Prototypical Implementation of a Decision-supporting System for

Operative Breast Cancer Therapy

Michael Dück

and Eberhard Beck

Department of Computer Science and Media, University of Applied Science Brandenburg,

Magdeburger Street 50, Brandenburg, Germany

Keywords: Decision Support, Breast Cancer, e-Health, BPMN.

Abstract: Based on the current edition of the German guideline on Screening, Diagnosis, Treatment and Follow-up of

breast cancer, we created a patient journey modelled in BPMN (Business Process Model and Notation V2)

serving as template for the development of a patient centered decision support system. This approach resulted

in two prototypical devices represented by a web-based information platform and a mobile application,

intended to support the decision support at the point of care. These early prototypes were discussed with a

clinical expert and the members of a regional breast cancer self-help group. The information gained by this

approach will be integrated in the further user centered design of the devices.

1 INTRODUCTION

The advent of computer assisted clinical decision

support goes back to the early 1960s (Shortliffe,

2018). Since then scientists and clinicians have

undertaken numerous efforts to create various

systems in order to improve the quality of clinical

decisions, enhance their transparency and increase the

number of guideline conform decisions, resulting not

only in patient centered decisions but also enabling

patients to engage in the process of shared decision

making (Middleton 2016, Beeler, 2014). Despite

these efforts, decision support systems are failing to

be introduced into daily routine for a number of

reasons. Among other reasons, a suspected negative

influence on the physician-patient relationship, the

extra time spent to utilize the system or that the

system could not be integrated into the routine

workflow, were named (Kilsdonk, 2017). On the

other hand, factors in favor of using computerized

decision support systems are seen in systems that e.g.

fit with routine care and provide recommendations at

the point of care (Kilsdonk, 2017). In order to address

at least some of these requirements we analyzed the

current guideline on Screening, Diagnosis, Treatment

and Follow-up of breast cancer of the German Cancer

Society (Wöckel, 2018). Based on our results we

https://informatik.th-brandenburg.de/

developed an early prototype of a clinical decision

support system, which should not only serve

clinicians but also support patients and their relatives.

2 METHODS

On the basis of the German S3 guideline on

Screening, Diagnosis, Treatment and Follow-up of

Breast Cancer, a patient journey as a process model

for breast cancer was extracted and modelled in

BPMN (Business Process and Model Version 2.0) as

described previously (Andrzejewski, 2015,

Andrzejewski, 2017). This resulted in the definition

of several important decision nodes, which were then

examined for their specific, decision relevant

parameters. These factors consisted of the Tumor

seize, axillary lymph nodes involved (N-status),

distant metastasis present or absent (M-status), the

tumor grading, the (clinical) breast to tumor relation,

estrogen and progesterone receptor status, the HER-2

status, the Ki-67 status and finally whether the patient

was pre- or postmenopausal (Fig. 1). In contrast to our

previous work, Ki-67 was newly introduced as

relevant decision factor in the 2018 edition of the

guideline, which forced us to redesign our process

models. The aforementioned factors were then used

Dück, M. and Beck, E.

Prototypical Implementation of a Decision-supporting System for Operative Breast Cancer Therapy.

DOI: 10.5220/0009164506830686

In Proceedings of the 13th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2020) - Volume 5: HEALTHINF, pages 683-686

ISBN: 978-989-758-398-8; ISSN: 2184-4305

683

to create a decision matrix, which however resulted in

more than 140,000 individual combinations.

Figure 1: Factors in the treatment of breast cancer.

We thus restricted our preliminary efforts to re-

generate our computer-based decision support system

to the factors necessary to make a transparent

decision concerning surgical treatment only. Tumor

size and the clinical breast to tumor-relation were

identified as the two essential factors concerning

decisions on surgical treatment. In the following short

matrix, the possible outcomes for four of the eight

different possible combinations are shown.

Table 1: Decision matrix for the surgical treatment.

