How AI & Telemedicine Are Helping to Detect Disease: Improving Women’s Health in Afghanistan

Friday, February 28, 2020

Over the past ten years, EPOS Health Management has observed a dramatic shortage of qualified medical specialists in Afghanistan. While engaged in hospital renovation and construction in the northern part of the country, the lack of doctors who are able to diagnose non-communicable diseases (NCD) such as cancer became apparent: there are only eight pathologists nationwide, responsible for a population of nearly 32 million inhabitants. The low capacities for early stage disease diagnoses regularly result in high treatment costs and even in preventable deaths.

Women are particularly disadvantaged in seeking cancer diagnostic support and treatment in Afghanistan, as cultural restrictions prevent them from moving freely or from consulting a doctor on their own. Breast cancer is a nationwide concern with one in eight women suffering from the disease at a rather young age compared to Germany, statistically accompanied with high mortality rates.

EPOS’ work in Afghanistan over the past decade was focused on piloting the use of telemedicine with a solution of artificial intelligence (AI) and deep learning added in 2019 for improved cancer diagnostics. We developed an integrated modular telemedicine solution that combines different medical cooperation interfaces via a multilingual electronic medical record (EMR) which has an expandable interface to AI diagnostic cancer cell interpretation. It further offers automated radiology and skin mole interpretation support.

This integrated solution is a joint initiative of EPOS Health Management, myCare2x, a web-based open source hospital information system and EMR, iPath, a case-based collaboration platform for telemedical second opinion, and the International Virtual e-Hospital which implements activities on telemedicine, telehealth and virtual education programs in developing countries. We also offer hands-free videoconferencing with the user-friendly technology developed by AMA Expert Eye.    

We have proven to carry out breast cancer diagnoses in Afghanistan in an efficient and reliable way through our intuitive multilingual web based platform with instantaneous translation into multiple languages. It serves to overcome acute shortages of specialized medical staff in remote locations or in conflict zones and offers support by a global network of medical experts. A pilot study tested women for signs of malignancy.

The AI algorithms are continuously trained by international health specialists and have proven to be able to efficiently identify benign and malignant breast cell images entered into the system, providing a reliable patient diagnosis. It allows the limited number of local human pathologists to obtain an automated second opinion and optimize their capacities on service delivery and treatment of complicated cases. The integrated telemedicine diagnostics tool also incorporates an online pre-examination form that registers structured documentation of clinical parameters for additional risk-grouping identification.  

With the support of integrated telemedicine and AI diagnostics, the number of breast cancer histopathology diagnoses has considerably increased, while the overall age of women seeking diagnoses has decreased from an average age of above 45 in 2009 to below 25 in 2019.

The benefits detected for of applying telemedicine with an AI interface for disease diagnosis include:

  • Doctors optimize their working time through faster and more accurate diagnoses, complemented by reliable artificial intelligence algorithms. This allows a) more patients to be diagnosed and eventually treated and b) doctors/specialists to focus on health care services such as advice and treatment.

  • Misdiagnoses by non-specialists decrease, thus reducing the number of unnecessary, missed or late applied interventions.

  • Inequalities in health are reduced. Smart solutions for distance support in healthcare provide remote rural communities with access to high quality health care services, which are usually only available in the main urban centers. Overall, access of patients to high quality services increases, promoting an inclusive approach which leaves less and less people behind.

  • In general, a higher number of suspected cancer cases can be diagnosed in the country itself, reducing the number of patients that travel to neighboring countries for diagnosis and treatment, linked to high costs which are only affordable to some parts of the population.

Many low to middle income developing countries are facing similar challenges of severe shortages of trained medical specialists, or suffer from restricted movement due to conflict, natural disasters, virus outbreaks or dangerous transportation infrastructure. They could benefit from an integrated solution of telemedicine and AI in a similar way.

View a slide presentation of the AI & Telemedicine Solution Pilot Study in Afghanistan here.

Amitava Sanyal is a consultant and project manager with expertise in market and feasibility analyses for health infrastructure, healthcare facility assessment and planning, with a special focus on eHealth and the digital transformation of health services.  His experience covers planning of the rehabilitation of over 50 hospitals of varied bed capacity, and three teaching hospitals. He is currently providing advice to several projects on strengthening healthcare facilities, their telemedicine capabilities and their EMR.

Michael Weinhara is a senior consultant in hospital management and hospital information systems with extensive experience in restructuring, streamlining and merging hospitals, development and implementation of health management information system (HMIS) for primary, secondary and tertiary level facilities, and the introduction, implementation and adaptation of telemedicine and international medical coding.