iiinova — where your digital self keeps evolving

Lucibee @lucibee.bsky.social · 16h

'Recovery' was always a big problem for the PACE trial, to the extent that they didn't actually mention it at all in the main trial paper in 2011. Discussion of it was relegated to a later paper in 2013, which also contained a metric tonne of fudge, because they knew they were in trouble over it.

1227

However, all this is akin to rearranging deck chairs (etc etc) when you have an unblinded trial of coached interventions with subjective endpoints. 🥁

Oh ha! I'd forgotten that they'd characterised Bowling's England health survey population (ages 16 to 85+) as a "UK working age population"!

But again you can't compare. Each point on the bimodal scale means you've got worse. On Likert scale, every point >11 means you've got worse. How tf do you compare 2 scores where you've "got worse" each time? And then call that an "improvement" because one worse was slightly less than the other?

3 points on the bimodal CFQ becomes 2 points on the Likert scale. Again, this is quite a difference, as 3 points on the bimodal scale can encompass up to 6 points on the Likert scale.

The benchmarks for improvement were shuffled into a "secondary post hoc analysis" and radically modified downwards to include as many participants as possible. 8 pts on the PF subscale is less than two clicks, as each question answer is awarded a minimum of 5 pts.

150

In the PACE trial publication (2011), the authors explained that they had ditched the requirement of 50% reduction in fatigue or 50% increase in PF score in favour of instead comparing group means.

I've already outlined the major flaws in the CFQ elsewhere. It is flawed as an outcome measure because it is not an absolute scale, so comparison between scores made at different times is fairly nonsensical.

150

However, by the time the trial was published, several changes had been made. They had switched from bimodal to Likert scoring for the CFQ. This makes it easier to amplify some changes. Bimodal was actually a slightly better way of recording "improvement", even though the CFQ generally fails here.

Lucibee @lucibee.bsky.social · 18h

Well, not really, no. They were the same measures, sure. But very different criteria were used indicate whether a participant had improved or not, and how they assessed the overall outcome of the trial. The protocol is shown below:

10. Measures.
10.1 Primary outcome measures.
10.1.1 Primary efficacy measures.
Since we are interested in changes in both symptoms and disability we have chosen to designate both the symptoms of fatigue and physical function as primary outcomes. This is because it is possible that a specific treatment may relieve symptoms without reducing disability, or vice versa. Both these measures will be self-rated.

The 11 item Chalder Fatigue Questionnaire measures the severity of symptomatic fatigue, [27] and has been the most frequently used measure of fatigue in most previous trials of these interventions. We will use the 0,0,1,1 item scores to allow a possible score of between 0 and 11. A positive outcome will be a 50% reduction in fatigue score, or a score of 3 or less, this threshold having been previously shown to indicate normal fatigue. [27]

The SF-36 physical function sub-scale [28] measures physical function, and has often been used as a primary outcome measure in trials of CBT and GET. We will count a score of 75 (out of a maximum of 100) or more, or a 50% increase from baseline in SF-36 sub-scale
score as a positive outcome. A score of 70 is about one standard deviation below the mean score (about 85, depending on the study) for the UK adult population. [50, 51]

Those participants who improve in both outcome measures will be regarded as overall improvers.

Lucibee @lucibee.bsky.social · 18h

The authors', in their reply, claimed that the outcome measures used in the trial analysis were the same as in protocol. www.bmj.com/bmj/section-... But was that the case?

bmj.com

150

Lucibee @lucibee.bsky.social · 18h

Of course, the eligibility criteria and the recovery criteria weren't the only changes from protocol (or pre-protocol). They changed the primary criteria too. Tom Kindlon noticed this, and tried to find out why. www.bmj.com/bmj/section-...

bmj.com

From the above graph, you can see how much the older ages will influence the mean PF score in the Bowling sample. It's also worth pointing out that the data spread (SD) also increases markedly by age. Thus White's reasoning looking increasingly flimsy.

One of the issues with the PACE data is that the only information we are given about the age distribution of the participants is the mean (38 years, SD 12). We don't even know the range, although we can assume the lower limit was 18. I think it's unlikely they included anyone over 65.

250

To get an idea of why age-matching is so important, these are the summary data by age bands that Bowling provided for her general population sample. Mean PF is used here as a descriptor, but you can see how it drops off with age.

Bar chart of mean physical function (PF) score by age shows that while PF is in the mid 90s until the 35-44 year age band, it then drops off quite sharply, down to just over 70 for 65-74 years, below 60 for 75-84 years, then below 40 for 85+ years.

Rather than age-match and sex-match, White has simply taken the summary data for males and females, added them together and divided by 2, to get the population mean score. You really can't do that! Particularly as the mean score on a highly skewed population is not a good measure here.

