Building а National Clinical Data Warehouse

Ustun Yildiz

and Ayse Keles

1,2

Agency of Health Informatics, Ankara, Turkey

Yildirim Beyazit University, Ankara, Turkey

{ustun.yildiz, ayse.keles}@saglik.gov.tr

Keywords: EHR, Clinical Data Warehouse, Data Privacy, Health Transformation Program

Abstract: Turkey has created an e-health vision along with the Health Transformation Program. In this framework,

standard coding systems such as standard definitions of the institutions, databank of healthcare providers,

standard disease, drug and medical supplies classifications have been developed and a national clinical data

warehouse (Health-NET) was established. Health-NET is an integrated, safe, fast and expandable health

information system which aims to improve efficiency and quality of health services by collecting all kinds of

data produced in the health institutions in line with the standards and generating information adequate for all

stakeholders of the collected data.

1 INTRODUCTION

Access to healthcare is regarded as a fundamental

element of social development. While classic theories

of development explain development in terms of

many socio-economic and cultural indicators such as

GNP per capita, level of industrialization and

employment, and consumption level of primary

goods and services, today, it is explained primarily in

terms of indicators related to the access to education

and healthcare. At the end of 2002, the status of the

Turkish health system made it necessary to undertake

radical changes in many areas from service delivery

to financing and from human labor to information

systems. In order to improve the quality of service in

these areas, Turkey has gone through a Health

Transformation Program (HTP) started as early as

2003(OECD, 2014)(Akdag, 2008)(Akdag, 2011). As

a result of the health infrastructure rehabilitation

efforts, Turkey has observed a rapid decline in the

Under-5 Mortality Rate (U5MR) from 29 in 2003 to

7.7 in 2011 per 1000 births (T.D., 2010) Similarly, the

life expectancy at birth has increased to 75 in 2011

from 70 in 2004. Consequently, the satisfaction with

the government provided health services was

measured 75.9% in 2011 comparing to 39.5% in

2003(WHO, 2012). In the context of HTP, the

effective collection and use of nation-wide Electronic

Health Records (EHR)s became a primordial goal. In

parallel with the development of the relevant

legislative framework, a national clinical warehouse

(CDW) that collects EHR and other operational data

from all health organizations in the country had been

established in 2012 and became operational. As

analogous to the worldwide approach of putting the

health information at the center of decision processes

along with the patient, the data collected from the

healthcare facilities are used to structure and manage

the new health infrastructure of the country (De Mul,

2012)(Yoo, 2014). Generally speaking, CDW is used

to empower traditional application software in order

to analyze public health behavior and support several

different decision workflows such as clinical quality

improvement (Weiner, 2012), pay-for-performance(

Van Herck, 2010) and evidence-based medicine

systems (Sacklett, 2000). The integration and use of

diverse healthcare data from various sources into the

same clinical repository is a challenging problem

when implemented at a nation-size level. The

problem becomes even more complex when Hospital

Information Systems (HIS) that create the collected

data are autonomous and implemented with different

technologies. The current HIS market in Turkey

includes more than 150 private firms that implement

their customized software in more than 1500

healthcare facilities operating with more than 200.000

beds capacity (Kose, 2013). In this paper, we share

our experience in implementing CDW and its use in

the critical decision processes of the Ministry of

Yildiz U. and Keles A.

Building Ð

r National Clinical Data Warehouse.

DOI: 10.5220/0005890000930097

In Proceedings of the Fourth International Conference on Telecommunications and Remote Sensing (ICTRS 2015), pages 93-97

ISBN: 978-989-758-152-6

Health of Turkey. We summarize the design and

operation processes, the software architecture, its use

in Health Transformation Program and the lessons

learned.

2 BACKGROUND

In this section, we describe some of the main aspects

of a large-scale CDW that needs conceptual and

technical considerations.

