Monday, July 30, 2007

A list of mistakes

Our little girl turned four the other day. I see Sofie, recovered from James' birth four weeks ago, and James, eating like the golden pig he is, thriving. Caitlin's first months – actually her first couple of years – went much worse. And Sofie took over two years to recover.

Not long after Caitlin was born, the high-risk pregnancy specialist at Tromsø hospital asked me for a list of the mistakes that I saw in Sofie's treatment. I wrote over 4,000 words back. Maybe he was hoping for a lever to institute some changes. Maybe he thought it might make me feel better.

If he was aiming for change - from our later experiences, and what we've heard from others, it didn't seem to do any good. Maybe posting it will give an idea of just how badly pre-eclampsia can be handled (managed seems an inappropriate verb here).

The photo is of Sofie and Caitlin. If anyone can offer a good explanation for the petechiae on Sofie's chest, I'd be interested to hear it. It must have been something done in the few hours she was in ICU when I wasn't allowed in. Best I've heard is that it's from a Valsalva, implying (as best I can work out) that she had an abnormal heartbeat at some point.

The letter follows below the fold..............................


Thank you very much for giving me the opportunity to raise my concerns regarding the manner in which Sofie’s treatment was handled at the University hospital of Northern Norway (UNN).

[SNIP some introductory stuff]

Obviously, medicine is not my area, so I may make some egregious mistakes in the comments that I provide below. If so, I’ll apologise up front, and trust that you will points these out to me. I have no intention of being antagonistic in what I write, but I also want to be explicit, so again, I apologise up front if my Australian bluntness appears as rudeness. I have no desire to be rude. My main reason for writing this is to help reduce the likelihood that others will go through what Sofie and I have been through, or worse.

In the time since it became obvious that Sofie was pre-eclamptic (now over a month ago), and particularly since Sofie had her seizure, I’ve read up a bit on eclampsia / pre-eclampsia / HELLP, and the papers that I’ve read are listed at the end of this letter. Where necessary I refer to them in the text. I’d be very surprised if I hadn’t missed some of the important literature, and I’m sure you can point me towards important work that I’ve missed, if necessary.

[SNIP some more introductory stuff]

You may think it appropriate to pass this letter on to other members of medical staff in the obstetrics unit for their consideration. I am very happy to leave that decision with you, and if you think it necessary, please feel free to do so.

It seems to me that Sofie having an eclamptic seizure after being diagnosed with pre-eclampsia three weeks previously means that the onus is on the director and members of medical staff of the obstetrics unit to demonstrate how this was not a case of dangerous incompetence. I’ve divided my questions into a series of dot points, broken into what I see as conceptually distinct areas.

Monitoring prior to Caitlin’s birth (I think I have the time lines correct, but if I’ve made mistakes, my apologies, and I’m sure they will be corrected when checking through Sofie’s records).

Sofie first was referred to hospital after a high blood pressure reading on a Friday, her week 34 day 4 (34/4). She was admitted for observation 34/5, and at that time [Dr A] (the German doctor who saw her, who I presume is a senior resident? I think I’ve got his name right) discussed delivering Caitlin relatively quickly. Sofie was discharged from hospital on the Monday (35/0). She saw medical staff (whose names I forget) on the following Wednesday and Friday, and had blood tests done on both these days. She then saw [Dr B] on the following Monday (36/0). [Dr B] discontinued blood tests. At the time, I queried why the tests were discontinued. My recollection is that the general thrust of [Dr B]’s answer was that she was a highly experienced obstetrician, and in her judgement the tests were unnecessary, as blood tests only revealed when pre-eclampsia was advanced, and breakdown products were becoming obvious. Monitoring requires a comparable time series, and discontinuing blood tests seems to me to introduce a break in the time series that requires justification. Do you consider discontinuing blood tests at this point of a patient’s monitoring a suitably risk-averse approach when treating pre-eclampsia.

[Dr A] tested Sofie for hyperreflexia and clonus, but to the best that either Sofie or I can remember, he was the only person to do so prior to Sofie’s eclamptic seizure. Is our recollection correct? If so, why did other medical staff not carry out this basic, simple but potentially very informative test?

A minor comment: Sofie’s illness clearly showed a two-day cycle. One day she’d be quite ill, the next day she’d be a little better, the following day she’d be a little worse than the first day. Perhaps the best way to describe it is a sine wave with a two-day period, but with a generally downward trajectory. I don’t know if this is news, or useful, but perhaps it might be some help (I’ve seen nothing on this in the reviews I’ve read).

When blood tests were actually carried out, what blood tests were carried out? Were tests oriented towards assessing the possibility of HELLP syndrome carried out? I provide a series of dot points of tests for which I’d be curious to know the values. If it is unethical to provide me with these (I guess it is unethical without Sofie’s consent), we can discuss what to do about this when we meet. I’ve taken my list of tests from the Wiltin and Sibai review in Hospital Physician in February 1999. I’d done so for the following reasons [SNIP]. Sibai is clearly a player in studies of pre-eclampsia /HELLP; the timing of the paper (1999) means that it’s not so old as to be completely out of date, but not so new that people could not be aware of it; the journal is not obscure; what’s provided in the review is more or less a recipe for blood tests for HELLP diagnosis. So, what were the results over time for Sofie’s levels of :

Lactate dehydrogenase
Serum aspartate amino transferase
Lactate dehydrogenase [yes, a typo, I know]
Platelet count?

And did she have abnormal peripheral smears at any time?

At least twice, it has been suggested to me that Sofie’s pre-eclampsia was not “severe”. This is given in justification for the delay in inducing Sofie. How can this definition of not “severe” be rationalised as a reason to not proceed with induction earlier than it was? I’m not talking about the day before Sofie was induced; I’m talking about from 37/0 or appreciably earlier. If it is because Sofie’s blood pressure was not sufficiently high, then the implication is that “severe” pre-eclampsia occurs only when blood pressure cannot be kept under control with medication (which is not a definition that I’ve seen written anywhere, if it exists in the refereed literature I’d be interested to see it). I recollect seeing at least one systolic blood pressure reading of Sofie’s of over 170 anyway. So, even though on medication for hypertension, Sofie had a blood pressure that went over 170 systolic at least once, and over 110 diastolic more than once as I recall (if my recollection is incorrect, I’m happy to be corrected). She also had headaches, complained of visual disturbances to nursing staff at least once, and had some photophobia at least once that I can recall (although no-one ever asked us about that). Also, Caitlin demonstrated an abnormal (retarded) pattern of foetal growth in the latter stages of Sofie’s pregnancy. (As I’m sure you are aware, foetal size estimates that fall inside the 95% confidence intervals of a nonlinear regression of foetal growth are not necessarily the same as normal foetal growth). How is this not “severe” pre-eclampsia?

The UK eclampsia study reported by Douglas and Redman (1994) makes it obvious that there is not necessarily a linear (or predictably nonlinear) relationship between “severity” of pre-eclampsia and likelihood of an eclamptic seizure occurring in all cases. The study does not make clear that failing to deliver a patient (who is past 34 weeks) whose pre-eclampsia is clearly getting progressively worse is in some way not increasing the risk to both mother and foetus. I haven’t seen anything in the literature that I’ve read to suggest this, but if it exists, I’d be happy to view the paper(s). There is a logical difference between attempting to manage for an eclamptic seizure that comes with no warning (Broughton Pipkin’s (1995) thunderbolt) and allowing a case of pre-eclampsia to continue several weeks after initial diagnosis and when past 34 weeks of pregnancy. In the first (the thunderbolt), the likelihood function of mortality or morbidity over time is unpredictable (perhaps chaotic), in the second; the likelihood function is almost certainly exponential.

It has been argued to me twice now that what goes on with pre-eclampsia is a matter of risk management (although not necessarily put in exactly those terms). I accept this. However, I haven’t seen anywhere in the literature that suggests a suitably risk-averse approach is to stop monitoring via blood tests, nor to allow a patient to progress to the point where even to the untrained eye (and nose) can detect that she has become seriously ill (see 2b, below) before starting induction. To then stop induction for reasons that seem to relate more to what staff felt like doing than for the best interests of the patient also does not seem acceptably risk averse to me.

