the

users at work. The system estimates the hours that

each user sleeps every night and when they are below

a specified threshold, the user is informed in the

morning and asked to sleep earlier the following

night. Moreover, if the average sleep hours for several

consecutive days is low, the user is informed about

the possibility of accumulated fatigue due to

repeating poor sleep and is recommended to adjust

his/her long term go-to-bed strategy.

The implicit or explicit users’ response to a

recommendation is recorded and employed by the

system for deciding whether a new similar

recommendation should be issued (or not).

3 PRACTICAL APPLICATION OF

THE FRAMEWORK

In order to enable the technical implementation of the

recommendation framework presented in the

previous section, the following aspects have been

considered:

i) multiple streams from different sensors

are exploited for the continuous and real-time analysis

of user-centred and contextual information as these

may reveal important properties of user state, actions

and interactions with the environment; ii) the different

modalities from sensor measurements are combined to

achieve a more robust detection; iii) unobtrusive

monitoring via a privacy-by-design approach is in the

centre-point, exploiting a minimum setup of low-cost

sensors and in parallel adopting privacy and security

mechanisms in data ingestion, management and

communication for improved user acceptance.

The sensor streams allow to detect low level events,

which are further correlated with past user-data, and

higher-level information about possible conditions or

restrictions, and subsequently trigger MiMos and the

respective recommendations. The system relies on a

wide range of sensor feeds and devices to capture and

exploit multimodal information:

Wellness Sensor. The Garmin Vivoactive-3

smartwatch is used to collect heart-rate, beat-to-beat

intervals, the number

of steps, and 3D accelerometer

information with a dedicated smartwatch app.

SmartWork 2021 - 2nd International Workshop on Smart, Personalized and Age-Friendly Working Environments

414

Environmental Sensors. A custom solution that has

been developed to enable the use of multiple third-

party sensors installed in a sensor box in various

places at work. They are used for capturing

environmental data such as ambient temperature,

relative humidity, barometric pressure, wind speed,

air quality, illuminance and noise level.

Visual Sensors. Visual information is captured

through a monocular camera installed on a mobile

harbour crane. The camera is connected directly via

GigE connection to a PC supporting all relevant

processing close to the original source. This camera

shares a common reference frame with the crane, as

described in Lourakis et al. (2020) visual sensors are

used to monitor moving containers, and extracted

information is correlated with workers location data

to trigger proximity alerts. In the manufacturing case

passive stereo vision is exploited across each

workstation to detect stressing body postures.

Speech Sensors. The human voice encapsulates

linguistic and paralinguistic information, which

relates to a speaker’s current states, traits and well-

being. Workers utterances may be captured through

the microphone embedded in the Xiaomi Mi 9

smartphone used by the platform users. The

smartphone plays a key role, being the primary means

of interaction between the users and the system.

Positioning Sensors. Positioning information is

transmitted through the smartphone multi-GNSS

chipset by simultaneously tracking the signals

broadcasted from all operational satellites (i. e. GPS,

GLONASS, Galileo and BeiDou), providing

increased robustness, reliability and spatial coverage.

3.1 Manufacturing Use Case

For detecting worker fatigue in the manufacturing

case, the smartwatch is used to continuously collect

heart rate (HR) and heart rate variability (HRV) data

of each worker. By comparing the actual HR and

HRV values with the worker’s personalised “normal”

values the system detects a user fatigue MiMo.

Subsequently, the Recommendation Engine (RE)

examines the worker’s context (at work or not), the

average fatigue state of the team, and the worker’s

sleep hours the previous day. Depending on the case,

the RE either associates the fatigue with low sleep

quality and notifies the worker to adjust his/her sleep

schedule, or if the last night sleep time was sufficient

it recommends the worker to take a micro break.

Worker safety mainly targets the repetitive

actions and body postures that stress workers’ body

parts and may potentially lead to injuries and MSD.

Taking advantage of the visual sensors, the system

detects workers’ postures in real-time. Postures are

correlated to risk indices and if repeated for a longer

period, the system immediately alerts users to change

posture and perform stretching activities for

compensation of musculoskeletal strain as soon as

possible.

