New data from UK primary care for 2025–2026 reveal a striking divergence: antibiotic prescribing rates fell sharply in telehealth appointments while holding steady in face-to-face visits. The pattern, drawn from routine NHS digital records, suggests that the mode of consultation itself—not just clinical need or guideline adherence—shapes how often general practitioners reach for the prescription pad. For those concerned about antimicrobial resistance, the finding opens both an opportunity and a puzzle.

Telehealth Appointments Show Sharply Lower Antibiotic Prescribing Rates

According to NHS England's quarterly prescribing dashboard, remote consultations—phone and video—recorded an antibiotic script rate roughly 15% lower in 2025–2026 than the previous year. In-person visits, by contrast, showed no significant change over the same period. The absolute numbers are modest—a few fewer scripts per hundred consultations—but consistent across regions and practice sizes.

What might explain the drop? One plausible driver is reduced diagnostic uncertainty. In a remote encounter, clinicians cannot examine the throat, ears, or chest directly. Without clear physical signs of bacterial infection, many practitioners may default to a more conservative approach, especially for acute respiratory infections where antibiotics are often unnecessary.

Another factor is patient expectation. Patients who choose a phone call may be less invested in obtaining a prescription than those who attend the surgery. They may simply want reassurance or advice on symptom management. The remote format may also make it easier for clinicians to suggest a wait-and-see approach without the pressure of a face-to-face encounter.

Yet the same data also show that the overall proportion of consultations conducted remotely has stabilised at roughly 30–35% of all GP appointments, down from the pandemic peak but still substantial. That means the prescribing drop in that channel has a measurable effect on total antibiotic use in primary care—a sector that accounts for roughly 80% of all human antibiotic consumption in the UK.

Why In-Person Visits Resist the Downward Trend

The in-person prescribing rate, stubbornly flat, demands its own explanation. One possibility is defensive medicine. When a patient is physically present, the clinician may feel more pressure to act—to do something visible. Handing over a prescription is a tangible intervention, whereas watchful waiting can feel like doing nothing, even when it is the evidence-based choice.

Time pressure also plays a role. A typical GP appointment in England lasts ten minutes. In that window, explaining why antibiotics are not indicated, discussing symptom trajectories, and agreeing on a safety netting plan takes time that many practitioners simply do not have. Writing a script is faster. The same time constraints apply in remote consultations, but the absence of a physical exam may nudge the balance toward restraint.

Patient expectation is stronger in person. Research from the University of Bristol suggests that patients who attend the surgery are more likely to expect a prescription, and clinicians are adept at reading those expectations. Even when GPs judge antibiotics unnecessary, they often prescribe to maintain the therapeutic relationship—a phenomenon well documented in the literature as 'acquiescent prescribing'.

Finally, no new NICE guidance on acute respiratory infections has been issued in the past two years. The baseline rate of antibiotic prescribing in UK general practice has been declining slowly for a decade, but the pace has plateaued. The in-person channel, where most prescribing still occurs, appears to have reached a floor that is hard to lower without structural changes to consultation length or decision support.

The Evidence Base for Watchful Waiting and Delayed Scripts

The idea that many antibiotics can be safely withheld is not new. A 2023 Cochrane review of delayed antibiotic prescribing for respiratory infections found that it reduces overall antibiotic use by roughly 40% without increasing complication rates. Patients given a delayed script are also less likely to reconsult for the same illness.

Safety netting advice—telling patients which symptoms should prompt a return visit—is effective in trials. In the UK, the National Institute for Health and Care Excellence (NICE) recommends offering a delayed prescription or a no-antibiotic strategy with clear safety netting for most acute respiratory infections. Yet real-world uptake remains patchy.

Telehealth may inadvertently enable a safer version of watchful waiting. Without a prescription in hand, the patient is forced to observe their symptoms and decide whether to recontact. The remote channel also makes it easier to schedule a follow-up call a few days later—a key component of the delayed-prescribing strategy. But the logistics of remote follow-up are not yet standardised across practices.

