[SELV/SIIL] Forecasting & Estimation: Estimating Vaccine Needs

Target release	Vaccine Module
Epic	Forecasting and Estimation
Document status	DRAFT
Document owner	Vidya Sampath
Technical Lead

Goals/Scope

This functionality seeks to provide users of the system a way to estimate vaccine product need across a series of variables. Successful acceptance of this functionality would involve the following:

Estimation of product is done considering at least the following factors in its calculation (target population for that particular vaccine, buffer stock level, wastage rate level, number of doses in the vaccination schedule) and potentially historical consumption, CCE functioning status, CCE capacity

Background

Effective vaccine management requires accurate vaccine forecasting and needs estimation, as well as adequate stock ordering that follows the delivery schedule, in conformity with cold chain capacity. This page focuses on the former and adequte stock ordering is captured under the "delivery/supply requirements" pages.

Forecasting/Needs estimation can be calculated using:

Target population provided by MoH, usually based on census projections
Buffer stock and acceptable wastage rate policy set by MoH/EPI Program
- Can different countries have different wastage amounts? If so, to what level (do disticts people have different ones from national?)
Previous consumption (more likely to be considered at national level but ideally should be reviewed by lower levels as well as part of data analysis/performance improvement cycles);
- Are there preset formulas for this? What level of flexibility is needed?
Available CCE capacity and functionality status
Time period (40 days, two weeks, 1 month, quarter, year, multi-year, etc)

Assumptions

Users will primarily access this information in close conjecture with the features in the "delivery/re-supply domain"
In push based SCs, users will reference this information when estimating how to package deliveries to an area and load the cold box/distribution truck
In pull based SCs, users will reference this information when estimating how much to request for their next order
In both scenarios, it is likely that while users will "View" this information online, they would also take a print out of the page to show in meetings, with supervisors etc so that needs to be considered

User Stories

#	Title	User Story	Label	Importance	Notes
1	Calculate Needs Estimates for Facilities in the system	As a district EPI supervisor and/or as an Intermediate Vaccine Store Manager, I want to calculate the needs estimates for all the health facilities under my responsibility by using a series of variables (See list under "background" above) so that I can ensure the stock in my district depot is enough to meet their needs via next scheduled re-supply and so I know when to place a requisition order from the level above for additional stock.	SELV does part of this	Must Have	The accuracy of vaccine forecasting and needs estimation depends upon both the level of implementation (national, district or local service delivery) and the time period of estimation used (month, year, multi-year). Whichever method is applied, the accuracy of the estimation will depend upon the quality of data used; this is a pressing issue in the iSC space, and which we will cover under the "Reporting" Domain when looking at ways of reporting the "In-Full Order Delivery" indicator. From the list above, the available CCE capacity is least consideration because there is usually sufficient storage at facility level, but functionality status is important to be included.
2	Calculate Needs Estimates for Storage Points in the system (outside of HCs)	As an intermediate vaccine store manager, I want to calculate needs estimates for all the storage points dependent on me (district depots, zonal depots, etc) using the variables listed under "Background" so that I can ensure the stock in my warehouse is enough to meet their needs via next scheduled re-supply and so I know when to place a requisition order from the level above for additional stock.	SELV does part of this	Must Have	Point 1 from above also valid here Avail CCE capacity is relevant here, because depots might not have sufficient storage for vaccines needed all the facilities under them so the level above needs to factor that into calculations on how much to send.
3	Dynamic calculation of needs estimates	As a district EPI supervisor or Intermediate Vaccine Store Manager, I want to have a dynamic way of calculating needs estimates where I can turn on or off the variables listed above to see which ones have a better accuracy for predicting correct estimates so that I can inform MoH policy and guidelines	SELV does not do this	Big debate	This is related to point 1 made above. The discussion in the iSC world is that instead of relying on static census figures to calculate needs, which ALL agree is wrong, maybe we can test a bunch of variables and over time determine which ones for that country is doing a good job in estimating need. The one that gets most agreement is using previous consumption (look at last X months of usage, look at same time last year, same time last two years etc) data to inform needs for this time period.

Diagrams

Determine the need:

Preset formula from WHO

Catchment size for facility (concern around this because census is old so their is concern)
Wastage rate for that vaccine (set by WHO)
How often a child should be vaccinated
Safety stock that needs to be there

Dependencies

Description	Link

Open Questions

Below is a list of questions to be addressed as a result of this requirements document:

#	Question	Status
1	From Tech/Brandon: How much is OpenLMIS intended to be the "planning tool" that people use to plan these amounts? If the planning process and the formulas and policies are highly variable by country and program/funder, then maybe it's better if the planning happens outside OpenLMIS, and OpenLMIS simply gets informed about the target amounts/ideal stock amounts.	Open
2	Can we take the code from OpenLMIS 1 & 2 and bring that into v3?
3	Are lead times a thing? Estimating that you'll be out of stock in X days and that it will take Y days to receive new shipment.