COLLABORATIVE IT PLATFORM FOR RARE DISEASES

Dragan Jankovic, Tatjana Stankovic

University of Nis, Faculty of Electronic Engineering, Aleksandra Medvedeva 14, 18000 Nis, Serbia

Branimir Todorovic

University of Nis, Faculty of Science and Mathematics, Visegradska 33, 18000 Nis, Serbia

Keywords: Rare diseases, Collaborative IT platform, Web services, Data mining, OLAP, Medical data analyses,

Electronic healthcare record.

Abstract: For a long time the rare diseases have not been in the "focus" of pharmaceutical companies and research

because of potentially lower wages and the fact that very few institutions have a representative set of data

necessary for quality research. Unfortunately, patients suffering from rare diseases are left in the margins of

many societies, their drugs are usually not on the “positive lists” of insurance organisations and their price is

extremely high. The number of rare diseases is between 6000 and 8000 and the estimated number of cases is

about 5%, i.e. about 250 million. This paper presents Collaborative IT platform model for rare diseases by

reviewing four important aspects: creating a national register of people suffering from rare diseases that can

potentially grow into an international; establishment of a central repository for rare diseases with a

collections of medical data characteristic for rare diseases, with modern data analysis tools in order to create

better conditions for scientific research in the field of rare diseases, where some tools would be oriented to

help doctors to more easily and with less cost came to proper diagnosis; improving living conditions and

treatment of patients by forming a set of virtual patient’s associations to exchange experiences and find

useful information; to create conditions for better education of medical workers and patients. The proposed

platform is the subject of the project that we apply to the call for proposals of the Ministry of Science and

Technological Development (MSc&TD) in the Republic of Serbia for a period 2011 - 2014.

1 INTRODUCTION

EHR (Electronic Healthcare Records) has had great

public attention lately all over the world (Kukafka,

2007). As the amount of collected electronic medical

data increases everywhere, the health-care services

and supporting industry are making efforts to

identify better ways to use this data for patients care

(Ford, Menachemi, Phillips, 2006). Ideally, data is

collected in a real time, can support point-of-care

clinical decisions, and, by providing instantaneous

quality metrics, can create the opportunities to

improve clinical practice as the patient is being

cared for (Michael, Holl, Badawi, Riker, Silfen,

2010).

However, all segments of the health-care

industry are plagued by many challenges that have

made it a latecomer to business intelligence and data

mining technology (Wickramasinghe, Schaffer,

2006). For example, the adoption of electronic

medical records is delayed in many countries,

integration between different medical information

systems (MIS) is poor, and there is a lack of uniform

technical standards. There is a poor interoperability

between complex medical devices and MIS. Until

basic technical infrastructure and well designed

clinical applications are implemented through the

health-care system, data aggregation and

interpretation cannot effectively progress

(DesRoches, 2008).

Unfortunately, all these facts affect the area of

rare diseases tracking even more than other

healthcare related processes. Rare diseases occur

with significantly less frequency than common

diseases. But, their frequency should not reduce the

professional attention that is given both to diseases

and to patients suffering from them. Unfortunately,

practice often shows just the opposite: patients are

309

Jankovic D., Stankovic T. and Todorovic B..

COLLABORATIVE IT PLATFORM FOR RARE DISEASES.

DOI: 10.5220/0003171003090314

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2011), pages 309-314

ISBN: 978-989-8425-34-8

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

usually marginalized by the society. The number of

medications used for rare diseases treatment is

relatively small, many of them are not in the list of

medicine that government pays for, and their price is

often even ten times higher than the price of

medications for common illnesses. This situation

may be justified by the fact that the interest of

researchers and pharmaceutical companies

especially in these diseases and these drugs was

minor, when compared to the interest in common

diseases. Rare diseases are too specific to study, to

find the appropriate causes of disease (most of rare

diseases are genetic), treatment methods, and the

appropriate preparations. Researching difficulties are

certainly the result of a small amount of relevant

data in this area, in research centres separately. The

only goal of this paper is to emphasize the problem

of patients suffering from rare diseases, and to

propose one way to help those patients, their

doctors, and the researchers interested in the area of

rare diseases. Our suggested solution introduces one

model of Collaborative IT (CIT) platform for rare

diseases (Fig. 1).

