A meta-analysis for medical health records

Decision-making related to public health relies heavily on concrete evidence for resource allocations, policymaking guidelines, and medical intervention. Statistical methodology like meta-analysis forms a powerful tool to generate actionable insights from heterogeneous dataset related to public health. Electronic Medical Records (EMRs), including the clinical data, medication history, test results from laboratories, and medical history of patients, are all managed and preserved by EMRs. An EMR is primarily used to track patient care, establish objectives, document treatment, and assess the outcomes of that care. Private Health Information (PHI) from the patient is included in these medical records, which might reveal the patient’s identity. Thus, in order to prevent the disclosure of PHI, it is necessary that PHI phrases be disguised and protected before being made available to the general public. A synergetic integration of EMRs into a meta-analytic framework with Big Data and other complementary methodologies has been explored in this chapter with more focus on de-identification (de-ID) methods. The key challenges were data variability issues, subgroup analysis, Big Data integration, and Artificial Intelligence (AI), which insisted systematic reviews. De-identifying medical records safeguards patient privacy by deleting identifiable information. The manual and automatic de-ID methods and tools are also discussed. The chapter also covers the drawbacks and difficulties of using AI to the de-ID process. Eclectic applications of de-ID of EMRs are discussed. We propose a thorough review of the prior art in de-ID methods for EMRs.