Bad History....The Cartwright Report

With any historical material as a whole you have to rely that the account given by the historian is an accurate account and reflect the historical record fairly. In most cases, this is done and there have been many excellent writings on any particular topic. The problem is when they don't it can be difficult to tell with revisionist historical accounts which is the factual material and which isn't or whether the topic is represented fairly. This has happened with New Zealand material, particularly because our written history is fairly short and in some cases like with the claims about "Ancient Celtic New Zealand" it's wholly off base - there is no evidence of human habitation much prior to the Kaharoa eruption in 1314AD (+/- 12 years) that coincides with the arrival of the Maori people in New Zealand. The signs appear just before and just after. Getting on to more recent history, does anyone remember the Mason Report into Psychiatric care and hospitals in Auckland that came out in late 1988? That was probably more critical of the health care system and those administering it and was the subject of court injunctions and was published with portions redacted. The Hansard record is an interesting read.

The Cartwright Inquiry that also ran in 1988 wasn't half of it, while overall our health care system has worked well there have been times it's failed. In the case of the Cartwright Inquiry there was poor management of research projects, lack of informed consent for patients and a number of other problems besides. Those things needed to change and even now there are probably areas that could be improved. It wasn't until 1996 that there was a patient code of rights:

• The right to be treated with respect;
• The right to freedom from discrimination, coercion, harassment, and exploitation;
• The right to dignity and independence;
• The right to services of an appropriate standard;
• The right to effective communication;
• The right to be fully informed;
• The right to make an informed choice and give informed consent;
• The right to support;
• Rights in respect of teaching or research; and
• The right to complain.

This was a fair few years ago now and many people would have forgotten or maybe not even followed this in the first place. It's important to remember though that those patients in the system often had their rights disregarded. In the past, patients were meant to comply with health care professional’s directions (this could be termed the "Biomedical Model").
Today, it's patient centred care. The big flaw in the case of National Women's Hospital being there was a lack of informed consent, with patients being enrolled into research without their consent. "Consent" in general could mean that the doctor simply presented their preferred treatment and it was only if the patient objected that any further discussion was made of alternative options (often patients didn't feel they could object but that's another story). Often patients weren't given adequate explanation when undergoing treatment, with them only being told that they had such things as "odd cells". Clinical freedom or autonomy meant doctors were free to decide what they thought best with little or no interference, which in this case meant peer review failed despite attempts to address the issue.

The Cartwright Inquiry meant many changes, and there have been a number of articles which have been critical of the Inquiry over the years. Then in 2009 this book was published, causing debate to come up again: Confused offering adds little insight.

On reading the book there are statements like "...for clinical research it only required consent "if at all possible, consistent with patient psychology." (p67) and "Green did not breach the 1964 Declaration of Helsinki, which included a strong exemption for patient consent in therapeutic research." (p69)

In full this says: "If at all possible, consistent with patient psychology, the doctor should obtain the patient’s freely given consent after the patient has been given a full explanation. In case of legal incapacity, consent should also be procured from the legal guardian; in case of physical incapacity the permission of the legal guardian replaces that of the patient." I don't think that is a strong exemption at all, and certainly Green's patients were capable of understanding any reasonable explanation of their medical care and giving their consent to treatment or participate in research so there was no need to withhold from that from them. It has the effect of excusing the lack of consent, when it was clear it was required and it was only if it wasn't possible (i.e. the patient was incapacitated in some way) that it could be waived. This is an issue, as it appears that the treatment of the topic would lead the reader to see the situation in a particular way, and that would be what the author wishes to present rather than an unbiased assessment of the Inquiry. That isn't the only questionable part, and there has been quite some debate in various venues.

This has also been written about in NZ Skeptic, but the original article fails to look at whether those claims made in the book are valid. There are also some errors, like at the start it says it is considering "...whether or not the Inquiry made a fair and just assessment of the current (1988) internationally accepted management of carcinoma-in-situ of the cervix (CIN3) but the Inquiry spent a lot of time making an assessment of the internationally and local accepted management of CIS as it was in 1966. This has the unfortunate effect of implying that the standards of an different medical era applied during the Inquiry. In addition, the discussion of McCredie et al appears to be done from the abstract, and it's unfair to complain that terms weren't defined in that. Definitions of terms were in the full paper.