The proposed platform is the subject of the

project we apply to the call for proposals of the

MSc&TD in Serbia. Another purpose of this paper is

to find potential partners for an international project

with the same subject and goals and a wider set of

users from different countries.

Figure 1: CIT platform for rare diseases illustration.

HEALTHINF 2011 - International Conference on Health Informatics

310

The existence of a project DILS (Delivery of

Improved Local Services) is very important for the

success of our project (http://www.dils.gov.rs).

DILS main objectives are: to increase the capacity of

institutional actors and beneficiaries in order to

improve access and the efficiency, equity and

quality of local delivery of health, education and

social protection services (see

http://web.worldbank.org/external/projects/main?pa

gePK=64312881&piPK=64302848&theSitePK=409

41&Projectid=P096823). One of the DILS

components is establishment of the information

systems that connect local service providers (health

centres, schools, centres for social work and non-

governmental organizations) with the relevant

ministry and allow efficient and transparent

financing of services, service delivery and inter-

agency information sharing. 157 health centres in

Serbia will be provided by computer equipment and

relevant MIS. Therefore, our platform for rare

diseases will have sufficient IT base for successful

work.

2 RESEARCH SIGNIFICANCE

This research will represent continuation of our

previous work (http://medisc.elfak.edu.rs, Rajković,

Janković, Tošić, 2009, Rajković, Janković,

Stanković, 2009) on development of different MIS.

Working on both ambulatory and clinical MIS, we

have offered many data retrieving forms to medical

professionals as well as connection to laboratory

devices, different reporting and data analyzing tools.

Our systems are primarily oriented to fit the needs of

Serbian public healthcare, but designed configurable

so they could be applied anywhere. Mentioned

solutions are in process of adoption within medical

facilities in Southern and Eastern Serbia. Population

of this region is about 1.7 million people, which

ensures representative population for data retrieval.

Since we have realized that rare diseases had more

complex medication process we have decided to

start research in this area, having in mind wider

sociological significance.

Research and activities that took place during the

modelling of the platform are important from several

aspects: scientific research, national, social, and

educational.

2.1 Scientific Research Aspect

Our first goal is to establish central repository for

rare diseases that would be, in the beginning, used in

the Republic of Serbia. The establishment and

operation of a central repository will provide, in a

relatively short time, critical amount of data

necessary for quality research to a number of

researchers and research centres dealing in the area

of rare diseases. Without the existence of a central

repository, with variety of relevant medical

information about patients suffering from the rare

diseases, research is performed on the small amount

of data from real patients, in separated centres,

which can potentially reduce the quality and

effectiveness of research. Initially, central repository

will import data retrieved by previously developed

medical information systems that are in active use in

Nis region. Next step will be defining data

interchange standards (under the authority of the

Serbian Ministry of Health and National health

insurance) to allow other medical information

systems to share data with central repository. Also,

legal point of view must be included to ensure that

data exchange process will not intrude patients’

privacy. As a side repository effect, the complete list

of patients suffering from rare diseases (that does

not exist at the moment in Serbia) will be formed.

A very important point of the research aspect is a

novelty in the approach to data analyses.

Specifically, power software tools will be developed

for analysis of data entered by patients themselves.

Similar approach is mentioned in (Michael, Holl,

Badawi, Riker, Silfen, 2010), but not against the

wide spectrum of medical data entered by patients.

2.2 National Aspects

By establishing a rare diseases central repository, we

will create a suitable location that offers different

kinds and pieces of information useful for patients.

