How Computational Medicine Technologies Can Improve Health Delivery in Developing Countries

Fig. 1: Harnessing emerging technologies and analytics to increase healthcare availability (11)

Computational medicine, an emerging field that seeks to improve patient care and accessibility through technology and data, promises to revolutionize healthcare systems around the world with its breadth of applicability (1). In the last several years, we have seen the advent of advanced wearable sensors, computational genome analysis, machine learning for disease detection, and health informatics models that bring clarity to medical mysteries. In the future, the potential for applications of these technologies will only continue to grow. These novel methods cut costs associated with traditional medical treatments, while still increasing efficacy and rates of successful patient outcomes (2, 3). In developing countries where health care systems struggle to treat all patients thoroughly and equitably, these technologies can close the harmful gap in patient care and community health.

While healthcare systems in developing countries have made significant public health gains, such as controlling communicable diseases, there is much room to improve in hospital-based care. For example, the WHO recommends “at least 10 medical doctors per 10,000 people to ensure adequate coverage at the primary care level,” but Africa, on average, reports only about three medical doctors for every 10,000 people (4). The primary issues in developing countries’ healthcare systems are insufficient human resources (doctors, nurses, and other healthcare professionals), limited institutional capacity and infrastructure (below-global-average hospital beds per 100,000 population members), and patients’ ability to seek care (5). Because building infrastructure around education and accessibility to healthcare is a long-term process, the best short-term path forward for these developing countries is to invest more in outpatient care. The novel medical technologies that currently exist and are in development can be applied on a broad level to address human resource and infrastructure shortcomings.

Wearable technologies can have a massive impact. Technological capabilities have expanded beyond inpatient ECG metrics to address a wide array of sensing and intervention modalities. These use cases range from monitoring blood glucose levels with intradermal skin patches to tracking body movements using gyroscopes and accelerometers. In poverty-dense regions of the world where healthcare resources are limited, these sensors can provide high-risk patients with hands-off yet effective care.

Researchers have looked at specific cases of wearable sensors applied in various poverty-stricken areas in the world. For example, of the nearly 9 million people who live in Mexico City, about 3 million have prediabetes. And on a broader scale, roughly 4 million people in Mexico who have Type II Diabetes do not know it, emphasizing the urgency of the problem. In a large urban area like Mexico City, a single wearable sensor composed of an activity recorder and glucose level monitor could be offered to all individuals at risk of Type II Diabetes. This would provide patients with complete access to their medical data and greater control over their care. If this approach is implemented, an estimated 2 million people in the city may be prevented from developing Type 2 Diabetes, with annual savings of over $1 billion in total healthcare costs. Ideally, governments and NGOs would work to ensure these savings support the average consumer through public hospital systems and small outpatient clinics. While building sensor networks like this on a large scale requires a significant upfront investment, it is important to appreciate that the financial return on investment for intervention-based wearable sensors can exceed $25 for every $1 invested (6), not including the additional public health benefits.

Artificial intelligence and machine learning models compose another field within computational medicine with significant potential in diagnostics. Recent studies have shown that certain ML algorithms can detect breast cancer from mammogram images at an accuracy comparable with trained radiologists (7, 8). Similar models have been applied not only to medical imaging contexts but also time series data, such as ECG and EEG, and crowdsourced medical data to personalize care and improve rates of favorable treatment outcomes. Applying these models in countries where health care labor forces significantly lack in numbers relative to population would mean earlier medical interventions, and consequently lives saved. Artificial intelligence is still limited in scope and application, especially within the medical domain where false negatives bear heavy consequences. Regardless, the ability of these models to detect and diagnose various medical conditions is undeniable, as is their potential impact on human health.

If financial and logistical challenges for initial investments can be overcome, computational medicine technologies can save millions of lives worldwide in the next several years (9). This process starts with policy makers, hospital administrators, and public health officials around the world advocating for the allocation of funds to computational medicine applications. Extrapolating the recent trajectories of these new technologies suggests that associated overhead costs (hardware for sensors, infrastructure for data storage) will continue to decrease. In the short-term, increasing awareness of these novel innovations holistically is the best way to expedite applications in the parts of the world where they are needed most. As access to these technologies grows, so too will the ability of patients to take an active and effective role in their own care due to lowered barriers to quality care.

Sources

  1. https://www.bme.jhu.edu/research/research-areas/computational-medicine/#:~:text=Computational%20Medicine%20aims%20to%20advance,diagnosis%20and%20treatment%20of%20disease.
  2. https://www.ajog.org/article/S0002-9378(20)30885-1/fulltext
  3. https://dashtechinc.com/how-machine-learning-help-to-cut-costs-spent-on-treatment-and-care/
  4. https://www.livemint.com/science/health/india-has-strong-credentials-to-meet-africa-s-health-system-requirements-11626692999838.html
  5. https://www.apha.org/policies-and-advocacy/public-health-policy-statements/policy-database/2014/07/23/09/09/strengthening-health-systems-in-developing-countries#:~:text=The%20health%20systems%20in%20countries,of%20services%2C%20absence%20of%20community
  6. https://www.scirp.org/journal/paperinformation.aspx?paperid=76486
  7. https://healthitanalytics.com/news/deep-learning-may-detect-breast-cancer-earlier-than-radiologists#:~:text=The%20results%20showed%20that%20the,standard%20detection%20in%20many%20cases
  8. https://www.ibm.com/blogs/research/2019/06/ai-models-radiologist-level-accuracy/
  9. https://www.healthworkscollective.com/how-technology-can-change-healthcare-in-developing-countries/#:~:text=Technology%20can%20make%20a%20major,record%20keeping%20and%20transmission%20easier.
  10. Erikkson, J., et al. (1999) Prevention of Type II Diabetes in Subjects with Impaired Glucose Tolerance: The Diabetes Prevention Study (DPS) in Finland. Study Design and 1-year Interim Report on the Feasibility of the Lifestyle Intervention Programme. Diabetologia, 41, 793-801.
  11. https://pixabay.com/photos/x-ray-mri-ct-scan-6841384/