This paper investigates the use of machine learning models for detecting haemoglobin (HB) levels and screening for anaemia in clinical healthcare. It employs metrics like precision accuracy, recall, delay, deployment cost, and scalability levels to compare models such as support vector machines, random forests, logistic regression, and deep learning architectures. The paper introduces the aggregated HB and Anaemia Detection Rank (HBADR) for efficient model analysis. Motivated by the need for efficient and accurate HB detection and anaemia screening, the study aims to address the limitations of traditional methods which are often labour-intensive and time-consuming. Experiments are conducted on a large dataset from various clinical settings, assessing models based on their ability to accurately predict HB levels or anaemia cases. The paper also considers the practicality of deploying these models in real clinical settings. The HBADR is proposed as a tool to facilitate comprehensive model evaluation, allowing healthcare practitioners and researchers to compare models and select the most suitable for their needs. The research has broad applications, enabling clinicians to select effective machine learning models for HB detection and anaemia screening, saving time and resources. The HBADR can also guide healthcare organizations and policymakers in adopting specific models. Additionally, the study contributes to the advancement of automated diagnostic systems and patient care in the field of machine learning in healthcare. It explores use cases like screening programs for high-risk populations, remote monitoring of HB levels, and integration into electronic health record systems. In conclusion, this study provides important insights and recommendations for applying machine learning models to HB detection and anaemia screening in clinical settings, leading to improved patient outcomes.