OPS.21:4.2 - Choose evidence that could distinguish improvement from a better score

Describe an observable good result and a plausible case in which the proposed indicator improves but that result does not. A higher completed-job count can coexist with unaccepted work, repeated failures or delayed complex jobs. Retain the observation that distinguishes these cases: acceptance, later return, case mix or waiting of the affected cohort. OPS.15 develops the observation account when its population or meaning is missing.

State what is counted, whose cases are included, when observation begins and ends, and how unfinished work remains visible. Choose the horizon from the recipient’s need and the time in which action could change the result. An annual reporting date can remain while an urgent operating response needs a shorter interval.

Select only observations that can change interpretation or action. An early indicator can prompt intervention; a later result can test whether the intended benefit obtained. Where a numerical proxy is inadequate, describe the result to be examined and the evidence and competent judgement needed to recognize it. Do not invent a score to replace an unavailable professional criterion.