Antivenom has worked for a hundred years, and in the Amazon what decided who lived was how far away it was kept

The rest of the day

47 more stories on this beat.

Each with its own sources. None of these is a link to the story above.

1. 02

   A nurse on the phone matched the clinic

   In a US trial called EMPallA, 1,283 people aged 50 and over with advanced cancer or failing organs were randomly given one of two kinds of palliative care after an emergency visit. Palliative care treats symptoms and comfort rather than the disease itself. One group got a nurse telephoning every week or two for six months. The other was offered a monthly appointment with a specialist. 68% stayed with the phone calls; 53% of the clinic group attended even one visit. [15]

   Nurse phoning: stayed with it68%

   Clinic: came to one visit53%

   Why it matters — Quality of life improved by about the same amount in both groups, so the telephone version was not better. It was simply easier to reach, and specialist palliative staff are far too few for the number of people who need them. [15][16]

2. 03

   Child deaths fell by half, then slowed

   The United Nations published new estimates in March showing 4.9 million children died before their fifth birthday in 2024, including 2.3 million newborns. Deaths of under-fives worldwide have more than halved since 2000. Since 2015 the rate of improvement has slowed. It was the first time the report broke down the leading causes of under-five deaths in full, and most of them are preventable with cheap, proven treatment. [49]

   Why it matters — Most of these deaths are preventable with treatments that already exist and cost little. What is missing is access to decent care, which the report names directly. [49]

3. 04

   Village workers treated malnourished children

   In rural Somaliland, researchers ran a cluster-randomised trial called SETiPlus. Community health workers, who already treat childhood diarrhoea and pneumonia in their own villages, were also trained to find and treat acute malnutrition. Children found in the intervention villages were more likely to start treatment than children in the comparison villages. But the workers' diagnosis was often wrong, so some children were treated who did not need it. [51]

   Why it matters — Fewer cases turned up than the researchers expected, so the trial could not prove that overall coverage rose. It lands on families who live hours from the nearest clinic, which is where most untreated malnutrition sits. [51]

4. 05

   Nepal's child nutrition worsened after aid cuts

   A survey reported in June found child malnutrition in Nepal at what its authors called alarming levels, after foreign aid for nutrition programmes was cut. Nepal had been one of the clearest successes in the field: deaths among under-fives there fell 72% between 1996 and 2022. The programmes that produced that fall were largely paid for from outside the country. [50]

   Why it matters — Under-five deaths in Nepal had fallen 72% between 1996 and 2022. Those gains rested on money that can be withdrawn in a single budget round. [50]

5. 06

   Calcium and vitamin D did not prevent fractures

   A review in The BMJ pooled 69 randomised trials covering 153,902 adults who were not already on osteoporosis drugs. Osteoporosis is bone thinned to the point where it breaks easily. Vitamin D alone made no difference to the risk of any fracture across 36 trials and 92,045 people. Calcium alone and the two combined showed little or no effect either, and none of them reduced falls. The authors rated most of this evidence as high or moderate certainty. [28]

   Why it matters — Almost a third of adults over 65 fall at least once a year, and half of women will break a bone at some point from a low-impact injury. This is the largest test yet of whether the supplements prevent that. [28]

6. 07

   Exercise cut the injuries that falls cause

   A meta-analysis pooled 29 randomised trials of exercise programmes for older people living at home, covering 6,367 participants. Exercise cut all fall-related injuries by 13%, injuries needing medical attention by 29%, severe injuries by 33% and fall-related fractures by 44%. The trials ranged across balance work, strength training and mixed programmes. Heterogeneity between the studies was low to moderate, meaning the trials broadly agreed with each other. [29]

   Why it matters — This is the comparison that matters against supplements: the pill did nothing for fractures and the exercise cut them by nearly half. Both were tested in older people living in their own homes. [28][29]

7. 08

   An app cut leaking by 61 per cent

   A German trial called DINKS enrolled 194 adults, all assigned female at birth and all leaking urine at least once a day. They were randomly assigned to an app called Kranus Mictera on top of usual care, or to usual care alone. The app taught pelvic floor exercises, bladder habits and relaxation over 12 weeks. Leaking episodes fell by an average of 61% in the app group and by 1.7% in the usual-care group. No relevant treatment-related side effects were reported. The trial was funded by the app's maker. [20]