Adverse environmental conditions occasionally

occur in the indoor environment of a manufacturing

industry, e.g. when the A/C system is malfunctioning,

when noise or dust is temporarily high. The system

continuously monitors the environment to identify

conditions that may cause worker’s discomfort, or

increase accident risks. In such cases, it notifies

workers to protect themselves from the extreme

conditions and their managers in order to, if possible,

fix the problem.

3.2 Maritime Logistics Use Case

Worker fatigue in the port case can be detected for a

single worker or a group of workers and the

recommendations may differ depending on the

situation. For single worker fatigue events the

recommendation is for the worker to take a micro break

or fix his/her sleep schedule. When fatigue events

occur for multiple workers, a recommendation is sent

to the foreman and the crane operator to slow down

operations or grant an extra short break for recovery.

Worker safety/accident prevention: Taking

advantage of the real-time users’ location monitoring

and the state-of-the-art computer vision that

accurately tracks containers during transfer, the

distance between the individual users and the moving

container is continuously estimated. If this distance is

below a safety threshold (i.e. users into a hazardous

zone), the system immediately alerts users to move

away from the container and eliminate the risk of an

accident.

Adverse environmental conditions affect workers

differently in the maritime logistics environment. As

the dock-workers operate in the open space, the

system monitors the environment to identify

conditions that may cause worker’s discomfort, or

increase accident risks. The system blends

information from the temperature, hygrometer and

wind speed sensors, to compute the heat and chill

index, i.e. how the current weather feels for the

workers (Anderson et al., 2013). These “Feels-like”

values are used to detect situations that may require

the intervention of the system. In such a case, the RE

Micro-moment-based Interventions for a Personalized Support of Healthy and Sustainable Ageing at Work: Development and Application

of a Context-sensitive Recommendation Framework

415

retrieves the list of workers and the supervisor in their

workplaces, and issues recommendations for short

breaks, resting, or hydration, in case of adverse

weather. Especially in the case of extremely high

wind at the port, crane operators are notified to seize

operations in order to minimize accident risk.

4 EVALUATION AND RESULTS

In order to evaluate the sustAGE intervention to the

two domains, we engaged 23 participants (port

operations =15; manufacturing = 8) who volunteered

to use the system for a four weeks pilot period. The

average age for the overall sample was M=54 with a

SD=3,79 (port operations: M =53,5; SD =4,32;

manufacturing: M = 54,8, SD= 2,53). For the

evaluation of recommendation-based intervention,

we distinguish three different types of criteria: i) the

recommendations’ acceptance, ii) the

recommendations’ impact and iii) the subjective

impact on workers’ workload.

For the acceptance of recommendations, we rely

on the explicit user feedback on each

recommendation, which is expressed with an

“accept” or “ignore” option button that workers can

press in any recommendation. For the impact we

evaluate whether the recommendations have

modified the workers’ training habits during the pilot

period, i.e. if they have performed the gamified

cognitive training and if they engaged in physical

activity (walking) more than before. In addition, we

examine whether workers have adopted better sleep

habits in order to avoid fatigue events and enhance

overall well-being. Finally, we assess the users’

perspective on system-induced workload and the

positive or negative interference of sustAGE in the

work process, we used the NASA TLX questionnaire

that captures the subjective task-related workload on

six different scales: mental demands, physical

demands, temporal demands, quality of performance,

effort and frustration.

4.1 Maritime Logistics/Port Operations

The overall recommendation acceptance in the

maritime logistics scenario is higher, reaching 88%

(Figure 5

Figure 5

). The acceptance breakdown to

different recommendation types, shows that there is

one recommendation (i.e. 𝑅_ℎ𝑒𝑎𝑟𝑡_ℎ𝑒𝑙𝑝: ask if help

is needed) that collects the majority of user rejections.

This recommendation is either based on a

misconception of the recommendation message by

the workers or by an incorrect setup of the high heart

rate levels, which are probably set lower than they

have to be, resulting in oversensitivity for assistance

requests (false-positives). If we leave this

recommendation aside, the overall acceptance rate for

all other types rises to 95.5%.

Figure 5: Recommendations’ acceptance rates.