This lead me to write a response (unedited version). A response to the article was published in the next edition which was a bit of a surprise given that while the original article may have dealt with the book it wasn't mentioned in the article, although it was referenced. A response was written, but publication was refused with the suggestion that online publication be sought. It seems that the request to get it online has been done and I've received a brief statement from the author via the editor (who I'd asked to forward the article) but it's not readily available so here is the short version of "Orthodoxy? - revisiting The Cartwright Report (as submitted but with a correction and responses) and the long version, with extra material that was edited out for length before submission. I'm sure I've seen the comment before, but can't think of where I saw it but it seems to be that the author doesn't grasp the concept of preventative medicine, where early detection and intervention could prevent major problems down the track. That's probably why pathologists were positive about smear tests as they were part of an era where, for the first time, this could be instrumental in preventing both mortality and morbidity. Vaccines were part of that too.

I managed to get hold of a copy of Sandra Coney's book "The Unfortunate Experiment" on the topic, which was written before the results of the Cartwright enquiry were released. Green in the mid-sixties clearly had the belief that CIS was harmless, he is cited as saying to a GP in a letter that CIS and invasive cancer were two totally different diseases, and that "cases of invasive cancer appeared among women who had abnormal cervical cytology in much the same way as any large group of women followed because they had a broken leg." I'd forgotten about the "flying uterus" case. It's often been said that Green was trying to save women from hysterectomies, but in this case he didn't. Green, in his proposal, indicated that he already had patients being observed with persistent positive smears. In addition around 1960, at least 5 women or as many as 16 had hysterectomies based only on visual inspection and smear results with no biopsy being taken for confirmation. The book states that Green wrote that "serial section studies of lesions thus undisturbed by biopsy are being made." The intention being to observe intact lesions to study in the lab. This could have the effect of either over-treating and removing a healthy organ or alternatively if there was invasive cancer, the treatment could be more complicated and have more risks as this involved local irradiation and then hysterectomy and with the uterus removed the rods could not be inserted. Normally, tissue would be examined in the lab right after the surgery but in this case, some of the uteri went on a trip with Green to New York. Some months later a cable arrived reporting that one of them was an incidence of invasive cancer. This not only complicated treatment, but it delayed it and possibly could cause the spread of the original cancer. Despite further treatment the women died two years later at the age of 34. Then between 1963 and 1965 the rate of cone biopsy at the hospital markedly increased as Green did another special study aimed to correlate smear findings and histology. Some of the reports said there was only a finding of cervicitis (inflammation) that didn't require such radical treatment. The hospital policy of the time required that a series of abnormal smears was needed before proceeding with considering further treatment - either all A3 smears or three A2 smears. That, combined with the other studies that required repeated visits and interventions for follow-up purposes for many years didn't save anyone from anything. That they "saved" some women from a procedure or could treat cancer after it developed doesn't apply.

There is also an response from Phillida Bunkle that might be of interest to readers and a response from Clare Matheson that appeared in the New Zealand Medical Journal.

Trackback URL for this post:

