I have spent a significant portion of my career working with health data: surveillance data, programme monitoring data, evaluation data, administrative health data. One thing I have learned is that data is not self-interpreting. The meaning of a number depends entirely on understanding what it is measuring, how it was measured, and what was happening in the environment when the measurement was taken.
The most dangerous data errors I have encountered were not caused by fraud or negligence. They were caused by competent analysts interpreting technically sound data without adequate context.
A district with a high case notification rate may be detecting disease efficiently. Or it may have a stronger health worker reporting culture than neighbouring districts. Or it may have recently received a training intervention that temporarily increased vigilance. Or it may simply have more facilities that report. Each of these explanations has completely different implications for programme response.
A district with a low case notification rate may be genuinely low burden. Or it may have a weak surveillance system that is missing cases. The difference between those two interpretations is critical, and the data alone cannot tell you which is true.
This is why programme data should always be interpreted by people with field knowledge, not just people with analytical skills. The field context provides the interpretive framework that makes data meaningful. Without it, sophisticated analysis applied to surveillance or programme data can produce conclusions that are technically rigorous and operationally wrong.
In practice, this means building review processes where analysts and field staff look at data together. It means creating feedback loops where field staff see the analyses generated from their data and can flag when the interpretations do not match what they are observing. It means resisting the pressure to produce clean, simple stories from data that is inherently complex and contextual.
Data is an asset. But like most assets, it can be misused. The misuse of public health data through decontextualised interpretation has led to resource misallocation that has cost lives. That is a strong reason to take the interpretive discipline seriously.