Data such is common symptoms of diseases, specific

symptoms, medical and general advices, lists of

medicaments, trends in medical practice,

geographical distribution, and the success rates of

different treatment under different circumstances

will be maintained by applications based on the

repository. This will bring national (even

international) importance to a repository. Such a

repository should provide access to all researchers in

this field, as well as clinicians who participate in the

project, to store the data they have on their clinics.

One goal is to enable all doctors to store the relevant

data in the central repository, whenever they detect

rare diagnosis.

Another crucial national aspect is bringing

systemised set of data directly to the patients, in

their own language. There are many different

COLLABORATIVE IT PLATFORM FOR RARE DISEASES

311

sources about rare diseases on the Internet, but there

are just few South Slavic languages sites, having

pure information about rare diseases.

The existence of repository will allow

appropriate planning of the activity, primarily in the

Ministry of Health and National health insurance

organisation, and then in the medical institutions.

That way rare diseases data repository should

become a significant referent centre, which should

be recognizable in the region, and we hope even in

Europe.

2.3 Social Aspects

Serbian government has made significant efforts

during the last decade to educate people in the health

issues area, but those suffering of rare diseases are

still treated, unfortunately, as marginalized group of

citizens. Having CIT Platform for rare diseases,

patients would be able to actively participate in their

own treatment in different ways. They will have

better status in society along with better ability to

independently organize themselves, and the

motivation to do additional efforts to decrease the

consequences of their illness and healing where

possible. They will be able to contact and share

experiences with other people, find information

about the disease, medications, therapies, medical

institutions that deal with their disease, research

centres, pharmaceutical companies, the latest

scientific developments and other useful links.

In addition to the project significance (which will

over time get more and more important) goes the

fact that the estimation of the number of rare

diseases patients (about 5%) is very significant and

it grows. Only in Serbia the number of those patients

reaches 400000, about 30 million in Europe, i.e. 250

million around the world! As science progresses, the

number of rare diseases will probably increase, even

though it has not been small yet.

Because of these figures, pharmaceutical

companies nowadays increase attention directed to

this segment. Their income from this area becomes

more important. We can expect that their further

engagement will grow, and then, medical data

collected and stored in repositories can be essential

for them, both for planning and for the research

itself, in the area of new medications. The problem

of rare diseases is known and marked as extremely

important all over the world. The number of

projects, international associations, organizations

and sites dealing with this stuff grows every day

(www.rarediseases.org, www.raredisease.org.uk,

www.eurordis.org, http://www.crdnetwork.org, etc.).

In Serbia, some medications for rare diseases

lately became a part of government’s so-called

“positive list”, which means that patients can get

them free of charge. Unfortunately, only few of

them are on positive list currently.

3 CENTRAL REPOSITORY

The number of patients suffering from rare diseases

is not significant in health-care facilities; there are

diseases with one or zero patients per facility. The

problem with rare diseases is that critical amount of

data needed for research and education is spread all

over the world. Also, due to the procedural,

administrative and other problems those data cannot

be gathered easily from different clinics located in

different countries which are pretty interesting from

researchers’ point of view. For this reason we are

commencing development of this platform, having

leading idea to easy gather medical data about rare

diseases as much as possible at one place.

The first step in implementing CIT platform for

rare diseases would be the classification of diseases

and gathering significant medical information about

every entity (name, type, common and specific

symptoms, possible treatments, and many more).

That would be much easier part of the research. The

harder one is gathering medical data about patients

suffering from these diseases. The idea how to

gather significant medical data about particular cases

is described later. This process must be performed in

order not to harm patients’ privacy and make their

lives even more complex.

3.1 Collecting Rare Diseases Data

We will try to define several use cases how to

collect data in proper way, both from medical and

legal point of view. Our intention is, again, to collect

more relevant medical data without making patient’s

personal data directly connected and exposed in

public. Possible cases are:

 Physicians directly involved in rare disease

medical treatments will be able to enter medical and

demographic patients data (with the respect of

patients privacy) when they discover new case of

rare disease (through the corresponding web

application). They will leave symptoms, used

medicaments, results of laboratory analyses,

recommended therapies, results of therapeutic

treatments, etc.

