Information and Communication Technologies (ICTs) have made significant impact on healthcare industry in the globe. Its adoption and use, has transformed the way healthcare services are delivered and monitored. Influenced by this transformation, MOHCDGE have formulated an eHealth Strategy 2013-2018 to manage these massive shifts from traditional HMIS to electronic means. The strategy highlights the important aspect to consider when automating health facilities. One of the key areas targeted for transformation is the use of ICT to make health facilities more effective and efficient.
However, health information Systems (HIS) in developing countries have been characterized as fragmented, with a number of subsystem operating independently (Shidende, 2005). The fragmentation nature of health systems has been with other factors due existing of multiple initiative to computerize healthcare systems, each initiative serving its own data need goals. Therefore, integration is in essence an initiative to rectify fragmentation. Integration of health care information systems have been ongoing efforts to strengthening quality of care by ensuring proper management of healthcare resources through information-based management system. Due to fragmentation nature of HIS, various efforts have been employed to ensure coherent and standardized HMIS (Nyella, 2015)
With respect to Health Information Systems, integration refers to the act of bringing data and systems together from disparate subsystems to be viewed as one in a certain perspective. As per Park, (2004) integration combine together the data and process of sub-systems into one system, and dealing with aggregation of subsystems cooperating so that the system is able to deliver the overarching functionality and ensure that sub-systems function together as a whole system (Frye & Jones, 2010; Loutas, Kamateri, & Tarabanis, 2011). In HIS, integration further looking on the processes and routine to capture, process and analyses wide range of data, for both patient and aggregate across different health programs to be used by policy maker and healthcare providers for planning, resource allocation and decision-making purposes. HIS Integration does not only imply computer systems, some integration is paper based reporting system to electronic systems (Sahay, 2012). For example, looking on integrated HMIS system, it integrates all health data collection forms from all health programs (vertical program and ministry lead programs) and reported to single electronic system i.e. DHIS2.
This paper reviews some integration approach used in health information systems, drawing literature experiences form developing countries.
Integration vs Interoperability
Integration and interoperability in health information systems have been touted with ambiguities. For the purpose of this study, integration is understood as the process of joining distinct HIS systems in a way that they appear as whole in a certain perspective (Monteiro, 2003). The systems are not necessarily interconnected to each other, but as long as to a certain point they appear to share common understanding. Integration is slightly different from interoperability, as according to Braa & Sahay, (2012) interoperability is understood as the ability to exchange data between the two or more systems.
According to Toohey et al., (2012), inteoperabilty refers to the interconnection of system application so that one application can access and write data of another application. Also, Interoperability as defined by Sabooniha et al., (2014) is the ability of an information system to use services and data from another information system to achieve a specified task in a given context and provides continuous exchange of information between collaborating information systems. As indicated on Figure 1, in integration, doesn’t matter what means of data sharing, it can be paper based or electronic. For example, an integrated warehouse because contains aggregate data from different sources such as TB, HIV, Malaria program and HMIS data. These data share platform and available at one point, but the TB and HIV or Malaria systems are not interoperable. But interoperability deals with electronic way as a strategy to achieve integration, in that case it requires all systems to share data electronically.
Abbildung in dieser Leseprobe nicht enthalten
Figure 1: Integration vs Interoperability (Source: Braa & Sahay, 2010)
Horizontal, Vertical and Star Integration
According to Toohey at el., (2012) there are three basic Health Information systems Integration Approaches: Horizontal integration, Vertical integration and Star integration as illustrated on figure 2 and 3. Horizontal integration approach deals with integrating system functionality which are within the same level or between different systems supporting different functional areas of a business (Wangler, 2000). A shown in Figure 2., a typical example in health systems is integrating different modules of the hospital management system, such as integrating registration module to billing, then billing to laboratory, and laboratory to diagnosis and treatment. The aim of horizontal integration is to automated transaction operations of the organization.
On other hand, Vertical integration is about joining two systems of different levels following the line of command (Monteiro, 2003; Sahay, 2012; Smith, Madon, Anifalaje, Lazarro-Malecela, & Michael, 2008). Vertical integration is between systems on different control and managerial levels of an organization. As indicted on Figure 2.2, the typical example in health systems is integrating the hospital systems with the district health information system or integrating district system to national data system. The aim of vertical integration is to make data availability for managing services offered at lower level.
Abbildung in dieser Leseprobe nicht enthalten
Figure 2: Horizontal and Vertical Integration (Author’s own work)
Star integration deals with integrating each information systems to a central warehouse (Xiao, Qiu, & Zhou, 2013). This type of integration is sometimes known as inter-organizational integration (Wangler, 2000) i.e. integrating between systems of different organizations, as illustrated on Figure 2.3. Star integration is more like horizontal integration but there different is on the way implementation is done, unlike star integration, horizontal integration is not necessary each system is directly connected, sub-systems can be integrated indirectly.
Abbildung in dieser Leseprobe nicht enthalten
Figure 3: Star integration (Author’s own work)
This study uses a vertical integration approach, because it deals with two systems at different levels of operations i.e. on how electronic medical records at facility (lower) level can aggregate data and report data to DHIS2 at HMIS (high) level. Currently, the DHIS2 is accessed at the district level, where each month all health data from the facility are submitted to the district data office on a summary form and the entered into the electronic database system. Integrating data reporting between these systems will reduce human error to compile and record data into manual form and also reduce the redundant of data collection.
