JAMIA Open, 5(2), 2022, 1–6

https://doi.org/10.1093/jamiaopen/ooac038
Application Notes

Application Notes

Smart Scheduling (SMASCH): multi-appointment

scheduling system for longitudinal clinical research stud-
ies
Carlos Vega 1, Piotr Gawron1, Jacek Lebioda1, Valentin Grouès1,
Piotr Matyjaszczyk2, Claire Pauly1,3, Ewa Smula1, Rejko Krüger1,3,4,
Reinhard Schneider 1, and Venkata Satagopam 1

1
e du Luxembourg, Esch-sur-Alzette, Luxembourg, 2Institute of Computing
Luxembourg Centre for Systems Biomedicine, Universit
Science, Poznan University of Technology, Poznan, Poland, 3Parkinson Research Clinic, Centre Hospitalier de Luxembourg,
Luxembourg, Luxembourg and 4Transversal Translational Medicine, Luxembourg Institute of Health, Strassen, Luxembourg

Corresponding Author: Dr. Venkata Satagopam, Luxembourg Centre for Systems Biomedicine, Universite du Luxembourg,
L-4365 Esch-sur-Alzette, Luxembourg; [email protected]
Received 6 December 2021; Revised 11 April 2022; Editorial Decision 3 May 2022; Accepted 10 May 2022

ABSTRACT
Objective: Facilitate the multi-appointment scheduling problems (MASPs) characteristic of longitudinal clinical
research studies. Additional goals include: reducing management time, optimizing clinical resources, and se-
curing personally identifiable information.
Materials and methods: Following a model view controller architecture, we developed a web-based tool written
in Python 3.
Results: Smart Scheduling (SMASCH) system facilitates clinical research and integrated care programs in Lux-
embourg, providing features to better manage MASPs and speed up management tasks. It is available both as
a Linux package and Docker image (https://smasch.pages.uni.lu).
Discussion: The long-term requirements of longitudinal clinical research studies justify the employment of flexi-
ble and well-maintained frameworks and libraries through an iterative software life-cycle suited to respond to
rapidly changing scenarios.
Conclusions: SMASCH is a free and open-source scheduling system for clinical studies able to satisfy recent
data regulations providing features for better data accountability. Better scheduling systems can help optimize
several metrics that ultimately affect the success of clinical studies.

Key words: data integration, multi-appointment scheduling problems, web services

C The Author(s) 2022. Published by Oxford University Press on behalf of the American Medical Informatics Association.
V
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unre-
stricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. 1
2 JAMIA Open, 2022, Vol. 5, No. 2

LAY SUMMARY
Longitudinal clinical research studies require conducting various assessments over long periods of time. Such assess-
ments comprise numerous stages, requiring different resources defined by multidisciplinary research staff and aligned
with available infrastructure and equipment, altogether constrained by time. While it is possible to manage the alloca-
tion of resources manually, it is complex and error-prone. Efficient multi-appointment scheduling is essential to assist
clinical teams, ensuring high participant retention and producing successful clinical studies, directly impacting patient
throughput and satisfaction. This application note presents Smart Scheduling (SMASCH) system, a web application for
multi-appointment scheduling management aiming to reduce times, optimize resources and secure personal identifiable
information. SMASCH is present in multiple clinical research and integrated care programs in Luxembourg since 2017,
including Dementia Prevention Program, the study for Mild Cognitive Impairment and gut microbiome, and the National
Centre of Excellence in Research on Parkinson’s disease which encompasses the study for REM sleep behavior disorder
and the Luxembourg Parkinson’s Study.

