The research team, led by Joanne Reekie and Bette Liu, exploited Australia’s extraordinary national data infrastructure to answer two linked questions: whether the country’s zoster vaccination program actually reduced disease at the population level, and how well the current recombinant zoster vaccine protects the people it targets. Using the Person Level Integrated Data Asset managed by the Australian Bureau of Statistics, the investigators linked records from the Australian Immunisation Register, the Pharmaceutical Benefits Scheme, the Medicare Benefits Schedule and death registrations, with data available through 30 June 2025. Because zoster diagnoses are not captured directly in these registries, the team used dispensed prescriptions for antiviral drugs specifically indicated for zoster treatment, including aciclovir, valaciclovir and famciclovir, as a surrogate marker of disease.

The ecological arm of the study tracked quarterly antiviral prescription rates across the adult population from January 2014 to June 2025, spanning the entire history of Australia’s national zoster vaccination effort. That effort began in November 2016 with a single dose of the live-attenuated zoster vaccine, known by the brand name Zostavax, offered to people aged 70 with catch-up for those aged 71 to 79. The live vaccine was never an option for immunocompromised people because it contains a weakened but living virus. In June 2021 the adjuvanted recombinant zoster vaccine, Shingrix, became available in Australia, and in November 2023 it replaced the live vaccine on the National Immunisation Program, with free eligibility extended to everyone aged 65 and over and to adults aged 18 to 64 taking selected immunocompromising medications.

The prescription data told a striking temporal story. Rates of zoster antiviral dispensing among 70 to 79 year olds fell from 3.84 per 1000 population in the final quarter of 2016 to 2.61 per 1000 by early 2018, coinciding with the introduction of the live vaccine. After the recombinant vaccine joined the national program in late 2023, declines appeared in every eligible age group: rates in 65 to 69 year olds fell from 3.16 to 2.32 per 1000, in 70 to 79 year olds from 3.07 to 2.30 per 1000, and in those aged 80 and over from 3.79 to 2.81 per 1000 by mid-2025. Crucially, prescription rates among adults aged 18 to 64 remained essentially stable across the entire eleven-year observation window, a pattern consistent with the declines being driven by vaccination of the targeted older groups rather than by some population-wide shift.

Not every trend mapped neatly onto program milestones. Between late 2021 and mid-2023, prescription rates declined across all age groups, a period that corresponds to no change in zoster vaccination policy and to only limited private uptake of the newly available recombinant vaccine. The authors suggest the COVID-19 pandemic and reduced healthcare-seeking behaviour during that period as a plausible explanation. Rates in younger adults returned to pre-2021 levels by late 2023, while rates in those aged 65 and over continued downward, a divergence that aligns with the expanded recombinant vaccine eligibility. The data also carry a reassuring message about childhood chickenpox vaccination: despite hypotheses that reduced natural boosting with wild-type varicella might raise adult zoster risk, prescription rates among 18 to 49 year olds stayed low and stable throughout the study period.

To measure vaccine effectiveness directly, the team constructed two cohorts. The first comprised 4,488,937 adults aged 65 and over on 1 November 2023, the day the recombinant vaccine entered the national program. Most, 64.6 percent, had no record of zoster vaccination at baseline, while a third had received the older live vaccine, mostly those aged 70 and above who had been eligible under the earlier program. During a median follow-up of 607 days, 71,985 individuals, or 1.6 percent of the cohort, had a record of zoster antiviral treatment. Uptake of the new vaccine was substantial but incomplete: among those without prior recombinant vaccination, 42 percent received at least one dose during follow-up, and 35 percent completed the two-dose series by 30 June 2025, with a median interval of 98 days between doses.

The effectiveness estimates in this older cohort were solid. The incidence rate of zoster was 12.84 per 1000 person-years among unvaccinated individuals, compared with 4.94 per 1000 person-years among those who had received two doses of the recombinant vaccine. After adjustment for age, gender, general practitioner visit frequency, comorbidities, prior zoster treatment and immunocompromising medication use, two-dose effectiveness against treated zoster was 63 percent, with a single dose providing 54 percent. The older live vaccine performed considerably worse: 37 percent when given within the previous five years and only 20 percent when given more than five years earlier, a waning pattern that mirrors international evidence and underpins the global phase-out of the live vaccine. Effectiveness of the recombinant vaccine was consistent across the 65 to 69, 70 to 79 and 80-plus age strata, and rose slightly to 69 percent in a subgroup analysis restricted to people with no prior zoster diagnosis.