Information Systems Interoperability Methods
In achieving interoperability, there are substantial technical issues to be considered, these technical issues are termed as semantics (Loutas et al., 2011). Semantic deals with the definition of metadata to be accessed by interconnecting systems. There are about three typical semantic interoperability methods for semantic interoperability of information systems, these are Mapping, Intermediary and Query-Oriented methods.
Mapping-Based Interoperability Methods
The Mapping-based approach attempts to construct drawing between semantically related information sources. Mapping is usually accomplished by constructing a federated or global schema and by mapping it with the participating local schema (Loutas et al., 2011). In our DB Example, with mapping we define a global schema which take in the same attribute names of all two databases, then link it to the local dBs’ This is possible for two different subsystems developed with similar protocol and languages, example DHIS2 and eHMS, both implemented in client – server architecture and using web API. The drawback of the mapping approach is that it is not designed to be independent of particular schemas and application (Loutas et al., 2011). It still relies on the local schema for its global schema to function as intended. Explicit representations of semantics of data sources can help resolve the problem associated with interoperability when design mapping between them.
Intermediary-Based Interoperability Methods
The Intermediary-based approach depends on the use of intermediary mechanisms to achieve interoperability. These intermediaries are somehow called as mediators, agents, ontologies and etc. The intermediary use either mapping knowledge or domain-specific knowledge to coordinate various autonomous information sources (Loutas et al., 2011). In most cases ontologies are used to share standardized vocabulary or protocols to communicate with each other. In health information system several mediators are available to coordinate and facilitate sharing of information, example OpenHIE project, EDI project and so forth. The advantage with this approach is the ability to accumulate knowledge from different ontologies in a domain specific way. But, although this approach is theoretically valid, it’s practically infeasible to develop and maintain an ontology in autonomous, dynamic and heterogeneous databases due to inherent complexity of the knowledge domain (Loutas et al., 2011), this is because the formalism used to represent the ontology becomes complicated for the wide application when representing the complex conceptualization.
Query-oriented Interoperability Methods
The Query-oriented approach is based on the interoperable languages, most of it are either declarative logic-based languages or extended SQL. This involve developing query that can span several databases, but the databases should have a high order expression that range over both data and metadata. The main drawbacks of this approach it places too heavy burden on the users of the systems, by requiring them to understand each of the underlying local databases (Loutas et al., 2011). Because the query driven depend on the understanding between the DBs involved and the user need to understand their underling languages and construct a query that span both DB. Therefore, this approach requires users to engage in the detection of semantic conflicts and resolve them, since the approach doesn’t provide support for identifying semantic conflicts.
HIS Integration Technologies
After looking on methods used in semantic interoperability in information systems domain, now the study confines in the healthcare domain. In order to support different perspective of HIS integration, various approaches have been proposed. These include data exchange standards and protocols (Toohey et al., 2012), middleware technologies, unified models (Shen et al., 2004), domain specific standards and medical coding (Shaw, 2005). These approach goes with the contemporary approaches such as semantic integration as explain in section 2.6. Toohey et al., (2012) state that despite of these distinct approaches, interoperability relies on the common agreement between the participating information systems and the components (boundaries) involved. There are several technical HIS integration solutions developed under those approaches, which are Message-Oriented Integration, Application-Oriented Integration, Middleware-Oriented Integration and Coordinated- Oriented Integration.
The message-oriented integration refers to the use of different messaging standards to exchange data between health information systems (Shidende, 2005). This involves the use of text messages or multimedia message protocols through HIS databases, for instance is using APIs and EDI (Toohey et al., 2012). The message – Oriented Integration uses peer to peer communication between clients sharing the data. Example, DHIS offer a module to allow sharing of data from remote areas through mobile phones sms. Also, some HIS at regional level have implemented message-based approach to automate the medical processes such as patient admission, transfers, prescription, ordering of medicine and equipment and sharing diagnoses results. Messaging integration approach provide effective way to solve the basic integration challenges. Examples of message-based approach integration in HIS is seen in the use of DICOM messages, HL7 standards, XML DTD and HL7 CDA documents.
In application-based approach, it involves constructing a layer of defined application on top of the existing applications so as to support the flow and exchange of information between them. The layered application combines both application and their processing through engine and workflow manager and integration sever (Toohey et al., 2012). Before using the application-oriented approach is best to understand the organizations applications and processes that need to be integrated. The application-based integration takes in the already existing medical standards, technical profiles, specification of actors and transaction as ontologies. Ontology in this framework include information systems or component part of information systems that provide act, produce, operate and manages the needed information to exchange. The main goal of application-oriented integration is to ensure all information needed for decision making is available, accurate and complete on time for the users of HIS. Example of application-based approach project is the IHE (Integrated Health Information Exchange) workflow and SDMX-HD framework developed by WHO.
Middleware-Oriented integration involves the use of intermediary layer. It defines the set of services, interfaces and shared methods which support the entire information exchange between the two systems (Shidende, 2005). This approach provides the platform/infrastructure for sharing the functional services and shared information. This approach reduces the need for replication of methods and data in several systems to be integrated. Since the healthcare organization require collection of disparate users that perform different task and require sharing of common data set and use of common protocol and standards, this approach eliminate such complexity by offering a common platform for each disparate system. Although, this approach may require change in legacy systems to adapt a common infrastructure. The OpenHIE, CORBAmed and PICNIC projects are examples of middleware-oriented integration approach. For CORBA standard HISA is constructed in three layers each responding to specific design, function and operation of information system, the middle layer act as center to all other layer, supplying an infrastructure for other application to be connected.