Graph panel from Bowling et al. (1999) showing a histogram of the physical functioning subscale for the whole dataset. The peak of the normal plot will show where a calculated 'mean' value will lie (~85). But as you can see, the median value (where at least half the sample lie) is 100.

170

Bowling's data was from a health survey, and she comments that those with long-standing health issues were therefore likely to be over-represented in the data. Also, the demographics are likely different. Bowling's sample was older and had a more equal sex distribution.

Checking your browser - reCAPTCHA

While they've made extensive use of Bowling et al.'s paper, they have made quite a few mistakes in how they've used the data, and ignored the caveats that are highlighted. The first error is to claim that Bowling's sample was a "general working age population", because it isn't.

pubmed.ncbi.nlm.nih.gov

The extract below from the Recovery paper (2013 - pmc.ncbi.nlm.nih.gov/articles/PMC...) shows their workings. However, there are quite a few things that are wrong with their reasoning, as many many others have tried to point out over the years.

Extract from White et al. (2013):

We changed our original protocol's threshold score for being within a normal range on this measure from a score of ⩾85 to a lower score as that threshold would mean that approximately half the general working age population would fall outside the normal range. The mean (s.d.) scores for a demographically representative English adult population were 86.3 (22.5) for males and 81.8 (25.7) for females (Bowling et al. 1999). We derived a mean (s.d.) score of 84 (24) for the whole sample, giving a normal range of 60 or above for physical function.

The main lump of fudge is how they (mainly Prof Peter White it seems) justified use of the lower recovery criteria of just 60 points on the SF-36 PF subscale. As far as we can tell, this was done very much post hoc, and uses some extraordinary mental and statistical gymnastics. Buckle up!

170

It's from the RAND Short Form (SF)-36 physical function subscale. This is the version they used in the trial:

SF-36 physical function sub-scale (as used in the PACE trial).

The following questions are about activities you might do during a typical day. Does your health limit you in these activities? If so how much.

Tick boxes for "Yes, limited a lot"=0 pts, "Yes, limited a little"=5 pts, "No, not limited at all"=10 pts.

Activities:
1. Vigorous activities, such as running, lifting heavy objects, participating in strenuous sports.
2. Moderate activities, such as moving a table, pushing a vacuum cleaner, bowling or playing golf.
3. Lifting or carrying groceries.
4. Climbing several flights of stairs.
5. Climbing one flight of stairs.
6. Bending, kneeling or stooping.
7. Walking more than a mile.
8. Walking several hundred yards.
9. Walking one hundred yards.
10. Bathing or dressing yourself.

Full score is out of 100 points.

110

This gives rise to the scenario whereby a participant *could* enter the trial with a score of 65 and be classed as "recovered" with a score of 60 at the end of the trial. To be fair, this never happened. No-one fell into that category. But the fact that it was a *possibility* was a problem.

190

Only, this time, there was another problem. Although the original eligibility criteria for entering the trial was a physical function score of ≤60 (see TMG minutes below), this had been revised upwards to ≤65 to aid recruitment by the time the protocol was being written.

Extract from the PACE Trial Management Group (TMG) meeting on 7 May 2004:

The SF-36 Physical Functioning Sub Score. This scale is rated on a 0-100, 100 representing perfect and healthy functioning. The healthy adult working population score a mean of 90 with a standard deviation of 15. Those who have a chronic disease or depressive illness score 70 with a standard deviation of 20. After much discussion we agreed that patients would need to score 60 or below to enter the trial and a score of 70 or more would constitute a positive outcome. It was agreed that we would use the proportion of participants who had a positive outcome in both their fatigue and physical functioning score as another primary outcome measure. We would then be able to determine the percentage of people who improve on both measures.

PACE trial protocol: Final version 5.0, 1 February 2006.

Section 7: Selection of Participants.

7.2.2 Eligibility Items
3. The participant meets operational Oxford research diagnostic criteria for CFS.
4. The participant's Chalder Fatigue Questionnaire score is 6 or more.
5. The participant's SF-36 physical function sub-scale score is 65 or less.
6. The participant will be aged at least 18 years old.

170

However, the physical function criteria of ≥85 meant that virtually no-one met the criteria at the end of the trial, so not only did they drop the concept of 'recovery' for the main trial paper, but they massaged it heavily when it came to the actual 'recovery' paper. ≥85 became ≥60

291

The trouble started in the trial protocol itself, which was published in 2007, while the trial was underway. They clearly defined what it should be, and that should have been that.