   App plus usual care61%

   Usual care alone1.7%

   Why it matters — Pelvic floor training is the recommended first treatment for leaking and is often never offered. An app costs one download and does not need a physiotherapist's appointment. [20][26]

8. 09

   Training in pregnancy prevented leaking later

   A trial across nine hospitals in China randomly assigned 764 women expecting a first baby to a supervised pelvic floor programme from 28 weeks of pregnancy, or to usual care. The pelvic floor is the sheet of muscle that holds the bladder outlet shut when belly pressure rises. Six weeks after birth, 8.7% of the exercise group leaked urine on effort against 13.9% of the control group. Muscle strength scores were more than twice as likely to be good. [21]

   Why it matters — Muscle training is the recommended first treatment for leaking during pregnancy and after it. This trial gave it before the problem started rather than after. [21][27]

9. 10

   A common condition got a new name

   Polycystic ovary syndrome is being renamed polyendocrine metabolic ovarian syndrome. The Mayo Clinic, one of the largest hospital groups in the United States, now uses the new name and its initials, PMOS, throughout its patient pages. The old name points at cysts, but the dark circles on a scan are follicles, the fluid-filled sacs an ovary makes anyway. The World Health Organization says 10 to 13% of women have the condition. [37][36]

   Why it matters — The name told women the problem was their ovaries. The condition also involves insulin resistance, higher male-type hormones and long-term diabetes risk, which is what the new name is trying to say. [36][37]

10. 11

    Four kinds of PCOS found in 11,908 women

    Researchers clustered nine clinical measurements from 11,908 women with polycystic ovary syndrome and found four reproducible subtypes, checked against five international groups of patients. One is driven by high male-type hormones, one by obesity, one by a hormone-carrying blood protein and one by two reproductive hormones. Followed for 6.5 years, the groups diverged sharply: the hormone-driven type had the highest risk of losing a pregnancy in the middle months, the obesity type the worst metabolic complications and the lowest live birth rate. [38]

    Why it matters — Women given one label have been given one treatment plan. These four groups had different pregnancy outcomes and different diabetes risk, which is an argument for splitting the diagnosis. [38][41]

11. 12

    A drug for schizophrenia's thinking problems failed

    Luvadaxistat, an experimental drug for the memory and concentration problems that come with schizophrenia, missed its target in a phase 2 trial called ERUDITE. 203 adults already on antipsychotic medicine were randomly given placebo or one of two doses for 12 weeks. On the main measure of thinking, neither dose beat the dummy pill; both were very slightly worse. About a third of participants in each group reported a side effect. [43]

    Why it matters — Antipsychotics reduce hallucinations and delusions reliably and do much less for thinking and memory. This was one of the serious attempts at that gap. [43][47]

12. 13

    451 trials on the quiet symptoms of schizophrenia

    A meta-analysis in Molecular Psychiatry pooled 451 randomised trials covering 42,566 people, all aimed at what psychiatrists call negative symptoms: loss of drive, loss of pleasure, withdrawal from other people and flattened speech. The authors worked out how big an effect has to be before a clinician would notice it, then measured which treatments cleared that bar. Antibiotics, combined psychosocial programmes, antidepressants, physical activity, brain stimulation and immune-system drugs did, usually added on top of antipsychotics. [42]

    Why it matters — These are among the most disabling features of schizophrenia and there are no agreed treatment guidelines for them. The things that cleared the bar were mostly added on top of antipsychotics rather than replacing them. [42][47]

13. 14

    A new antipsychotic beat placebo in 359 adults

    LB-102, a new drug that blocks three receptor types in the brain, was tested in 359 adults during an acute episode of schizophrenia. Over four weeks, the 50 mg and 75 mg doses beat placebo on the standard symptom scale, and on every secondary measure the trial recorded. The drug was reported as generally safe and tolerated at all doses over that short period. It is a chemical relative of amisulpride, an older antipsychotic. [44]