A trial published in The Lancet Infectious Diseases in 2024 tested a telehealth-delivered delayed-prescribing protocol for sore throat and found it reduced antibiotic use by a third compared to usual care. The authors noted that the approach required careful scripting for triage staff and a reliable callback system—both of which are still variable in routine NHS practice.

Antibiotic Resistance Trends Add Urgency to the Gap

The stakes are well known. UK Health Security Agency data show that bloodstream infections with E. coli resistant to key antibiotics have risen steadily over the past five years. The WHO lists antimicrobial resistance (AMR) among the top ten global public health threats, estimating that 1.27 million deaths were directly attributable to bacterial AMR in 2019.

Primary care is the main driver of antibiotic selection pressure in humans. Even small reductions in prescribing can, over time, slow the emergence and spread of resistance. The telehealth prescribing drop offers a natural experiment: if the lower rate persists and is not offset by higher prescribing elsewhere, it could contribute to a measurable impact on resistance patterns at the population level.

But the relationship is not simple. Lower prescribing in one channel could be offset if the same patients later reconsult in person and receive antibiotics anyway. The current data do not track individual patients across modalities, so the net effect on total antibiotic exposure is uncertain. Modelling studies suggest that a 10% reduction in primary care prescribing could reduce resistant infections by 5–10% over a decade, but those estimates depend on sustained change.

What the Telehealth Prescribing Data Cannot Tell Us

Routine data have blind spots. The prescribing dashboard does not include linked patient outcomes—whether patients who did not receive antibiotics later developed complications, required hospitalisation, or reconsulted. Without those links, it is impossible to say whether the lower prescribing rate in telehealth is safe or whether it represents missed diagnoses.

There is also the risk of misclassification. A remote consultation coded as an upper respiratory tract infection may be a viral illness for which antibiotics are never indicated, but the same code could mask a bacterial sinusitis that would benefit from treatment. The data cannot distinguish.

Selection bias is another concern. Sicker patients may be more likely to choose in-person care, and those patients may genuinely need antibiotics more often. If that is the case, the lower prescribing rate in telehealth simply reflects a healthier patient pool, not a change in clinician behaviour. The data do not include severity scores or comorbidity indices to adjust for this.

Confounding by season and practice list size also muddy the picture. Winter surges in respiratory infections disproportionately increase in-person visits, and larger practices may have different prescribing cultures. A proper interrupted time series analysis, ideally with practice-level randomisation, would be needed to confirm a causal effect of telehealth on prescribing rates. That work has not yet been done.

Practical Levers for Clinicians and Commissioners

While the evidence base matures, there are practical steps that practices and commissioning groups can take. One is to embed delayed-prescribing defaults into telehealth templates. If the electronic health record prompts the clinician to offer a delayed script or a no-antibiotic plan for common respiratory infections, the default behaviour shifts toward stewardship.

Training phone triage staff in safety netting language is another low-cost intervention. A 2025 study from the University of Oxford found that a brief training module for receptionists and nurses reduced antibiotic prescribing in subsequent remote consultations by 12%. The key was giving staff simple phrases to use when patients asked for antibiotics: 'The doctor will want to check your symptoms first' rather than 'You'll need a prescription.'

Local antibiograms—summaries of local resistance patterns—can guide empirical choices when antibiotics are indicated. Many GPs still prescribe amoxicillin or doxycycline without checking local resistance data, which are often available on the practice intranet but rarely consulted. Integrating antibiogram data into the prescribing interface could help.

Audit and feedback cycles that compare a practice's remote versus in-person prescribing rates—and benchmark against peers—have been shown to narrow the gap. The Quality and Outcomes Framework (QOF) already includes indicators for antibiotic stewardship, but they do not distinguish by consultation mode. Adding a modality-specific target could accelerate change.

Trade-offs and Unintended Consequences

While the lower prescribing rate in telehealth is encouraging, it may come with trade-offs. One concern is diagnostic delay: a patient with a bacterial infection who is managed remotely might not receive timely antibiotics, leading to progression of illness and potentially avoidable hospitalisation. A 2024 analysis from the University of Manchester estimated that for every 100 fewer antibiotic scripts written in remote consultations, there could be roughly 1–2 additional hospital admissions for pneumonia or sepsis, though the margin of error is wide and the numbers are debated.

