A litany of medical mistakes

From serotonin syndrome to six years of misdiagnosis and mistrust

Source: NightCafé

Many future Psyverses will be based around the theme of medical mistakes. Psyverse #2 will be specifically about how independent national voluntary reporting systems, used in the aviation industry, could be applied to the psychiatric system. Before we get there, however, I thought I would tell you the story of my severe reaction to an antidepressant. And the many mistakes that were made.¹

---

Mistake #1 may have been the advice from my GP to take an antidepressant. It was late 2015, a year into my PhD. An unfortunate series of events had sprouted blackened vines of depression, dragging me deep into dysfunction. In a desperate bid to escape, I tried meditation, eating healthily, exercising, quitting drinking, and university counselling. In my head, I thought taking an antidepressant was the last remaining option. I did not know that other forms of counselling existed. I didn’t know psychotherapy could go on for longer than six sessions.²

I was prescribed the SSRI³ Sertraline at the starting dose of 50 mg once daily. On the first day, I experienced the typical side effects: slight worsening of depression, a little nausea and a lack of appetite. During the second and third days on Sertraline, numbness replaced depression. It was strong. So strong that stressful events like going to a restaurant with my mum and disabled sister, with the accompanying glares, ceased to influence me. I thought this was brilliant (well, a fuzzy version of brilliant). I could return to my normal life.

After the fourth 50mg dose, I felt an “energy” build inside my head while washing dishes. Overcome by a strong urge to find safety, I dropped a half-washed dish into the sink and rushed to the sofa. A few seconds later, I snapped into the foetal position.

A violent tremor spread through my limbs, torso, and head, growing stronger with each moment. A viscous, lava-like ooze expanded outwards from the centre of my brain. It felt physical, like a real substance swallowing neurons. Just before it captured my outermost brain matter, I closed my eyes and waited to die.⁴

My first thought when I opened my eyes was, “Oh good, I’m not dead!” Shock, as I found out, affects different people in different ways. My version was not running around like a headless chicken, or sitting stationary in frozen silence. I moved mechanically in eerie calm. The only thoughts that occupied my mind during the next few hours were related to figuring out what on earth had just happened.

In conversation with my mum on the phone, we determined that I had just experienced a panic attack. We were wrong. Before that day, I had never experienced anything close to a clinical anxiety symptom. Therefore, I had no frame of reference.⁵

The next mysterious definitely-not-a-panic-attack jolted my eyes open during deep sleep. Due to the unnerving disorientation of waking up in the middle of one of these realistic nightmares, I cannot say for certain if it was worse than the first attack. It was at least just as bad.

My legs were shaking so violently this time that I became physically unable to walk. To pass the time, I grabbed my phone from the bedside table and rang the NHS 111 service.⁶

The person on the other end of the line instructed me to go through the checks for a stroke. Arms in the air, etcetera. When I passed these checks, they advised me to go back to sleep and call the GP in the morning. Then they hung up. This was Mistake #2.

I looked down at my still-shaking pair of legs. “Ah”, I thought, “probably should’ve led with the whole non-functioning legs thing”.⁷

There was no way in hell I was going back to sleep. After waiting fifteen minutes to gain control of my legs again, I gingerly hobbled to my housemate’s door, woke him up, and asked if he could give me a lift to the hospital. He did, no questions asked.

By this point, I had figured out that I was experiencing something worse than a panic attack. But this meant I didn’t have the language to describe what had happened to me. When I reached the front desk of the hospital emergency department at around 1 am, I picked the closest condition I could think of: “I think I’ve had a seizure…”

The receptionist asked me to take a seat in the waiting area as the nurses were all busy. I turned to see around fifty empty chairs. I didn’t give it a second thought. In an underfunded and understaffed NHS, I believed the receptionist.

“If this was a seizure, you would have choked on your tongue” a nurse opined ten minutes later. Mistake #3 was a doozy. Part one was pre-judging me as a patient before I gave her a proper patient history. Part two was making the egregious error of assuming all seizures were Tonic-Clonic.⁸ I didn’t know at the time, but there are many types of seizures that present in many different ways.⁹ For instance, in some types of seizures, like myoclonic seizures, sufferers usually don’t pass out.¹⁰

The nurse proceeded to make Mistake #4. She told me that getting on to antidepressants was tough; I needed to get through the initial side effects before the medication would start to work.

