Lot Quality Assurance Sampling Measures Performance
What originated in the 1920s as a quality control methodology in the manufacturing sector is proving to be quite useful in the public health context, and increasingly constructive to MEASURE Evaluation’s in-country work. Lot quality assurance sampling (LQAS) assesses the quality of a product and can be used to see how health indicators are performing.
According to the World Health Organization, LQAS was initially used to determine whether a batch, or lot, of a product met the desired specifications. Manufacturers took a sample of the product and defined how much risk they were willing to take for not inspecting each item. Then they accepted or rejected the entire lot based on these risks. Many manufacturers began to prefer the LQAS approach because it did not require that every single item be checked for defects. The only outcomes in this type of sampling are “acceptable” or “not acceptable,” and there are no varying levels of unacceptability.
In the 1980s, the method was adapted for the public health context.“The same concept has been taken to measure how a program indicator is performing,” Stephanie Watson-Grant, Country Portfolio Manager at MEASURE Evaluation, explained in an interview about the methodology’s use in Liberia earlier this year.
LQAS is considered to be a relatively rapid and inexpensive approach to data collection in lieu of traditional surveys. It allows for small sample sizes and more frequent sampling than standard probability surveys. In the LQAS application, a pre-defined area is sampled. If that sample comes up with an indicator deemed as performing acceptably, that indicator as a whole is deemed as acceptable. If not, then the indicator is not performing acceptably. “It really gives you an idea of whether or not a program is meeting its objectives and targets,” Watson-Grant said. “And it lets you target efforts and perhaps resources to program areas that are not meeting the acceptable level.”
The standard LQAS sampling frame includes five supervision areas or “lots” and 19 random sample points in each supervision area. A typical way of applying this sampling frame in a community-based health study would be having five supervision areas in a county and a simple random sample of 19 households (sample points) in each county, for a total of 95 total interviews per county.
Janine Barden-O’Fallon used the methodology with the Child Indicator Survey in Kenya. She spoke of the methodology’s advantages in a phone interview. “It’s designed to assist decision makers at a local level by giving them monitoring information,” she said. “They can do it every year. So, for example, if there’s a five-year project, they can do this every year to get an idea of how they are doing and to change things in response before the end of the project. Decision makers and program implementers can get that information rather quickly and in time to actually impact current programming.”