Just before Caitlin’s birth

I know that we have discussed that induction was stopped on Friday, but there’s an issue of process that I would like to raise. You are the high-risk pregnancy specialist for the unit, yet your decision (direction? order?) to induce was overridden. How was this decision to override reached, and who was responsible for the decision? It seems to me that the process by which your decision was overridden must be investigated. Whoever made the decision should have the point clarified that it is not only the decision, but the process by which the decision was reached, must be rectified in the future. That is, it seems to me that it’s not just that the induction was delayed, its how the induction was delayed that is important. I realise this is more of an issue for the director of the obstetrics unit at UNN, but as you’re the high-risk pregnancy expert, this may create more problems for you in the future also.

On the Friday morning that initial induction was started, I could tell that Sofie was clearly extremely ill. Her perfusion (clear in her eyelids) and her smell indicated to me that she was much more ill than she had been previously. If I, with no training other than first aid can pick this, how was it missed so badly by nursing or medical staff that induction was ceased on that day?

I left Sofie round 2100 (I think) on the Saturday evening when she was induced. Sofie tells me that she was left alone for much of the night (I arrived just before 0500 on the Sunday morning, after she called me). Over those roughly 8 hours, Sofie spent time in her room with only another patient, or in the corridor alone, and when she went to the toilet she locked the door, as no-one told her not to do so. She was not under constant surveillance by medical or nursing staff during this time. I believe that this is appallingly poor management of her over this period, given the risk of eclamptic seizure that she faced over that time. Had she been struck by an eclamptic seizure during the night, I suspect that a coroner’s inquest would agree with me. I believe that you must be made aware of this, as this seems to me to be a blunder that must never occur again.

In the few hours after Caitlin’s birth

I’ll preface these comments with a couple of quotes from Walker’s (2000) review: “Delivery is the ultimate cure for pre-eclampsia, but most maternal deaths occur post-partum”; and “Continued close monitoring is required after delivery”, both on page 1264. I believe that Sofie’s monitoring post-partum was demonstrably inadequate, for the following reasons:

In the time from when I arrived at around 0500 until Sofie had her seizure sometime around 1500-1600, [Dr B] visited Sofie on one occasion, to check the swelling of her right labia (labium?). Given the risks associated with inducing Sofie as late as she was, and the need for close monitoring stressed in the literature, why was no doctor present at birth and on a regular basis for the hours thereafter? If the answer is a lack of staff, this would imply that future cases of pre-eclampsia must be dealt with in a manner that is more risk-averse for the mother.

Sofie had no injuries from falling as a result of her seizure. This is due to me. When Sofie got out of bed, I always accompanied her, not because anyone told me to, but because it seemed sensible to me. When Sofie started shaking prior to going stiff as a board, I recognised that she was having an eclamptic seizure, supported her as she fell, and then laid her on the ground as she started convulsing. [Nurse C], the nurse who was then responsible for observing Sofie, responded magnificently to what was going on, organised a diazepam injection instantly, and instructed Sofie’s mother to hit the call button. [Nurse C] was (in my opinion) the best nurse who cared for Sofie during her time at UNN, and I am very glad that she was present when Sofie had the seizure. I am left with a rather unpleasant memory of Sofie having a seizure in my arms. For me, this is a very small price to pay for knowing that, thanks to me, Sofie did not suffer greater injury than from the seizure itself. The issue here is that reliance of a patient’s partner to perform initial first aid for a life-threatening illness is rather unusual hospital policy.

We have discussed in general why magnesium sulphate was not used, but I still question the wisdom of not considering magnesium sulphate prophylaxis post-partum, particularly when Sofie did not urinate for several hours after giving birth. If the answer is “it’s not done in Norway”, this implies that magnesium sulphate prophylaxis s not permissible in Norwegian hospitals under circumstances when a woman is at extreme risk of eclampsia. If I am correct and this was coupled with poor early monitoring, inadequate monitoring during birth, and not delivering a pre-eclamptic patient until 37/6, then this seems to me to be a rather odd approach to providing medical care. If the argument is that patients on magnesium sulphate prophylaxis require monitoring in an intensive care unit (ICU) - well, in this case, the ICU was almost empty that day, so why not send her there? If the manner in which Sofie was treated is indicative of how dangerously pre-eclamptic women are treated at UNN, then more women are going to end up in ICU on magnesium sulphate– why not get them there before, rather than after, an eclamptic seizure?

Sofie’s care in the ICU.
I have one issue regarding this, for your information, as you’re the high-risk pregnancy specialist and this may be useful background for you.

