pop culture

The Truth about Method Failure

“And then we’re going to need some urine to make sure you’re not pregnant.”

With any patient between the age of 14 to 54, this line is typically how I conclude my initial assessment of an Emergency Department patient. Unless a woman is there with an eye problem or physically does not have a uterus, I (and the provider) insist upon urine pregnancy tests.

“But I can’t be pregnant” is the usual response.

“How do you know?” is mine.

Answers vary from not having sex since last menstrual period (physically impossible – OK) to not having a male partner (also physically impossible – OK) to just not thinking it’s possible (and you have cramping, one-sided lower abdominal pain with a late period? Yeah, go pee now so I can make sure you don’t have a life-threatening ectopic pregnancy please!).

What is striking me as more and more disturbing is how women with an intrauterine device (IUD) or tubal ligations think it is physically impossible for them to become pregnant. Even women on the pill or with an implant like Nexplanon give me confused, blank stares when I insist on urine. “You reduce your risk significantly,” I say, “but it can still happen.” If the blank stares continue, I elaborate and say I have seen women on these all contraceptives have positive pregnancy tests (true story).

But most disturbing to me as of late is how health care providers think it is impossible for a woman to get pregnant while properly using a contraceptive. Just a few weeks ago, I had to quickly call the CT technologists after they took a female patient of childbearing age without asking me. CT contrast is dangerous to a growing baby, and despite all my gripping that it is dangerous that CT cannot tell if a urine pregnancy is done and should call me before taking a patient, CT takes my patients all the time without asking. (Pet peeve #372 at work.)

“We haven’t gotten a urine pregnancy yet,” I explained to the technician.

“She’s got an IUD,” the technician replied.

“You can still get pregnant with those,” I explained.

“Oh, really? I thought it was impossible.”

“Nope. They reduce the risk, but it can still happen. I’ve seen it.”

“Wow. Ok. Good to know.”

“Yeah. I’ll call you when we get a urine preg.” (Which was negative after a long drawn out battle to get a sample.)

WHAT!? Of course she can still be pregnant!

Sure, this is a CT technologist, so he or she has a bachelor’s or associate’s degree and took a licensing test. I don’t know how much education they get on pregnancy. Sure. But I work with physicians, advanced practice providers, and nurses who express disbelief when a female patient who is on some kind of contraceptive comes back with a positive urine pregnancy test. They often conclude the patient was not adherent enough to the regimen when it comes to pills, diaphragms, and condoms. But when it comes to implants, IUDs, and tubal ligations, they’re often in disbelief.

I’m not. I’m not ever in shock when a contraceptive fails. I console the shocked patient, yes, but here’s the thing: any kind of birth control reduces risk of pregnancy, but it never eliminates it. 

One of my dear friends in town had a method failure pregnancy. She and her husband are very excited and happy to be pregnant. Their baby girl is just sooner than they expected. This is not uncommon of unintended pregnancies. Finer and Zolna (2011) estimate that 49% of pregnancies are unintended, of which 29% were mistimed, 19% were unwanted, and 43% ended in abortion.

The catch is that she was using natural family planning (NFP). Two methods concurrently, actually. No provider she has talked to is shocked she got pregnant. I’m guessing most are ranting about the ineffectiveness of the rhythm method (which most providers think is the same as natural family planning) behind closed doors.

As she was talking to me about a month ago, she expressed how in her NFP group boards on social media, she was shocked at how many other woman stated they had method failure pregnancies as well. I was not shocked.  Why? Because here’s the thing: every kind of birth control reduces risk of pregnancy, but it never eliminates it. 

Why the shock at the failure of the pill, condoms, diaphragms, IUDs, and tubal ligations, but rolling of the eyes at the failure of natural family planning? I think it is because we have been habitually fed a lie that effective birth control eliminates the need for abortion because it eliminates unplanned pregnancy.

This is a half-truth.

