Trial discontinuation, non-publication and non-contactable investigators

One in five surgical randomised controlled trials are discontinued early, one in three completed trials remain unpublished, and investigators of unpublished studies are frequently not contactable. This represents a waste of research resources and raises ethical concerns regarding hidden clinical data and futile participation by patients with its attendant risks. To promote future efficiency and transparency, changes are proposed to research governance frameworks to overcome these concerns.

[http://www.bmj.com/content/349/bmj.g6870]

PSA test for prostate cancer screening does more harm than good

Some doctors are better than others. And some are pathetically behind the curve. Why did the doctor I visited order a PSA test given that 1. I'm asymptomatic and 2. PSA screening has been shown to do more harm than good?

We know that the sensitivity and specifities of screening tests are imperfect. We also know the prevalence of the diseases they're trying to detect tends to be low. Hence, screening tests have high false discovery rates (the ratio of false positives to total positives). We also know the perils of overdiagnosis and overtreatment--detecting and treating abnormalities/cancers that will never be the cause of death of the individual. Therefore the rationale for ordering some screening (as opposed to diagnostic) tests is dodgy at best.

[https://www.harding-center.mpg.de/en/health-information/facts-boxes/psa]

Prostate cancer screening and innumeracy

"Research has shown that numeracy among the American public is shockingly low." Wouldn't be surprised if "shocking" innumeracy isn't confined to the United States.

Many laypersons do not understand that a control group is needed in order to evaluate the effectiveness of a medical intervention or test. By attending primarily to cell A and by giving insufficient attention to cells C and D, many members of the general public will not comprehend a central basis for the task force’s negative opinion of the PSA test.

This insufficient attention to disconfirming information is the good old fallacy of confirmation bias.

Given a 4ng/ml cutoff, PSA screening has a laughable sensitivity of only 21%. And even with a specificity of 94%, the false discovery rate still comes out to a whopping 81% for a cancer prevalence of 6.3%. That just means that out of every 5 men who test positive, 4 were incorrectly tagged. Epic fail.

Now I read in some cancer website that 1 in 7 men will have prostate cancer. Let's round that up to 15%. If we take that as our base rate (prevalence), we obtain a false discovery rate of 61.8%. So out of every 5 we will still be erroneously labeling 3 as having cancer.

Now that's just the false positives--those who don't have cancer. What about the true positives--those who actually have some type of prostate cancer? PSA does not distinguish between aggressive cancers and those that progress very slowly. The latter need not be treated, will probably not cause discomfort or some disability in the patient's life, and will almost surely not be the cause of his death. But the fact that cancer has been discovered almost certainly will prod both patient and doctor to fiddle around with it further, calling for more tests and even (invasive) treatment of sorts.

[http://citrixweb.mpib-berlin.mpg.de/montez/upload/PaperOfTheMonth/arkes.gaissmaier.2012-1.pdf
]

Gawande's The Checklist Manifesto and reduction of surgical complications

Just finished Dr. Atul Gawande's The Checklist Manifesto: How to Get Things Right. It is so well written, moving and gripping that I would've gone through it in one sitting had the rest of life been kind enough not to butt in. In 2007 to 2008 a team of doctors and researchers, including Gawande, under the auspices of the World Health Organization performed a study that involved hospitals and surgical teams in eight developed and third world countries (USA, Canada, New Zealand, UK, India, Jordan, Philippines, Tanzania) to test the effectiveness of a simple 2-minute, 19-item surgical checklist the team had developed. In both his book and in a TED vid (starts at around 14:00) Gawande shares the incredible and almost unbelievable results they obtained: a drop in the number of deaths by 47% and in overall major complications by 36%.

Imagine being able to cut mortality rate in half! At very little cost and with no invasive procedures, drugs and side effects. Given how checklists have been effective in a number of areas including the airline and construction industries, it stands to reason that medicine could benefit from having checklists in place to make sure vital items/steps are not missed or incorrectly performed.

And now for the nitpick. First and quite clearly this was not a randomized, blinded clinical trial so certain biases may have affected the results. As Gawande himself pointed out, among other things, he was concerned the Hawthorne effect may have kicked in (but as he says the fact that observers were in place during the baseline data collection phase ought to have canceled out any effects resulting from that effect.) Secondly, and specifically with how Gawande presented the study findings, the double digit numbers can be misleading and can inflate the benefits of the intervention (the checklist). And this is because Gawande chose to impress us by providing the relative risk reduction and not the absolute numbers or percentages. For those figures I had to search for their published study. And so in their NEJM paper the results they obtained were summarized as follows:

The rate of death was 1.5% before the checklist was introduced and declined to 0.8% afterward (P=0.003). Inpatient complications occurred in 11.0% of patients at baseline and in 7.0% after introduction of the checklist (P = 0.001). 