The recommendations have a direct impact on the

amount of training (physical and cognitive) that the

workers performed during the pilot. More specifically

the average daily number of steps of dock workers

increased by 700 steps within one month (a 25%

increase in relation to the average number of steps in

the first week). The average number of completed

cognitive game-based cognitive training sessions, and

the respective average high score gradually increased

during the study.

Finally, the NASA TLX results have shown only

one statistically significant difference with respect to

the use of the recommender system. More precisely,

a main effect of the system’s use on quality of task

performance for both groups (crane operators and

dock workers) has been observed (F (1,14) = 6,089,

p=.027), something that indicates that the system may

have had a positive impact in the respective aspect.

However, this issue needs further elaboration in order

to specify the perceived benefits for performance

stated by the users. With respect to the rest of the

workload scales, it seems that the system does neither

positively nor negatively affect workload.

Considering the short phase of the evaluation and the

novelty of the recommender system for the port

workers, this point can be viewed in a positive

manner, as working with sustAGE was not perceived

as obtrusive or negatively affecting the work process.

SmartWork 2021 - 2nd International Workshop on Smart, Personalized and Age-Friendly Working Environments

416

4.2 Manufacturing

The first results on recommendations’ acceptance

from the manufacturing domain pilot demonstrate

that the recommendations are well accepted with an

overall acceptance rate of 75%. The analysis per

recommendation type shows that the majority of heart

rate and fatigue related recommendations have been

accepted (both by shift managers and line workers),

with a cumulative acceptance rate over 90%. Some

technical issues related to a geofencing mechanism

that detects when a worker is at work, led to the

issuing of false recommendations concerning the

arrival at work.

Figure 6: Average daily number of steps for the

manufacturing pilot.

We also had a high ratio of rejected

recommendations for performing physical training

(aerobic activity/walking) after work. This is mainly

because many users may draw a clear line between

work and their personal life and private time.

Therefore, although they accept the use of the

sustAGE system in the working context, they

apparently prefer to “decouple” from it in the

afternoon when they are out of work and thus they

seem to be rather unwilling to follow the respective

recommendations, even if these can be beneficial for

the overall well-being.

Accordingly, the amount of physical activity has

shown a decreasing trend in this case. However, we

can see that workers in general perform more steps

than the shift managers and both walk more at

workdays than during weekends. The main reason

was that after the first week the users were

occasionally taking off smartwatches in the afternoon

hours, therefore the number of steps performed for the

rest of the day couldn’t be counted. This behaviour

became a standard practice for many users by the end

of the study, which is interpreted as an overall decline

in the average number of steps (Figure 6).

Figure 7: The perceived workload before and after the pilot

execution (Performance dimension: inverted scores).

Concerning the perceived system-related impact

on workload, the NASA TLX index for the

manufacturing case slightly decreases from 43,8 (pre-

evaluation) to 38,1 (post-evaluation) (Figure 7)

The difference between the overall workload scores

is not significant before and after using the sustAGE

system, so we can also infer that, similarly to port

operations, sustAGE does not pose an additional

impediment to work process or negatively affects the

perceived workload for users in the manufacturing

use case. Considering the additional technical burden

that a new system can bring to the workers, the

observed slight decrease in the perceived workload is

a positive indication for a beneficial effect of

sustAGE for supporting the working process.

Also similarly to the port environment, workers in

the manufacturing context have reported better

performance when using sustAGE, although the

difference is marginally non-significant

(F (1,7) = 4,065, p = .084 at a CI = 95%). Moreover,

this result becomes statistically significant if the

Confidence Interval is set to 90% (see figure 7).

Considering the small sample size, this result

indicates a clear tendency towards an increased

performance quality when supported by sustAGE.

Again, this issue needs further elaboration in order to

specify the perceived benefits for performance stated

by the users.

5 CONCLUSIONS AND

OUTLOOK

The initial results that we have from the two pilot

studies are promising. They show that the proposed

recommendation-based intervention can be beneficial

for the workers and implicitly for their employers,

given that all the participants adopt the proposed

Micro-moment-based Interventions for a Personalized Support of Healthy and Sustainable Ageing at Work: Development and Application

of a Context-sensitive Recommendation Framework

417

framework and the changes that it brings to their

routine. A more extended pilot in both cases, that is

scheduled as a next step of our work, will allow us to

implement the lessons learned from the first two pilot

studies and further improve the results of our

intervention.