I saw a resident doctor next who asked what had happened. I told them a shortened story, along with the symptoms I suffered, to the best of my ability. The doctor started a typical series of tests. Most came back pretty normal.¹¹ Then came the knee reflex test.¹²

As a naïve 23 year old, I had no idea what a “normal” reflex was supposed to look like. So, when I nearly kicked the resident doctor in the face after a light tap on the knee, I commented like an enthused child “Wow, it’s crazy how strong reflexes are!” then produced a dumb smile. She did not smile back.

Instead, I saw a deeply concerned glare directed at my knee. My face dropped immediately. After testing other reflexes on my body, producing similar limb-propelling results,¹³ the resident – with the once joviality in their voice changed to stern seriousness – stated that they needed to talk to the attending doctor and quickly exited the room.

The attending made Mistake #5 by not quickly rechecking my super-reflexes. Mistake #6 soon followed in a mostly one-sided conversation. “It is most likely just anxiety,” the attending remarked (paraphrasing). If the attending had taken a patient history, they might have realised that before taking Sertraline, I had never experienced a clinical anxiety symptom. At the very least, by allowing me to speak, they might have noticed my strange calmness.

I was sent home with a second occurrence of Mistake #3 – to keep taking the Sertraline. As this was a separate, independent incident, I’m calling this Mistake #7.

All the clinicians I saw that night made Mistake #8 – a missed symptom. A day or so later, my mum noticed my eyes juddering side to side. Ocular Clonus.¹⁴ Her brother had nystagmus¹⁵ growing up, which is why she spotted it.

Within a couple of hours, seven mistakes were made by four different clinicians. Quite the astonishing statistic. Each failed to recognise a potentially life-threatening syndrome. Unfortunately, this was not the end of the mistakes.

---

When I arrived home, I did not know whether I should take the next Sertraline tablet. My feeling was no, given the severe reaction, and I had already decided I was going to halve the next dose. But two clinical professionals had advised I continue. I wasn’t a doctor, and I knew my brain was in no state to make such an important decision.

I decided to call the university counselling service. I wanted to talk with my counsellor - someone I thought would be neutral. The administrator picked up the phone. They explained that it was not university policy for students to talk with counsellors outside of booked appointments. Through tears, I pleaded with the administrator to please give me two minutes, I could wait for a few hours, just two minutes. “No” was the response. But they could email me an e-leaflet containing strategies to deal with anxiety. “Okay”, I said dejectedly, and hung up. The first tip of the e-leaflet was to do a relaxing activity, like reading a book. I deleted the file.

Mistake #9 was soul-crushing. In my hour of need, the university counselling system was absent. Those two minutes could have been more helpful than the entire six-hour appointment structure.¹⁶

In the end, after talking with my mum, I decided to come off the Sertraline. When I rang my GP to inform him of my decision, I was asked three or four more times “Are you sure you don’t want to stay on the antidepressants?” Mistake #3 a third isolated time – which makes it Mistake #10. At every point throughout this story, I felt like clinicians were willing me to stay on the SSRI.

If I had listened to any of them, there is a possibility it would have resulted in my death. The signs and symptoms pointed to a serotonin syndrome – albeit a strange and rare case.¹⁷ Serotonin syndrome occurs when there is too much serotonin in the brain. It can be deadly. When John Mayer sings “I’ll be dreamin’ of the next time we can go into another serotonin overflow” in the song “Love on the Weekend”, he is either singing about a suicide pact or doesn't quite understand the ins and outs of brain chemistry.

I was experiencing mild to moderate symptoms, but had I continued to take the Sertraline at the same dose, one can imagine things becoming serious quickly.¹⁸

At the end of our conversation, my GP rounded off one of the most stressful and horrifying experiences of my life with Mistake #11. He confidently told me that once I came off the SSRI, the symptoms I was experiencing would disappear. I was relieved. For a few days.

The symptoms, while abating a little bit, did not disappear. In fact, some new ones arrived. For instance, severe panic attacks and hypochondria, and, well, it is easier to copy in the list I made at the time with the symptoms I was left with. I had experienced none of the following before taking the Sertraline¹⁹ :

• Nausea

• Palpitations

• Pupil dilation

• Dizziness

• Confusion

• Violent Shivering

• Sweating without cause

• panic attacks (approx. 10 per day)

• occasional very severe Panic Attacks²⁰

• difficulty concentrating

• anxiety attacks (very frequent and spread throughout the day

• loss of balance

• Twitches/facial tics

• “Brain Zaps”

• Numbness (extremities mainly)

• Hypersensitivity

• Hypochondria

• Insomnia

• Headaches of varying types (Dull, sharp and short, tension)

• Sharp pains all around the body (randomly come and go - they would trigger brain zaps)

• Diarrhoea

• Visual “fireworks”

• Hyper-reflexes

• Raised blood pressure

• Shortness of breath

• Fatigue

• Doom feeling and dropping sensation (associated with balance and anxiety attacks)

• Weakness in legs

• Restlessness

• Decreased appetite

• Emotionally and mentally all over the place

• Depression

• Generalised anxiety

And so began a six-year battle to find treatment.

---

During those six years, the mistakes did not end. It took six months for the serotonin syndrome to finally be diagnosed. I was prescribed Venlafaxine, an SNRI, ignoring a warning in my medical record that it should not be prescribed.²¹ In a care plan, my fourth psychiatrist rewrote my history based on a fifteen-minute phone call.²² I was given ten official diagnoses – only three of those were related to my condition. In Psyverse #0, I mentioned perhaps the biggest mistake of all, not diagnosing me with generalised anxiety disorder despite having clear symptoms. The most common mistakes were logistical and administrative errors. My lithium treatment was delayed by 42 days, mostly through administrative mess-ups.²³

Psyverse #0: The need for a psychiatric multiverse

My story is not unique. In medication errors alone,²⁴ it has been reported that an estimated 237 million occur every year in the UK.²⁵ Most of these, around 75%, are estimated to cause no harm to the patient - many are caught before they even reach the patient. An estimated 2% have the potential to cause harm,²⁶ with approximately 22,303 contributing to deaths.²⁷

The number of patients who die each year from medical mistakes is a controversial subject. It has been significantly overestimated in the past. In a 2013 Manuscript, it was claimed that 440,000 deaths due to preventable adverse events occur annually in the United States. This was followed by a 2016 commentary which claimed 251,454 deaths by medical error per year, making it the third leading cause of death in the United States. The 2016 commentary was widely reported in news articles and widely referenced in the academic literature. Unfortunately, both of these statistics were unfounded.²⁸ The actual figure is probably around 25,000 deaths by medical error per year in the US.²⁹

But from my experience, focusing on mistakes causing direct harm misses the point. A much greater problem is the lack of trust created by “non-harmful” mistakes. Every time a clinician did not know what medication I was on, an administrator did not pass on the correct information, or my medication was prescribed with incorrect instructions, trust was lost. I became paranoid not because of my anxiety condition – paranoia was not there at the beginning – but because of the additive nature of the small mistakes I kept experiencing. I knew that one of these mistakes could cause me harm.³⁰

Loss of trust between patient and practitioner means a loss of communication. I started to withhold information and accentuate certain symptoms to get the treatment I needed. To be clear, I didn’t start out this way. I started out as honest as I could possibly be. But when the honest information I provided led to misdiagnoses, misunderstandings and mismanagement, acts by clinicians became a threat to my health. The mistakes led to the expectation of a lack of competency.

As I will explore, there are many reasons for these mistakes occurring. There is no simple cause. In fact, health services have improved patient safety dramatically over the last twenty years. Nevertheless, the medical field is still playing catch-up. We are not near the end of the road yet.

---

1. Including mistakes by me…

2. I have thought that maybe if I had the chance to talk more in depth about my problems, perhaps I wouldn’t have needed the antidepressant, but this is a counterfactual. I don’t know what would have happened. I was very depressed. Taking an antidepressant seemed to be the logical option. No one else had any other suggestions.

3. Selective Serotonin Reuptake Inhibitor. They block the reuptake of serotonin by a presynaptic neuron, resulting in more serotonin in the space between neurons (synaptic cleft) and therefore increasing the likelihood of serotonin activating receptors on the postsynaptic neuron.

4. I want to make clear at this point that I am not against antidepressants, far from it. They help a lot of people. I believe it is important to consider all the risks of antidepressants, from common to rare. I don’t, however, want to leave readers considering antidepressants with the impression that what happened to me is common. I was extremely unfortunate.

5. In the days following my severe reaction to Sertraline, I subsequently developed severe panic attacks. So, with the benefit of hindsight, I can confidently say this was magnitudes worse than a panic attack.

6. It is an emergency call service like 999 (or 911 in the US), but for less severe or less time-sensitive events, to check whether something is an emergency or not.

7. On reflection, I did find it a bit strange that none of the tests involved the lower half of the body. According to the 111 service, I suppose the only thing they thought was worthy of an emergency was a stroke. I guess the other life-threatening conditions just aren’t important enough!