    Why it matters — Four weeks is long enough to show an effect on an acute episode and too short to show weight gain or movement problems, which are among the reasons people stop taking these drugs. [44][48]

14. 15

    Cutting the dose early: worse now, better later

    A Dutch trial followed 347 people who had recovered from a first episode of psychosis. Half were guided through reducing or stopping their antipsychotic; half stayed on the maintenance dose. In the short term the reduction group relapsed more often and rated their quality of life lower. From three years onwards, researchers rated their day-to-day functioning better. Medication use was similar in both groups from 12 months on. [45]

    Why it matters — Because the drugs ended up the same after a year, the authors think the later gain came from having tried a guided taper, not from the dose itself. [45]

15. 16

    Early care cut the slide into schizophrenia

    A trial across five centres in Egypt enrolled 1,000 people aged 16 to 40 who had just had a psychotic episode brought on by drugs, or a brief one. Half received usual care: standard antipsychotics, monthly appointments and case management. Half received that plus monthly family teaching sessions, weekly talking therapy and a low dose of risperidone with blood-level monitoring. After two years, 26.3% overall had gone on to a schizophrenia diagnosis; the intensive group's risk was 39% lower. [46]

    Why it matters — Cannabis-related cases had the highest rate of all, at 38.1%. The trial suggests the two years after a first episode are when the course can still be changed. [46]

16. 17

    Waiting beat dosing after a coral snake bite

    American coral snake antivenom stopped being made in 2003, and the shortage forced doctors in Florida to wait for symptoms rather than dose every bite. Researchers compared 301 cases reported to Florida poison centres between 1998 and 2021. Patients dosed straight away were about three times as likely to end up in intensive care, with no difference in breathing tubes needed or in deaths. Of 191 people given the antivenom, 38 had a reaction and eight of those were anaphylaxis. [7]

    Dosed immediately: went to intensive care75.2%

    Watched first: went to intensive care56.8%

    Why it matters — Antivenom carries its own risk of a serious reaction. When it is scarce, the question changes from how fast to give it to whether this patient needs it at all. [7][2]

17. 18

    Snake venom can cause a stroke

    A review pulled together 100 published studies describing 130 people who had a stroke after a snakebite. Vipers caused 96.4% of them, mostly Russell's viper in South Asia and lance-headed vipers in South America. Nine in ten of the patients had received antivenom. Clot-type strokes were commoner than bleeds, and 23.4% of the patients died. Just over a fifth of survivors were left with lasting weakness or paralysis. [6]

    Why it matters — Venom damages blood vessels and the clotting system, so a bite on the foot can end as brain damage. Cases came from 22 countries, led by India, Brazil and Sri Lanka. [6]

18. 19

    China counted 40,817 snakebite patients

    Researchers gathered hospital records on 40,817 snakebite patients across 11 provinces of southern China and the Yangtze basin, the first large multi-centre count the country has done. Almost half the cases came from three provinces: Jiangxi, Hunan and Guizhou, which are mountainous, wet, densely populated and farming-dependent. For more than half the patients, nobody recorded which snake had bitten them. [8]

    Why it matters — China is working towards the global goal of halving snakebite deaths and disability by 2030, and a country cannot stock the right antivenom without knowing which snakes are biting. [8]

19. 20

    Kenya found snakebites by asking during deworming

    Kenyan researchers added two questions about snakebite to mass drug administration campaigns already running for trachoma, schistosomiasis and worms. That reached 13.1 million people across 17 counties, about 28% of Kenya's population, between July 2022 and August 2023. Communities reported 4,667 bites in the previous year. Turkana county reported 412.9 cases per 100,000 people and Vihiga county 3.7, a difference of more than a hundredfold. [9]

    Why it matters — What people said about going to a health facility did not match what the facilities had recorded. Piggy-backing on an existing campaign cost almost nothing and covered a quarter of the country. [9]

20. 21

    Brazil's snakebite records miss the useful columns

    A study examined every snakebite notification in Brazil's national reporting system between 2007 and 2023. Sex and age were recorded almost every time. Education, occupation, race and the actual clinical signs of poisoning were recorded poorly or very poorly. Agreement with the separate deaths database was excellent overall, but weaker in the north, north-east, south-east and centre-west. [10]