 Web services that could collaborate with differ-

HEALTHINF 2011 - International Conference on Health Informatics

312

rent medical information systems will be developed

as a part of IT platform, to help clinics partners

provide electronic data from theirs EHR. This step is

very difficult to project, because of databases

heterogeneity in different MIS. To avoid bad data,

beside data recognition, human factor will have to

take place in this way of collecting data.

 Patients themselves will be able to leave their

data. This kind of data will be marked as ‘patient

left’, so it could be included or excluded from

research on researcher’s request.

3.2 Data Evaluation

All gathered data should be evaluated in order to

avoid bad or missing data for further researching.

Curtain procedures will be developed to check data,

and to filter or delete bad data. In some cases

procedures will add references to provide missing

referential data integrity. The set of procedures that

will be performed on gathered relevant data will

present “Data evaluation tool”.

Relevant data, in this context, is any medical

information relevant to the disease. It is possible to

develop various software tools for analytics, faster

diagnosis, conclusion based on data-mining

algorithms, and so on.

4 SOFTWARE TOOLS

Three kinds of software tools will be developed

inside of the platform: communication tools,

business intelligence (BI) tools and data mining

tools as described further.

4.1 Communication Tools

Communication tools are of importance both for

physicians and for patients themselves. Generally,

there is a possibility to use different tools for patient

to patient communication, patient to doctor, patient

to institution, and patient to patients association.

Communication tools will provide technical ability

to the patients to create "virtual associations for

certain types of rare diseases" through CIT platform.

If implemented in international level, the platform

would be localized to different languages.

The idea is every patient to have account, every

association to have account and responsible person,

and every institution or facility to have account and

responsible person. The relation between entities

will be precisely defined.

4.2 Business Intelligence Tools

BI in medicine can be referred to as clinical

intelligence (Wickramasinghe, Schaffer, 2006). In

our plan, this kind of tools represents the set of

software tools for analyzing data repositories that

can trigger periodically or on demand, and are

designed to find appropriate templates related to

specific diseases. Beside analysis tools, there can be

software tools for setting hypotheses about the

relationship of appropriate data and factors that

could potentially be of use to researchers dealing

with rare diseases.

These tools would be available to any medical

institution or university and its researchers who want

to participate in the project. Researchers should also

share their experiences, opinions, suggestions and

results between them on platform’s forums.

4.3 Data Mining Tools

Each patient record consists of medical history

symptoms, medical conditions, and various tests and

lab results. In order to categorize those symptoms,

and to recognize the existence of rare disease, we

will apply machine learning classifiers: decision

tree, support vector machine (Vapnik, 1998) and

Bayesian sparse trained logistic regression (Tipping,

2001).

Basically, there are two main tasks which will be

pursued during the building of these classifiers. First

task will be of course, to obtain as accurate as

possible classifiers which will be able to recognize

(based on symptoms and lab results) the presence of

rare disease. The second one, also important will be

to recognize which symptoms and lab (test) results

are relevant features for obtaining good recognition

accuracy. By selecting relevant test results which are

sufficient for accurate categorization (recognition) of

the disease we are able to cut the cost of the

potentially expensive tests and save the patients

from unnecessary physical exertion.

5 PLANNED ACTIVITIES

Brief overview of CIT platform planned activities

can be presented in two parts, as follows. Activities

related to data repository:

a. classification of diseases and data structure

modelling;

b. definition of information relevant to certain types

of diseases;

COLLABORATIVE IT PLATFORM FOR RARE DISEASES

313

c. central repository database designing;

d. design and creation of web application for

manual data entry by physicians;

e. design and creation of web services for

automatic data extraction;

f. definition of tools for data analyzing and

reporting (Data Mining and BI);

g. definition of tools for publishing the results of

analysis;

Activities related to collaborative platform:

h. portal design and implementation;

i. setting up an initial rare diseases data;

j. Portal promotion in national and international

level.