8. Previously known as Grand Mal seizures.

9. See this on types of seizures.

It may also interest you to learn what seizures are. I think many might have the impression that it is neurons firing randomly. But it is the opposite. From Grace Lindsay’s Book Models of the Mind:

What are neurons doing to create these strong signals during a seizure? They’re working together. Like a well-trained military formation, they march in lockstep: firing in unison then remaining silent before firing again. The result is a repeated, synchronous burst of activity that drives the EEG signal up and down over and over again. In this way, a seizure is the opposite of randomness  – it is perfect order and predictability.

I always allow myself a little chuckle when I hear the myth that we only use 10 percent of our brain. Using 100 percent of our brain (whatever that means) is not genius-level functionality, it is a seizure.

10. Myoclonic seizures cause a quick, uncontrollable muscle movement with no change in awareness. See here.

11. My blood pressure and heart rate were a tad high, and I had a slight temperature.

12. Note: A bit of a spoiler here, but unfortunately, despite the resident doctor being by far the most competent, they technically made Mistake #4.5. In the Hunter Serotonin Toxicity Criteria Inducible Clonus is checked before a reflex test. I was never checked for Clonus.

The reason this is only #4.5 and not #5 is because in order to carry out the Hunter Serotonin Toxicity criteria, you first have to suspect Serotonin Syndrome. At this stage, I’m not sure the resident doctor suspected this yet.

It should have been checked after the results of the knee reflex test, but this would’ve required the resident to challenge the attending. Easier said than done. (I plan to talk about culture in a later Psyverse.)

13. This symptom is called hyperreflexia.

14. Ocular Flutter & Opsoclonus are terms I see too. I’m not entirely sure if they are the same thing, or if each term describes something slightly different? I do know that the terms describe a different collection of vectors for saccades. If a medical practitioner could help me out here, I would be grateful!

15. Nystagmus and Ocular Clonus are different. See this video for an explanation.

16. If you would forgive my dark humour as a result of this episode, I sometimes quip that university counselling, and other short-form counselling, are “six sessions and a funeral”.

I believe there is not enough time to learn the skillsets needed to cope with mental health problems - not even CBT. I’m struggling to think of any skill in which you can achieve a reasonable degree of competence in six one-hour sessions. You cannot speedrun counselling.

I believe the university counselling six-session structure can be counterproductive, as they were in my case. They produced a false sense of security.

(To my younger readers, the dark joke above is based on the film title “Four Weddings and a Funeral”.)

17. Serotonin syndromes are not commonly thought to occur at such low doses of an SSRI.

18. I am mindful I might be scaring people into not wanting to take SSRI’s. Which isn’t what I want. There was something different about my brain compared to the vast majority of other people. Unfortunately, I don’t know what that was. Perhaps I had a weirdly high or weirdly low number of serotonin receptors (and something went wrong with up/downregulation), or one part of my brain wiring resulted in a traffic jam of serotonin, which then cascaded to the rest of my brain. Perhaps I had too few monoamine oxidase enzymes to break down serotonin.

Ultimately, I’ll probably never know. Nevertheless, it is important for people (clinicians and patients) taking/prescribing SSRI’s to know the signs and symptoms of Serotonin Syndrome.

Note that many webpages on serotonin syndrome will say something like this (from Mayo Clinic):

Although it's possible that taking just one drug that increases serotonin levels can cause serotonin syndrome in some people, this condition occurs most often when people combine certain medications.

Which may not be accurate. According to a 2024 review by Simon, Torrico and Keenaghan:

[S]erotonin syndrome reports to The United States Food and Drug Administration Adverse Event Reporting System revealed that nearly half of serotonin syndrome cases involved using a single serotonergic agent. [citing Culbertson et al. and Scotton et al.]

19. Apart from depression, and some of the milder symptoms like headaches I had experienced before - but only acutely. It was now chronic.

20. These are copied from old notes. I fell into a habit of capitalising “Panic Attacks”, because in common discourse, panic attacks have come to mean “anxiety attacks”. Which are much milder in nature. I capitalised it to emphasise that mine were not the everyday variety.

21. At the dose I was taking, it was essentially a weak SSRI, slowly but surely worsening my condition over the three years I was on it. I never noticed for three reasons:

1. The decline was exceptionally slow

2. I put a lot of work into making cognitive adaptations (through counselling, meditation, etc.), which gave the illusion of improvement. Improved functioning through better management of symptoms does not equal treatment. I realised this too late.

3. I tried to come off a couple of times, however, my condition would worsen rapidly once I started. I mistook this to mean the medication was doing something when, in fact, it was the start of withdrawal symptoms (when I finally came off the medication, the withdrawal symptoms were very bad).

22. The call was about increasing the dose of lithium (I was on the lowest dose at the time).

But, to my horror, it was used as an assessment. I was not told this. The care plan was a surprise through my letterbox.