The second cohort examined 239,002 adults aged 18 and over who were immunocompromised by virtue of taking selected immunomodulatory or immunosuppressive medications for at least two consecutive months, ranging from moderately to highly immunocompromising therapies. This group fared distinctly worse on both counts. More than 80 percent were unvaccinated at baseline, and by the end of follow-up only 22 percent had completed two recombinant vaccine doses, though the figure concealed a sharp age divide: 44 percent among those 65 and over, who were eligible for free vaccination regardless, versus just 10 percent among those aged 18 to 64, whose eligibility for the moderately immunocompromising categories only opened in September 2024. Overall zoster incidence in this cohort was 17.59 per 1000 person-years, and two-dose effectiveness was 43 percent, with one-dose effectiveness of 36 percent and the live vaccine providing a mere 16 percent.

Most striking was the age gradient within the immunocompromised cohort. Two-dose effectiveness was 48 percent among those aged 65 and over but only 22 percent, with a confidence interval approaching zero, among those aged 18 to 64. The authors offer several non-exclusive explanations. Younger immunocompromised patients may genuinely mount weaker vaccine responses because they receive more potent immunosuppressive regimens; the study found that 41 percent of immunocompromising medications dispensed to the younger group were classified as highly immunocompromising, compared with 31 percent in the older group. Alternatively, the surrogate outcome may perform differently across ages, since previous Australian research suggests general practitioners are less likely to prescribe zoster antivirals to younger patients, which could dilute the measured signal. The 43 percent estimate also sits below the roughly 60 percent efficacy reported in trials that focused on narrower immunocompromised populations such as transplant recipients and people with haematological malignancy, underscoring the value of studying a broad, heterogeneous medicated population.

The authors are candid about the limitations inherent in registry-based surveillance. Vaccination reporting to the Australian Immunisation Register only became mandatory in 2021, so some live-vaccine recipients may have been misclassified as unvaccinated, biasing effectiveness estimates downward. The antiviral prescription outcome may capture suspected as well as confirmed zoster, and milder cases that went untreated would have been missed entirely, both of which would push estimates toward the null. The ecological trends, while compelling, are not a direct measure of incidence because repeat prescriptions across different months were counted again. Even so, the convergence of two independent analytical approaches, population-level prescription declines timed to program changes and cohort-based effectiveness estimates, provides mutually reinforcing evidence that targeted zoster vaccination is working. The clear message for policy is twofold: the recombinant vaccine delivers meaningful protection to older adults, but immunocompromised people, who face the highest risks, receive weaker protection and are vaccinated at suboptimal rates, demanding greater provider awareness, stronger uptake efforts and continued monitoring of how long protection lasts in this vulnerable group.

Subject of Research: Population-based evaluation of herpes zoster vaccine impact and effectiveness in older and pharmacologically immunocompromised adults in Australia

Article Title: The epidemiological impact and effectiveness of herpes zoster vaccination among older and pharmacologically immunocompromised adults: a population-based observational study

Article References: Reekie, J., Stepien, S., Li-Kim-Moy, J., Simpson, A., Macartney, K., & Liu, B. (2026). The epidemiological impact and effectiveness of herpes zoster vaccination among older and pharmacologically immunocompromised adults: a population-based observational study. The Lancet Regional Health – Western Pacific, 74, Article 101982. https://doi.org/10.1016/j.lanwpc.2026.101982

Image Credits: AI Generated

DOI: 10.1016/j.lanwpc.2026.101982

Keywords: herpes zoster, shingles, recombinant zoster vaccine, Shingrix, vaccine effectiveness, immunocompromised, older adults, Australia, National Immunisation Program, postherpetic neuralgia, varicella zoster virus, pharmacoepidemiology

