A farmer in Amroha district brought his wife to a small medical centre when her labour pains started. The family was told she would deliver normally. The man who ran the place had no medical qualification. He admitted her anyway, put her on a drip, and when she got worse he said she needed an operation and made a phone call. The man who turned up from a village nearby had no qualification either. They operated at around midnight, without asking anyone to consent. The baby lived.

Getting from that night to a sentence took nine years, and the case nearly fell apart on the way. The centre was sealed, the FIR named three people, and investigators dropped the first man's wife for want of evidence. Then the prosecution's own witnesses began leaving. Twelve were examined in court, six turned hostile, and one of the six was the husband who had carried his wife in that night.

Both men, now in their thirties, were convicted of culpable homicide not amounting to murder, cheating, and offences under the Indian Medical Act. Bail ended with the verdict and they went to jail from the courtroom. Four days later each of them got life.

Hand over a card you have used for ten years and watch the front desk push it back. Since 16 September, 99 hospitals, dental clinics and diagnostic centres across Delhi, Noida, Greater Noida, Ghaziabad, Gurgaon, Faridabad and Sonipat cannot register a CGHS patient, admit one, or process a fresh pre-authorisation. Dr B.L. Kapur Memorial Hospital on Pusa Road, the hospital that trades as BLK-Max, is number 69 on the annexure. Umkal, Medstar, SG and Jeevan Jyoti are on it too.

Anyone already in a bed is protected, and the order is blunt about it: no refusal of care, no discharging them early. The patients with a problem are the ones holding half a workup from somewhere that is no longer on the panel.

The order is dated 22 September, six days after the cut-off it enforces, and it alleges nothing whatsoever about the standard of care at any of them. What actually went wrong is duller, and it is why every one of these 99 can be back on the panel without changing a thing about how they practise.

Read →

The governor of Irkutsk called an emergency meeting on a Friday and said there was a suspected case of what Russian rules call a particularly dangerous infection. Then two days of official silence, and the story got away from everybody. A hospital ward shut and its lab stopped work. The governor of the next region along wrote that a woman had died of plague, then edited the post to say she possibly had. A nearby town warned people off travelling there, then deleted the warning. A national channel pulled three reports, one of which had printed her name. On Sunday the health watchdog said her samples showed nothing connected to her work and nobody around her was ill. On Monday the Kremlin told everyone to stop speculating. No official source has confirmed that she died.

India buried something like 12 million people to plague between 1898 and 1918, and is not on the list of countries where it circulates today. What is still worth knowing is clinical. Pneumonic plague is curable with two old, cheap antibiotics, and it kills almost everyone in whom it is caught late. From the first symptom you have roughly one day. The reason that window is so short has nothing to do with the bacteria.

Read →

Forty children were randomised to fluoxetine or nortriptyline in 2006, and the result went into an Iranian medical journal. The 2016 Lancet meta-analysis that put fluoxetine into national guidelines for childhood depression counted that trial. So did the 2020 Lancet Psychiatry update. The 2021 Cochrane review left it out by accident, having excluded comparisons against the class nortriptyline belongs to, and concluded the advantage over placebo was small, unimportant and probably no better than any other antidepressant. Three analyses, one drug, three answers, and the disagreement sits entirely in those forty children.

The margin that trial reported is one a co-author of the new paper likens to asking people to choose between “chocolate or faeces”. Nothing else in paediatric psychiatry comes near it, and re-running 2016 and 2020 without it makes fluoxetine's edge over placebo disappear. The researchers emailed the trial's authors repeatedly. They posted a preprint naming the study. Nature asked for comment. Nobody has answered any of it.

The journal's publisher has now asked its editors to look into the paper, and says it expects that to take a while given what Iran is currently dealing with. So the one study propping up a prescribing recommendation in dozens of countries waits on a review in a country that cannot easily run one. Andrea Cipriani, who co-authored both Lancet analyses, agrees the data are problematic and says no rule existed then that would have let his team drop it. Cochrane wrote one since. Nobody is asking you to stop prescribing, only that the committees look again, and there is a second thing wrong with those forty numbers that is harder to explain than the effect size.

Read it on Nature →

Know a colleague who should be reading this? Forward this email to them.

Forwarded this issue? Subscribe free →