A blood test that measures the protein p-tau217 can identify cognitively healthy older adults who are likely to develop Alzheimer’s disease within five to ten years, according to a study of nearly 2,700 people tracked for up to a decade. Presented at the Alzheimer’s Association International Conference in London and published in JAMA, the research found that symptom-free individuals with very high p-tau217 levels had an estimated 78 per cent chance of progressing to cognitive impairment within ten years. The biomarker reflects the build-up of tau tangles in the brain and, researchers say, could one day be used much like cholesterol tests are used to gauge heart-disease risk.

The finding arrives as artificial intelligence begins to offer non-invasive pathways for early detection of cardiovascular illness. At a cardiology meeting in Jakarta, clinicians from Singapore and South Korea detailed a retinal-analysis system called Dr Noon that uses fundus photographs and AI algorithms to predict atherosclerotic cardiovascular disease risk without blood draws or radiation. Separately, an Indonesian cardiologist presented NAVI-HF, an AI-enhanced stethoscope that analyses chest sounds to detect residual lung congestion in heart-failure patients. In a study of 246 acute heart-failure patients, the device showed 86 per cent accuracy, 91 per cent sensitivity and 82 per cent specificity compared with lung ultrasound. Both tools remain in the research-validation phase and are positioned as decision-support aids, not replacements for clinical judgment.

These advances contrast sharply with the real-world consequences of medical technology that reaches patients without adequate scrutiny. An investigation by The Sydney Morning Herald and 60 Minutes revealed that the CellAED portable defibrillator, approved in Australia in 2021 without human clinical evidence, was pulled from the UK market after safety concerns. In one incident, a unit delayed a shock for roughly 50 seconds, then discharged through a pregnant rescuer. The device had been distributed to ambulance services and community volunteers; both Victorian and Canberra ambulance services have since withdrawn it. The case has focused attention on Australia’s Therapeutic Goods Administration, which relied on a private German assessor’s approval without independently reviewing clinical data.

Patient voices, too, remain a weak link in the diagnostic chain. Nigerian media recount cases of “medical gaslighting” in which doctors dismissed a woman’s seafood allergy and a new mother’s knowledge that eating would normalise her soaring blood pressure. In Melbourne, a coronial finding into the death of a 48-year-old man who was mistakenly given another resident’s antipsychotic medication ran to just six pages, prompting disability rights lawyers to question the depth of the inquiry. Viewed from Jakarta, where AI-based screening is being promoted as a way to strengthen preventive care, the Australian and Nigerian experiences underscore that even the most sophisticated diagnostics will underperform if regulatory systems and clinical listening do not keep pace.

The next milestones to watch are the publication of longer-term validation studies for the p-tau217 blood test in more diverse populations, and any regulatory guidance from the US Food and Drug Administration or the European Medicines Agency on incorporating such biomarkers into routine screening. In Australia, pressure is mounting on the TGA to reform its device-approval pathways, with a formal review likely to be announced in the coming months.