Before automating clinic admin, count the whole job
A practical workload audit for independent clinics: separate avoidable rework from essential review, and protect the people doing both.
United States: the work continues after submission
An insurance request may involve finding records, checking requirements, obtaining a clinician’s judgement and answering follow-up questions. Measuring only the time needed to submit a form misses much of that work. Before an independent clinic automates a task, it needs to understand where the task starts, who touches it and what counts as genuinely finished.
The AMA’s 2025 survey, published in May 2026, reported an average of 40 prior authorisations per physician each week, requiring 13 hours of combined physician and staff time. Its survey report also records perceptions of changing denial volumes. These are self-reports from 1,000 US physicians who regularly handle prior authorisations, across practice settings. They are neither a benchmark for every independent clinic nor evidence about UK workloads.
19 October 2026US physicians’ reported change in prior-authorisation denials
Perceived change over the previous five years, December 2025 survey
United Kingdom: protect capacity before promising savings
A UK clinic should build its own baseline around its actual referral, insurer, billing and clinical-record processes. It should not import an American insurance workload estimate into a staffing plan. The operational question is whether a proposed change leaves enough time for safe care, checks, training and the unexpected.
For providers carrying out regulated activities in England, CQC’s Regulation 18 guidance requires sufficient suitably skilled staff and necessary support, training and supervision. It also calls for staffing and skill mix to respond to changing needs. That is an important constraint on an automation business case. Scotland, Wales and Northern Ireland have their own regulatory arrangements; CQC guidance should not be presented as a UK-wide rule. Keep staffing assumptions provisional until local results justify changing them.
Count active work, waiting and work done twice
Start with one bounded process, such as preparing a referral pack. Over a representative fortnight, ask the people doing the work to record task volumes, hands-on minutes, interruptions, returns for missing information and time spent correcting errors. Include reception, administrators and clinicians. Keep waiting time separate: a request sitting with an insurer for three days is not three days of staff labour.
Record unfinished work as well as completed cases, and note unusual staffing or demand. Compare similar case types rather than rewarding teams for handling easier cases. A short daily team check can reveal what a timer misses: unclear instructions, duplicate entry, inaccessible documents or a decision that only a clinician can make. These observations may justify removing a redundant step before buying anything.
Make review a funded part of the workflow
If a tool prepares a draft, the trial should count the time needed to inspect its sources, correct it and approve it. Give the reviewer authority to reject the output and access to the underlying record. Clinical judgements and consequential patient communications need review by an appropriately qualified person. Set a clear route for uncertainty and a workable fallback when the system is unavailable.
For US HIPAA-covered practices, HHS cloud-computing guidance also matters before real patient information enters a proposed service. Where a cloud provider handles electronic protected health information on the practice’s behalf, the guidance addresses business-associate agreements and risk analysis. A demonstration is not proof that these requirements, or a clinic’s wider obligations, have been met.
Judge the result by total workload
Agree the test beforehand: total staff minutes per completed case, correction frequency, the age of outstanding work and work spilling beyond scheduled hours. Include training and supervision. Compare a limited pilot with the baseline, discuss the results with staff and stop or revise it if errors, delays or review pressure increase. Faster drafting can still leave the team with more work overall.
For a PRIS Care demonstration, use fictional data and bring one clearly described administrative task. Ask the team to walk through the actual workflow, including checking, correction and escalation, and verify what is currently supported. Decide what to test only after that conversation. The useful outcome is evidence that a particular process helps this clinic’s team while preserving safe human judgement.