Medical literature does show time and time again that contraception reduces unwanted pregnancies, which reduces the rate of abortions. Peipert and colleagues (2012), for example, studied the effect of offering free long-acting reversible contraceptive methods (LARC aka IUDs and implants) to 9,256 female participants at-risk for unintended pregnancy in the St. Louis region through the Contraceptive CHOICE Project from 2008 to 2010. The participants chose a variety of methods, a total of 75% selecting a LARC: levonorgestrel IUD – 46%, copper IUD – 12%, subdermal implant – 17%, oral contraceptive pills (OCPs) – 9%, contraceptive vaginal ring – 7%, Depo-Provera – 7%, and contraceptive patch – 2%.

Compared to a control group not offered the intervention in Kansas City and nonmetropolitan Missouri, the St. Louis group had a significant reduction in repeat abortions (P < 0.001), 4.4 to 7.5 per 1,000 per year during the study period compared to the national average of 19.6 per 1,000 per year.  The pregnancy rate for CHOICE participants was 6.3 per 1,000, significantly less than the rate in the United States which is 34.1 per 1,000.

Peipert et al. (2012) concluded, “Providing no-cost contraception and promoting the use of highly effective contraceptive methods has the potential to reduce unintended pregnancies in the U.S.” and “would prevent as many as 41% to 71% of abortions performed annually in the U.S.” Peipert et al.’s (2012) comprehensive proscpetive cohort study emphasizes the fact that every kind of birth control reduces risk of pregnancy, but it never eliminates it. 

Let’s look at some numbers. Assuming the pregnancy rate (6.3 per 1000) of the 9,256 CHOICE participants, there were 58.3 pregnancies each study year (2008-2010). Assuming the lowest rate of abortion (4.4 per 1,000) in the 9,256 CHOICE participants, there were 40.7 unintended pregnancies that resulted in abortion each study year (2008-2010). Therefore, despite most likely using LARC – a highly effective contraceptive method – the participants of the CHOICE program in the 3 years of study had a total of 174.9 pregnancies and 122.1 abortions. This means that a participant of the program had a 1.8% chance of becoming pregnant in the 3 years of the study period, despite being on a highly effective contraceptive method.

Yes, these numbers are significantly less than the national average (34.3 per 1,000 for teenage births and 19.6 per 1,000 for abortions). Yes, this program reduced unintended pregnancies and abortions. But here’s the other part of the half-truth we do not hear enough: no matter how effective the birth control, contraception alone will never end unintended pregnancies and subsequent abortions.

Let’s examine failure rates of some typical contraceptive methods:

Screen Shot 2018-07-07 at 9.04.34 AM
from Trussell, J. (2011). Contraceptive failure in the United States. Contraception83(5), 397–404.  It’s what the CDC uses when citing how effective each method is.

Looking at this data from Trussell (2011), I am annoyed that health care providers all of kinds think true NFP is less effective than other means. I mean, come on. 0.4% unintended pregnancy per year for the symptothermal method with perfect use and a greater unintended pregnancy rate per year for condoms, spermicides, and the sponge?

Most NFP rates in the literature include intended pregnancies, so it is difficult to find accurate data. Most available data is biased or outdated. For the Creighton Model (a mucus tracking method), Hilgers and Stanford (1998) reported a perfect use rate of 99.5% and a typical use failure rate of 96.8% in their prospective study while Fehring, Lawrence, and Philpot (1994) reported a method effectiveness rate of 98.8% in their prospective study. For the Marquette Method (mucus tracking with electronic hormonal fertility monitor), Fehring, Schneider, and Barron (2008) reported a perfect use rate of 99.4% and a typical use rate of 89.4% using a retrospective design. Frank-Herrmann et al. (2007) estimated a 1.8 unintended pregnancies per 100 in 13 cycles (aka about a year) with typical use of a symptothermal method in their prospective cohort study. (The prospective method is often noted to be superior to the retrospective method in cohort studies because looking back on intention can introduce unconscious bias. For these prospective studies, participants recorded their pregnancy plans at the beginning of a cycle instead of after it, which may lead to more accurate data.)

