How to Read WASH Evidence Without Overclaiming
This chapter turns the WASH score-versus-outcome distinction into an evidence-reading method. The focus is how to compare numbers from the Indonesian hospital study and the WHO/UNICEF JMP report without manufacturing a trend, and how to choose verbs that match what each design observed. The practical result is an audit that protects two ideas at once: monitoring can guide real improvement, and a useful minimum indicator should not be asked to prove more than it measured.
Seven questions before two percentages meet
A WASH percentage is never just a percentage. Before comparing two values, identify at least seven pieces of metadata.
What is the unit and denominator? The result may be the proportion of hospitals, all health-care facilities, a facility subgroup, or a population-weighted estimate. The Indonesian study analyzed 532 hospitals. A JMP global estimate combines country-level information and weights regional and global aggregates using population data because comprehensive counts of facilities are not available everywhere. Those are different denominators even when both results use the word “facility.”
What is the reference period? The Indonesian study used the 2019 Health Facility Research and outcome information from 2018. The JMP report gives estimates for 2025 in a 2026 publication. A year assigned to an estimate does not necessarily mean every observation was collected in that year. The JMP method can extend estimates beyond the most recent data point under defined rules, so “2025 estimate” should not automatically be read as “survey conducted in 2025.”
What definition is being applied? The Indonesian study classified proxy scores as poor, adequate, or good, with good above 75 percent. The JMP basic level is a standardized service ladder built from specified conditions. A proxy score can combine available items differently from a service-ladder threshold. The word “good” is therefore not a translation of “basic,” and neither label is self-explanatory without its indicator definition.
What is the threshold or limiting rule? A composite score may average or aggregate several items; the JMP composite service level can be constrained by the weakest available primary indicator. One method may show broad availability while another is held below its basic threshold by a missing essential component. A higher percentage under one rule is not automatically stronger evidence than a lower percentage under another.
How much of the relevant population is covered? The 2026 JMP report's global coverage differs by domain: the countries with estimates represented 52 percent of the world's population for basic water and waste management, 48 percent for sanitation, 46 percent for hygiene, and 40 percent for environmental cleaning. These coverage figures do not mean that the remaining facilities have no service. They mean the available evidence is not sufficient to produce the same estimate for them.
How was the number estimated? Is it a direct survey result, an administrative total, a census, a simple average, or a modeled estimate? The answer affects how much uncertainty and smoothing are present. A monitoring estimate can be the best comparable summary available while still being less direct than a facility-level observation.
What outcome is being measured?