2.1 Data Collection

The health institution that provides CDW with data is

composed by Level I Family Practitioners, Level II

Public and Private Hospitals and Level II University

Hospitals and Research Centers. It is obvious that the

data collection might not be achieved with 100\%

accuracy due to the complexity of data and the

widespread use of the system. Data collection

services are frequently updated due to the updates in

the data packages definitions. This continuous

process requires a tight coordination with Hospital

Information System (HIS) providers that will

implement client component for data upload to

collection services. Due the difficulties in the

integration process that needed continuous support, a

help desk has been established by the Ministry of

Health. The help desk provided regular data on the

amount of data collected by CDW, the amount and

type of upload errors to HIS firms. It has been

recognized that the success of integration capabilities

of institutions that belong to different levels are

different. Level I institutions were 99% successful in

uploading their data as described in the integration

kits of CDW public website. Level II institutions were

less successful in sharing their data. The main reason

behind the successful integration of Level I

institutions was based on the underlying business

model which makes possible the calculation of

practitioners salaries based on the collected data by

CDW. While Level I data collection was successful,

Level II data collection had not been as expected and

it was below 60% average for the first year. The main

reason for the lack of data was that Level II

institutions were not subject to any business model

implemented by CDW. The second year of the

implementation of CDW, the data collection rate was

increased to 77% as the parameters for Service

Quality Standards were started to be calculated with

CDW data. The integration of Level III research

institutions were even less successful given that they

were not managed by the Ministry but by independent

universities. During the third year, the integration of

Level II has improved by the increasing calculation

of different healthcare service parameters using

available CDW data.

2.2 Data Quality

It is important that we should not confuse the concept

of 'data quality' to the aforementioned concept of

'clinical quality data'. While 'clinical quality data' is

specific data based on clinical quality indicators

which help to understand the clinical quality of the

services provided to patients with specific diseases

e.g. diabetes, stroke etc. On the other hand, 'data

quality' is about the quality of any data collected by

CDW and based on certain criteria sets e.g. complete,

valid/correct, timely, without duplication (Kahn,

2012)(Arts, 2002). In this section, we detail the issues

and our solutions to improve the ‘data quality’ of the

‘clinical quality indicators’ collected for CDW. The

major issue that has been encountered in establishing

the data quality was the data wrongly packed that do

not fulfill data package acceptance rules. We

identified the main reason was the difficulties

associated with the establishment of data packages to

be sent to CDW. On the other hand, we identified HIS

users use different ICD-10 (WHO, 2004) codes for

certain diagnosis and diseases because HIS require

sophisticated data input interfaces for the latters. One

particular aspects that needs special attentions is the

geographic and temporal properties of the clinical

data. It is observed that public health indicators could

be misleading based on certain periods and on

locations (e.g. Temporary Refuge Spaces), these

problems are configured by expert knowledge.

2.3 Data Privacy

The privacy of EHR had been a high priority concern

in the implementation of CDW and the tools that

manage its data. In the collection process of EHR, the

definition of data packages was mainly defined by the

public health surveillance necessities and the

establishment of personal health records. The idea

behind the establishment of personal health records is

to support the continuity of healthcare and prevent

redundant services such as radiology. One other

advantage of the involvement of patients in the

structuring the records is the elimination of

inconsistencies. Turkish citizens have a unique and

publicly available 10 digit number. The use of this

number considerably facilitates the consolidation and

access of personal health records but in the same time

could be a major privacy concern.

Fourth International Conference on Telecommunications and Remote Sensing

Figure 1: Overview of Architecture

In order to overcome with privacy concerns, an

electronic consent form has been implemented in

personal health record system (MoH, 2015). Patients

can login and configure their preferences on the use

of their records. The preferences are overwritten only

in the case of emergency room services. It is

recognized that the collection private health

information and its computation after de-

identification of data still can be a privacy concern as

the identification of patient is possible with the

combination of different queries (Fernandez, 2013)

For certain diagnosis and diseases codes (e.g. HIV),

we have decided to collect the records without

identifiable information and a cryptic number without

publicly available citizenship number

3 SOFTWARE ARCHITECTURE

Figure 1 describes the overview of the architecture of

CDW with components related to data collection

through web services, storage and data analyze

modules. Messages coming from healthcare facilities

are collected under HL7 form by Load Balancer

component. Received messages are forwarded to an

available XML Gateway. In parallel, HL7 messages

are processed asynchronously by the JMS log queue.