PACE trial protocol definition of 'recovery':

Extract reads: Point 4. "Recovery" will be defined by meeting all four of the following criteria: (i) a Chalder Fatigue Questionnaire score of 3 or less [calculated on the binary (0,0,1,1) scale], (ii) SF36 physical function score of 85 or above, (iii) a CGI score of 1, and (iv) the participant no longer meets the Oxford criteria for CFS, CDC criteria for CFS or the London criteria for ME.

190

1227

No-one else seems to have mentioned it, and his IG acct is OK. Do we know if he is aware of it?

100

But if there's one thing we should have learnt during the past 7 years, it ought to be how to protect ourselves from an airborne infection. So 😷😷😷 They work!

162

...and someone with a pneumonic infection will be shedding bacteria at a much much higher rate - naturally. And once it's in the air, it is much easier to become infected by it. And if it's there to be breathed in, it will cause another predominantly pneumonic infection.

120

2. Route of transmission depends on the pathology in the host, and not necessarily on the strain of bacteria. Although pneumonic plague is rarer, that is because only 5% of bubonic infections progress to the lungs. *However* Once there, transmission becomes much easier, because people breathe.

230

Assessing the threat of Yersinia pestis harboring a multi-resistant IncC plasmid and the efficacy of an antibiotic targeting LpxC | Antimicrobial Agents and Chemotherapy

1. Multidrug resistance in Yersinia emerged in Madagascar quite a few decades ago, so this isn't something new. The more one uses antibiotics on an organism, the more likely it is to develop MDR. That's why the study of these bacteria is so important. journals.asm.org/doi/10.1128/...

journals.asm.org

Multidrug resistance (MDR) diminishes the antimicrobial arsenal that protects us against medically important bacterial agents, including those responsible for some of the worst known epidemics in huma...

Ben Mitchell @wilde-eep.bsky.social · 05/10/2026

A few thoughts about this plague outbreak in Russia. Some are commenting that maybe this is a new strain of Yersinia pestis that has been modified to be more easily transmissible, or to be multi-drug resistant. However, there are a few things to be aware of:

161

Reposted by Lucibee

A fearsome predator lurks in the lichens! This striking creature is the larva of a hoverfly - Epistrophe grossulariae. It mostly eats aphids. #FencepostOfTheWeek #hoverfly #epistrophe #lichen #naturephotography #argyll #scotland #uk

Photo of the top of a fencepost, covered in lichens and bryophytes. A small, lime green creature is visible on the right hand side.

photo of a translucent slug like creature perched on a lichen.

093

Do we know when this happened?

100

Although I'm now doubting whether he ever had a FB page... 😳

100

Ceri Turner 💚 😷 @ceri-turns.bsky.social · 05/10/2026

Yep. It's gone. And they've spiked searches too. That's seriously bad.

210

Reposted by Lucibee

Does anyone know what has happened to the Facebook account of @zackpolanski.bsky.social ? It appears to have disappeared. 🧐 Very concerning a few days before the Hackney By-election!

244

Hate to say this, but they seem pretty reasonable. That's what it costs to make things that last. Most clothes these days simply fall apart because they are so cheap. :( (See Vimes' Boots Theory)

000

Unclassified - A man from the future 2026 by Pet Shop Boys - BBC Sounds

Listened to this last night: Pet Shop Boys' tribute to Alan Turing - A man from the future. Premiered at the Proms in 2014, but rescored for a new album. www.bbc.co.uk/sounds/play/...

bbc.co.uk

Elizabeth Alker presents an exclusive chance to hear fresh material from Pet Shop Boys.

120

Indeed. Graham McP tried to get the step test data, but without success. He was only asking for the summary data. When (I think it was) Nick B managed to obtain it, it was detached from anything useful that would allow back-calculation of the relevant measures.

000

Spudislander45 🇨🇦❤️🌎💪🫡 @spudislander45.bsky.social · 04/10/2026

To clarify, I mean the co-authors, collaborators, reviewers, journal editors, statisticians, fellow researchers, respected critics, those deemed 'suitable' to request access to the data, those considered 'not vexatious', etc. Bcs we know requests have been made by those who aren't for much less.

100

Reposted by Lucibee

#CBC reporting on the extent of damage in #Gaza. #Israel is supposed to stop bombing the region as bodies and damage hold #evidence that must be maintained. #Cdnpoli #FreePalestine #CrimesAgainstHumanity #WarCrimes

911193

I'm not even going to try to predict what is going to happen given all the chaos so far! 🤣

010

And fall into the flawed trope that liars always lie (they don't), and that everything that looks suspicious is intentional. It's all absolutely fascinating!

000

But, of course, that's the beauty of the format. It is literally and absolutely unpredictable. Even when actors screw up in their utter excitement! I find it fascinating that everyone starts treating 'being a traitor' as a character trait, rather than a role to play.

110