Routine health surveillance data collection within the region faces major problems because it contains multiple inaccuracies and inconsistent information. The National Vector Borne Disease Control Programme of Punjab India achieved data cleaning success through a systematic approach in a research study. Science professionals built a logic model that used rule-based methods with semi-automated mechanisms to conduct interactive data screening and diagnosis before editing datasets. This methodology achieved successful cleaning and imputation of data for more than 96% of records in 2015 and 99% of records in 2016 which resulted in full data cleaning success in later years. The transformation process of raw data into analysis-ready datasets through such methods serves to improve the reliability of health research.

Advanced technology integration with machine learning (ML) provides promising opportunities to enhance data quality systems. The Machine Learning Data Quality Assurance (ML-DQA) framework demonstrates the capability of advanced technologies to enhance data quality. Across multiple ML projects dealing with over 247,000 patients the ML-DQA framework automated the identification process of data quality issues while streamlining the data quality check approach through unified systems. A typical project included 5.8 personnel from both clinical and data science backgrounds who altered or eliminated 23 distinct data points during their quality assessment process. Teams work together using a systematic process to demonstrate how technology and human experience need to align for maintaining reliable data integrity.

A Distributed Ledger Technology implementation in Bangladesh provides a modern approach to overcome data quality issues. A proposed blockchain platform establishes an integrated environment for healthcare providers in the public and private sectors which maintains security throughout the platform. Healthcare entities enjoy secure data sharing and digital agreements through an infrastructure that uses data immutability with smart contracts. The design of this system contributes to developing a national data warehouse while enabling systematic health data collection and analysis processes. Such an approach provides enhanced data security as well as privacy while enabling intersectoral healthcare collaboration which leads to better public health outcomes.

The evaluation of metadata emerges as a fundamental step for guaranteeing data quality because metadata provides necessary contextual information to raw data. The novel framework provides a systematic evaluation method for metadata quality assessment specific to epidemiological and public health research purposes. This framework encompasses general information, tools and technologies, usability, and management and curation. Different research testing conditions demonstrate that this framework effectively evaluates metadata quality to build more dependable healthcare research data.

Research accuracy improvement in Bangladesh along with other areas depends on achieving top-notch data quality standards. Researchers solving regional challenges can achieve this by implementing systematic data cleaning alongside advanced machine learning frameworks with secure blockchain data-sharing technologies. Better health data accuracy creates better health intervention delivery while enabling improved healthcare policies to foster healthier South Asian communities. The process of improving health is directly dependent on the quality of collected data.

Reference : 

1. https://www.jstor.org/stable/26948749 

2. https://pmc.ncbi.nlm.nih.gov/articles/PMC11555453/ 

3. https://bmchealthservres.biomedcentral.com/articles/10.1186/s12913-020-05322-2 

4. https://pmc.ncbi.nlm.nih.gov/articles/PMC8271208/ 

5. https://www.jmir.org/2024/1/e48294 

6. https://www.researchgate.net/publication/352537235_Data_Resource_Profile_Understanding_the_patterns_and_determinants_of_health_in_South_Asians-the_South_Asia_Biobank 

7. https://www.researchgate.net/publication/351980304_Big_data_and_predictive_analytics_in_healthcare_in_Bangladesh_regulatory_challenges 

Written By:
Jarin Tasnim Rafa

Assistant Content Lead, BIIHR