BACKGROUND
In the USA, 5.6% of the population is anemic; 1.5% has moderate-severe anemia. Globally, 1.62 billion people are affected by Hb diseases. Clinical assessment of hemoglobin levels, using the cyan-methemoglobin method is reliable, but the process is not portable, results are not immediately available, and this test is unaffordable, and cost-ineffective for most patients in low- and middle-income countries who might benefit. When medical facilities and financial resources are available, frequent repeated testing is less than convenient under this method. In the presence of serious illness, with demands for repeated testing, the delays in obtaining results and the associated blood loss are particular drawbacks of this testing method. In these multiple circumstances, the potential advantages of a non-invasive, point-of-care (POC) method for hemoglobin measurement are clear. There are currently commercial non-invasive POC tools available for non-invasive Hb measurements. Most of these tools have one or more of the following limitations:1) challenging data collection methods; 2) complex data analysis and feature extraction processes; 3) affordability and portability; and 4) lack of user-friendliness and costly external modules.
OBJECTIVE
Investigate several hemoglobin measurement techniques based on smartphone devices, which have encountered significant challenges in theoretical foundation, data collection methods and sensors, data-signal analysis processes, and machine-learning algorithms. Our objective is to We identify these issue, define specific recommendations for practical solution development, and offer a conceptual framework for a noninvasive hemoglobin level estimation system using different types of smartphones and cloud computing paradigm.
METHODS
Growing interest and potential low cost of non-invasive hemoglobin measurement solutions has encouraged their development for low-resource settings, where the use of smartphones has increased rapidly. In such settings, the smartphone offers the possibility of an affordable, portable, and reliable point-of-care tool with leveraging its camera capacity, computing power, and lighting sources. However, several hemoglobin measurement techniques based on smartphone devices have encountered significant challenges in theoretical foundation, data collection methods and sensors, data-signal analysis processes, and machine-learning algorithms. We address these issues to define specific recommendations for practical solution development. Finally, we offer a conceptual framework for a noninvasive hemoglobin level estimation system using different types of smartphones and cloud computing paradigm.
RESULTS
Based on the foregoing review and other considerations, we suggested methods for body site selection for signal acquisition, response or signal, signal processing, theoretical foundations, feature generation, and machine learning algorithm selection. We also propose a conceptual framework for a noninvasive hemoglobin level estimation system using different types of smartphones and cloud computing paradigm.
CONCLUSIONS
As a growing widely available computing platform, the smartphone offers an alternative, non-invasive point-of-care tool to traditional measurements of blood hemoglobin. We recommend fingertip as a data collection site due to easy access, use of three different NIR lighting sources, specific signal processing techniques and feature selection methods, and region of interest selection methods, for the optimal development of an accurate hemoglobin prediction model. We point out the theoretical foundation, which can be applied for the identification of several blood constituent levels non-invasively. We suggest a conceptual framework for a non-invasive Hb level estimation system using different types of smartphones and a cloud computing paradigm. Investigators need to consider the following issues before developing such a Smartphone-based POC tool: (1) Cost of the smartphone, the external device, reagents if needed, and training, internet, and cloud implementation. (2) Patient’s other physiological features. (3) Allowing the user to do multiple checks and be challenged with a minimal cognitive load. (4) Including the user’s location, sex, and age in the stored record. (5) Keeping the external device as optional so that a user can run a diagnostic without the device. (6) Creating an external device that is cost-effective, easily attachable, properly fit with the finger, and user-friendly.