Article published online: 2021-09-03

280
© 2021 IMIA and Georg Thieme Verlag KG

Public Health and Epidemiology Informatics:

Recent Research Trends
Gayo Diallo1, Georgeta Bordea1, 2, Section Editors for the IMIA Yearbook Section on Public
Health and Epidemiology Informatics
1
INRIA SISTM, Team ERIAS - INSERM Bordeaux Population Health Research Center, Univ.
Bordeaux, Bordeaux, France
2
Team ERIAS - INSERM BPH Research Center & LaBRI UMR 5800, Univ. Bordeaux, Bordeaux, France

Summary 1 Introduction Subject Headings (MeSH) descriptors was

used to retrieve relevant studies.The queries
Objectives:To introduce and analyse current trends in Public
The trend observed in recent years regarding targeted public health or epidemiological
Health and Epidemiology Informatics.
the increased use of digital technology in journal articles over the year 2020 that in-
Methods: PubMed search of 2020 literature on public health
healthcare and the recent evolution of ar- cluded medical informatics topics.
and epidemiology informatics was conducted and all retrieved
tificial intelligence (AI) and deep learning Returned references addressing topics
references were reviewed by the two section editors. Then, 15
technologies continue to have a positive of the other sections of the Yearbook, e.g.,
candidate best papers were selected among the 920 references.
impact on population-based studies. those related specifically to pandemics and
These papers were then peer-reviewed by the two section editors,
A review of research published last year Corona Virus (COVID), were excluded
two chief editors, and external reviewers, including at least two
showed that the increasing availability of from our search. Further, a cross-check
senior faculty, to allow the Editorial Committee of the 2021 Inter-
data from either the internet and social net- is performed with the 2021 International
national Medical Informatics Association (IMIA) Yearbook to make
works or more traditional health data sources Medical Informatics Association (IMIA)
an informed decision regarding the selection of the best papers.
such as Electronic Health Records (EHRs) is Yearbook Special Topic “Managing Pan-
Results: Among the 920 references retrieved from PubMed, four
were suggested as best papers and the first three were finally
fueling research on topics covering statistics, demics with Health Informatics: successes
selected. The fourth paper was excluded because of reproduc-
health informatics, and epidemiology [1]. and challenges” section for any overlaps in
ibility issues. The first best paper is a very public health focused
These topics continue to be key components themes that are not obvious.
paper with health informatics and biostatistics methods applied
of Public Health Data Science. The study was performed at the beginning
to stratify patients within a cohort in order to identify those at risk
The current synopsis reviews the sci- of February 2021, and the search returned a
of suicide; the second paper describes the use of a randomized
entific literature published in 2020 in the total of 920 references. Articles were sepa-
design to test the likely impact of fear-based messages, with domain of medical informatics with a rately reviewed by the two section editors
and without empowering self-management elements, on patient focus on the sub-field of public health and considering in the process the number of
consultations or antibiotic requests for influenza-like illnesses. epidemiology. The synopsis aims to identify citations of each reference. Articles were
The third selected paper evaluates the perception among com- new topics and trends within public health first classified into three categories: keep,
munities of routine use of Whole Genome Sequencing and Big as compared to previous years. Our review discard, or pending using the BibReview tool
Data technologies to capture more detailed and specific personal further describes the selection process fol- [2]. Then, the two lists of references were
information. lowed to choose the best papers that were merged, yielding 96 references that were
Conclusions: The findings from the three studies suggest that published in 2020. To this end, the section retained by at least one reviewer or classi-
using Public Health and Epidemiology Informatics methods could editors coordinated the reviewing process fied as “pending” by both of them. The two
leverage, when combined with Deep Learning, early interventions of the most representative research works, section editors jointly reviewed 96 references
and appropriate treatments to mitigate suicide risk. Further, they based on quality and originality. Special Topic section was performed in order
also demonstrate that well informing and empowering patients to double-check if any overlap between the
could help them to be involved more in their care process. two sections exists.
All of these papers were then peer-re-
Keywords 2 Methods viewed by editors and external reviewers.
Each paper was reviewed by at least five
Public health, epidemiology informatics, IMIA Yearbook
A comprehensive literature search was per- reviewers. Four papers were suggested at
formed using PubMed/MEDLINE database the end of this process. However, the fourth
Yearb Med Inform 2021:280-2
from the National Center for Biotechnology paper [3], although interesting for the field,
http://dx.doi.org/10.1055/s-0041-1726530
Information (NCBI). A large set of Medical was removed due to reproducibility issue.