The nurse who was on duty over night did not seem particularly concerned when Sofie’s monitor alarm went off. I recall three occasions when the monitor alarm went of multiple times. My recollection may be slightly amiss (that Sunday wasn’t the easiest day I’d ever had, and my memory might be a bit off). What was happening was that the magnesium sulphate was affecting Sofie’s heartbeat in such a way that her pulse doubled on the monitor. After the second set of alarms (or perhaps the third), the night nurse summoned the intensivist, who reduced Sofie’s dose of magnesium sulphate. I had to request the night nurse to check Sofie at least once during a series of alarms. On at least one occasion, the night nurse’s solution to Sofie’s raised pulse rate was to wake her, apparently assuming that her pulse was racing from a bad dream. I certainly did not get the impression that he understood that magnesium sulphate affected the heart, and that this can, on rare occasions, lead to the death of a patient. Had I been aware of this possibility at the time, I would have insisted on the intensivist checking Sofie each time that there was a series monitor alarms. Is it acceptable that a patient’s partner should get the ICU night nurse to investigate why a monitor alarm is going off? If the night nurse really did not understand the need to keep an eye on a patient’s cardiac output when the patient is on magnesium sulphate, is that nurse really competent to be working in an ICU? Perhaps this makes the reasons for my comment to your students about t-wave abnormalities and magnesium sulphate a little clearer.

After Sofie’s time in ICU
These issues are less crucial (that is, they were probably not life threatening), but they should be raised.

As you are aware, Sofie had a labial haematoma from delivery. [Dr B] investigated the swelling just after delivery and, as I recall, said at the time that it was not a haematoma, but was oedema. A few hours after leaving ICU, Sofie had a lot of pain when attempting to urinate (to the point where she could not urinate, and so reduced her fluid intake to compensate), and could feel that her labia was swollen. This caused her considerable pain and mental distress. It was approximately 11 hours after leaving ICU before a doctor saw Sofie, after I had made multiple requests to the nursing staff for a doctor to check her condition. The nursing staff started putting ice packs on the haematoma a while after Sofie initially complained of her pain. I see several issues associated with this episode where it seems to me that things could have been done better.

As I recall, [Dr B] initially misdiagnosed the haematoma as oedema. This, of itself, is probably not a big deal (although it does little to reassure me regarding her competence). Clearly there were many deliveries that Sunday, and I am sure that [Dr B] was very busy. However, given that Sofie developed (or had) HELLP syndrome, her low platelets were bound to make the haematoma worse than it might otherwise have been. Therefore, it seems sloppy to me that no member of medical staff instructed that the oedema / haematoma be checked at some stage.

Further to this, it was Sofie who had to point out that she had this problem. This is despite her recent move from ICU, and a guarantee that she would receive close monitoring, and had a predictable problem (see (i) above). Is this taking monitoring seriously enough?

Controlling Sofie’s blood pressure was clearly a major issue after she emerged from ICU. The (predictable) pain that she suffered from the haematoma is likely to have affected her blood pressure, isn’t it?

The same argument (as for iii) can be made for her mental distress arising from the haematoma.

It was something in the order of 7 hours from when Sofie first called a nurse to express her considerable pain until she saw a doctor. To me, this does not seem to fit the description of “close monitoring”. If the labour ward lacked the medical staff to deal with Sofie’s case, why wasn’t she left in ICU? They had the bed space there. I realise that this is not your responsibility, but if you’re the one dealing with high-risk pregnancies, then this issue will arise for you again.

Monitoring of Sofie’s blood pressure after her discharge from ICU was carried out extremely poorly. Again, this is really an issue for the director of your unit (I think), but you’re the person who deals with high-risk cases, when this matters the most. The incident not long before Sofie was discharged, when her blood pressure was recorded as 176/116 is the outstanding example of a systemic problem in the unit. Sofie’s blood pressure (if the measurement was to be believed) rose back up to where it was a trigger for watching for “severe” pre-eclampsia. As it is possible (albeit unlikely) that Sofie could have had a seizure more than a week after giving birth, it appears that this reading was treated in an extremely cavalier fashion.

Is it acceptable that, after the measure was made, Sofie made what was functionally a clinical self-diagnosis and decided that the measure must be wrong?

Is it acceptable that I had to point out that the reading needed doing again using a mercury sphygmomanometer?