Given the p values these results are statistically highly significant. Using the absolute risks above the absolute reduction in death is 1.5 - 0.8% = 0.7% and in overall complications ARR = 11% - 7% = 4%. In other words, 1 out of 143 inpatients was spared an untimely and needless death, and 1 in 25 spared major complications. These numbers I believe provide a much more realistic picture. Because death and complications are--fortunately--not that frequent (1.5% and 11%, respectively), quoting relative risk reductions has the tendency to mislead and to exaggerate the benefits (or risks).

Nonetheless, given that such a simple, inexpensive, apparently harm-free intervention does reduce major complications, particularly death, ought to be sufficient to spur the medical field to study checklists further, and implement, fine tune and improve the checklist that Gawande et al. have created.

Annual mammography did not reduce breast cancer mortality

Long term follow up of women (aged 40 to 59 years) showed no reduction in breast cancer mortality in those who had five annual mammograms.

During the entire study period, 3250 women in the mammography arm and 3133 in the control arm had a diagnosis of breast cancer, and 500 and 505, respectively, died of breast cancer. Thus the cumulative mortality from breast cancer was similar between women in the mammography arm and in the control arm (hazard ratio 0.99, 95% confidence interval 0.88 to 1.12).

[http://www.bmj.com/content/348/bmj.g366]

Screening > overdiagnosis > overtreatment > harm

There is no doubt in my mind that screening has, in general, been oversold, represented in some cases as a magic bullet that will save far more lives than it actually can. It’s been a relatively recent realization on the part of physicians that overdiagnosis is a real problem because it leads to overtreatment, which can cause harm. --David Gorski, oncologist

More physicians are now aware of the problem. But the general public must be informed for they're the ones who are harmed.

IMPROVE-IT trial of Vytorin

Missed this when it came out last November. The much awaited results of the 18,000-patient, 7-year, randomized, controlled, double-blind, mulitcenter study of Vytorin are out. Ezetimibe (in conjunction with simvastatin) isn't a complete lemon after all as the ENHANCE study showed (although that trial used a surrogate endpoint rather than outcomes of real interest to patients and doctors).

The primary endpoint of IMPROVE-IT is a mishmash of cardiovascular events including death. The absolute risk reduction is 2%. But what's interesting to me is that there is zero reduction in mortality rates--ezetimibe doesn't reduce death rates. What it does clearly do is reduce LDL and the implication (as the researchers reveal) is that having an even lower LDL is better given the reduction in CV events compared to simvastatin alone.

The primary and secondary endpoints, being composites, do not paint a clear picture of what Vytorin exactly improves. Looking at the forest plot of the individual components of the composite reveals that there are only two--heart attack (MI) and ischemic stroke--which showed unambiguous risk reductions (p = 0.002 and p = 0.008, respectively). The improvements in both events are modest. For MI the absolute reduction is 1.7% while for ischemic stroke it's a mere 0.7%, both for a period of 7 years. In other words, we'd have to put 59 people on Vytorin for 7 years to reduce the number of heart attack incidents by 1 and the number of patients we'd have to give Vytorin, again for 7 years, to reduce ischemic strokes by 1 would be 143.

Moreover, as has been highlighted by some critics, coronary heart deaths (CHD), cardiovascular deaths (CVD), and all-cause mortality are the same for those on statin alone and on Vytorin (the 95% CI includes HR = 1). It's quite clear ezetimibe does nothing for mortality rates.

Ultimately, whether the benefits outweigh the cost (which includes harms and not just financial burden) will be a judgment call both by doctors and patients.

sources:
http://www.medscape.com/viewarticle/835030#vp_1
http://my.americanheart.org/idc/groups/ahamah-public/@wcm/@sop/@scon/documents/downloadable/ucm_469669.pdf

Simple checklists can save lives

A simple 2-minute preoperative checklist tested in the US, Europe, Asia, Africa. Brought complications down by 35% and death rate by 47%.

Saquib et al. systematic review of screening tests

Here's the latest blow to (most?) screening tests:

Screening for diseases that can lead to death typically does not prolong life substantially; a few screening tests may avert some deaths caused by the disease being screened, but even then it is difficult to document an improvement in overall survival.

The only tests which showed benefit in terms of disease-specific mortality are "ultrasound for abdominal aortic aneurysm in men, mammography for breast cancer, and fecal occult blood test and flexible sigmoidoscopy for colorectal cancer."

The authors conclude: "Among currently available screening tests for diseases where death is a common outcome, reductions in disease-specific mortality are uncommon and reductions in all-cause mortality are very rare or non-existent."

I think that last point is the most crucial--that practically no screening test affects all-cause mortality. 

Ethicists who don't have the public and patients' welfare in mind

Had we participated in a trial would any of us say the following:

When I signed up for that clinical trial, I was putting my health at risk and exposing my life to examination only for the benefit of one pharmaceutical company’s sales or one professor’s publication record. I never intended for any independent scholars to be able to double-check whether the drug actually works.

Would we ever be part of a trial that said in the fine print:

You’re signing a form that guarantees that our paid analysts and ghostwriters will write about this data only if it is in our financial interest. Thanks to your signature, we will be able to convince 'ethicists' that you didn’t want anyone to double-check our analysis, and we will thus be able to misrepresent the trial more effectively.

To ethicists who say it is unethical to re-examine trial data to make sure the analysis and conclusions are correct, perhaps you should divulge any and all conflicts of interests that you may have, specially connections to and gifts and remuneration from companies and research groups involved in clinical trials. Let's begin the ethics investigation with you.

[http://blogs.bmj.com/bmj/2015/01/30/stuart-buck-sharing-data-from-past-clinical-trials/]

Jorge Soto's cancer detection kit

Early detection would be beneficial only if the cancers detected are those that will either eventually kill us or cause us suffering. If it cannot distinguish between nonprogressive, gradually progressive and aggressive cancers, early detection can lead to harm via overtreatment. Secondly, no detection method is perfect. Unless the test has 100% sensitivity and specificity, it will produce false positives, and the less prevalent the disease the worse the test will fare. 

Soto says @9:35 their microRNA test is "accurate"  but accuracy can mean different things. Does he mean that their microRNA test will, for instance, correctly tell us 95 out of a 100 times that cancer is present when given blood samples from patients who we know have cancer? If so then he's talking about the test's sensitivity. I'm inclined to believe this is what he means by accuracy. He might also be defining accuracy in another statistical sense: as the ratio of correct results to the total number of people tested: accuracy = [true positives + true negatives] / population.

Either way, the accuracy figure would be bloated and deceiving, particularly if he's using "accuracy" in the second sense as the example computation below shows. I've taken the liberty of giving Soto's test a very high sensitivity and specificity. As can be seen accuracy increases even as the prevalence of the disease decreases! That's the way to mislead with statistics.

In contrast to statisticians, the "accuracy" that doctors and patients are interested in are the positive and negative predictive values, which are, respectively, the probability of having the disease if the test result comes out positive and the probability of not being sick if one tests negative. And because different cancers have different prevalences, the rarer the disease the more prone to false positives this and any test will be. And so in the example, if the type of cancer is such that one out of every 100 people has it we see that PPV = 49%, meaning if we test positive then we might as well just flip a coin to find out if we have cancer or not, because for every 1000 people who test positive 510 don't.

"Surrogate end points in clinical research: hazardous to your health"

Hear, hear!

A valid surrogate end point must both correlate with and accurately predict the outcome of interest. Although many surrogate markers correlate with an outcome, few have been shown to capture the effect of a treatment (for example, oral contraceptives) on the outcome (venous thrombosis). As a result, thousands of useless and misleading reports on surrogate end points litter the medical literature. New drugs have been shown to benefit a surrogate marker, but, paradoxically, triple the risk of death. Thousands of patients have died needlessly because of reliance on invalid surrogate markers. Researchers should avoid surrogate end points unless they have been validated; that requires at least one well done trial using both the surrogate and true outcome. The clinical maxim that 'a difference to be a difference must make a difference' applies to research as well. Clinical research should focus on outcomes that matter.

[http://www.ncbi.nlm.nih.gov/pubmed/15863552]

Rilotumumab bites the dust

Yet another cautionary medical tale. 

At the beginning of 2014 they said that "several new drugs have recently shown promise for the treatment of gastric cancer, including rilotumumab, ramucirumab, and anti-HER2 agents." And that they hoped "with these new treatments, we will move the needle forward on survival." 

Phase 2 trial indeed showed that rilotumumab increased survival rates. So they went on with Phase 3. With the larger trial it didn't take too long (fortunately) for rilotumumab to show its true colors. The number of deaths in the experimental arm actually went up versus the control. And so the trial was promptly discontinued.

Never count your chickens until they've run the gauntlet and passed phase 3 testing.

Wonder what lies ahead for that vaunted new antibiotic Teixobactin.

Pretty dismal figures from a lung screening study

In "Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening," the authors state that "a total of 96.4% of the positive screening results in the low-dose CT group and 94.5% in the radiography group were false positive results [emphasis added]." Had to read that line several times just to make sure.

The authors also state that there was a 20% relative risk reduction in mortality in the CT vs the radiography group. Besides the fact that translates to an absolute risk reduction of a measly 0.062% (1 in 1,613 screened people), there doesn't seem to be any mention of how a non-screened group would fare or had fared.

To their credit the authors do note that: 

In addition to the high rate of false positive results, two other potentially harmful effects of low-dose CT screening must be mentioned. Overdiagnosis, a major source of controversy surrounding low-dose CT lung-cancer screening, results from the detection of cancers that never would have become symptomatic.... The other harmful effect, the association of low-dose CT with the development of radiation-induced cancers, could not be measured directly, is a long-term phenomenon, and must be assessed in future analyses.

Harm not

Really frustrated at how some doctors don't seem to take into account the potentially serious (and even fatal) side effects of a number of the medications they prescribe. And I daresay practically no physician takes into account or even knows the adverse interactions and combined effects of drugs when the patient is under polypharmacy. 

Primum non nocere. First, do no harm. If only physicians actually took that tenet to heart.

Taming our fear of cancer

I found this a helpful way of visualizing the development of cancer and taming our crippling fear of one of the most dreaded diseases. At every milestone of its development Dr. Gilbert Welch clearly notes that the body has the potential to overcome the abnormal growth.

Unfortunately, what Dr. Welch did not put in this diagram (which, however, he does explain very well in other parts of the book) is the fact that having an invasive cancer does not necessarily lead to death from that cancer. For one, the cancer may be nonprogressive or a slow progressing type such that it will never be the cause of death. Or the cancer may even regress without treatment.

Death from the cancer can happen if the cancer is of the aggressive type and if left untreated (if a reliable treatment is in fact available) and metastasizes.

[diagram shamelessly pilfered from Should I Be Tested for Cancer? Maybe Not and Here's Why by H. Gilbert Welch]

Living with cancer

Look hard enough, biopsy enough tissue, and we will find cancer in every elderly and in the majority of the middle-aged population. To age and grow old is to live with cancer. But not necessarily become sick or die from it.

Cochrane review of screening mammography finds no reduction in all-cause mortality

In the Cochrane review using only the best randomized trials, among women who had undergone screening, death from breast cancer went down, although the 95% confidence interval for the effect includes relative risk (RR) = 1.0 which simply means the finding is not statistically significant, that it is possible there is no difference in breast cancer mortality between the two groups (RR = 0.90, 95% CI 0.79 to 1.02).

But here's the rub. Even as screening apparently reduces the risk of dying from breast cancer, the all-cause mortality (death from all causes) was the same for those who didn't undergo screening and those who did. In other words women's chances of dying is the same whether they get screened or not. Needless to say no woman in her right mind would gleefully queue up for screening knowing that her chances of dying from breast cancer would be reduced only to have her chances of dying from something else increased (including death from the consequences of radiotherapy and surgery to treat breast cancer).

Gøtzsche concludes screening mammography doesn't work

"Screening causes breast cancer." I take that as Peter Gøtzsche's brand of humor. Needless to say screening mammography doesn't actually turn normal cells into cancerous ones. What Gøtzsche means is that screening by its nature detects abnormalities and cancers which are not destined to ever harm or kill. That is, screening causes overdiagnosis. And overdiagnosis almost always leads to overtreatment (doctors will opt to do something rather than nothing, because we currently can't distinguish between early stage cancers that will become life threatening and those that won't). Because the cancer being treated isn't destined to be aggressive, overtreatment harms the patient (unnecessary lumpectomy or mastectomy, radiotherapy with its suite of harms, psychological harms, all the well known nasty effects of chemotherapy, etc.).

[image source: https://www.youtube.com/watch?v=8A9xuLmUHcQ]

The Swiss Medical Board does not recommend screening mammography

Screening advocates (including those who stand to lose financially, professionally or politically from discontinuation of screening), most certainly did not take this report soberly and sitting down: "The report caused an uproar and was emphatically rejected by a number of Swiss cancer experts and organizations, some of which called the conclusions 'unethical.' ”

The Swiss Medical Board's report was made public on February 2, 2014 (www.medical-board.ch). It acknowledged that systematic mammography screening might prevent about one death attributed to breast cancer for every 1000 women screened, even though there was no evidence to suggest that overall mortality was affected. At the same time, it emphasized the harm — in particular, false positive test results and the risk of overdiagnosis. For every breast-cancer death prevented in U.S. women over a 10-year course of annual screening beginning at 50 years of age, 490 to 670 women are likely to have a false positive mammogram with repeat examination; 70 to 100, an unnecessary biopsy; and 3 to 14, an overdiagnosed breast cancer that would never have become clinically apparent.5 The board therefore recommended that no new systematic mammography screening programs be introduced and that a time limit be placed on existing programs. In addition, it stipulated that the quality of all forms of mammography screening should be evaluated and that clear and balanced information should be provided to women regarding the benefits and harms of screening.

Medical wisdom from over half a century ago.

No one need be afraid of the word cancer. In clinical practice to say that a person has cancer gives as little information about the possible course of his disease as to say that he has an infection. There are dangerous infections that may be fatal and there are harmless infections that are self-limited or may disappear. The same is true of cancers. Cancer is not a single entity. It is a broad spectrum of diseases related to each other only in name. It is no harder to live with these various types of cancer, no worse to die with them than with many other diseases.

Dr. George Crile ends his article saying that "[L]ife is change and growth. Cancer is a part of life."

[Life, October 31, 1955, p.128-130]

Fish oil still a dud

The 2004 Cochrane systematic review found no evidence that fish oil alters mortality rate or cardiovascular events. Apparently since then an overwhelming majority of "rigorous studies" still couldn't find any.

Wonder how many millions of dollars people all over the world have wasted on this supplement.

Table of benefit and harms of screening mammography

One of my beefs with this table is that it doesn't provide a column for number of all-cause mortality averted with screening against nonscreened women (over the specified time frame). Nevertheless, the number of false positives and the number of overdiagnosed lesions (most of which will probably be treated and therefore turn women into patients and harmed) should give women pause when it comes to screening. Hence, well tabulated data such as these must be made available for women to make an informed choice about screening. 

[http://www.cancer.gov/cancertopics/pdq/screening/breast/healthprofessional/page8]

Alendronate reduces risk of fracture

Hmmm, looks like alendronate does reduce the risk of fractures among postmenopausal women with osteoporosis. The Cochrane review says:

The best estimate of what happens to women that have already been diagnosed with low bone density or have already had a fracture in the bones of their spine:

Fracture of the spine

- 12 out of 100 women had a fracture when taking a placebo
- 6 out of 100 women had a fracture when taking alendronate

Fracture in the hip or wrist
- 2 out of 100 women had a fracture when taking a placebo
- 1 out of 100 women had a fracture when taking alendronate

Fractures in bones other than the spine
- 9 out of 100 women had a fracture when taking a placebo
- 7 out of 100 women had a fracture when taking alendronate

General health checks don't reduce morbidity or mortality

Nope. Looking for disease in asymptomatic adults (but perhaps excluding those >65 yrs old) does not decrease the risk of cardiovascular or cancer death, but could lead to overdiagnosis and overtreatment.

"General health checks involve a contact between a person and a healthcare professional to identify signs, symptoms, or risk factors for disease that were previously unrecognised. They are combinations of screening tests, few of which have been adequately studied in randomised trials. For example, although the benefits and harms of treatments for conditions such as hypertension and diabetes have been extensively studied in randomised trials, screening asymptomatic people for these conditions has not."

...

"We did not find an effect on total or cause-specific mortality from general health checks in adult populations unselected for risk factors or disease. For total mortality, our confidence interval includes a 5% reduction and a 3% increase, both of which would be clinically relevant. However, for the causes of death most likely to be influenced by health checks, cardiovascular mortality and cancer mortality, there were no reductions either."

[http://www.bmj.com/content/bmj/345/bmj.e7191.full.pdf]