IMIA Yearbook of Medical Informatics 2021

 281
Public Health and Epidemiology Informatics: Recent Research Trends

Indeed, the described study use a propri- Table 1 Best paper selection of articles for the IMIA Yearbook of Medical Informatics 2021 in the section ‘Public Health and Epidemiology
etary tool and the dataset is not available Informatics'. The articles are listed in alphabetical order of the first author’s surname.
for reuse. Finally, three papers, respectively,
[4], [5] and [6] were selected as best papers
by the Yearbook Editorial Committee (Table Section
Public Health and Epidemiology Informatics
1). A content summary of these selected
papers can be found in the appendix of  Degeling C, Carter SM, van Oijen AM McAnulty J, Sintchenko V, Braunack-Mayer A, Yarwood T, Johnson J, Gilbert GL.
this synopsis. Community perspectives on the benefits and risks of technologically enhanced communicable disease surveillance systems: a
report on four community juries. BMC Med Ethics 2020;21(1):31.
 Roope LSJ, Tonkin-Crine S, Herd N, Michie S, Pouwels KB, Castro-Sanchez E, Sallis A, Hopkins S, Robotham JV, Crook DW,
Peto T, Peters M, Butler CC, Walker AS, Wordsworth S. Reducing expectations for antibiotics in primary care: a randomised
experiment to test the response to fear-based messages about antimicrobial resistance. BMC Med 2020;18(1):110.
3 Results  Zheng L, Wang O, Hao S, Ye C, Liu M, Xia M, Sabo AL, Markovic L, Stearns F, Kanov L, Sylvester KL, Widen R, McElhinney
DB, Zhang W, Liao J, Ling XB. Development of an early-warning system for high-risk patients for suicide attempt using deep
Beyond the high-quality of research pub- learning and electronic health records. Transl Psychiatry 2020;10(1):72.
lications and their contributions, several
research themes were prioritized based on
their potential impact for future research.
These themes cover diverse aspects in-
cluding the interpretation of deep-learning
approaches, assessing benefits and risks of 4 Conclusion Methods Inf Med 2015;54(2):135–44.
3. Eletreby R, Zhuang Y, Carley KM, Yagan O, Poor
using linked data in disease surveillance, Our analysis confirmed once again the gener- HV. The effects of evolutionary adaptations on
spreading processes in complex networks. Proc
open access aspects of epidemiology data- al trend of increased interest in Public Health Natl Acad Sci U S A 2020;117(11):5664–70.
sets, addressing gender bias in medical and Epidemiology Informatics (PHEI) in re- 4. Zheng L, Wang O, Hao S, Ye C, Liu M, Xia M, et
informatics, and optimisation of messages cent years, even beyond the specific context al. Development of an early-warning system for
in online public health campaigns. These are of an ongoing pandemic. Access to big data, high- risk patients for suicide attempt using deep
still open but pressing research questions trust in artificial intelligence, privacy, and learning and electronic health records. Transl
Psychiatry 2020;10(1):72.
that bear witness to sometimes conflicting minority bias are some of the most pressing 5. Roope L, Tonkin-Crine S, Herd N, Michie S,
goals in public health informatics that are research questions in this area with a shift to Pouwels K, Castro-Sànchez E, et al. Reducing
related to access to Big Data, trust in arti- online investigation that empowers patients expectations for antibiotics in primary care: a
ficial intelligence, privacy, minority bias, to be more involved in their care process. randomised experiment to test the response to fear-
and a general shift to online investigation. based messages about antimicrobial resistance.
BMC Med 2020;18(1):110.
Overall, the findings from the three Acknowledgements 6. Degeling C, Carter S, Oijen A, Mc Anulty J, Sin-
selected studies suggest that using Public The PHEI section editors would like to thank tchenko V, Braunack-Mayer A, et al. Community
Health and Epidemiology Informatics meth- the reviewers for their valuable contribution perspectives on the benefits and risks of technologi-
ods could leverage, when combined with in evaluating and selecting the best papers cally enhanced communicable disease surveillance
Deep Learning, early interventions and ap- for this section of the 2021 IMIA Yearbook. systems: A report on four community juries. BMC
Med Ethics 2020;21(1):31.
propriate treatments to mitigate suicide risk.
Further, they also demonstrate that when Correspondence to:
people are well-informed, here using a jury References Gayo Diallo
proxy [6], they are likely to support routine 1. Cossin S, Rodolphe Thiébaut R. Public Health INRIA SISTM, Team ERIAS
collection, linkage and use of administrative and Epidemiology Informatics: Recent Research INSERM Bordeaux
and pathogenomic data for the purposes of Trends Moving toward Public Health Data Science. Population Health Research Center
Yearb Med Inform 2020: 29(1):231– 4. Univ. Bordeaux
public health research. In addition, empow- 2. Lamy JB, Séroussi B, Griffon N, Kerdelhué G, 146, rue Léo Saignat
ering patients could help them to be involved Jaulent M-C, Bouaud J. Toward a formalization F-33000 Bordeaux, France
more in their care process. of the process to select imia yearbook best papers. E-mail: [email protected]

