Incident 150: Swedish Contraceptive App, Natural Cycles, Allegedly Failed to Correctly Map Menstrual Cycle

Description: Some women using the contraceptive app, Natural Cycles, reported unwanted pregnancies, revealing its algorithm's difficulties in mapping menstrual cycles.
Alleged: Natural Cycles developed and deployed an AI system, which harmed Natural Cycles users and Women.

Suggested citation format

Lam, Khoa. (2018-07-21) Incident Number 150. in McGregor, S. (ed.) Artificial Intelligence Incident Database. Responsible AI Collaborative.

Incident Stats

Incident ID
Report Count
Incident Date
Sean McGregor, Khoa Lam


New ReportNew ReportNew ResponseNew ResponseDiscoverDiscover

Incident Reports

Natural Cycles was hailed as a stress-free, hormone-free contraceptive. Then women began reporting unwanted pregnancies.

Last summer I had an abortion. Statistically unremarkable, yes, but mine wasn’t because of a split condom or a missed pill. I was four months into a tense relationship with a much-hyped Swedish “digital contraceptive”, a smartphone app called Natural Cycles. I had spent my 20s on the pill, but hated not knowing whether my emotional state was down to artificial hormones or not. My boyfriend and I had been together for eight months, and I was desperately seeking something new, something that wouldn’t make me feel so anxious.

That’s when the adverts started following me around on social media: glowing women reclining in Scandi bedrooms, all pale grey sheets and dappled light, brandishing basal thermometers and telling me how great it felt to “get to know yourself better”. Natural Cycles’ ads promised the “world’s first contraceptive app”, something “natural, hormone free & non-invasive”. I could start using it without a two-week wait for a doctor’s appointment and so, in a fug of hormones and frustration, I bought a subscription. I was sold on shiny promises, a sleek user interface and the fact that a former Cern physicist, Elina Berglund, was at the company’s helm. But four months in, it failed. Berglund helped discover the Higgs boson; but it turns out her algorithm couldn’t map my menstrual cycle.

Femtech, or female health technology, is going through a boom phase, with an estimated $1bn of investment raised worldwide in the last three years. Apps such as Clue, Dot, Glow and Spot On are all popular period trackers, but Natural Cycles is the only one certified as contraception. In 2017, it was approved for use across the EU, getting the green light from the German inspection and certification organisation, Tüv Süd.

How does it work? It comprises an app, an annual subscription of about £60, and a thermometer accurate to two decimal points (free in the post). You input your temperature as soon as you wake up, and the app makes predictions about your fertility each day: green for “go have unprotected sex”, red for “not unless you want a baby” (you can also use the app to plan a pregnancy). No hormones, no implant and, supposedly, no stress. It has its own language: users are known as “Cyclers”, and useful information is available via a “Cyclerpedia”. It seems as easy as ordering a takeaway or a taxi from your phone; of course there’s an app for fertility, too.

Natural Cycles has now registered more than 700,000 users from more than 200 countries, 125,000 of them in the UK. But its certification as contraception is under review in Sweden, where the company and its married co-founders are based. In January, a major Swedish hospital reported that 37 of the 668 women who had sought an abortion there between September and December 2017 were using Natural Cycles as their sole birth control, and the Medical Products Agency of Sweden began to investigate [*See Update below]. Natural Cycles has responded that the number of pregnancies is proportional to the registered number of Swedish users and “in line with our expectations”; but as someone who didn’t report my own pregnancy last year, keeping it secret even from my parents, I wonder how many more there have been.

It wasn’t the stigma that kept me quiet, or the sadness, though that trailed me all summer like the sinister melody of an ice-cream van. It wasn’t the fact that being 28, in a stable-seeming relationship and game for motherhood in a couple of years, I lacked an explanation other than precarious finances and a relationship just shy of its first anniversary (those are excellent reasons). No, my silence was because I felt colossally naive. I’d used the app in the way I do most of the technology in my life: not quite knowing how it works, but taking for granted that it does. Speaking to others who bought the app as contraception (about 75% of Natural Cycles’ total user base, according to its CEO), it seems that many feel the same.

I spoke to Amy, 29, who was fed up with hormones when she started using the app as her sole birth control. Three months later, she was pregnant, a “massive shock”. Though she admits she may have made a mistake, she can’t pinpoint the error. “You’re told all you need to know is yourself. I believed in it the same way I did the pill and thought I did everything right.” Having already booked her wedding, she went ahead with the pregnancy, giving birth weeks before she walked down the aisle. “It’s supposed to make you feel like you have more control, but in fact it did the opposite: when I fell pregnant it felt like a decision was taken out of our hands. It wasn’t how we’d have planned it, and I don’t recommend weddings two weeks postpartum, but I’m lucky it was something we wanted in the long run.”