Reducing the number of visits also impacts participant retention,

OBJECTIVE especially in elderly participants. Due to the special needs of some
We introduce Smart Scheduling (SMASCH) system, a web-based participants, both PD studies provide a mobile team to conduct
tool designed for longitudinal clinical studies requiring recurrent examinations at different recruitment hubs across the country,
follow-up visits of the participants. Longitudinal clinical research thanks to a specifically equipped van. MCI-BIOME study is similar
studies are often ran on one-stop shop setups in which participants to the previous 2, taking its participants from NCER-PD and PDP
only need to visit the clinical research facility few times to undertake cohorts.
all required assessments with different specialists in the same place. RBD study started with a large scale survey to determine the inci-
Participants of such studies usually undergo a series of follow-up dence of RBD in the Luxembourgish population and increase the
appointments with a reduced clinical visit frequency (eg, yearly). chances of having a sizable group to assess the link between RBD
SMASCH assists clinical teams on the management of sensitive in- and PD. More recently, CON-VINCE study involved a wider de-
formation, required items for conducting assessments, and schedul- ployment of sampling collection facilities, requiring the development
ing participant appointments. Although it does not provide full of data connectors to bulk import PII from recruited participants
automatic scheduling of individual appointments, it automatically (see Figure 2 in reference3). CON-VINCE study required a much in-
creates follow-up visits for the participants once the previous visit is tensive data and sample collection, with a bi-weekly frequency for
finished. Generally, such visits must be consistently spaced in time the first 2 months (totalling 5 visits).
to prevent misalignment or time lag. In this context, we refer to visit A more recent project includes the EU-funded ORCHESTRA
as the date interval in which all required assessments need to be con- project (https://orchestra-cohort.eu), which aims to establish a
ducted to be considered an individual time-point for the clinical EU-wide cohort based on already available and new cohorts from
studies. both the EU and non-European countries (eg, Congo, Argentina)
including SARS-CoV-2 infected and noninfected individuals of
Background and significance all ages and conditions for a retrospective evaluation and pro-
Marynissen and Demeulemeester1 reviewed multi-appointment spective follow-up. SMASCH roles include resource assignment
scheduling problems in hospitals (MASPHs) and the different meth- (both infrastructure and staff), scheduling, communication, and
odologies available in the literature to find the optimal assessment the management and storage of PII from Luxembourg partici-
paths for patients. Their work shows different MASPH scenarios pants.
such as general hospitals or facilities for diagnostic tests.
In this sense, SMASCH is present in multiple clinical research
and integrated care programs in Luxembourg since 2017, including Data protection and privacy
Dementia Prevention Program (PDP, http://demence.lu); the study Above studies share a common interest to safeguard sensitive infor-
for Mild Cognitive Impairment and gut microbiome (MCI-BIOME); mation and PII, requiring strong data protection to ensure data pri-
and the National Centre of Excellence in Research on Parkinson’s vacy but at the same time the possibility to exchange
disease (NCER-PD, https://parkinson.lu),2 which encompasses the pseudonymized data with scientific collaborators within the frame-
study for REM sleep behavior disorder (RBD, https://rbd.lu) and the work of the General Data Protection Regulation (GDPR) of the
Luxembourg Parkinson’s Study. European Union (EU). Thus, PII and clinical data are managed by 2
Moreover, SMASCH contributed during the recent COVID-19 separate systems. Namely, all pseudonymized clinical data are
National survey for assessing Viral spread by Non-affected Carriers stored in a dedicated REDCap4 instance, while SMASCH is in
(CON-VINCE study)3 organizing appointments and results of charge of securing PII, providing functionalities to track data modi-
SARS-CoV-2 PCRs and prevalence tests, as well as securing personal fications and access permissions (see “Features” section).
identifiable information (PII). Together with the GDPR, national and institutional (eg, hospi-
The challenge for NCER-PD program was to schedule Parkin- tal, clinics) regulations constrain the space of solutions suitable for
son’s disease (PD) study participants in yearly visits, optimizing the the aforementioned purposes. In particular, should any commercial
number of appointments with different specialists from the clinical solution be employed during such studies, it would have to comply
research team to happen in a certain time interval but without be- with GDPR requirements. Importantly, as a result of the GDPR and
ing too intense for the elder participants. For example, clinical data the recent SCHREMS ruling,5 many institutions are reconsidering
and sample(s) collection, analysis and other examinations should their use of cloud solutions and devising new policies, which may re-
happen closely to be considered as a single time-point in the study. sult in ruling out cloud services for sensitive health data. This topic
JAMIA Open, 2022, Vol. 5, No. 2 3