User compliance and willingness to use the method is also a big factor in the efficacy of NFP methods. A prospective randomized control trial of two different NFP methods – ovulation and symptothermal methods – from the late 70s in Los Angelos by Wade et al. (1981) had a pregnancy rate of 13.7 per 100 for the symptothermal method and 39.7 for the ovulation method. However, Wade et al. (1981) also reported high user dropout rates during training and poorer compliance compared to other studies, which could skew the data. Trussell and Grummer-Strawn (1990) also reported significantly higher rates of pregnancy for participants with poor compliance to the method, a rate of 84.2% per year, which is remarkably close to the 85% chance pregnancy per year with no method use reported by Trussell (2011). (Which should surprise no one because NFP is all about not introducing sperm to an egg without any other factors to reduce fertility, so a percentage that is mere decimals away from the percentage using no method makes intuitive sense.)  Analyzing data from various countries from the World Health Organization, Trussell and Grummer-Strawn (1990) ultimately had a 3.4% failure rate for perfect use and 22.5% for imperfect use, noting that NFP is “extremely unforgiving of imperfect use.” (Which, again, makes intuitive sense because the method does not introduce any chemicals or physical barriers to reduce a couple’s fertility, thus the N in NATURAL family planning.)

So, basically, it’s difficult to find a reliable and current method failure rate for any NFP method. After an extensive-ish search of the literature, the most reliable and most current rate I could find was the Frank-Herrmann et al. (2007) rate of 1.8% per year. NFP studies are old, have methodology issues, and the people using the method don’t like it and aren’t compliant. Oh, and providers do not know enough about it, as Pallone and Bergus (2009) pointed out in their overview of the various methods and notable methodology issues in the medical literature.

However, just after I was typing and ranting about NFP effectiveness studies, literally a systematic review of effectiveness rates of NFP was just published by Peragallo Urrutia et al. (2018) in Obstetrics and GynecologyThey estimated first year pregnancy rates on various NFP methods to be the following for typical use (Figure 2) and perfect use (Figure 3):

Screen Shot 2018-08-13 at 8.51.37 AM.png

Screen Shot 2018-08-13 at 8.43.16 AM

Let’s compare and examine failure rates of typical contraceptive methods, including LARCs:

Screen Shot 2018-07-07 at 9.12.59 AM.png
from Sundaram, A., Vaughan, B., Kost, K., Bankole, A., Finer, L., Singh, S., & Trussell, J. (2017). Contraceptive Failure in the United States: Estimates from the 2006–2010 National Survey of Family Growth. Perspectives on Sexual and Reproductive Health49(1), 7–16. http://doi.org/10.1363/psrh.12017

Not included in this table is method failure rates for female sterilization aka tubal ligation. Gariepy and colleagues (2014) estimated that for laparoscopic sterilization, there is a 0.33–0.85% failure rate in the 1st year and 0–1.62% in the 10th year while for bipolar (both tubes) coagulation, there is a 0.03–0.42% failure rate in the 1st year and 0–1.81% in the 10th year. Date and colleagues (2014) reported 180 cases of pregnancy after tubal ligation at their institution in India, 54 pregnancies due to a tuboperitoneal fistula (creation of a new connection) and 25 cases of spontaneous recanalization (the tube spontaneously re-connected). The authors even had one case of spontaneous recanalization 20 years after the procedure. 

OK, health care providers and patients. LARC has a 1.4% failure rate in 12 months (Sundaram et al., 2017), and the symptothermal method has a 1.8% failure rate in 13 cycles, so about 12 months, depending on cycle length (Frank-Herrmann et al., 2007). And the Marquette Monitor had a 0.0% failure rate with perfect use or 2.0% with typical use in the first year (Peragallo Urrutia et al., 2018). Can someone please explain to me exactly how NFP much less effective than a LARC? And can someone please tell me how sterilization is perfect?

With all this data in mind, it is clear that contraception – even sterilization – always carries a risk of pregnancy. Therefore, every kind of birth control reduces risk of pregnancy, but it never eliminates it. 