    Why it matters — Antivenom is distributed on the basis of these records. The regions where the data agree least are the ones with the longest journeys to a hospital that stocks it. [10][1]

21. 22

    In northern Ghana, 2,684 bites in five years

    Researchers reviewed the records of 2,684 snakebite patients seen between 2016 and 2020 at two referral hospitals in northern Ghana, in the Upper West and North East regions. Snakebite is among the five commonest emergencies there. About 90% of cases were caused by snakes whose venom attacks the blood, and 91% of patients were admitted to a ward. Swelling, severe pain and bleeding were the commonest signs on arrival. [11]

    Why it matters — Treatment practice in the region had almost no published clinical data behind it until now, which makes it hard to tell whether the national guideline is being followed. [11]

22. 23

    India's snakebite bill falls on the poorest

    A community survey led by frontline health workers covered 25 districts in 11 Indian states and a population of 60 million for a year. It recorded 7,094 snakebites and a death rate of 0.33 per 100,000 people. Nearly half the deaths happened outside a hospital. 86.4% of victims sought hospital care and 60.2% received antivenom, of whom 17.2% had a side effect. Average out-of-pocket cost was 6,500 rupees, and 47.5% of victims lived below the poverty line. [4]

    Private hospital27,400 rupees

    Public hospital3,900 rupees

    Why it matters — The same treatment cost 27,400 rupees in a private hospital and 3,900 in a public one. For a farming household that gap decides which building you go to, and how long you argue about it first. [4]

23. 24

    A man let snakes bite him 200 times

    Tim Friede, a former window cleaner in Wisconsin, spent nearly 20 years letting some of the world's deadliest snakes bite him on purpose, to build antibodies in his own blood. The attempt nearly killed him, put him in a coma and almost cost him a leg and his fingers. Centivax, a Californian vaccine company he now works for, is using his antibodies to build an antivenom that covers many species at once. [13]

    Why it matters — Antivenom today must roughly match the snake that bit, and several hundred species can seriously wound a person. A near-universal one would mean a clinic stocking one product instead of a shelf of them. [13][2]

24. 25

    Old drugs blocked viper venom in the lab

    Researchers tested drugs originally developed for other purposes against the venom of Bothrops, the lance-headed vipers responsible for most serious bites in Central and South America. Marimastat and DMPS, which block the enzymes that break down tissue, and varespladib, which blocks another venom enzyme, neutralised venom from every species tested in laboratory assays. The tissue-enzyme blockers also protected against clotting and bleeding effects in further laboratory tests. [14]

    Why it matters — These are laboratory and animal results, not patients. What makes them interesting is that a small molecule can be a tablet in a health post, where antivenom needs a fridge and a drip. [14][1]

25. 26

    A computer flag doubled palliative referrals

    A meta-analysis pooled seven randomised trials covering 125,666 patients, all testing whether an automatic flag in the hospital record could identify people who need palliative care. The flags roughly doubled palliative care consultations for patients without cancer and raised them more than fivefold for cancer patients. Documentation of do-not-resuscitate decisions rose by 22%. Effects on hospice use and on deaths in hospital were marginal. Families reported no change on the psychological measures the trials used. [18]

    Why it matters — Randomised trials had already shown that this care helps; what was missing was the referral. Automating the referral changed the referral and little else. [18][16]

26. 27

    Who never turns up to the palliative clinic

    Researchers looked at the 603 patients randomly assigned to the specialist clinic arm of the EMPallA trial and asked who actually attended. About 54% came to at least one visit. People without a high school education were less than half as likely to come as graduates, and so were people who needed considerable help with daily life. Older patients, women and people who had never married came back for fewer follow-up visits. [19]

    Why it matters — A clinic appointment is not a neutral offer. The people least likely to take it up had less schooling, needed more help with daily life, and had less social support. [19][15]

27. 28

    Early palliative care in brain cancer

    A German trial called EPCOG randomly assigned 217 people newly diagnosed with glioblastoma, the most aggressive brain cancer, to standard care or standard care plus palliative care from within four weeks of diagnosis. Six university hospitals took part. Quality of life at six months favoured the palliative group but did not reach statistical significance. The control group happened to live longer, and once the analysis allowed for that, the quality of life result did become significant. [17]

