MIT Sloan Health Systems Initiative

Prize-winning Research Reduces Transportation Inefficiencies

Problem 

Professor Jónas Jónasson and PhD student Emma Gibson’s research focuses on transporting medical samples from geographically dispersed healthcare clinics to central laboratories in Malawi. The system comprises more than 700 clinics and 10 laboratories in Malawi’s 27 districts. For the purposes of this work, the researchers focused on 51 facilities and three laboratories in three of the country’s districts. While the setting is specific, their research could be applicable to other resource-constrained systems with transportation difficulties, limited data, and operational inefficiencies.

In Malawi medical sample transportation is provided by Riders for Health through a national network of motorcycle couriers. Riders for Health – Malawi has a fleet of 86 vehicles (79 of which are motorbikes), and employs 111 staff, 80 of whom are riders, who cover two million kilometers a year. Riders for Health transports both the samples and the paper-based test results among the healthcare clinics, district offices, and central laboratories.

The current system suffered from inefficiencies due to a fixed pick-up schedule. Drivers would report to one of the district hubs and then head out to the healthcare clinics. Each clinic was visited once or twice a week regardless of need. Clinics with a high demand for sample transport services to the central laboratories were treated similarly to less busy sites. While samples wouldn’t be left behind for lack of pick-up capacity, sometimes high-volume clinics had samples waiting several days for transport to the lab. Approximately 30% of Riders’ visits to the healthcare facilities were unnecessary and there were long delays returning paper sample results to the clinics.

What’s New 

Jónasson’s and Gibson’s research goal was to redirect the drivers to reduce the average turnaround time for results to reach the healthcare clinics. They modeled and tested an intervention that accounted for demand variability by replacing the fixed schedule with one that mapped the best routes daily, so visits were scheduled in response to the current demand at each facility.  This system, the Optimized Sample Transportation (OST) system, is comprised of two components: a novel data-sharing platform to monitor incoming sample volumes at healthcare facilities, and a comprehensive optimization model that generates daily transportation schedules.

The OST system does not require new hardware at healthcare facilities, a software installation or download, or even electricity. Communication between the different players takes place on healthcare workers’ mobile phones via a text-based tool that was already in common use. The researchers purposely wanted to use what was already available. Using existing tools and organizations may have also helped with implementation and acceptance by the healthcare clinics and laboratories. 

Jónasson and Gibson began by testing various optimization algorithms on a simulated version of the sample transport system. Following the successful test, the researchers rolled out the new system in three districts. To judge the success of the system, the researchers focused mainly on transportation delay reduction, delays that are related to the transportation of samples and results. They counted delays as the number of days that the samples or results wait for someone to pick them up each time they need to move to a new location. They focused on transportation delays since the sooner results are returned to patients, the sooner they can receive appropriate medical care.

Healthcare clinic personnel were concerned that, with the new system, some sites would be left out if the data or calculations were wrong. To address this issue, the new system was rolled out with conditions and guarantees in place. For the first stage, every healthcare site was promised that no more than seven days would pass between Riders for Health pickups. Even with this constraint, unnecessary trips to healthcare facilities were reduced from 24% without the new system to 10% with the guarantee of less than a week between visits.

Initially, the researchers were concerned about their new platform disrupting the longstanding operations at Riders for Health. Fortunately, the transporters engaged enthusiastically with the research team and provided valuable feedback. Riders made the platform more responsive by making changes on the fly in response to roads being closed or flooded, and to changes in demand from the healthcare clinics.

Results

The OST successfully reduced delays returning results to patients. Prior to the system implementation, with the original fixed routs, the average transportation delay per sample was about 6.44 days. After the OST system was operational with optimized routes, the average delay decreased to around 4.83 days per sample. The overall average reduction of 1.61 days is a reduction of 25%. 

Reduction in delays also resulted in fewer unnecessary trips. As the research team relaxed the guarantees and allowed more flexibility in the optimization, the number of unnecessary trips decreased from about 25% to about 2%. This means that when the couriers were using fixed routes, around 1 in 4 visits to healthcare facilities were unnecessary. By 2020, use of the OST system resulted in a ratio of unnecessary trips to around 1 in 50.

Impact 

The results from the implementation of the new system in three districts demonstrate the feasibility of improving sample transportation throughout Malawi. And this work may be adapted to other resource-limited setting regardless of country.  More generally, the OST system could be used to connect rural locations that don’t have existing technology for monitoring demand and inventory, regardless of location.

Connection and efficiency are the main point of the research. However, this work may also be applied to other supply chains that require quick responses, such as fresh food deliveries to rural areas or other healthcare applications such as the delivery of medications, vaccines, and medical supplies.

The OST system requires advanced data systems and optimization models; however, it is designed to facilitate communication across very rural areas without the need for computers or other technology at each facility. It’s a high-tech solution designed to operate effectively in a low-tech setting. 

With this research, Jónasson and Gibson won the INFORMS Institute for Operations Research and the Management Science (INFORMS) 2020/21 M&SOM (Manufacturing & Service Operations Management) Practice-Based Research Competition, which focuses on practical research with a real-world impact. This article is based on their paper, Redesigning Sample Transportation in Malawi Through Improved Data Sharing and Daily Route Optimization. Click here to view Emma Gibson’s presentation of their research.