The review identifies three core challenges that have defined traditional pain management for decades. First, assessment depends on patients describing their own suffering, a process vulnerable to emotional state, cultural background and cognitive function, and effectively unusable for infants, dementia patients, the critically ill and postoperative patients who cannot self-report. Second, conventional imaging lacks the sensitivity to detect many pain-related structural changes: plain X-rays miss early osteoarthritis and soft tissue lesions, while MRI, despite excellent soft tissue resolution, struggles with functional pain and is expensive and time-consuming. Third, treatment selection relies on clinical experience and guidelines without individualized prediction, leaving roughly 30 to 40 percent of patients failing to respond adequately to their initial regimen. The result is prolonged suffering, repeated medication adjustments, rising costs and heightened risk of adverse drug reactions. Deep learning, the authors contend, offers an end-to-end pathway from symptom identification to mechanism analysis, extracting latent pain biomarkers from multi-source heterogeneous data.

The most technically rich portion of the review concerns objective pain assessment, where deep neural networks are being trained to quantify suffering from signals that patients cannot suppress. Computer vision models analyze facial micro-expressions such as frowning and squinting; speech systems capture changes in vocal tone, pitch jitter and spectral energy; and physiological pipelines integrate electroencephalography, skin conductance, heart rate variability and respiration. The performance figures are striking. A neonatal convolutional neural network recognizing pain from infant facial expressions achieved 91 percent accuracy with an area under the curve of 0.93, while a three-branch network analyzing newborn cries reached 96.77 percent accuracy in binary classification, using only about 2.6 percent of the parameters of VGG16. In adults, a spatial-temporal attention LSTM network classified postoperative pain into three levels from facial landmarks with 86.6 percent accuracy, and an autoencoder-LSTM model fused motion capture with surface electromyography to detect protective behaviors, improving over single-modality baselines by 38.5 percent.

Physiological signals have proven equally fertile ground. A framework called PainAttnNet, built on transformer architectures with multiscale feature extraction, classified pain intensity from electrodermal activity with 85.56 percent accuracy on the BioVid dataset. A dual-branch spatiotemporal model processing scalp EEG in children distinguished pain from non-pain states with 87.83 percent accuracy, and, notably, visualization of electrode contributions showed that accuracy remained at 84 percent even when only nine electrodes were retained, a finding that could dramatically simplify data collection in pediatric settings. A hybrid BiLSTM-support vector machine pipeline classified postoperative pain intensity from electrocardiographic signals at 84.14 percent validation accuracy, while bidirectional LSTMs applied to functional near-infrared spectroscopy achieved 90.6 percent accuracy across four pain intensity categories, outperforming unidirectional variants by 5 to 8.4 percentage points. Resting-state frontal EEG biomarkers have likewise been proposed for grading chronic neuropathic pain severity, moving the field closer to objective clinical translation.

Multimodal fusion, however, emerges as both the field’s greatest promise and its most sobering cautionary tale. Because any single signal can be lost in real clinical environments, obscured by oxygen masks, sedation, motion artifacts or equipment failure, fusing facial, vocal and physiological streams offers redundancy and robustness. In neonatal postoperative pain assessment, a decision-level voting fusion of facial expressions, body movements and crying maintained strong performance even when a quarter of each modality’s data was randomly removed, with the fused area under the curve of 0.868 clearly surpassing the best single modality at 0.774. Yet the review is candid that fusion is not a universal win: in real postoperative wards, single-modality models, particularly those using respiratory rate at 88.24 percent balanced accuracy, consistently outperformed multimodal fusion, which was degraded by motion artifacts, asynchronous acquisition and environmental noise rarely encountered in laboratory datasets. The authors call for cross-modal pretraining, medical knowledge graphs and event-driven fusion strategies to close this gap.