T-Status T1 T2 T3 T4

BTR G G G G

Neo - - Op

Treat BCT BCT BCT

BCT /

(Mas

)

Radio + + + +

Opt

Adju + + + +

The abbreviations in the table in order of appearance are: T-

Status: the size of the tumour;

BTR: Brest-Tumor-Relation, the size of the breast in relation to

the tumour;

G: good or rather favourable;

Neo: neoadjuvant system therapy;

Opt: optional;

Treat: surgical treatment;

BCT: breast conserving surgical

therapy;

Mast: mastectomy;

Radio: radio therapy;

Adju: adjuvant system therapy.

Based on these combinations and the resulting

actions to be recommended, we created a prototype of

a rule-based decision support system. A decision

support system is defined by Keen und Scott-Morton

(Keen, 1978) as:

"Decision Support Systems (DSS) represent a

point of view on the role of the computer in the

management decision-making process. Decision

support implies the use of computers to:

• Assist (managers) in their decision processes in

semi structured tasks

• Support, rather than replace, (managerial)

judgement

• Improve the effectiveness of decision-making

rather than its efficiency".

Our decision support system is aimed to assist

patients in a complex process of medical decisions by

delivering recommendations with very little input.

This rule-based system was primarily implemented as

an app, which advocates the respective operative

intervention, once the relevant factors are entered.

The app was built with the MIT App Inventor as a fast

prototype and to use with the treating doctor for a

shared decision making in a consultation meeting.

Also, a website with the same functions, built in

JavaScript and HTML was designed. After the

development of these prototypes, they were presented

to a health care professional and a former breast

cancer patient for a preliminary evaluation of the

system. As a second step, we presented the

application and the website at the annual meeting of

the regional cancer self-help group

Berlin/Brandenburg.

3 RESULTS

To generate the best decision support for breast

cancer patients, the precise diagnosis and all relevant

clinical and personal factors which may affect the

choice of treatment should be known. A newly

diagnosed patient, who in our view is most likely in

need of a decision support system, will not

necessarily be aware of all these data. By focusing on

the factors mentioned above which were derived from

the S3 guideline we created the prototype of an

evidence-based decision support platform. Our aim

was to test if it was possible to translate the medical

guideline into a digital rule-based system, which is

able to create relevant suggestions based on the

information given by a patient. In Fig. 2 the patient

journey, starting with the decision conflict between a

mastectomy or a breast-conserving operation as

initial surgical step, is visualized. In the first instance

the tumor size is used to define the disease as early

stage cancer or locally advanced tumor. Combined

with the factor “breast to tumor relation”, this results

either in a strait forward decision concerning the type

HEALTHINF 2020 - 13th International Conference on Health Informatics

684

of surgical intervention or raises the question of a

primary (neo-adjuvant) Chemotherapy. The latter

could be applied in cases of locally advanced tumors

or in early breast cancer with an unfavorable breast to

tumor relation.

These respective decisions are based on the

decision matrix shown in Tab. 1. The further process

model then describes the consecutive steps of

adjuvant chemo and/or radiation therapy once the

operative intervention was carried out.

Figure 2: Process of the breast cancer treatment.

Figure 3: Screenshot of the app.

Most of the modelled tasks like the “perform

breast-conserving therapy” are further detailed in

various sub-processes, which however for the sake of

clarity are not shown.

Based on these models and the previously

mentioned decision matrix, two prototypes of a

decision support system were developed. Figure 3

shows the app. The app was primarily designed in

order to serve as a mobile device available for the

evidence-based decision support at the point of care.

4 EVALUATION

In order to improve these first iterations of the

applications, two interviews were conducted with a

medical expert and a person affected by breast cancer.