is further discussed in the Supplementary Material. Besides, institu- tion, it ensures restricted access to sensitive information and
tional regulations (especially public institutions) often prohibit the destructive operations. Closely related to the previous feature, prov-
transfer of personal data outside their infrastructure without special enance allows audit trail of critical modifications (eg, marking a par-
agreements. ticipant as deceased) or PII export.
Section 2 of the Supplementary Material provides further infor-
mation on SMASCH alternatives. However, all of them are closed Information export
source cloud-based solutions, with the grand majority of them only Whenever the authorized clinical team exports the list of subjects or
offering paid plans. the appointments to an Excel file, SMASCH records the user and
date of the action in the system log. Of course, such an option can
be disabled or limited to certain members of the clinical team with
MATERIALS AND METHODS authorization to use this feature. This feature enables a user-friendly
SMASCH follows a model view controller (MVC) architecture, a approach to export information for further use such as importing it
popular software design pattern that divides the program logic to other medical systems or statistical tools for data analysis. The
into 3 interconnected elements. MVC is often enforced by several clinical team can select which fields to export, deciding whether to
web frameworks. In SMASCH, the model is an abstraction of the export PII or not. Very often, strict regulations call for different
database model and manages the application state and data- approaches involving manually supervised import/export of
related logic between the program and the relational database information between systems to prevent unnecessary access to
management system (RDBMS). The view comprises the presenta- sensitive patient/participant data in electronic health record (EHR)
tion layer that shows the application state to the user, managing databases.
the user interface logic of SMASCH. Finally, the controllers inter-
face the model and view components to handle the user actions, Notifications
retrieving data from the model and rendering the response A notification menu (see Figure 2B) warns the users about incom-
through the views. plete tasks (eg, missing appointments, unfinished visits) and reminds
SMASCH is written in Python 3, using Django web framework them about approaching appointments and visits in advance.
for the back-end (ie, model and controllers). For the front-end
functionalities comprising the views, SMASCH employs Jinja2, a Redcap consistency checks
web template engine providing evaluation of Python-like expres- To ensure record consistency between separated systems, SMASCH
sions to enrich the rendered views. Additionally, several JavaScript implements a series of checks on demographic data (eg, birth date,
libraries (eg, jQuery) drive the user interface. SMASCH can be vital status) using a restricted REDCap token with limited access,
accessed with browsers supporting ECMAScript 2015 (see Table 1) thus unable to access medical or clinical information. Should a mis-
and is licensed under the open-source license AGPLv3. SMASCH match occur between systems, a warning would be issued in the no-
code is openly available in the GitLab server from the Luxem- tification menu.
bourg Centre for Systems Biomedicine at https://doi.org/10.17881/
8wvb-mt36.
Sample collection summary
Paginated views and application programming interfaces
Different appointment types (eg, sample collection) can be registered
(APIs) allow managing very large cohorts with scant machine
with their corresponding items (eg, saliva kits). With this informa-
resources. The minimum requirements (see ”Deployment” section)
tion SMASCH sends regular e-mails to biobanks and relevant pro-
allow SMASCH to handle cohort sizes up to 100K participants
viders to help forecasting stock requirements for sample collection
without issues. Bigger cohorts can be easily handled with increased
kits or organizing pickup of collected samples.
resources.

User location awareness

RESULTS For a tailored user experience and faster usage, members of the clini-
cal team can select which locations (eg, clinics, hospitals, biobanks)
Close collaboration with clinical teams helped developing features they work on, filtering the daily planning (see Figure 2D) to the
aiming to reduce times of recurrent and time-demanding tasks appointments happening on their locations.
such as generation of mail letters or finding key information.
Below we describe the available features and describe the deploy-
Mail templates
ment procedure.
To facilitate output of recurring documents, SMASCH supports
uploading document templates in different languages. SMASCH can
Features register the participant spoken languages, filtering the templates ac-
Permissions and provenance cordingly whenever a document is issued and automatically gener-
SMASCH provides flexible and fine grained user permission man- ates the appointment letters and visit vouchers in the language
agement of the system features. Together with 2-factor authentica- spoken by the study participant.

Table 1. ECMAScript 2015—ES6 browser support

Browser Chrome 58 Edge 14 Firefox 54 Safari 10 Opera 55

Support since January 2017 August 2016 March 2017 July 2016 August 2018
4 JAMIA Open, 2022, Vol. 5, No. 2

Custom fields ments on cloud, as such requirement is independent of SMASCH im-

Adaptability is an essential requirement for rapid evolving scenarios. plementation.
SMASCH supports custom fields of different types (eg, text, dates)
to further describe a subject, visit, or appointment.
DISCUSSION
Stable and well-maintained software is key for the careful manage-
Extract, transform and load
ment of long-term projects. In this sense, SMASCH provides a solu-
Extract, transform and load pipelines (ETLs) enable the implemen-
tion for MASPHs in the context of longitudinal clinical studies. The
tation of connectors both to and from other systems. Currently,
MASPH is defined as the problem of scheduling patients who need
such ETL procedures can pull information from different survey sys-
appointments on a subset of hospital or clinical resources by gather-
tems such as RedCap or Alchemer (https://www.alchemer.com/),
ing all stakeholders in the scheduling process.1 In this sense, the
helping to import and update subject and appointment information.
scheduling process comprises both combination appointments and
Currently, SMASCH ETLs are developed as custom modules in
appointment series. The former entails multiple appointments for a
Python.
single patient planned on the same day so that a patient requires
Such ETLs enable as well interfacing systems storing EHRs.
fewer visits. The latter consists of multiple (recurring) appointments
However, in practice, very often strict privacy regulations call for
that may span several weeks or months.6
an intermediary to import/export records in CSV/TSV format in a
In turn, the long-term needs of clinical studies justify the employ-
supervised manner. Such measures are taken to avoid unnecessary
ment of well-maintained methods, frameworks, and libraries such as
access to sensitive data by other tools and limit the space for vul-
Python and Django which have a proven track of continuous
nerabilities and potential data breaches. Nevertheless, such politi-
updates to rely on. At the same time, the high-level abstraction pro-
cal decisions are beyond our control, but condition how SMASCH
vided by Python and the adaptability of Djangos MVC framework
should interface with other systems, which often implies not
enable a fast and iterative software development life-cycle that is
benefiting from the potential automatic and unsupervised capaci-
able to cope with the rapidly evolving needs of the clinical team, re-
ties of certain systems in favor of other nonfunctional require-
ducing response times to new scenarios. The views provided
ments such as security and privacy. Notwithstanding, SMASCH is
by SMASCH quickly summarize the relevant information, and at
looking forward to improving this feature (see “Future work”
the same time, are properly split to enable fine-grained permission
section).
management.
The Supplementary Material and SMASCH documentation pro-
Similarly, SMASCH modular design allows interoperability with
vide further information regarding the setting of such ETLs.
other systems through connectors, enabling its participation in
multi-institutional projects that require importing and exporting in-
formation from and to different systems (see “Extract, transform
Deployment
and load” section). For instance, CON-VINCE study required the
Our production deployments of SMASCH commonly employ Linux
development of data connectors to bulk import PII from recruited
machines, using a dedicated encrypted partition for PostgreSQL to
participants. Figure 2 from reference3 depicts the role of SMASCH
store the system’s database. However, these deployment choices are
in the data flow from CON-VINCE. In this case, the PII from the
not mandatory and SMASCH is able to work with other databases
recruited participants was sent from the survey company (TNS-Ilres)
as well. The clinical studies where SMASCH is involved are long-
to SMASCH, where it is securely stored and accessed only by the
term projects requiring equally long-term software support. For this
corresponding clinical team.
reason the development team of SMASCH maintains its unit test
coverage around 90%. Installation can be done via Debian package
as described in Figure 1. Alternative installation methods (eg, Future work
Docker) are also available. For SMASCH, a virtual machine with a SMASCH future is open to improve the usability and interopera-
single CPU core and 2GB of RAM are the minimum requirements. bility with other services. Upcoming features include SMS
To secure access to SMASCH, HTTPS should be enforced by the reminders for participant appointments and the support for Light-
web server (eg, NginX, Apache). weight Directory Access Protocol (LDAP) to facilitate multi-
In contrast to cloud-based only solutions, the provision of project access to the workers. Particularly, LDAP authentication is
SMASCH source code and its corresponding Docker container facil- already available in the next release candidate. Additionally, fur-
itates the deployment of SMASCH in any infrastructure or institu- ther support for ETL procedures is currently under development
tional data center (eg, hospitals or clinics) that may require the use aiming to increase interoperability with other data sources
of SMASCH. And at the same time, it allows for remote deploy- through a dedicated ETL API.

Figure 1. Installation of SMASCH in Debian based Linux systems.

JAMIA Open, 2022, Vol. 5, No. 2 5

Figure 2. SMASCH interface and functionalities. Image (A) displays the appointment calendar with the initials indicating workers availability. Appointment colors
indicate different appointment types. (B) Notification menu showing different warnings and reminders. (C) List of the study subjects (ie, participants) showing in-
formation and visit status. (D) Daily planning with the participant assessments on top which can be dragged and dropped to the calendar. Workers are sorted
from left to right based on their availability. On mouse-over, the workers and participant names show a tool-tip box indicating the spoken languages to help
choosing a suitable worker. Sidebar menu seen in A has been cropped out from C and D for the sake of clarity and space.

