Supply chain system for rural health facilities
Electronic Information Management System (eIMS). A open-source, comprehensive system to track and manage health commodities (i.e. drugs, syringes, condoms) from the central to rural health facility level
Supply chain for health is hard in lower-to-middle-income-countries (LMICs), and most public health systems experience “stock outs” (i.e. patient with X can’t receive medicine for X) of key medical supplies at facilities, as ordering between facilities and warehouses is often cumbersome.
It's also difficult to forecast inventory and manage contracts with suppliers (ex. pharmaceutical companies). Many times, a stock out occurs because someone has placed a drug order against a supplier contract which has expired!
Few countries with free public health systems have real-time access to data at health facilities. Many employ pen and paper with monthly reporting. Infrastructure (internet, power) is hard and tech interventions (complex tablet apps, SMS reporting) often return subpar data quality.
CHAI had launched two prior tablet versions in Nigeria and Mozambique, so I wanted to use their lessons learned when starting our design phase.
Our initial field tests confirmed that the Nigeria application was too complex for users, which was why Mozambique had opted for a simpler UI. But Mozambique was highly customized to their stock cards, so we opted to explore a similar approach that was more tailored to Uganda’s paper forms.
The Ministry of Health in Uganda mandates that health facility workers log stock transactions on a stock card. Additional forms are used for orders, delivery receipt, and monthly stock counts.
So I used these forms as the basis for our UI design, while incorporating some of the lessons from other rural mobile/ tablet applications (ex. big buttons, navigation hierarchy limited to 1-2 steps max, limit number of usernames/ passwords given out to simplify account management).
We had a week of field tests 6 hours north of Kampala, in Soroti district. I put all the versions into Invision as clickable prototypes, and then asked health workers to complete certain tasks like processing an order or logging a single outbound item.
It wasn’t a perfect process, but we quickly saw that the stock card approach was a necessary step to have high usability. Additionally, we were surprised to learn that an omnipresent navigation bar – at the top of the screen instead of the bottom (since it was sometimes hit by dragging fingers or palms) – was a better approach.
Each night I iterated the prototypes a bit based on the day’s learnings. We tested the versions with the lowest level health facility workers (one-man shop facilities where most workers had never seen or used a touchscreen device) up to larger regional hospital admins.
These visits informed both the user interface and the overall pilot and implementation planning. We gathered data on the most prevalent (and reliable) telecom networks, solar and grid power availability, and manual as-is user processes.
We opted for tablets because we witnessed health care workers moving around a great deal while processing stock transactions, but a mobile phone would be easier to lose / steal and past inventory-focused technology pilots indicated poor data quality on cell phones (rare that more than 10-15 data points were entered before it became obvious that health workers typed the same, incorrect entry to speed up the input process). And there are a surplus of solar options that allow for relatively cheap charging (but not necessarily the consistent power supply needed for desktops).
After we settled on the front-end design and business process requirements, we designed a pilot in one district and continue to monitor and tweak usability based on a number of inbuilt monitoring & evaluation tools.