The recommendations extracted from these

interviews were applied to further improve the

prototypes before we presented them on the annual

meeting of the regional cancer self-help group

Berlin/Brandenburg. While the website was assessed

as appropriate for making an informed decision, the

app however raised major concerns. The approach to

have a fast and mobile decision support system at

hand, was viewed as a useful development. However,

it was stated, that without a consultation with the

treatment overseeing physician or another health care

professional, the information given by the system

could lead to confusion and insecurities, which would

counteract to the expected result.

5 CONCLUSIONS

Based on the various tasks and decision nodes

extracted from the S3 guideline on Diagnosis,

Treatment and Aftercare of Breast Cancer we were

able to map the complete clinical process, starting

with the first suspicious findings and ending with the

completion of adjuvant treatment utilizing BPMN.

The complexity of the information provided in the

guideline, however, constrained us in the first

instance to restrict our efforts in developing a

computer-based decision support system to the tasks

related to the decisions concerning the operative

intervention. Utilizing the TNM classification and the

breast-to-tumor relation, we created a decision

matrix. Based on these rather preliminary results, two

applications were developed in parallel, a website that

serves as a breast cancer information platform and an

app that represents only the decision matrix. The

website contains not only information about breast

Prototypical Implementation of a Decision-supporting System for Operative Breast Cancer Therapy

685

cancer, but also user-controlled information paths that

allow patients to navigate freely according to their

individual information needs and thus receiving

specific, patient centered information. These

pathways lead the users to the decision matrix in

which they can enter their individual clinical data

resulting in a suitable recommendation.

In summary, we developed two prototypical

systems, which were designed to support the

decision-making process in breast cancer. These

prototypes, however, will have to be tested by

independent experts on the basis of a larger number

of realistic case reports. By applying the best

available evidence for the diagnosis and treatment of

breast cancer, represented by the S3 guideline we are

convinced, that our systems could not only help to

increase the number of guideline conform decisions

in the near future, but could also help to improve the

patients knowledge and thus assist in shared decision

making.

REFERENCES

Shortliffe EH, Sepúlveda MJ. Clinical decision support in

the era of artificial intelligence. JAMA. (2018) Dec

4;320(21):2199-2200. doi: 10.1001/jama.2018.17163.

Middleton B, Sittzig DF, Wright A. Clinical decision

support: a 25 year retrospective and a 25 year vision.

Yearb Med Inform (2016) Suppl. S103 -16,

http://dx.doi.org/10.15265/IYS-2016-s034.

Beeler PE, Bates DW, Hug BL. Clinical decision support

systems. Swiss Med Wkly (2014) 144:w14073,

doi:10.4414/smw.2014.14073.

Kilsdonk E, Peute LWP, Jaspers WM. Factors influencing

implementation success of guideline-based clinical

decision support systems: A systematic review and gaps

analysis. Int J Med Inform. (2017) Feb;98:56-64. doi:

10.1016/j.ijmedinf.2016.12.001. Epub 2016 Dec 5.

Wöckel A, Albert S, Janni W, Scharl ., Kreienberg R,

Stüber T. The screening, diagnosis, treatment and

follow-up of breast cancer. Dtsch Arztebl. Int. (2018)

115 (18) 316 – 323. doi: 10.3238/arztebl.2018.0316.

Andrzejewski D, Tetzlaff L, Beck E, Haeusler N. The

transparent representation of medical decision

structures based on the example of breast cancer

treatment, Conference paper. International Conference

on Health Informatics (BIOSTEC), 2015. Lissabon.

Andrzejewski D, Breitschwerdt R, Fellmann M, Beck E,

Supporting breast cancer decisions using formalized

guidelines and experts decision patterns: initial

prototype and evaluation. HealthInfSciSyst, (2017)

5(1):12. doi: 10.1007/s13755-017-0035-8.

Keen P, Scott Morton M. Decision Support Systems: An

Organizational Perspective. Addison-Wesley Inc.

(1978) https://link.springer.com/chapter/10.1007%2

F978-3-663-09143-1_9

HEALTHINF 2020 - 13th International Conference on Health Informatics

686