Method failure is merely probability. It cannot always be due to a method, to the user(s), or any other factor. Method failure is probability

If you’re flipping coins, eventually you’re going to get heads. But coins are supposedly fair with a 50% chance of either option. Obviously contraception is not just 50% effective. But even if you’re flipping a weighted coin where the weight is set so you have a 99% of getting tails, after enough flips, probability would dictate eventually you’re going to get heads.  Same with contraception. If my contraception is 99% effective a year with typical use, even though the probability is weighted against me getting pregnant, with enough metaphorical “coin flips,” I have a high probability of method failure.

Let’s say I’m an 17 year old girl who goes to my primary care doctor because I decide to start having sex with my boyfriend. I’m put on an OCP or “the pill” because let’s also say I don’t plan to start a family until I’m at least 27. Me and said high school boyfriend end up in a monogamous long-term relationship eventually resulting in marriage with fairly regular intercourse.  I have 10 years of wanting to be sexually active but not pregnant, and I’m on that good ol’ OCP from my PCP. What’s my risk of pregnancy? Using Sundaram et al.’s (2017) 7.2% failure rate per year for the pill, over the course of 10 years with standard use, I have a 56% chance of becoming unintentionally pregnant during the 10 years I’m with my significant other.*

So, here’s where the lie of proper birth control eliminates abortion really frustrates me. Frost, Darroch, and Remez (2008) from the Guttmacher Institute estimate that the average American woman spends about 3 decades avoiding pregnancy and only a few years actively achieving pregnancy. So, if I’m the average American woman who for 30 years does not want to become pregnant, with every contraceptive method – even highly effective ones like implants, IUDs, and tubal ligations – I have a collective high probability of experiencing a method failure pregnancy during my lifetime.

Getting pregnant depends on a lot of factors though. This high probability of unintended pregnancy changes if I have underlying, undiagnosed infertility issues, or if I am not sexually active every cycle or during the fertile part of my cycle. This probability is also dependent on my partner not having fertility issues. Nevertheless, if I plan to be sexually active for 30 years without getting pregnant using contraception, it is really unrealistic to never expect to have a method failure pregnancy.

Using Sundaram et al.’s (2017) data, with a LARC (implant/IUD) I would have a 1.4% failure rate per year, so over 30 years, that’s a 35% chance of unintended pregnancy. Using Gariepy et al.’s (2014) data, with a tubal ligation I would have a 1.62% failure rate over 10 years, so in 30 years with the lowest possible failure rate for laparoscopic sterilization, I would have a 6.3% chance of unintended pregnancy. With the highest possible failure rate of 0.85% per year for laparoscopic sterilization, I would have a 23.2% chance of unintended pregnancy.*

This is where the logic of “good” contraception education and abortion baffles me:

Getting every American woman on some sort of birth control will not stop unintended pregnancies. 

Getting every American woman on some sort of birth control will not stop abortions. 

Data on this from the United States is old as in older than me. The best match and most recent study I could find linking contraception and abortion is from the 1980s by Sophocles and Brozovich (1986). The authors found that 19% of patients seeking abortions in the United States between 1982-1984 were using contraception consistently and correctly. The rest either did not use anything or used a method incorrectly or inconsistently. But 19% were using birth control as they should have!

Looking at Spanish abortion and contraception data, Dueñas et al. (2011) noted an overall increase use of contraception and an overall increase in abortion from 1997-2007. Other studies like that of Bajos et al. (2014) utilizing French data do not demonstrate the same correlation. However, every country studied in what I could find even if the abortion rate decreased with overall contraception use, it was never eliminated. 

(Sadly, most modern literature about contraception and abortion has to do with giving women contraception post abortion. My favorite was an article literally called, “Why after 50 years of effective contraception do we still have unintended pregnancy?” (It’s called math, science, and an imperfect way of preventing pregnancy, people.))

Getting every American woman on some sort of birth control only reduces her chance of pregnancy and does not eliminate it. 

Unfortunately, teaching every American about the risks of sexual intercourse with birth control and trying to delay first sexual encounter will probably not stop it either. A Cochrane review of interventions for preventing teenage pregnancy by Oringanje et al. (2016) noted that educational interventions alone have a low likelihood of preventing teenage pregnancy.

So, what is a health care provider to do? The truth is, I don’t know.

I really fear the day I need to counsel my female patients of child-bearing age about sexual intercourse and pregnancy. It terrifies me to have a conversation about risk of pregnancy because I don’t want to come off as judgmental. I don’t care who in the world is having sex. I care that people are in good, healthy relationships. I care that people who decide to have heterosexual intercourse realize there is always a chance of pregnancy unless the woman is lacking a uterus. I care that children are loved and cared for.

With that in mind, I cannot look at data about contraception failure rates and still think contraception is the perfect method for preventing pregnancy that it’s made out to be. I also cannot look at regular condom use and its effectiveness in preventing HIV and other diseases and think it is the perfect answer to sexually transmitted diseases either. As Genuis (2008) wrote, “The ongoing assertion that condoms are “the” answer to this escalating pandemic [of sexually transmitted infections] reminds me of Einstein’s words, “The definition of insanity is doing the same thing over and over again and expecting different results.”” Just the same, the ongoing allegation that effective contraception will eliminate the need for abortion is the definition of insanity. The data does not support this claim at all, though it does validate that contraception can reduce it, as noted in the Peipert et al. (2012) study.

NFP already has a lot stacked against it, and I would be remiss if I did not mention good ol’ psychological tendencies that come into play as well. Confirmation bias is the psychological tendency to perceive new information in a way that supports what the person has already concluded. If I think the world is a mean, horrible place, I’m likely to see someone kicking a puppy as confirmation of that and forget about how someone helped an old lady cross the road.

Parmley (2007) noted in her experimental study for her dissertation that confirmation bias plays a role in health care diagnostics. About half of studied clinicians displayed confirmation bias, even after receiving information about it. Therefore, it should surprise no one that despite new information, a health care provider who thinks little of NFP in general finds a method failure pregnancy from NFP validating of his or her existing opinion while simultaneously finding a method failure pregnancy from the pill or an IUD shocking. Providers and patients ought to really look at raw numbers and study methodologies before forming an opinion about its efficacy, but confirmation bias about NFP is rampant. I cannot tell you how many times I have sat through a lecture making impressions into my tongue and cheeks, trying to hold back while I’m seriously taught 1950s information about the “rhythm method” aka “natural family planning.”

Yes, effective contraception reduces the risk of unintended pregnancy. I am not arguing that at all, and the Peipert et al. (2012) study demonstrated it well. However, to say that effective contraception will eliminate abortion is a blatant mistruth. It’s time we bring this well-hidden lie to light and recognize the basic probabilities of method failure. Even low-probability events are likely to happen over an extended period of times.


Bajos, N., Le Guen, M., Bohet, A., Panjo, H., Moreau, C., & the FECOND group. (2014). Effectiveness of family planning policies: The abortion paradox. PLoS ONE9(3), e91539. http://doi.org/10.1371/journal.pone.0091539

Date, S. V., Rokade, J., Mule, V., & Dandapannavar, S. (2014). Female sterilization failure: Review over a decade and its clinicopathological correlation. International Journal of Applied and Basic Medical Research4(2), 81–85. http://doi.org/10.4103/2229-516X.136781

Dueñas, J., Lete, I., Bermejo, R., Arbat, A., Pérez-Campos, E., Martínez-Salmeán, J., . . . Coll. (2011). Trends in the use of contraceptive methods and voluntary interruption of pregnancy in the Spanish population during 1997-2007. Contraception., 83(1), 82-87.

Gariepy, A. M., Creinin, M. D., Smith, K. J., & Xu, X. (2014). Probability of pregnancy after sterilization: A comparison of hysteroscopic versus laparoscopic sterilization. Contraception90(2), 174-181.

Fehring, R. J., Lawrence, D., & Philpot, C. (1994). Use effectiveness of the Creighton model ovulation method of natural family planning. Journal of Obstetric, Gynecologic, & Neonatal Nursing23(4), 303-309.