Carrying on with "Population

Carrying on with "Population based screening" p89 In the introduction it is stated not only did McIndoe disagree with Green on the treatment of CIS, but McIndoe was disheartened by his questioning of the value of screening as well as causing 'strife' with community health professionals working on a national programme. Professor Skegg with other others had published recommendations for Population based screen and the following year stated that 'Cervical screening is one of the few effective tools that doctors have for preventing cancer.' Unfortunately the report or that statement isn't discussed at all, as the next sentence moves on to Skegg's coining of the term 'the unfortunate experiment at National Women's Hospital'. The term was used in a letter to the NZMJ in 1986, in response to letter by Green who was disputing the statement that 'those who questioned the value of cytology screening and the invasive potentiality of intraepithelial neoplasia have now been clearly shown to be wrong.' Green had referred to articles published in 1979 and 1978. Skegg responded, pointing out the articles were questioning yearly smears, not three yearly smears as was being advocated. This point is acknowledged, but then the statement is undermined by stating that "...they also outlined the controversies about the effectiveness of pap smears and questioned their contribution..." The point is conceded, but it carries on without examining whether the criticism is applicable and returns back to the claimed 'controversy'. Needless to say, not every debate deserves the label of controversy but in order to paint the picture of Green being right, it's labelled as such. Skegg also said Green had cited 'new articles from Time and Science seven years ago'. Bryder apparently doesn't get this may be sarcasm about the age and the source of the articles. One of the articles can be found here:,9171,946170-1,00.html The debate on yearly tests vs three yearly tests show that less frequently is better and it's possible that doing more often can lead worse outcomes, as there will be a greater number of questionable abnormalities that will need investigation and treatment and false negatives as well. Most of these changes will resolve without treatment. It's quite clear that the authors are not debating the smear itself but the problems associated with annual smear, and point out the effectiveness is obscured as they use opportunistic screening, which means at-risk groups miss out. The debate had moved on from establishing effectiveness which had generally been confirmed to more operational issues.

The fact is with population based screening, a much clearer idea had appeared since 1960 and Skegg roughly dated in the Cartwright Report p198 a consensus on the subject emerging around 1977. It's not easy to understand the position of those who were fine with opportunistic screening but opposed mass screening, as after all mass screening is merely opportunistic screening which has been organised. The paradox is that NWH was one of leading exponents of the use of cytology, but opposed screening. Next up: what the Cartwright Inquiry said.

Moving on with the

Moving on with the introduction, there's very little of what Cartwright actually said, and of on what evidence she based those statements, and there seems nothing else in this chapter on that outside of covering some submissions to the Inquiry. The introduction states it will consider the claim that doctors were pitted against the rest of the world on the subject of population based screening, however there isn't any defining of what we are looking at (such as the difference between mass screening vs opportunistic) here and this becomes important later.

"The Cartwright report stated that, despite a worldwide consensus on its effectiveness, the importance of cervical screening 'has been consistently undermined by the confused statements from Dr Green and some of his colleagues at National Women's Hospital...Even in 1987 [they] Believed that they were right and the rest of the world was wrong in it's assessment of the benefits of a mass screening programme.' Cartwright also asserted that teaching at National Women's Hospital had hindered the implementation of a nationwide screening programme by devaluing the importance of the screening test to students."

You'll notice some ellipses there, normally you use them to show a gap when quoting a paragraph to get the point across without long quotations. In this case, there is about two pages of material between one statement and the other as there is a whole chapter on cervical screening, and relevant material in the teaching and ethics and even patients rights sections. What the report actually said was that in theory screening should be value if most abnormalities found will progress to cancer if left, and if any treatment is effective in preventing this progression. It won't be shown effective unless most of those at risk have the test, the results are read correctly and there is proper treatment. A paper by EG Knox 'Cervical Cytology: A Scrutiny of the Evidence' (1966) is introduced which reviewed the evidence and said that information to determine effectiveness wasn't present at the time. This author had rejected use of RCT's (and these weren't done in the 50's either for similar reasons) as impractical and unethical. The report states (p198):

"It has been obvious from listening to Dr Green's evidence and reading his papers and public comments on cervical screening, that he shared similar doubts to those expressed by Knox in that 1966 paper. His scepticism might have been justified in the 1960s. However, the emerging evidence of progression from CIS to invasion, not only in his own work at [NWH] but also from studies undertaken in other parts of the world, appears to have been disregarded by him or not fully understood. Since the 1960s a far clearer picture has emerged. Centres with well organised screening programmes have shown substantial falls in the incidence and mortality rates from cervical cancer. Evidence of this sort was regarded by expert witnesses as demonstrating the value of cervical screening. Professor Skegg considered that a consensus on the positive effectiveness of population based screening started to emerge in 1977. In [NZ], however, the importance of these findings has been consistently undermined by the confused statements from Dr Green and some of his colleagues at [NWH].