More specifically, an increased number of MiMos

recognized by the system produced a high number of

recommendations, which in general were accepted

positively by users. However, we had an indication

that the number of recommendations should be

slightly reduced, focusing mainly on those that are

absolutely necessary to improve safety, health and

productivity of employees, and sporadically

addressing secondary aspects.

The difference between the perceived workload

aspects of the working task with and without the use

of sustAGE and the respective work-related

recommendations was found to be not significant.

This means that the system did not negatively affect

the perceived workload especially considering some

technical issues with regard to the loss of connection

and the recommendations’ delivery.

The sustAGE recommendations framework and

its supporting ecosystem with the sensors, the

processing modules, and the delivery and feedback

mechanisms will allow to maximise the impact of the

intervention. Moreover, it will provide ageing

workers with a solution that empowers them to keep

their competencies, skills and fitness at a high level

and remain healthy and active in the workforce

throughout the foreseen time-span until retirement

(but also after retirement), setting the preconditions

for and supporting sustainable work and overall

healthy ageing.

ACKNOWLEDGEMENTS

This work has received funding from the European

Union’s Horizon2020 research and innovation

programme under the grant agree-ment No. 826506

(sustAGE) (https://www.sustage.eu).

REFERENCES

Anderson, G. B., Bell, M. L., & Peng, R. D. (2013).

Methods to Calculate the Heat Index as an Exposure

Metric in Environmental Health Research.

Environmental Health Perspectives, 121(10), 1111–

1119. https://doi.org/10.1289/ehp.1206273

Anguera, J. A., Boccanfuso, J., Rintoul, J. L., Al-Hashimi,

O., Faraji, F., Janowich, J., Kong, E., Larraburo, Y.,

Rolle, C., Johnston, E., & Gazzaley, A. (2013). Video

game training enhances cognitive control in older

adults. Nature, 501(7465), 97–101. https://doi.org/

10.1038/nature12486

Antão, P., Calderón, M., Puig, M., Michail, A.,

Wooldridge, C., & Darbra, R. M. (2016). Identification

of Occupational Health, Safety, Security (OHSS) and

Environmental Performance Indicators in port areas.

Safety Science, 85, 266–275. https://doi.org/10.1016/

j.ssci.2015.12.031

Ballesteros, S., Prieto, A., Mayas, J., Toril, P., Pita, C.,

Ponce de LeÃ

n, L., Reales, J. M., & Waterworth, J.

(2014). Brain training with non-action video games

enhances aspects of cognition in older adults: A

randomized controlled trial. Frontiers in Aging

Neuroscience, 6. https://doi.org/10.3389/fnagi.2014.00

277

Belin, A., Graveling, R., Bongers, P., Hagen Mikkelsen, S.,

Crawford, J. O., Davis, A., Cowie, H., Dupont, C.,

Dixon, K., European Agency for Safety and Health at

Work, Milieu Ltd, Toegepast natuurwetenschappelijk

onderzoek, COWI, & Institute of Occupational

Medicine. (2016). The ageing workforce implications

for occupational safety and health: A research review.

Publications Office.

Boenzi, F., Digiesi, S., Mossa, G., Mummolo, G., &

Romano, V. A. (2015). Modelling Workforce Aging in

Job Rotation Problems. IFAC-PapersOnLine, 48(3),

604–609. https://doi.org/10.1016/j.ifacol.2015.06.148

Cezar-Vaz, M., de Almeida, M., Bonow, C., Rocha, L.,

Borges, A., & Piexak, D. (2014). Casual Dock Work:

Profile of Diseases and Injuries and Perception of

Influence on Health. International Journal of

Environmental Research and Public Health, 11(2),

2077–2091. https://doi.org/10.3390/ijerph110202077

Cezar-Vaz, M. R., Bonow, C. A., Almeida, M. C. V. de,

Sant’Anna, C. F., & Cardoso, L. S. (2016). Workload

and associated factors: A study in maritime port in

Brazil. Revista Latino-Americana de Enfermagem,

24(0). https://doi.org/10.1590/1518-8345.1347.2837

Colcombe, S., & Kramer, A. F. (2003). Fitness Effects on

the Cognitive Function of Older Adults: A Meta-

Analytic Study. Psychological Science, 14(2), 125–

130. https://doi.org/10.1111/1467-9280.t01-1-01430

Gajewski, P. D., & Falkenstein, M. (2015). Lifelong

physical activity and executive functions in older age

assessed by memory-based task switching.

