A framework for implementation of Syndromic surveillance in low resource settings: a case study of facilities in Eldoret town

Abstract

Background: Syndromic surveillance utilizes aggregated patient-level data on signs and symptoms, and other discernible health indicators apparent at the beginning of infection, to identify possible disease events before confirmed diagnosis can be made. Electronic Medical record Systems (EMRS) can improve surveillance through the use of such data, as has been shown in developed countries. Despite the increasing use of EMRS in developing countries, there is no clear, documented evidence of their use for, and contribution to syndromic surveillance. Further, no guidelines on optimal approaches for syndromic surveillance implementation in such low resource settings exist. Objectives: The study sought to: 1) assess the existing surveillance approaches used by various health facilities; 2) assess the readiness of the implemented EMRS at the facilities for syndromic surveillance activities; and 3) propose a framework for syndromic surveillance system implementation. Methods: This study was conducted in 2 phases. Phase I: 36 purposively selected health workers (facility administrators, system administrators, Public Health Officers, information officers, clinicians, physicians and nurses) across 6 health facilities (MediHeal, St Luke, Fountain, Huruma, Pioneer and MTRH) were interviewed for assessment of current surveillance approaches as well as assessing EMRS’ readiness for syndromic surveillance. Thematic analysis was used to identify the prevailing approaches of disease surveillance, the challenges faced as well as the gaps and opportunities that exist for the implementation of syndromic surveillance with the existing EMRS. Phase II: Findings of the surveillance and readiness assessments were then used to inform the design of the surveillance framework based on the real-time outbreak detection system framework. The results: Passive surveillance is primarily applied in Eldoret and is countered with a number of challenges, including untimely data acquisition and reporting, data inconsistencies and incompleteness, insufficient financing for surveillance activities and suboptimal staffing. In terms of readiness for syndromic surveillance implementation, the assessed systems were; un-interoperable, low on technical staff for systems support, inadequately financed for facilitation of additional functioning. On the other hand, there was adequate infrastructure, connectivity, policy and data standards, though not satisfactorily implemented. Based on the findings, a syndromic surveillance framework was designed whose components included a syndromic surveillance system composed of an HL7 listener, a detection engine/algorithm, a database for storage of information parsed from EMRS and a Geographical mapping tool. The system is linked to EMRS of interest via a Virtual Private Network (VPN) for secure transfer of patient level data. Data from the EMRS is parsed unto a standards normalization platform and then generated into an HL7 message which is sent over the VPN and then received by the HL7 listener for classification and analysis. Conclusion: The electronic health record environment in Eldoret still has barriers to syndromic surveillance implementation. Addressing of the observed gaps and development of adequate guiding principles is very pertinent to ensure adoption and implementation. Recommendations: Implementation of the framework needs to be done and the implementation evaluated to get a clear understanding of how best the syndromic surveillance framework can fit within the workflow of the health care facilities in low resource settings.