So what she is saying is that by the 1980s, much more evidence had been gained, and few doubted the value of screening with Dr Jordan quoted as saying on that page "I think the debate is not whether it should be set up but how."

The second portion of the quote is p199: "In spite of world authorities' broad agreement on the value of a screening programme, key gynaecologists in the Postgraduate School of [O&G] still dispute the evidence. In an editorial, 'Cervical Human Papilloma Virus Infection and Colposcopy' Professor Bonham, Dr Green, and Professor Liggins criticised a paper by Pagano and colleagues from the Royal Women's Hospital, Melbourne. It had recommended the mandatory use of colposcopy in the management of and follow-up of all patients whose smears suggest HPV infection. The authors say:

"Therefore, before we accept for universal application the recommendations of Pagano and colleagues, would it not be wise to review more critically the conflicting data on current screening programmes, discrepancies in epidemiological data and cohort trends, the role of tobacco and viruses (whatever happened to Koch's postulates?) and, above all, to seek more objectively the natural history of cervical cancer?"

Professor Bonham described this editorial as having been written with 'flair', implying 'provocatively'. I did not draw that inference. Even in 1987 these authors believed they were right and the rest of the world was wrong in its assessment of the benefits of a mass screening programme."

In addition, the editorial quoted Bethwaite which fixed the value of a woman's life at $NZ420,000 based on annual screening and used this question the level of expenditure on a disease that causes 0.4-0.5% of deaths is NZ/Australia. The author did not recommend annual screening and the current NZ recommendations and those of the author were based on 3 yearly screening, at $86,000 per life. The report states "It is disturbing to see this selective use of published data."

So I've flicked to a random

So I've flicked to a random page in the book A History of the "Unfortunate Experiment" and want to check on what the claims are. On page 100 ("Population-Based Cervical Screening" which does not cover New Zealand screening at all) it says "Thus, there were lively local and International debates about the costs and benefits of cervical screening around the time of the Cervical Cancer Inquiry. Green and his colleagues were not pitted against the rest of the world as claimed by Cartwright. This debate continued beyond 1988. A 2005 World Health Organization publication listed the 'hazards' of a screening programme. These effectively summarised the concerns voiced by Green and his colleagues, and show that more than 20 years after Green's retirement they were still regarded as valid...." The summary includes psychological consequences, misunderstanding the meaning of positive test, misunderstanding a negative smear as no risk rather than no risk, false positives and negatives, overtreatment, complications of treatment, opportunity costs and impact of incidental findings during screening.

These things are valid to consider as possible negative impacts and to use to work towards good communication so as to communicate a realistic picture to patients and work towards resolution of the other problems, but the impression given is that screening is still contentious and that these concerns are major ones. However, this is a very large report - (with the quoted portion found on page 214 with further discussion of the implications of the points) and it over-emphasised this and ignores such things as the problems of opportunistic screening (p219) which is pretty much all that existed in 1988 and beyond because of opposition to setting up an organised programme - these have the same or similar problems as listed previously, but it can be added to that it isn't cost effective, the groups being screened are skewed (as found here, middle class women that attended doctors got smears regularly, but those at higher risk because of age and so on generally missed out), it can create sporadic work flow and less proficiency, there can be greater chance of harm due to overscreening, it can be difficult to ensure quality, and women screening positive may not have easy access to further diagnostic and treatment services. It's likely that there were more issues caused by resistance to the idea of screening, and Green used the results of oppotunistic screening to claim that the New Zealand experience showed that screening wasn't useful but as pointed out in the Inquiry, if you screen the wrong groups and then don't treat the one's you find appropriately then it's likely that efforts will be futile and there won't be the desired reductions in morbidity and mortality. False positives and false negatives can be a problem, but they can be managed with the right systems to be at a minimum and in addition, it's always been that if a smear proves positive that I can tell that repeat smears are ordered prior to going further with intervention to ensure that the result isn't a chance finding and this would help prevent false positives leading to overtreatment.