Was it acceptable that the duty doctor did not think to query the manner in which the blood pressure was measured, rather than have Sofie and I do his or her job?

What was the duty doctor doing that was so important that he or she had to issue instructions by telephone, rather than personally visit Sofie? It should be possible to extract this information from the records for that evening. (This is not a rhetorical question).

Following from (b) above, there are important issues here regarding monitoring in general for the unit (again, back to the point about you being the high-risk specialist). I know a lot about monitoring, albeit in the context of time series of estimates of animal abundance. Without going into detail (which I can do if you want), an unreliable data point is worse than no data point at all when monitoring. I’m sure you’re aware of this, but the importance of this seems to need to be stressed to medical and nursing staff in the unit.

There are some generalities that I think need to be raised as well. Yet again, these may be issues for the unit director, but as you’re the high-risk pregnancy specialist these may be bigger problems for you than for anyone else in the unit.

[Dr A's] decision to induce Sofie as early as possible after 35/0 (which seems to have been his decision?) was rescinded. Given what followed, revoking this decision looks like a bad judgement call on someone’s part. It seems to me to be worth checking the process by which this decision was revoked. It may be possible to determine if there are ways in which this decision-making process can be improved. This may be really important.

I was struck by the way in which many of the nursing staff seems to be looking for resolution of listed symptoms rather than of an illness itself. For example, some nurses would take a series of blood pressure measurements until they got one that they were happy enough with, and go with that measurement. When Sofie could not urinate post-partum, she was encouraged to force herself to urinate, as urination, supposedly, would indicate that her risk of eclamptic seizure was reduced. Although it was not put as explicitly as that, Sofie was encouraged to make an effort to urinate, as if by urinating slightly, the risk of eclampsia would vanish. To me, this indicates a mindset that concentrates on listed symptoms, rather than that real indicators that risk associated the physiological process of concern were diminishing. I suspect that this problem is pretty common in hospitals around the world. My other guess is that you’re only too familiar with what I’m talking about, but if someone like me, with no medical training, can see it, then perhaps its worth someone like me pointing it out.

There’s a final big picture issue regarding multiple doctors seeing a patient (that again, of which I guess you’re only too aware): who is responsible for the problems that I’ve described? This assumes that I have a reasonable case regarding at least some of the questions / comments above. Some issues seem systemic and so a matter for the director of the unit, some seem issues that individual medical staff need to consider. Obviously, I don’t have any answers on this (other than that I presume there are legal precedents in Norway that provide direction)- I just think I should raise this issue.

In your opinion, in how many of the dot points above do I need to be correct before my unhappiness with the manner in which Sofie was treated can be described as reasonable? It would seem to me that for some of them (e.g. 1d,e; 2c; 3b; 4a), the answer is one only, but then perhaps I’m biased.

Finally, I’d like to thank you for the manner in which you treated Sofie.


Yours sincerely

References I’ve read through (for your information)

Broughton Pipkin, F. 1995. The hypertensive disorders of pregnancy. British Medical Journal 311: 609-613.

Dekker, G. and Sibai, B. 2001. Primary, secondary and tertiary prevention of pre-eclampsia. The Lancet 357:209-215.

Douglas, K.A. and Redman, C.W.G. 1994. Eclampsia in the United Kingdom. British Medical Journal 309: 1395-1400.

Higgins, J.R. and de Swiet, M. 2001. Blood-pressure easurement and classificationin pregnancy. The Lancet 357: 131-135.

Magee, L.A., Ornstein, M.P. and von Dadelszen. 1999. Management of hypertension in pregnancy. British Medical Journal 318: 1332-1336.

O’Hara Padden, M. 1999. HELLP Syndrome: recognition and perinatal management. American Family Physician. .September 1999, downloaded from:

Witlin, A.G. and Sibai, B. 1999. Diagnosis and management of women with Hemolysis, Elevated Liver Enzymes, and Low Platelet Count (HELLP) Syndrome. Hospital Physician February 1999; 40-49.

Walker, J.J. 2000. Pre-eclampsia. British Medical Journal 356: 1260-1265.

Waterstone, M., Bewley, S. and Wolfe, C. 2001. Incidence and predictors of severe obstetric morbidity: case-control study. British Medical Journal 322: 1089-1094.

No comments: