Study's Life Extension Claim for The Pill Adds Up to Only 4-1/2 Hours

by Heidi Stevenson

14 March 2010

The new study that's currently being hyped in the new media as showing that birth control pills improve life expectancy is one of the most twisted pieces of obfuscation I've ever seen. The claims reported in news media do not match the claims made in the British Medical Journal's report. The study itself is the epitome of junk science and the use of statistics to obfuscate, and it's an example of why you should ask who paid for it.

So, Who Paid For It?

The list of funders of the Royal College of General Practitioners’ Oral Contraception Study⁽¹⁾ includes these pharmaceutical giants:

• Schering AG
• Schering Health Care
• Wyeth Ayerst International
• Ortho Cilag
• Searle

As ever, when Big Pharma money is involved in research, you can expect it to be skewed in favor of their drugs—and this one shows just how bizarre the twisting of results can be.

What's a Woman Year?

One the first things to jump out is the term woman year. It's the basis of all data in the study. It refers to the total number of years the women were followed. So, if Mary signed up at age 20 and died at age 25, she counts for only 5 woman years. If Jane signed up at age 31 and died at age 57, she counts for 26 woman years. Together, Mary and Jane count for 31 woman years.

Thus, women who die early get less weight in the study.

It's worse than that, though. Women were signed up at different ages and times. Here are the recruitment ages of women in the study, along with whether they ever used or never used oral contraceptives:

Age at Recruitment	Never Used	Used	Notes
<30	8,922	18,323	Oral Contraceptive Users 10,483 (36.39%) were added after age 30. Never Used Oral Contraceptives 8,384 (48.45%) were added after age 30.
30-39	6,605	8,712
40-49	1,747	1,748
>50	32	23

Nearly all the women in the study were white. Almost no blacks or browns were included. That's a major flaw. It does not represent the vast majority of women.

Rate of Death Per 100,000 Woman Years?

All of the study's results are reported on the basis of woman years. The tables show results in terms of 100,000 woman years. Keep in mind that these woman years represent only the years that women were in the study.

A woman who died at the age of 26 and was registered for the study at the age of 24 counts as only 2 woman years in this study. Only 2 of her 26 years counted. A woman who entered the study at age 20 and lived to age 58 would have a weight 19 times greater than the woman who died at 26: 58 - 20 = 38, 38 years in the study is 19 times more than 2 years.

Reporting the number of women who die at a certain age and comparing that to the total number of women in a study is meaningful. Reporting deaths as a risk per 100,000 woman years a study was performed is an idea that could only have been developed by researchers trying to obfuscate the truth.

At least 24% of the enrolled women were lost to follow up. The authors concede that "...the full dataset contains approximately two thirds of an estimated 1,800,000 woman years of observation that would have occurred if nobody had been lost to follow-up." This alone is enough to declare the study invalid.

In the year 2007⁽²⁾, an earlier report from the study stated that the results showed a mortality rate 20% less than the general female population because women with chronic conditions were not enlisted. Don't women with chronic conditions count?

Then, there's the fact that length of contraceptive use was not tracked. If a woman took an oral contraceptive for any length of time, she was counted as a user, and her experience was given the same weight as a woman who took them for twenty years. The only information given about length of usage is that the average was 44 months. Since the study went to lengths to assure that issues of smoking and social class were taken into account, this stands out as a glaring error, one that should result in significant questions about its sincerity.

What Are The Death Rates Reported?

Even using the absurd basis of deaths per 100,000 woman years, the study's claimed results still don't pan out as reported.

Age at Death	Never Users (#Deaths)	Ever Users (#Deaths)	ARR*
<30	6	25	2.85
30-39	50	133	1.06
40-49	159	368	0.93
50-59	393	812	0.82
60-69	569	916	0.87
>70	570	610	0.87

Keep in mind that 24% of the women in the study are not represented by these numbers. Then consider that the study's authors are saying that women who have taken the pill and live through childbearing years are less likely to die, but it's at the expense of the deaths of young women taking the pill.

It gets even more complex than that, though. These figures are all jury-rigged. They've been extrapolated from a concept of woman years, which makes little sense in the real world. They also contradict earlier results produced when GPs were providing data. These results, taken from the study's own tables, are of deaths reported by GPs in women under age 45 from all causes, from cancer, and from circulatory diseases:

	Time since last oral contraceptive use:
	Never used (ARR)	Current and <5 years (ARR)	5-9 years (ARR)	≥10 years (ARR)
All causes	1.00	1.08	1.76	0.97
All circulatory diseases	1.00	2.04	3.08	2.31
All cancers	1.00	0.82	1.91	0.39

This information given by GPs for women up to age 45 doesn't seem to correlate with the information given for the study as a whole.

Finally, the authors say that the overall results are that women who take the pill are at lower risk of dying by a factor of 12/100 (ARR of .88), the overall risk of dying from circulatory disease is 14/100 lower (ARR of .86), and of dying from cancer is 15/100 lower (ARR of .85).

Skewed Results?

These results must be skewed. There can be only one reason for using a calculation of death rate according to woman years, and that is to obscure the real story told by this study. The other flaws described above clarify even more the absurdity of the study's results.

To even further muddy the waters, the results were rather coyly summarized as:

The estimated absolute reduction in all cause mortality among ever users of oral contraception was 52 per 100,000 woman years.

For all their twiddled figures, the best they could come up with was lengthening life by 52/100,000 of a year. That is about 19/100 of a day. All the ballyhoo about this study adds up to a claim that an average of approximately 4½ hours of life is provided by taking birth control pills!

Just think of all the juryrigging that had to be done to come up with a pitiful result like that.

Do you have an opinion about this? Click here to comment!

References: