Cervical Cancer Mortality Before and After the Pap Smear

The CDC claims that the decline in cervical cancer mortality is largely the result of pap testing. [1] There is no proof of this—no control trials, hard data, or studies to account for changes in other risk factors; simply a claim that correlation is causation. While HPV has remained relatively constant many cervical cancer risk factors, like smoking [2][3], DES use [4][5][6][7], number of children born to a woman [8][9], age at first birth [10], etc., began to decline concurrently with the introduction of, and expansion of participation in, screening. None of the claims made that screening is the cause of the decline in cervical cancer have accounted for these changes in risk factors. Additionally, there was a large increase in hysterectomies [11] overlapping the decline in cervical cancer mortality, which the CDC ignores. As women without a cervix cannot get cervical cancer, total hysterectomy patients are at no credible risk of cervical cancer mortality. [12]

The only outcome study we have, performed by Dr. Angela Raffle in the UK, [13] demonstrates an extreme lack of benefit from screening: a doctor and nurse would need to perform screening for 35 years straight to save one life. This suggests that the cause of the decline in cervical cancer rates could not be due to screening, or even primarily because of it. The mortality reduction reported by Dr. Raffle due to screening is simply too small to account for the majority of decline in cervical cancer mortality.

We collected historical mortality data for cervical cancer for every country for which it was available through the World Health Organization's IARC. [14] We then excluded all countries that did not have a screening program or that did not have a consistent screening program identifiable in peer-reviewed literature. We also excluded countries for which a definitive start date for a cervical cancer screening program could not be identified. Of 194 countries, 24 met all criteria and were included.

Since no control trials exist and the only well designed cervical cancer screening outcome study ever conducted [13] doesn't prove that screening has the benefits the CDC claims it does, we decided to graph the data from multiple countries before and after screening was introduced. If screening is the cause of falling cervical cancer mortality, we would expect (1) that mortality would only begin to fall following the introduction of screening and (2) mortality should fall fairly rapidly following the countrywide rollout and population uptake of those programs. Neither of these outcomes was observed in the data.

The rate of change was calculated by subtracting the 100k crude mortality rate from the prior year's 100k crude mortality rate. Before and after rates of change were averaged to determine if mortality fell faster following the introduction of screening. Negative numbers mean mortality declined. Positive numbers mean mortality increased. No real difference, or a worsening in mortality, was observed for 17 of 24, or 71%, of countries plotted. 7 countries showed a faster decline in mortality after screening was introduced. 14 countries showed no real difference after screening was introduced, defined as a change of +/- .1 deaths/100k, or < 1 death per 1,000,000 women annually. 3 countries showed faster declines in mortality before the introduction of screening using our method of comparing rates of change. It will be obvious to readers looking at the graphed data that several countries actually show increased mortality following the introduction of national screening programs. This means more women died of cervical cancer after screening was introduced in those countries than died before it was introduced. This possibility did not occur to us during initial study design and our method doesn't account for it. This outcome obviously doesn't support the idea that screening is lowering cervical cancer mortality in those countries.

The change in mortality before and after screening did not exhibit the dip we would expect if cervical cancer screening were very effective. As an example of what we would expect to see were screening and treatment highly effective [15] is a graph of what polio mortality looked like both before and after an effective intervention was introduced. If screening was as effective as claimed, we would expect a similar dip in mortality rates following its introduction. Such a dip doesn't occur. Instead, we see a minor effect in only some countries. This is the result we would expect from an intervention that is not very effective at the population level. This tends to support the conclusion reached by Dr. Raffle relating to the limitations of screening. It does not support the claim put forth by the CDC that "this decline is largely the result of many women getting regular Pap tests[...]" [1] .

Given this data, we believe that factors (confounders) other than cervical cancer screening are likely the primary cause for the decline in cervical cancer mortality.

Please note that exact start dates for screening in the US are less certain than in other countries. This appears to be mainly due to the fact that screening was invented in the US. We know screening was occurring opportunistically in the US at very low rates back to its introduction as a medical procedure in the 40s-50s, we also know that screening was introduced as a federal program in 1990, which is the date we try to use for most countries as screening tends to be highly sporadic before national efforts occur. In terms of substantial uptake of screening in the US, it occurred earlier than 1990 due to doctors mandating screening to access birth control. In the US legalization of birth control likely marked the start of uptake, which occurred in 1965 following Griswold v. Connecticut. It is probable that enough women were participating in screening by 1970 that, were screening the cause of falling mortality rates, it would have been detectable in the data. With that said, you will note that cervical cancer mortality has been falling in the US since the first data point in 1950, which precedes any organized screening effort or widespread adoption of screening within the US.