Neuropsychologia, 73, 195–207. https://doi.org/

10.1016/j.neuropsychologia.2015.04.031

Green, C. S., & Bavelier, D. (2003). Action video game

modifies visual selective attention. Nature, 423(6939),

534–537. https://doi.org/10.1038/nature01647

Guerreiro, M., Serranheira, F., Cruz, E. B., & Sousa-Uva,

A. (2017). An analysis on neck and upper limb

musculoskeletal symptoms in Portuguese automotive

assembly line workers. International Journal of

Occupational and Environmental Safety, 1(1), 59–68.

https://doi.org/10.24840/2184-0954_001.001_0006

Ilmarinen, J. (2006). The ageing workforce—Challenges

for occupational health. Occupational Medicine, 56(6),

SmartWork 2021 - 2nd International Workshop on Smart, Personalized and Age-Friendly Working Environments

418

362–364. https://doi.org/10.1093/occmed/kql046

Ilmarinen, J., & Ilmarinen, V. (2015). Aging and work

ability. In Facing the challenges of a multi-age

workforce: A use-inspired approach (S. 177–207).

Routledge.

Lourakis, M., Pateraki, M., Karolos, I.-A., Pikridas, C., &

Patias, P. (2020). Pose estimation of a moving camera

with low-cost, multi-gnss devices. The International

Archives of the Photogrammetry, Remote Sensing and

Spatial Information Sciences, XLIII-B2-2020, 55–62.

https://doi.org/10.5194/isprs-archives-XLIII-B2-2020-

55-2020

Otto, A., & Battaïa, O. (2017). Reducing physical

ergonomic risks at assembly lines by line balancing and

job rotation: A survey. Computers & Industrial

Engineering, 111, 467–480. https://doi.org/10.1016/

j.cie.2017.04.011

Sardianos, C., Varlamis, I., Chronis, C., Dimitrakopoulos,

G., Alsalemi, A., Himeur, Y., Bensaali, F., & Amira, A.

(2020). The emergence of explainability of intelligent

systems: Delivering explainable and personalized

recommendations for energy efficiency. International

Journal of Intelligent Systems, n/a(n/a), 1–25.

https://doi.org/10.1002/int.22314

Smith, P. J., Blumenthal, J. A., Hoffman, B. M., Cooper,

H., Strauman, T. A., Welsh-Bohmer, K., Browndyke, J.

N., & Sherwood, A. (2010). Aerobic Exercise and

Neurocognitive Performance: A Meta-Analytic Review

of Randomized Controlled Trials: Psychosomatic

Medicine, 72(3), 239–252. https://doi.org/10.1097/

PSY.0b013e3181d14633

Strobach, T., & Karbach, J. (Hrsg.). (2021). Cognitive

Training: An Overview of Features and Applications.

Springer International Publishing. https://doi.org/

10.1007/978-3-030-39292-5

Turnbull, P. & International Labour Office. (2011). An

international assessment of training in the port sector.

ILO.

Walters, D., Wadsworth, E., & Bhattacharya, S. (2020).

What about the workers? — Experiences of

arrangements for safety and health in global container

terminals. Safety Science, 121, 474–484.

https://doi.org/10.1016/j.ssci.2019.09.017

Yakub, N. W., & Sidik, S. M. (2014). Prevalence and

Contributing Factors of Job Strain among Crane

Operators in a Port Container Terminal in Malaysia.

10, 8.

Micro-moment-based Interventions for a Personalized Support of Healthy and Sustainable Ageing at Work: Development and Application

of a Context-sensitive Recommendation Framework

419