IMIA Yearbook of Medical Informatics 2021

282
Diallo et al.

Appendix: Content Summary diagnosed mental health conditions or

pain, to have suicide attempts, treated by
Degeling C, Carter SM, van Oijen AM
McAnulty J, Sintchenko V, Braunack-Mayer
of Selected Best Papers 2021 psychotropic medications, and have open A, Yarwood T, Johnson J, Gilbert GL
IMIA Yearbook Section Public wounds or injuries due to unspecific rea-
sons. The findings of the study enable early Community perspectives on the benefits
Health and Epidemiology interventions and appropriate treatments to and risks of technologically enhanced
communicable disease surveillance
mitigate suicide risk.
Informatics systems: a report on four community juries
Roope LSJ, Tonkin-Crine S, Herd N, Michie
Zheng L, Wang O, Hao S, Ye C, Liu M, Xia BMC Med Ethics 2020;21(1):31
S, Pouwels KB, Castro-Sanchez E, Sallis A,
M, Sabo AL, Markovic L, Stearns F, Kanov Hopkins S, Robotham JV, Crook DW, Peto T, The paper focuses on assessing how routine-
L, Sylvester KL, Widen R, McElhinney DB, Peters M, Butler CC, Walker AS, Wordsworth S ly using Whole Genome Sequencing (WGS)
Zhang W, Liao J, Ling XB Reducing expectations for antibiotics in
and Big Data technologies to capture more
detailed and specific personal information
Development of an early-warning system primary care: a randomised experiment to could be perceived among communities in
for high-risk patients for suicide attempt test the response to fear-based messages two demographically different Sydney mu-
using deep learning and electronic health about antimicrobial resistance nicipalities and two regional cities in New
records BMC Med 2020;18(1):110 South Wales, Australia (western Sydney,
Transl Psychiatry 2020;10(1):72 Wollongong, Tamworth, eastern Sydney).
The aim of this study is to test the likely Four community juries were created by
The paper of Zheng et al. reports a study impact of fear-based messages, with and recruiting participants at each study site
focusing on tackling suicide attempt. It is without empowering self-management ele- by an independent professional research
a public health focused paper combining ments, on patient consultations or antibiotic service. The aim was to elicite the views
health informatics and biostatistics methods requests for influenza-like illnesses, using of well-informed community members on
and which are applied to stratify patients a randomised design. To do so, they relied the acceptability and legitimacy of making
within a cohort in order to identify those on the use of an on-line survey of adult pathogen WGS and linked administrative
at risk of suicide. They developed an ear- members of the UK general public. They data available for public health research
ly-warning system for high-risk suicide proceed to the randomisation of patients, using this information in concert with data
attempt patients through the design and to receive one of three different messages linkage and machine learning to enhance
implementation of a population-based risk about antibiotics and Antimicrobial Resis- communicable disease surveillance sys-
stratification surveillance system. A Deep tance. As of results, they have been able to tems. Participants across all four events
Neural Network (DNN) model was trained demonstrate that empowering patient could strongly supported the introduction of data
for the prediction, stratification, and cali- help to better involve them in their care pro- linkage and pathogenomics to public health
bration. Then a Local Interpretable Mod- cess and make a better decision regarding research under current research governance
el-agnostic Explanations algorithm was Antibiotics related issues. The study con- structures. This study demonstrates that
utilised to interpret the risk stratification cludes that while fear-only messages could when public is well informed, here a jury
results. As of result, a total of 117 features be effective in public campaigns to reduce proxy is used, they are likely to support
were significant in the predictive model; inappropriate an- tibiotic use, they should routine collection, linkage and use of ad-
the DNN on the EHR-based data, enabled be combined with messages empowering ministrative and pathogenomics data for the
finding that suicide attempts patients were patients to self-manage symptoms effec- purposes of public health research.
more likely to be gin: age groups of 6–54, tively without antibiotics.

IMIA Yearbook of Medical Informatics 2021