My psychiatrist exaggerated and fabricated parts of my story (for instance, he insinuated I was drinking heavily at university when this wasn’t true). Some parts were factually incorrect - he said I had taken olanzapine when I hadn’t. I was mischaracterised too - he said I was relying too heavily on medication and not trying hard enough at counselling (despite having over 250 hours of counselling over the previous four years of many different modalities). Medication was the only option left.

23. I was in a lot of pain, and this prevented me from trialling other treatments.

24. According to the NHS resolution page, medication errors are

an error in the process of prescribing, preparing, dispensing, and administering, monitoring or providing advice on medicines. Medication errors can occur at many steps in patient care, from ordering the medication to the time when the patient is administered the drug.

25. This is based on a 2020 study. My gut instinct tells me that the number seems too high.

It is important to acknowledge that there is significant uncertainty, primarily due to a lack of data. From the limitations section of the study:

Source studies were generally conducted in small numbers of English centres…. Estimates of the total number of errors represent the sum total of errors at each stage rather than the errors that actually reach patients.

This study only considers medication errors under the responsibility of healthcare professionals and care staff, without including errors in administration and monitoring by patients and their caregivers…. We had to assume that the number of items prescribed in primary care equated to the number dispensed, which will have led to an underestimate of prescribed items, and any estimates of associated errors […].

[W]e were not able to make direct links between errors and harm, or what proportion of errors occurring at different stages of the medicines use process reached patients, and what proportion of those errors reaching patients caused actual harm. Therefore, the estimates of error prevalence are generated from completely separate data from the data used to generate estimates of harm. We have had to assume that the errors we have estimated to occur will lead to the burden that we have estimated will occur […].

A major, necessary, assumption in the estimation of the burden was that definitely avoidable ADEs constitute harm from errors. Estimated burden only included short-term costs and patient outcomes, as we had no data on burden of errors managed in care homes, and therefore it is likely to be an underestimate. Some key source studies from which the burden of errors was estimated were at least 10 years old, or from non-UK countries in scenario analyses.

It is also important to note that (I think) repetitions (like a typo in a repeat prescription) and minor errors are included in the estimate, too.

For instance, for my repeat Bupropion prescription, the pharmacist mistook the XL version of Bupropion for the SR version. The XL version is supposed to be taken once per day with SR twice per day. So my instructions were to take the XL tablets twice per day, each spaced eight hours apart, instead of two tablets together once per day (150mg *2). This was never corrected.

26. Which is still over 4 million medication errors.

27. This number is probably high, considering it accounts for around 10% of hospital deaths. Based on preventable death percentages (see also here), the number is more likely to be around 3 to 5 thousand.

28. [1] K. G. Shojania and M. Dixon-Woods, ‘Estimating deaths due to medical error: the ongoing controversy and why it matters: Table 1’, BMJ Qual Saf, p. bmjqs-2016-006144, Oct. 2016, doi: 10.1136/bmjqs-2016-006144.

[2] B. L. Mazer and C. Nabhan, ‘Strengthening the Medical Error “Meme Pool”’, J GEN INTERN MED, vol. 34, no. 10, pp. 2264–2267, Oct. 2019, doi: 10.1007/s11606-019-05156-7.

29. Based on the approximate 3 - 5% preventable death error rate from:

[3] H. Hogan, R. Zipfel, J. Neuburger, A. Hutchings, A. Darzi, and N. Black, ‘Avoidability of hospital deaths and association with hospital-wide mortality ratios: retrospective case record review and regression analysis’, BMJ, vol. 351, p. h3239, Jul. 2015, doi: 10.1136/bmj.h3239.

[4] T. Rogne et al., ‘Rate of avoidable deaths in a Norwegian hospital trust as judged by retrospective chart review’, BMJ Qual Saf, vol. 28, no. 1, pp. 49–55, Jan. 2019, doi: 10.1136/bmjqs-2018-008053.

Though it is important to note that determining the exact statistics is quite hard. What constitutes a preventable death due to medical error? There will usually be multiple factors, and concluding that medical error was the most prominent one is not straightforward. Especially with hindsight bias.

30. I am not alone in this conclusion. See:

[5] M. Schlesinger and R. Grob, ‘When Mistakes Multiply: How Inadequate Responses to Medical Mishaps Erode Trust in American Medicine’, Hastings Center Report, vol. 53, no. S2, pp. S22–S32, 2023, doi: 10.1002/hast.1520.

(Also, regarding the paranoia not being there in the beginning, you can check my list of symptoms above.)