Another trade-off is the potential for increased reconsultation. If patients do not receive a prescription, they may call back or visit repeatedly, increasing overall workload. Some data suggest that remote consultations without a prescription have a higher reconsultation rate within 14 days compared to in-person visits that end with a prescription. However, much of that reconsultation is for reassurance and does not necessarily indicate clinical deterioration.

There is also the risk of widening health inequalities. Older adults, those with limited digital literacy, and patients from lower socioeconomic backgrounds are less likely to use telehealth. If the prescribing drop is concentrated among these groups—because they are more often managed remotely—then the stewardship benefit may be unevenly distributed, potentially exacerbating disparities in infection outcomes. A 2025 report from the Health Foundation highlighted that practices in deprived areas have higher antibiotic prescribing rates overall, and any intervention that differentially affects these populations must be monitored carefully.

Finally, there is the question of clinician satisfaction. Some GPs report that remote consultations are more stressful when it comes to antibiotic decisions, as they worry about missing something. This could lead to burnout or defensive prescribing in the opposite direction—overprescribing to avoid risk. The net effect on stewardship may depend on local culture and support systems.

International Comparisons and Lessons

The UK is not alone in seeing a telehealth-prescribing gap. Data from Australia's Medicare Benefits Schedule show a similar pattern: during the peak of telehealth expansion in 2020–2021, antibiotic prescribing per 1000 consultations fell more sharply for telehealth than for in-person visits, and the gap has persisted as telehealth use stabilised. Australian researchers have attributed part of this to the inability to perform point-of-care tests remotely, such as C-reactive protein testing, which are more commonly used in face-to-face care to rule out bacterial infection.

In the United States, a 2024 analysis of commercial insurance claims found that antibiotic prescribing rates for acute respiratory infections were roughly 10–15% lower in telehealth visits compared to office visits, but the difference was smaller for children and for patients with chronic conditions. The US data also showed higher rates of broad-spectrum antibiotic use in telehealth, possibly because clinicians lacked local resistance data and chose broader coverage empirically.

Canada's experience offers a cautionary note: a 2023 study from Ontario found that while overall antibiotic prescribing declined during the pandemic, the proportion of prescriptions written for urinary tract infections actually increased in telehealth, likely because of the lack of urine culture confirmation. This highlights that the telehealth effect is not uniform across all infection types—it depends heavily on the diagnostic process.

These international examples suggest that the UK's prescribing gap is not unique but also not automatic. It results from a combination of factors: a relatively mature telehealth infrastructure, a strong stewardship culture in primary care, and a payment system that does not incentivise volume. Systems that lack these features may see a different pattern.

The Broader Lesson for Infectious Disease Management

The UK telehealth prescribing data underscore a truth that implementation science has long recognised: the medium through which care is delivered matters as much as the content of guidelines. A well-written recommendation to avoid antibiotics for acute bronchitis will have little effect if the consultation format pressures clinicians to prescribe.

Telehealth is not inherently good or bad for antimicrobial resistance. It can amplify stewardship when designed well—with decision support, safety netting defaults, and follow-up protocols—or it can create new pathways for inappropriate prescribing if left to ad hoc practice. The current data suggest a net benefit, but that could change as remote care becomes more sophisticated and patients become more accustomed to requesting prescriptions online.

The UK experience may inform other high-income health systems that are expanding telehealth. Countries like the US, Australia, and Canada have seen rapid growth in virtual primary care, often without the infrastructure for antibiotic stewardship that the NHS has built over decades. If they can replicate the UK's prescribing patterns in the remote channel, the global impact on antibiotic use could be substantial.

The next step for researchers is to link prescribing data to diagnostic coding and patient outcomes. Without that, the debate will remain speculative. But for now, the message for clinicians is clear: the next time you pick up the phone, you may be doing more for antimicrobial stewardship than you realise. As always, this article is for informational purposes and does not constitute personalised medical advice. Consult a qualified healthcare professional for decisions about your health.