OPAT

The standing claim

OPAT exists to deliver equivalent clinical outcomes with the patient at home rather than in a bed. But the sharper, more recent insight is that the greatest value often comes one step earlier: not giving intravenous antibiotics at all where oral treatment is non-inferior. “Good” in this domain is therefore two things at once — a well-governed service for those who genuinely need IV, and the discipline to right-size who that is.

Where practice stands

The model is unusually well-codified. The UK Good Practice Recommendations specify a minimum four-role multidisciplinary team (a treating clinician, an infection specialist, a specialist nurse and an antimicrobial pharmacist), documented patient-selection criteria, at-least-weekly monitoring while on IV antimicrobials, and formal governance (Chapman et al., 2012; updated 2026). NHS England’s April 2025 guidance to integrated care boards sets expectations for prompt MDT assessment, regular review and round-the-clock access to advice, and NHS England presents OPAT as substantially cheaper than equivalent inpatient care and a large source of avoided bed-days. Those are service estimates; we have not been able to check the figures at source, so we do not repeat them here. The knowledge is not the constraint; provision is uneven across systems — a capacity and equity gap, not an epistemic one.

What good looks like

Three things define it, on the evidence:

1. The governance standard. The BSAC recommendations and the IDSA 2018 guideline converge: a defined MDT, mandatory infection-specialist review before OPAT, structured patient selection, and serial laboratory monitoring (IDSA makes infection-specialist review and serial monitoring strong recommendations).

2. Right-sizing IV — the OVIVA pivot. According to PubMed, the OVIVA trial (Li et al., 2019 — a 26-NHS-hospital randomised non-inferiority trial, n=1,054) found oral antibiotics non-inferior to intravenous for bone and joint infection (treatment failure ~13% oral vs ~15% IV; non-inferiority met), with far fewer IV-catheter complications (9.4% vs 1.0%) and around £2,740 saved per patient with no difference in quality-adjusted life-years (Scarborough et al., 2019). Follow-up was one year, so late recurrence cannot be excluded. So best practice increasingly means avoiding IV where oral suffices, with NICE NG15’s 48–72-hour review of IV antimicrobials, and national IV-to-oral switch criteria, as the upstream filter that decides who reaches OPAT at all.

3. Agent selection for the patients who do need IV. Choosing once-daily, OPAT-friendly agents matters, but convenience is not the same as equivalence, and the evidence here has moved against it. A 2022 meta-analysis of twelve retrospective cohort studies found no statistically significant difference between ceftriaxone and standard care for meticillin-sensitive Staphylococcus aureus bloodstream infection, with wide confidence intervals (Alsowaida et al., 2022). A 2026 meta-analysis of eleven studies (2,568 patients) found ceftriaxone associated with significantly higher 30-day mortality (OR 3.33, 95% CI 2.17–5.10), though not at 90 days, and its authors conclude that routine use, especially as initial therapy, is not supported (Maraolo et al., 2026). Both rest on observational studies, so neither is decisive, but the more recent and more specific signal points towards harm: ceftriaxone for this infection is a question for a randomised trial, not a default.

The exemplar is the UK itself, chosen on trial outcome rather than reputation: OVIVA, the trial that reframed practice, ran in 26 NHS hospitals, and the UK good practice recommendations draw international input (the 2026 update includes co-authors from the US, Australia and France). Australia’s Hospital-in-the-Home programmes and the US model are the obvious comparators; we have not yet assessed their outcome data, and the US model’s different reimbursement structure limits how far it transfers to a tax-funded system.

Through the lenses

Through the VAT lens, an intravenous line is often exchange-value activity — visible, billable, reassuring — that can be avoided without loss of cure; the use value is the cure achieved with days spent at home and fewer line complications. This is McGilchrist’s caution in miniature: do not mistake the intervention (the line, the activity) for the thing it was meant to serve (the patient cured). Cross-domain, OPAT and oral-switch are admission-avoidance levers that connect directly to acute frailty (D01) and depend on community nursing capacity.

Confidence and limits

OVIVA is high-quality randomised evidence (open-label, with blinded adjudication of outcomes) — but specific to bone and joint infection; it should not be over-generalised to all infections (endocarditis behaves differently). The UK and IDSA documents are consensus and guideline standards, built on literature review and expert consultation rather than trials of the service model. The ceftriaxone evidence is observational and now conflicting, with the newer synthesis showing a short-term mortality signal. NHS England’s cost and bed-day claims are service estimates, not trial outcomes. And the central limit is structural: the model is far more complete on paper than in uniform delivery — the capacity gap is real.

What we are watching

Uptake of the IVOS criteria (how much IV is safely avoided upstream); the equity of OPAT provision across systems; and whether the oral-switch evidence base extends robustly to infection types beyond bone and joint.