The acronym ePAR stands for electronic Paediatrics Admission Records. This is an electronic format of the Paediatric Admission Records (PAR) and the Newborn Admission Records (NAR) developed by the Child Health and Newborn Group. CHNG is a group of clinical researchers led by Mike English, a paediatrician who has devoted his career in research on child health in the tropics, working with the KEMRI-Wellcome Collaborative Research Programme. The PAR/NAR were developed as a resource to improve the quality of inpatient hospital care accorded to children admitted to hospitals in under-resourced settings of sub-Sahara African countries. The PAR/NAR paper-based forms have been subjected to rigorous clinical evaluation. They effectively assist clinicians elicit essential aspects of clinical history, examination and treatment of the most common problems experienced by children in low income countries and thus can impact on the quality of health care given to children in public hospital settings typical to these countries. These forms also in help collecting a complete set of routine health information that is vital for clinical governance, clinical audit, child health policy and research.
Although there have been isolated efforts to develop electronic medical records (EMR) in Kenya, the very few that so far exist in public health settings are disease-centric, not adapted for use by clinicians at the point of care and have no real-time linkage with the laboratory and complimentary clinical disciplines. Most endeavours to develop EMR have concentrated on developing Enterprise Resource Programmes (ERP) that are prinicipally deployed for human and financial resources management with primordial functionalities for clinical work such as basic discharge reports, listing of morbidities and mortalities and so on. There has been no development of systems that can be used by clinicians or laboratory information systems linked to clinical operations in public settings. Typically, the demand for such systems is virtually non-existent despite the massive advantages of efficiency, safety and reproducibility they can deliver. In particular, ICT and clinical practice present two diametrically different disciplines, which, unlike ICT and engineering, are totally not taught to either of the two professions to extents that can lead to creation of innovative ICT solutions for clinical and laboratory practice in developing countries.
The opportunity to develop an electronic medical records system ideal for paediatrics clinical, laboratory and research use presented itself with the need to establish a surveillance to monitor vaccine preventable bacterial diseases and to evaluate the effectiveness of a new vaccine to be introduced in Kenya among children in 2010. Any choice of a representative site for such work in Kenya would necessarily yield one without standard clinical, microbiological or documentation practice. The choice of Thika Level 5 Hospital necessitated the establishment of systems to ensure good clinical and documentation practices and standardisation of microbiological services. A traditional clinical and laboratory surveillance system would be very heavily-reliant on paper forms that are cumbersome to fill and store but also not amenable to the early detection of systematic errors as data entry takes place several weeks after encounter with patients. Such systems are prone to numerous data entry errors and cannot yield sufficiently updated data for rapid decision making in settngs where skilled ICT manpower and departments are non-existent. Having anticipated this problem in 2006, three years to the establishment of such a site, we set to develop an EMR that would be deployed contemporaneously with clinical and laboratory practice by staff within hospital settings for all type of clinical problems.
We developed the first integrated paediatric clinical EMR in Sept 2006 but was never deployed at the Kenyatta National Hospital for which it was written. This may partly have been because of the inability to show examples of such successful endeavours elsewhere, the perception that greater health problems than can be surmounted by adoption of ICT exist in the hospital setting and a lack of clear drive or policy to adopt electronic health records, which upto now are seen purely as a reserve for research. With continued background study of EMR and programming, it became evident that adoption of a maintainable and extensible way of writing programmatic code was necessary. The system was moved onto the new Zend Framework PHP Model View Controller (MVC) architecture introduced in 2006. As the ZF MVC rapidly developed to introduce modular programming two more versions of the original EMR were written with the first practical implementation of the system being undertaken during the third version in April 2009. During development time, the EMR was adapted for a research question to evaluate the use of evidence by medical students through simulation of clinical clerking, decision-making and management.
The third version of the paediatric EMR was based on the PAR/NAR forms. The system automated paediatric clinical history taking, diagnosis, prescription, diagnosis, clinical notes generation and discharge. These funtionalities were linked with a laboratory information system that encompassed all inpatient paediatric and adult inpatient and outpatient specimens processing and ward/bedside diagnostic work-up including rapid HIV-testing and rapid blood sugar. Basic clinical and laboratory dashboards to gauge the performance of blood cultures and lumbar puncture performance among others were set up. Rather than introduce a new clinical workflow, we used our experience in clinical service and research to computerise the daily clinical activities and digitise all the forms used in these settings with minimal inteference to clinical work processes. We particularly wanted the system to be a real-time tool for clinical and laboratory use and not a sophisticated secondary data entry tool.The successful deployment of the system within a provinicial level hospital where no computerised medical records existed and the staff had minimal to no experience with computers provided the proof that such a system was viable in under-resourced settings of sub-Saharan Africa hospitals with some basic supervision. The system was subsequently introduced at Embu Provincial General Hospital in Nov 2009 on the back of a successful introduction at Thika. It had then reached a point where minimal clinical knowledge was needed for implmentation purposes after a proof of concept at Thika.However, there is need to develop a system architecture amenable to collaborative development and one with a modular nature to allow the plug in of other modules developed for clinical, laboratory and community service and research purposes. The fourth phase of the ePAR has embarked on the creation a modular ePAR based using the web-based architecture that the system adopted from outset.
The strengthens of this system include the use at the point of care by clinicians, linkage with laboratory and community surveillance, modular code development making it easily extensible and the multiplicity of its use within hospital setting and population based surveillance. The web-based architecture of the system and the use of web 2.0 Zend Framework MVC has opened the system for easy integration with virtually all forms data handling and communication systems that find application in health informatics and social networking.