In the event, even including the 'hazards', the report concludes that "There is sufficient evidence that screening for cervical cancer precursors every 3–5 years between the ages of 35 and 64 years by conventional cytology in a high-quality programme reduces the incidence of invasive cervical cancer by 80% or more among the women screened. In women aged 25–34 years, screening at intervals of three years or less may have smaller impact. There is no evidence that screening annually in either age group results in much greater efficacy. Other forms of cytology screening using a validated system at the same ages and frequency can be expected to be as effective as conventional cytology. Efficacy of conventional cytology has been demonstrated only for squamous-cell carcinoma. Screening in well organized programmes is more cost-effective, with less harm due to overscreening and overtreatment, than opportunistic screening."

After this, the book states on p101 that "In 2007 Raffle and Gray commented that, 'In truth, most pre-cancer is transient, and screening therefore leads to major problems of overdiagnosis.' Moreover, they pointed out that 'the high rates of positive tests, most of which are transient minor cell changes, cause major anxiety unless you can change the universal belief that every abnormality must mean cancer.' These modern authors shared similar views to Green." (the pages cited are p20 and p25 of the book 'Screening: evidence and practice' which can be found on google books here - with the relevant pages available to read

That sounds rather surprising, and on looking at the book the authors don't actually have similar views to Green as what they are doing is presenting is a case study of screening in the UK 1950-2005 and discussing and illustrating important issues that screening faced over that time. The difference is that they don't overstress them or use them to argue that organised screening is useless or more harmful than no screening or opportunistic screening. They describe four phases (optimism, disillusionment, organization and realism) commenting that the example shows "a slow transition from a haphazard and largely ineffective activity begun on a wave of unrealistic optimism, to a well ordered programme that met objectives." They say it illustrates the types of problems that have to be solved if you want a screening to deliver a cost-effective public health programme. The first comment is referring to problems faced in the 1950's and that was the same problem Ostor described 1993 was talking about, what pathologists and doctors wanted to see was a steady progression from minor abnormalities through to cancer but the earlier stage of displasia didn't behave like that. There could be regression, and also a single cervix could show varying levels of abnormality that made it hard to predict what would happen. A similar thing happens with prostate cancer, many will be fairly benign, but some agressive and it can't be predicted which is which. That being said, it was clear the higher the grade of lesion, the more likely it was to persist and progress and with CIS it was clear that a significant proportion of cases (if not all given enough time)would continue to invasion. Doctors were aware of this too, Hulme and Eisenberg (1968) in "Carcinoma in situ of the Cervix in Connecticut" say that "When one understands and appreciates the "facts of life" concerning the diagnosis of CIS and related lesions, it becomes apparent that the actual diagnosis of severe dysplasia, CIS or CIS with questionable invasion is less important than an understanding of the natural history of premalignant cervical neoplasia and it's place on the spectrum leading to invasive carcinoma. It is of vital importance that the gynecologist treating patients with CIS or related lesions understand the limitations of the pathologist. A woman with severe dysplasia has the same disease as a woman with a diagnosis of CIS and to arbitrarily treat a patient on the basis of pathologic diagnosis alone belies an understanding of the nature of cervical neoplasia."

The second statement is in the section "organization" and is talking about later changes, where a proper network was formed and there were many improvements, but not all problems were ironed out. Other points made were that there are some cancer cases that cannot be prevented, and unless those limitations are recognised, staff can be inappropriately blamed for those cases. Their solution is simple, to set out screening can achieve and acknowledge what it cannot. They say the ideal would be to avoid those problems in the first place by gathering evidence and good planning with it being systematic and with good quality control. It's hard to get that evidence or do good planning of screening when there is widespread resistance and the method of getting smears (using opportunistic methods) can't demonstrate what the benefits would be if the entire at-risk population was screened and treated appropriately.