The second pillar of the review concerns intelligent structural identification, where convolutional and transformer-based models automatically segment the anatomical landscape of pain. Deep learning systems now detect lumbar spondylolisthesis from X-rays, quantify vertebral fractures through anchor-free keypoint detection with expert-level localization error of 0.92 millimeters and an AUC of 0.96, and grade intervertebral disc degeneration from MRI in real time using YOLOv5 architectures with over 95 percent classification accuracy. On the cervical spine, where vertebral similarity and complex anatomy make segmentation notoriously difficult, a 2D U-Net framework with superior-inferior labeling achieved Dice coefficients above 94 percent even on pathological data, and a transformer-based model reduced radiologist interpretation time for degenerative cervical MRI from up to 490 seconds to as little as 90 seconds, with the greatest benefit accruing to residents.

Nerve and needle localization extend this vision into interventional precision. Mask R-CNN-based systems segment the median nerve at the carpal tunnel from ultrasound without manual region selection, while U-Net variants track the vagus nerve in real time with over 90 percent recognition accuracy even in low-quality images, trained from mere bounding-box annotations. The dorsal root ganglion, a structure implicated in neuropathic pain but historically too small to segment automatically, has now been delineated in MRI using a meta-optimized nnU-Net framework, revealing genotype-related volume changes in a Fabry disease model. For ultrasound-guided nerve blocks, deep networks locate needle tips that are frequently invisible at steep angles: time-aware LSTMs combined with dynamic background subtraction recover weak tip echoes, and an optical-flow-enhanced YOLO variant tracks speckle dynamics of entirely invisible needles while cutting model parameters by 98 percent for real-time deployment, reaching sub-millimeter localization accuracy in robotic settings.

The third pillar maps AI onto specific pain conditions. For shoulder disorders, multimodal models fusing X-rays with clinical data rule out rotator cuff tears with 97.3 percent sensitivity, and 3D networks trained on more than 11,000 MRI studies classify full-thickness tears with AUCs as high as 0.99, outperforming experienced radiologists. In osteoarthritis, deep stacked ensembles grade knee severity at up to 99.71 percent accuracy, automated systems measure hip-knee-ankle angles 126.7 times faster than manual workflows, and a model called DeepKOA predicts structural and symptomatic progression over 24 to 48 months from multimodal MRI. Multiomic deep clustering has even identified three molecular subtypes of knee osteoarthritis that predict post-arthroplasty pain outcomes with AUCs of 0.84 to 0.88. For trigeminal neuralgia, machine learning on brain morphology predicted gamma knife surgery efficacy with 96.7 percent accuracy, and radiomics models now identify which patients will achieve durable relief from percutaneous balloon compression, lifting three-year pain-free survival in favorable subgroups from 51.1 percent to 86.4 percent. Machine learning models predicting postherpetic neuralgia from 23,326 real-world electronic health records, and LSTM networks forecasting cancer pain exacerbations hours before onset, illustrate the shift toward preemptive intervention.

The review closes with a bracing reality check. Most studies remain small, single-center and internally validated; when tested externally, performance routinely collapses, as when a sacroiliitis model’s sensitivity plummeted from near-expert levels to 56 percent. Data imbalance, annotation inconsistency, demographic bias and unmodeled anatomical variation pervade the literature, and explainable AI outputs often misalign with what clinicians actually need. Privacy risks from biometric pain data, unresolved liability frameworks, absent reimbursement mechanisms and the economic burden of deployment all stand between laboratory success and bedside reality. The authors argue that pain AI must now pivot from a method-driven race for benchmark accuracy to an evaluation-driven, utility-driven paradigm in which human-AI collaboration, longitudinal outcomes and patient-centered benefit define success. If that transition succeeds, the era in which suffering could only be described, rather than measured, understood and preempted, may finally be drawing to a close.

Subject of Research: Artificial intelligence applications in objective pain assessment, medical image analysis and treatment decision-making for chronic pain conditions

Article Title: Current state of research and future developments of artificial intelligence in pain diagnosis and treatment

Article References: Current state of research and future developments of artificial intelligence in pain diagnosis and treatment. (n.d.). https://doi.org/10.1186/s12967-026-08529-9

Image Credits: AI Generated

DOI: 10.1186/s12967-026-08529-9

Keywords: artificial intelligence, chronic pain, deep learning, pain assessment, multimodal data fusion, medical imaging, trigeminal neuralgia, osteoarthritis, postherpetic neuralgia, cancer pain, explainable AI, precision medicine