During its flow throughout the XML Gate way, the

Authentication happens with LDAP on OID. Next,

HL7 message is validated against XSD schema and

Schematron processes business rules (mandatory and

optional fields in data packages). After the validation,

HL7 messages are transformed to a local data format

to be saved in the database. The transformed

messages are also validated against XSD Entity.

Building а National Clinical Data Warehouse

Figure 2: Functions available on dashboard

Collected messages are transferred to Service

Bus. In the latter, the data is unpacked and the patient

ID is verified through an external ID verification

service (WS-Mernis). Figure 2 describes some of the

functions available to users.

4 CONCLUSIONS

This paper has presented a small set of lessons learned

from the establishment of a National Data

Warehouse. The main research problems that we took

the challenge on were: Data quality, data collection

and data privacy issues. Although the patient count

and basic computational problems were manageable

with relation data models and systems, the advanced

computation of patient records such as disease

correlation analysis, organization of cohorts for

evidence-based medicine applications require the use

of big data solutions.

REFERENCES

OECD, Oecd reviews of health care quality: Turkey

2014doi:http://dx.doi.org/10.1787/9789264202054-en.

URL /content/book/9789264202054-en

R. Akdag, Turkey health transformation program 2003-

2010, Tech. Rep. 839, Re- public of Turkey, Ministry

of Health (June 2011).

R. Akdag (Ed), Health transformation program in turkey

and primary health care services, Tech. Rep. 770,

Republic of Turkey, Ministry of Health (November

2008).

T. D. V. D. Lilia Jelamschi, Decline in the under-5 mortality

rate (u5mr) in turkey: A case study, Tech. rep.,

UNICEF,

http://www.unicef.org.tr/files/bilgimerkezi/doc/declin-

in-the-under-5-mortality-rate-u5mr-in-turkye-a.pdf

(March 2010).

W. H. O. R. O. for Europe, Successful health system

reforms: The case of turkey, Tech. Rep. WHO LIS

e96508, World Health Organization: Regional Office

for Europe (May 2012).

M. De Mul, P. Alons, P. Van der Velde, I. Konings, J.

Bakker, J. Hazelzet, Development of a clinical data

warehouse from an intensive care clinical information

system, Computer methods and programs in

biomedicine 105 (1) (2012) 22–30.

S. Yoo, S. Kim, K.-H. Lee, C. W. Jeong, S. W. Youn, K. U.

Park, S. Y. Moon, H. Hwang, Electronically

implemented clinical indicators based on a data ware-

house in a tertiary hospital: its clinical benefit and

effectiveness, International journal of medical

informatics 83 (7) (2014) 507–516.

J. P. Weiner, J. B. Fowles, K. S. Chan, New paradigms for

measuring clinical performance using electronic health

records, International Journal for Quality in Health

Care 24 (3) (2012) 200–205.

Fourth International Conference on Telecommunications and Remote Sensing

P. Van Herck, D. De Smedt, L. Annemans, R. Remmen, M.

B. Rosenthal, W. Sermeus, Systematic review: effects,

design choices, and context of pay-for- performance in

health care, BMC health services research 10 (1) (2010)

247.

D. L. Sackett, Evidence-based medicine, Wiley Online

Library, 2000.

R. Kose, Health statistics year book, Yearbook 978-975-

590 509-9, Agency of Health Research (2013).

M. G. Kahn, M. A. Raebel, J. M. Glanz, K. Riedlinger, J. F.

Steiner, A pragmatic framework for single-site and

multisite data quality assessment in electronic health

record-based clinical research, Medical care 50.

D. G. Arts, N. F. De Keizer, G.-J. Scheffer, Defining and

improving data quality in medical registries: a literature

review, case study, and generic framework, Journal of

the American Medical Informatics Association 9 (6)

(2002) 600–611.

W. H. Organization, et al., International statistical

classification of diseases and health related problems

(the) icd-10, Ph.D. thesis, World Health Organization

(2004).

T. M. of Health, E-nabiz personal health records system, ,

https://enabiz.gov.tr/Giris.aspx (2015).

J. L. Ferna ́ndez-Alema ́n, I. C. Sen ̃or, P. A ́. O. Lozoya, A.

Toval, Security and privacy in electronic health

records: A systematic literature review, Journal of

biomedical informatics 46 (3) (2013) 541–562.

Building а National Clinical Data Warehouse