    Why it matters — Mood and symptom problems improved most after the intervention ended. Caregivers showed no benefit, which is a result worth reporting because caregiver strain is one of the main reasons this care is offered. [17]

28. 29

    Twelve app trials for leaking, 1,647 women

    A meta-analysis pooled 12 randomised trials covering 1,647 women, comparing phone apps that deliver pelvic floor training with usual care. The apps cut the severity of leaking and reduced episodes by about two a day. Women using them were 42% more likely to stick with the exercises. Quality of life improved. A separate review of 18 studies and 2,308 people found the same pattern for severity and adherence, but no difference in quality of life or muscle strength, and rated its own evidence as low quality. [22][23]

    Why it matters — The two reviews agree on the thing the app is for, which is getting people to keep doing the exercises, and disagree on whether life gets better. [22][23]

29. 30

    Habits plus exercises beat either alone

    A Brazilian trial randomly assigned 110 women with an overactive bladder to three groups: bladder habit training alone, pelvic floor muscle training alone, or both together. An overactive bladder means a sudden urgent need to pass urine, often at night, with no infection behind it. After three months the combined group was 5.5 times more likely to be cured than the exercise group and 4.1 times more likely than the habit group. [25]

    Why it matters — Habit training teaches urge suppression and timed voiding; the exercises build the muscle. Each alone cured far fewer women than the pair. [25]

30. 31

    Biofeedback added to pelvic floor training

    A review pooled 21 studies of 2,373 women, testing whether biofeedback helps pelvic floor training for leaking on effort. Biofeedback means a sensor showing the woman on a screen whether she is squeezing the right muscle. The internal kind of sensor improved muscle strength, symptom severity, quality of life and the chance of cure, roughly doubling the odds of improvement. Cure rates were higher when a therapist supervised, sessions ran under ten minutes, and training totalled more than an hour a week for at least eight weeks. [24]

    Why it matters — Biofeedback exists because a woman cannot otherwise see whether she is squeezing the right muscle. The review found cure rates were highest where a therapist supervised the training, which is the part that needs a clinic. [24][26]

31. 32

    An app for falls cut the risk by a tenth

    A meta-analysis of nine randomised trials covering 3,437 older people living at home tested phone apps that coach balance and strength exercise. Over 12 months the apps cut the risk of falling by 11%, an absolute drop of 6.6%. The pooled fall rate did not fall significantly. Balance, strength and mobility improved moderately, fear of falling dropped, and no serious harm was reported. [30]

    Why it matters — The apps did not significantly cut the overall rate of falls, only the share of people who fell at all. The comparison that matters is against the supervised classes most older people never get to. [30][31]

32. 33

    Tai Ji Quan cut falls by a third or more

    A review of 27 randomised and controlled trials compared different kinds of exercise for preventing falls in people over 60. Tai Ji Quan, the slow Chinese movement practice, was associated with falls dropping by 31 to 58%. The Otago programme, a home strength and balance routine, cut them by 23 to 40%. Mixed strength and balance training cut them by 20 to 45%. The studies varied too much to pool into one number. [31]

    Why it matters — None of these needs a gym or special equipment. The ranking matters because a health service funding one programme is choosing between them. [31][29]

33. 34

    Hip fractures happen where care already is

    South Korean researchers looked at 170,095 people aged 65 and over who came to one of 23 emergency departments after a fall between 2011 and 2020. The chance that a fall broke a hip rose steeply with age: those aged 85 and over were 4.2 times as likely as the youngest group. Compared with falls at home, falls in medical facilities carried 1.6 times the odds of a hip fracture, and falls in nursing homes 1.2 times. [33]

    Why it matters — Places full of staff are not safer places to fall, because the people in them are frailer. In a separate study of 801 hip fracture patients, three quarters of the falls happened indoors, and those who fell at night had weaker grip. [33][35]

34. 35

    62,000 older Irish people fell badly in a year

    Researchers used the sixth wave of the Irish Longitudinal Study on Ageing, a representative sample of 2,299 people aged 70 and over, and scaled it to the national census. Almost 12% had a fall needing medical attention in 12 months, about 62,000 people. About 6%, over 32,000, went to an emergency department. Falls are already the commonest reason older people are admitted to hospital in Ireland. [34]

    Why it matters — The study was built to tell the health service how much falls care it needs to provide and where. The number is a demand forecast, not a statistic. [34]

35. 36

    Seven in ten Ethiopian children recovered

    A meta-analysis pooled studies published between 2015 and July 2025 on children under five treated for severe acute malnutrition in Ethiopia, including conflict-affected regions. 71.4% recovered. Anaemia, shock, HIV, tuberculosis, diarrhoea, vomiting and needing a feeding tube all made recovery slower. The authors say earlier reviews of the same question used outdated data and narrower regions. [52]

    71 of 100

    recovered out of every 100 children treated

    Why it matters — Severe acute malnutrition is behind about one in five deaths of children under five worldwide. Seven in ten recovering is what a working treatment programme looks like from inside. [52][53]

36. 37

    US malnutrition deaths climbed for a decade

    An analysis of US death records counted 158,117 deaths from malnutrition in adults aged 55 and over between 1999 and 2023. The rate fell by 8.4% a year until 2006, then rose by 22.5% a year from 2013 to 2021. The oldest group climbed fastest. Both men and women rose steeply from 2013 onwards. [54]

    Why it matters — The authors also found the gaps widening between racial groups and between town and country. This is one of the few numbers in today's reading moving the wrong way. [54]

37. 38

    India's poorest districts, measured

    Researchers used India's fifth National Family Health Survey to compare severe malnutrition in children under five in the aspirational districts, the areas the Indian government designates as most deprived, with the rest of the country. They measured severe wasting, severe stunting and severe underweight together, which earlier work had not done. All three were more common in the aspirational districts once household and parental factors were accounted for. [55]

    Why it matters — The aspirational districts are singled out for extra government attention precisely because they are the most deprived. Measuring all three forms at once shows which kind of shortfall each district has. [55]

38. 39

    Dialysis deaths predicted by three proteins

    Researchers measured 6,287 circulating proteins in 893 people on regular haemodialysis, a machine treatment that filters the blood when the kidneys have failed. More than half of such patients die within five years, and the usual risk factors do not explain it. A model built from just three proteins predicted five-year death far better than the standard clinical model. Pathways linked to a growth hormone signal stood out. [57]

    Why it matters — The usual clinical risk factors do not explain why more than half of these patients die within five years. Heart failure is one of the things going wrong, and admissions for it among Swedish dialysis patients rose 30% in a decade. [57][58]

39. 40

    Heart failure rose among kidney patients

    Sweden's renal registry tracked 54,705 people with advanced chronic kidney disease or on dialysis between 2011 and 2021. Hospital admissions for heart failure ran 3.7 to 6.8 times the rate in the general population, and rose over the decade: up 51% in stage 4 kidney disease, 36% in stage 5, and 30% among haemodialysis patients. Use of two standard heart drugs stayed flat and low. [58]

    Why it matters — Blockers of the renin-angiotensin system and beta blockers are standard treatment for heart failure. Prescribing of them held between about 44% and 70% across the decade while admissions climbed. [58]

40. 41

    Twenty years of home dialysis in Brazil

    A hospital in Botucatu, Sao Paulo, has used peritoneal dialysis for sudden kidney failure since 2004, a method that filters blood through the lining of the belly rather than through a machine and needle. Researchers followed 487 patients over 20 years. Two thirds needed drugs to support blood pressure and three quarters were on ventilators. The patients treated in the later decade were sicker, and the share whose bodies were breaking down their own tissue rose from 7.6% to 26.5%. [56]

    Why it matters — Peritoneal dialysis uses the belly lining as the filter, so it needs no dialysis machine and no treated water supply. The Brazilian unit ran it for 20 years on patients who were mostly ventilated and on blood-pressure support. [56]