CONCLUSION ule, increasing resource idleness. Conversely, overbooking can im-

prove participant access times but at the cost of increasing waiting
In this application note, we presented Smart Scheduling, a web-
times and staff overtime.
based multi-appointment scheduling system designed to handle
Although SMASCH did not have the opportunity to be specifi-
multi-appointment scheduling problems and management tasks
cally evaluated in these regards, SMASCH has proven useful in mul-
characteristic from longitudinal clinical studies.
tiple clinical research studies in which it has been actively involved
In general, scheduling systems can help optimizing a number of
since 2017. SMASCH features, such as its daily planning view, help
metrics that ultimately affect the success of clinical studies. Schedul-
the clinical team ponder the aforementioned factors and spot over-
ing systems impact the number of participants per consultation ses-
loaded or idle agendas for a better plan of the visits, appointments,
sion which affects participant waiting times and staff overtime.
and assessments. The flexible design of SMASCH allows for a fast
Similarly, the length of the appointment interval and the number of
and iterative life cycle with high adaptability to new scenarios re-
participants per appointment slot affect resource idle time and wait-
quiring rapid deployments, such as the recent CON-VINCE pro-
ing times. Other metrics such as participant overbooking must be
gram. Consequently, SMASCH is continuously updated to fit the
weighed too. For instance, “no-shows” may cause gaps in the sched-
6 JAMIA Open, 2022, Vol. 5, No. 2

needs of the respective clinical research facility teams, who im- ACKNOWLEDGMENTS
proved their response time and throughput thanks to a well-defined
We thank the clinical teams for their help and support. Special
set of features coupled with a user-friendly interface. We hope
thanks to: Claire Pauly (PRC), Anne Kaysen (PDP), Valerie
SMASCH features help to sustain high participant retention and as-
Schröeder (PDP), Guilherme Ramos Meyers (CON-VINCE), Clar-
sist clinical teams in the successful development of longitudinal clini-
issa Gomes (NCER-PD and CON-VINCE), Geraldine Hipp
cal studies.
(NCER-PD), Lars Geffers (NCER), and Lukas Pavelka (Centre Hos-
Finally, recent regulations such as the EU’s GDPR call for strict
pitalier de Luxembourg, NCER-PD), Wei GU, Christophe Trefois,
data accountability, which together with national and institutional
Pinar Alper. The Reproducible Research Results (R3) team of the
regulations requires further measures to ensure data privacy, espe-
LCSB is acknowledged for supporting the project.
cially in health-related settings. These regulations rule out the possi-
bility of cloud-based solutions and hinder outsourcing such
responsibilities. To date, we believe SMASCH is the only free and
open-source scheduling system for clinical studies able to satisfy the
above needs. CONFLICT OF INTEREST STATEMENT
None declared.

FUNDING
REFERENCES
This work was supported by the Luxembourg National Research
1. Marynissen J, Demeulemeester E. Literature review on multi-appointment
Fund (FNR) within the National Centre of Excellence in Research
scheduling problems in hospitals. Eur J Oper Res 2019; 272 (2): 407–19.
on Parkinsons disease with grant number: (NCER-PD [FNR/
2. Hipp G, Vaillant M, Diederich NJ, et al. The Luxembourg Parkinson’s
NCER13/BM/11264123]). Study: a comprehensive approach for stratification and early diagnosis.
Front Aging Neurosci 2018; 10: 326.
3. Snoeck CJ, Vaillant M, Abdelrahman T, et al. Prevalence of SARS-COV-2
infection in the Luxembourgish population: the Con-Vince Study. medRxiv
AUTHOR CONTRIBUTIONS 2020. doi:10.1101/2020.05.11.20092916.
VS, RS, and RK conceived the idea, and collected the functional and 4. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research
technical requirements. PG, JL, CV, VG, and PM developed electronic data capture (REDCap)—a metadata-driven methodology and
SMASCH and ES conducted software testing. VS supervised the workflow process for providing translational research informatics support.
J Biomed Inform 2009; 42 (2): 377–81.
project. CV, PG, CP, VS, RS, and RK wrote and edited the manu-
5. C-311/18, data protection commr v. facebook ire. ltd. & schrems; 2020.
script. All authors reviewed the manuscript.
[schrems ii], paras. 105, 162, at 31, 40. https://curia.europa.eu/juris/docu-
ment/document.jsf?docid=228677&text=&doclang=EN&pageIndex=0&
cid=1175857. Accessed May 13, 2022.
6. Hulshof PJ, Kortbeek N, Boucherie RJ, Hans EW, Bakker PJ. Taxonomic
SUPPLEMENTARY MATERIAL
classification of planning decisions in health care: a structured review of the
Supplementary material is available at JAMIA Open online. state of the art in OR/MS. Health Syst 2012; 1 (2): 129–75.