The most complicated phase in the CIT platform

project development is expected to be the phase e.,

because it implies collecting and importing data

from heterogeneous medical data sources.

Differences in medical standards between countries

will create special difficulties, which could be partly

overcome by involving international medical centers

in the very beginning of the project.

6 CONCLUDING REMARKS

Due to the project specificity and its national

(international) importance, precise analysis of the

evaluation plan for the return of investment is

relatively difficult. The very nature of the platform is

such that its’ result is significant at the national level

and in a broader perspective. For these reasons we

made a project application and funding request that

has been sent to Serbian government. Since we are

supported in this work partly by Nis Clinical Centre

and partly from National health insurance

organization we expect that our efforts will be

recognized as national interest.

At the same time there is a benefit immeasurable

financially, reflected in the satisfaction of patients,

and their restored sense that society cares about

them, and kind of returning them from the social

care margins. The other useful effect can be

achieved by including the pharmaceutical companies

in the platform through advertising their

manufactures, through the use of data collected and

payment for the service. A special aspect is the

possibility of forming a Balkan or even European

data centre to collect the data on rare diseases.

Certainly, the perspective is that after the project

achieves the planned results, it can be spread in

some international projects (for example within so-

me of FP7 calls).

There is also a factor directly immeasurable: the

patients themselves can access useful information to

reduce their cost of treatment and personal

problems. There is a great probability that the costs

of rare diseases diagnostic can be significantly

reduced, if we successfully develop such a platform.

If we involve medical doctors employed in the

public health in education and use of such CIT

platform, they can reduce the number of expensive

medical analyses for diagnostic in rare diseases.

Software tools can greatly assist rapid diagnosis

of rare diseases, of course after a period of data

collection in the repository, in order to create a

sufficient quantity of data to perform the conclusions

based on the knowledge base.

And at the end, software tools that will be

developed to analyze the data stored in the

repository can be used for many similar and

commercial databases. Also designed model of CIT

platform can be used in other social and public needs

(justice, sport, investment, etc.).

REFERENCES

Michael, M., Holl, R., Badawi, O., Riker, R.R., Silfen, E.,

2010. A Collaboration Between Industry, Health-Care

Providers, and Academia, IEEE Engineering in

Medicine and Biology, Volume 29, Number 2, pp. 18-

25.

Wickramasinghe, N., Schaffer, J. L., 2006, Creating

knowledge-driven healthcare processes with the

Intelligence Continuum, International Journal of

Electronic Healthcare, Vol. 2, No. 2, pp. 164-174.

Rajkovic, P., Jankovic, D., Tosic, V., 2009, A software

solution for ambulatory healthcare facilities in the

Republic of Serbia, in Proc. of HealthCom2009,

Sydney, Australia, IEEE, pp. 161-168

Rajkovic, P., Jankovic, D., Stankovic, T. N., 2009, An e-

health solution for ambulatory facilities, in Proc. of

ITAB 2009, Larnaca, Cyprus, IEEE, pp. 1-4

Kukafka, R., August 2007, Redesigning electronic health

record systems to support public health, Journal of

Biomedical Informatics Volume 40, Issue 4, Pages

398-409.

Ford, E.W., Menachemi, N., Phillips, M. T., 2006,

Predicting the adoption of electronic health records by

physicians: when will health care be paperless?, J Am

Med Inform Assoc 13, pp. 106–112.

DesRoches, C. M., at all, Electronic Health records in

Ambulatory Care – Survey of Physicians, The New

England Journal of Medicine, 2008, 359: 50-60.

Tipping, E. M., 2001, Sparse Bayesian Learning and the

Relevance Vector Machine, Journal of Machine

Learning Research, pp. 211-244.

Vapnik, V. N., Statistical Learning Theory, Wiley, 1998.

HEALTHINF 2011 - International Conference on Health Informatics

314