41. 42

    Eighteen drugs linked to PCOS reports

    Researchers mined the US Food and Drug Administration's adverse event database for reports of polycystic ovary syndrome appearing after medication, covering 2014 to 2024 and 1,516 cases. Eighteen drugs came up as statistically over-represented, spanning respiratory medicines, antipsychotics and anti-seizure drugs. Valproic acid and olanzapine were among them. Most reports appeared more than a year after the medicine was started. [40]

    Why it matters — A signal in a reporting database is not proof of cause; people report what they notice. Two of the drugs named, olanzapine and valproic acid, sit in the drug classes that are the mainstay of psychiatric treatment. [40][47]

42. 43

    Weight-loss drugs studied in PCOS

    A scoping review examined the newer weight and diabetes drugs in polycystic ovary syndrome: semaglutide, the dual-acting tirzepatide, and retatrutide, which acts on three hormone receptors and is still seeking approval. All three classes improved weight and insulin sensitivity more than the standard combination of metformin and a hormone pill. Insulin resistance is thought to drive much of the condition. [39]

    Why it matters — A scoping review maps what has been published rather than pooling it, so this is a picture of where the research has gone, not a verdict on any drug. [39][41]

43. 44

    An hour's delay after a cobra bite

    A study in Thailand reviewed 69 people bitten by monocled cobras in central and southern regions. Cobra venom attacks the junction between nerve and muscle, which shows first as drooping eyelids and can end as paralysis of the muscles that breathe. Arriving at hospital more than an hour after the bite was common in younger adults and was linked to more severe nerve damage, including more drooping eyelids and more transfers to other hospitals. Antivenom was given to a third of the patients. [5]

    Why it matters — Of the 23 people who received antivenom, 9 had an adverse reaction. That is the risk that kept the medicine inside hospitals in the first place. [5][1]

44. 45

    Seventy-five years of the same target

    A review marking 75 years since chlorpromazine, the first effective antipsychotic, notes that the core mechanism of these drugs has barely changed: they block a dopamine receptor called D2. They reliably reduce hallucinations and delusions and do much less for thinking and motivation. Clozapine remains the only drug that works when others have failed. The old split between typical and atypical antipsychotics is now seen as unhelpful. [47]

    Why it matters — The first genuinely different mechanism to reach patients is xanomeline-trospium, which acts on muscarinic receptors instead, and was reported to cut symptom scores by 8.4 points more than placebo. [47][48]

45. 46

    Health workers in four countries cannot get antivenom

    The Global Snakebite Taskforce, a campaign group working to cut snakebite deaths, surveyed 904 medical professionals in India, Brazil, Indonesia and Nigeria, the four worst-affected countries. In India, 99% reported difficulty giving antivenom, blaming poor infrastructure, short supply and too little training. Nearly half of all those surveyed said delays in treatment had led to amputations, extra surgery or lifelong trouble walking. India's federal government puts its own snakebite deaths at about 50,000 a year, roughly half the world's total. [3]

    Why it matters — This is the same shortage the Amazon programme was built to fix, reported by the people holding the syringe. The barriers named were storage, supply and training, not the price of the medicine. [3][1]

46. 47

    A third of bite victims went to a healer first

    A hospital in Raichur district, Karnataka, reviewed 366 snakebite cases seen between January 2020 and April 2023. A third of them, 121 cases, involved a venomous snake, most often the common krait, whose bite is often painless and can kill overnight. A third of all victims had tried a traditional remedy before coming to hospital. Their average age was 29. [12]

    Why it matters — India records an estimated 35,000 to 50,000 snakebite deaths a year. In Thailand, arriving at hospital more than an hour after a cobra bite was linked to more severe nerve damage. [12][5]

47. 48

    Practising real tasks steadied the over-75s

    A review pooled 20 trials of task-oriented training for people over 60, comparing it with conventional exercise. Task-oriented training means practising the actual movements of daily life, such as standing up from a chair or stepping over a kerb, rather than doing separate strengthening exercises. Balance scores, the time taken to stand and walk, walking speed and confidence all improved more than with conventional training. In people aged 75 and over the gains were large enough to be noticeable in daily life. [32]

    Why it matters — People get better at what they actually practise. Most of the exercise programmes competing for the same funding train strength and balance away from the situations in which people fall. [32][31]