Fehring, R. J., Schneider, M., & Barron, M. L. (2008). Efficacy of the Marquette Method of natural family planning. The American Journal of Maternal/Child Nursing33(6), 348-354.

Finer, L. B., & Zolna, M. R. (2011). Unintended pregnancy in the United States: Incidence and disparities, 2006. Contraception84(5), 478–485. http://doi.org/10.1016/j.contraception.2011.07.013

Frank-Herrmann, P., Heil, J., Gnoth, C., Toledo, E., Baur, S., Pyper, C., … & Freundl, G. (2007). The effectiveness of a fertility awareness based method to avoid pregnancy in relation to a couples sexual behaviour during the fertile time: A prospective longitudinal study. Human Reproduction22(5).

Frost, J. J., Darroch, J. E., & Remez, L. (2008). Improving contraceptive use in the United States. Alan Guttmacher Institute, (1), 1-8.

Gariepy, A., Creinin, M., Smith, K., & Xu, X. (2014). Probability of pregnancy after sterilization: A comparison of hysteroscopic versus laparoscopic sterilization. Contraception., 90(2), 174-181.

Genuis, S. J. (2008). Are condoms the answer to rising rates of non-HIV sexually transmitted infection? No. British Medical Journal336(7637), 185.

Hilgers, T. W., & Stanford, J. B. (1998). Creighton Model NaProEducation Technology for avoiding pregnancy. Use effectiveness.

Oringanje, C., Meremikwu, M. M., Eko, H., Esu, E., Meremikwu, A., & Ehiri, J. E. (2016). Interventions for preventing unintended pregnancies among adolescents. The Cochrane Library.

Pallone, S. R., & Bergus, G. R. (2009). Fertility awareness-based methods: another option for family planning. The Journal of the American Board of Family Medicine22(2), 147-157.

Parmley, M. C. (2007). The effects of the confirmation bias on diagnostic decision making.

Peipert, J. F., Madden, T., Allsworth, J. E., & Secura, G. M. (2012). Preventing unintended pregnancies by providing no-cost contraception. Obstetrics and Gynecology120(6), 1291.

Peragallo Urrutia, R., Polis, C. B., Jensen, E. T., Greene, M. E., Kennedy, E., & Stanford, J. B. (2018). Effectiveness of fertility awareness–based methods for pregnancy prevention: A systematic review. Obstetrics and Gynecology. doi: 10.1097/AOG.0000000000002784

Sophocles, J. A., & Brozovich, E. M. (1986). Birth control failure among patients with unwanted pregnancies: 1982-1984. The Journal of Family Practice22(1), 45-48.

Sundaram, A., Vaughan, B., Kost, K., Bankole, A., Finer, L., Singh, S., & Trussell, J. (2017). Contraceptive failure in the United States: Estimates from the 2006–2010 National Survey of Family Growth. Perspectives on Sexual and Reproductive Health49(1), 7–16. http://doi.org/10.1363/psrh.12017

Trussell, J. (2011). Contraceptive failure in the United States. Contraception83(5), 397-404.

Trussell, J., & Grummer-Strawn, L. (1991). Further analysis of contraceptive failure of the ovulation method. American Journal of Obstetrics and Gynecology165(6), 2054-2059.

Wade, M., McCarthy, P., Braunstein, G., Abernathy, J., Suchindran, C., Harris, G., . . . Uricchio, W. (1981). A randomized prospective study of the use-effectiveness of two methods of natural family planning. American Journal of Obstetrics and Gynecology., 141(4), 368-376.

*I’m not a statistics major. It took some looking through my statistics notes, using online resources, and this handy geometrics calculator. Basically, probability of an event is not as simple as adding each probability together. There’s a crazy formula involved, and I called pregnancy a “successful” event and did not include the probability of having a previous successful event. The probability for each lower slightly if there is a previous successful event. (If anyone wants to comment on the statistics I used, have at it. My brain hurts from figuring this much out. This is why if I ever do research I will hire someone to do the statistics.)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s