Marie, 30, first heard about the app when she saw an Instagram post about it (search for Natural Cycles and you will find hundreds of posts by influencers telling you how it changed their lives). “I didn’t spot the hashtag at the very end of the caption which said that it was a sponsored post,” she says. She had been taking Yasmin, a commonly prescribed contraceptive pill, for six years when she made the switch, hoping that the app would be a reliable and easy alternative. A year into a relationship, and eight months into using Natural Cycles, Marie realised she was pregnant. She had an abortion that proved traumatic, contributing to the breakdown of the relationship and leading her into what she describes as “a pit of despair”.

She didn’t want to tell anyone about it. She’d had an abortion once before, when a morning-after pill didn’t work, but this time she felt ashamed: “I felt like I’d acted alone in the decision to use the app and had been overly trusting. But I was also angry that I’d been treated like a consumer, not a patient.”

Like Marie, I didn’t go to my GP before I switched to the app, probably because I subconsciously knew he’d advise against it. In many ways he knows me better than any algorithm can. He put me on the pill at 18 because I had an irregular cycle. I later learned I had polycystic ovary syndrome, which I now know makes me a terrible candidate for Natural Cycles, because my ovulation is unpredictable and erratic.

A year earlier, before I’d heard of the app, I had been to see a gynaecologist to discuss birth control, thinking I wanted a non-hormonal coil fitted. It was the first time a medical professional had helped me to truly understand the extent of my options. She drew me a set of coordinates and plotted each option available (no app got a mention) to show me the benefits and drawbacks. Spotting v cramps, depression v maintenance, long- v short-term.

I’d read grim things about the hormonal vaginal ring – a widely shared article about a young, fit woman who died after a blood clot – but agreed, based on what she felt would suit me best, to try it. We laughed at how it’s impossible to research any birth control online without encountering horror stories. I told myself I would trust a professional and cease my Googling as it only induced anxiety; but after a few paranoid weeks wearing compression socks to avoid blood clots, I was done.

None of the posts on my social-media feed suggested that being a “Cycler” would be such a frustrating, often daunting commitment. One paid-for post I saw featured a still life of a puppy, a pair of on-trend headphones, a self-help book and a thermometer, with a 250-word caption starting with “5 things I need in the morning. Cuddles from Bee [the dog], tea, music, positive quotes and the first thing I do when I wake up – my Natural Cycles thermometer.” But I found that taking your temperature regularly is not so easy. The number of times I leapt out of bed bleary-eyed and needing to pee, then realised I hadn’t first taken my temperature, meant I started waking up in the middle of the night to pre-emptively urinate, panicked about missing my measuring window in the morning. On the pill, it didn’t matter if I’d just woken up, was lying down or standing up when I took it. With Natural Cycles, the slightest motion seemed to count. It was comedic until it became tragic; I got pregnant when the predictions of fertile and infertile changed back and forth in one day, turning from green to red, after I had unprotected sex.

I now know that the ideal Cycler is a narrow, rather old-fashioned category of person. She’s in a stable relationship with a stable lifestyle. (Shift-workers, world-travellers, the sickly, the stressed, insomniacs and sluts be advised.) She’s about 29, and rarely experiences fevers or hangovers. She is savvy about fertility and committed to the effort required to track hers. I could add that her phone is never lost or broken and she’s never late to work. She wakes up at the same time every day, with a charged phone and a thermometer within reach.

“From the information provided by Natural Cycles, I expected that my body temperature would follow a clear pattern and that I would be able to pinpoint five days in every four-week cycle that I was fertile,” says Lucy, 32. She switched from the pill after becoming concerned about an increased risk of breast cancer, after one of her friends was diagnosed. “I did feel like I was getting to understand my body better, but soon realised that I can’t pinpoint when I wake up each day. Some mornings I stir at 5am, roll over and try to sleep for another hour or two, sometimes I toss and turn from 2am to 6am and then fall asleep, and so on.” Her readings were erratic. “I couldn’t see a pattern and this undermined my confidence. After using Natural Cycles for three full cycles, I found I was still having eight to 10 red [ie possibly fertile] days per cycle.” After four months, she decided it was no better than using a calendar and went back on the pill.

No form of contraception is 100% effective; most are assessed according to two metrics: typical use and perfect use. “Typical” reflects a margin for human error; “perfect” is when it’s used absolutely correctly. With perfect use, Natural Cycles scores as 99% effective, with just 1% of women becoming pregnant. With regular use, according to clinical studies carried out by the company (self-selecting, rather than randomised control trials), that drops to 93%. This is often cited by the company as favourable compared with the pill (91% effective with “regular use”). But, unlike the pill, you’re not covered for every day of the month. You have to abstain or use other contraception on fertile days. And in the first few months, as the app “gets to know you”, these are pretty near continuous.

When I talk to Raoul Scherwitzl, the CEO and co-founder of Natural Cycles, he is charming and sincere and calls at precisely the appointed hour, not a second early or late. “My wife and I represent a typical user-couple,” he says. “Elina was on hormonal birth control for 10 years and we knew we wanted children, but in a couple of years. We both had PhDs in physics and [Elina was] working at Cern, dealing with messy, fluctuating data, trying to look for the Higgs boson, which is basically looking for a signal amid noise. We started applying the same statistical methods to pinpoint my wife’s ovulation amid her varying temperatures. We read up on the literature and developed an algorithm which our colleagues started using, too. We were running it on the Cern servers and then using Google spreadsheets. We saw it as an unmet need. There was a lack of choice and we wanted to innovate in an important field.”

I tell Scherwitzl that, though the need is real, after purchasing the app, caveat after caveat revealed itself. I didn’t know it would take months to become reliable. “The algorithm is cautious by design,” he explains. “It gives red days unless it’s sure.” I tell him how I got pregnant, when the predictions changed after I’d had sex. Scherwitzl empathises (“I am sorry to hear that”) but says that as well as “downsides”, there is “a huge upside with all the happy users” and that “the most important thing is to use protection on red days: it relies on that”.Has the company adapted its communication strategy to reflect the experience of users who have become pregnant? “At the core, our messaging has always been grounded on facts but we do evolve what we say. We used to state the 93% figure but without the right context, so there were certain expectations on the product.” The 93% figure comes from existing users responding to the company’s calls for participants, research that has been criticised by a reproductive health expert as “inappropriate and misleading”, and more like “marketing research” than a medical study. But Scherwitzl insists the data is robust, and preferable to a medicalised control test. “There are pros and cons to that type of study, and in our view this reflects the world better.”

What about the targeted advertising? Isn’t it strange to get social-media influencers (one prominent Swedish blogger is now an investor) to promote a medical product? He doesn’t think so. “Social media allows us to control the narrative because there’s lots of misinformation out there, not just with us but with every type of birth control. We can also target the right age group.”

The investigation by Sweden’s medical watchdog is now six months in, and has started its second phase, reviewing marketing material, past clinical studies and fresh user data. Reports of unwanted pregnancies have not, however, had any negative effect on the business.

In a 2016 interview I wish I’d read, Scherwitzl’s wife and business partner Elina Berglund described her ideal user as a woman who is planning to have children at some point, and who would like a break from hormonal contraception before trying.

Kristina Gemzell Danielsson, a professor of obstetrics and gynaecology at the Karolinska Institute in Stockhom was part of a team of researchers hired by Berglund pointed out that it’s not a good option for women who want to entirely avoid a pregnancy, she said. But somehow this message has got lost in Natural Cycles’ marketing; this is very much not what the word “contraception” means to me.

Indeed, on the section of the Natural Cycles website aimed at medical professionals there is a “decision tree” for doctors considering prescribing the app as birth control. Is the patient over 18? Is she satisfied with her current birth control? If the answers to those questions are yes and no, then the third is: would she be “devastated” to get pregnant within the next year? If the answer is yes, the doctor is told not to prescribe the app. Perhaps those questions should be compulsory when you click through on a Natural Cycle link.

Instead, the app assumes the intimate voice of a trusted doctor, mixed with the sort of gamified messaging you find on other apps. You might get an update saying, “Nice curves! You have a nice and smooth temperature curve with small day-to-day variations. Keep up the good work!” The company’s social media is peppered with hashtags such as #yourcyclematters and #wakeupmeasuregetup. The perkiness is grating – even the thermometer bears the slogan “Good morning!” – and can be pressurising, too. One woman I spoke to who bought Natural Cycles to try to plan a pregnancy told the company she wanted to leave after six months, as the daily tracking was too stressful. She emailed to ask for a break, “for emotional wellbeing”. A customer service manager responded to say she could cancel, and reactivate when she wanted, but that “I took a quick look at your data, and in terms of ovulation everything looks good!”, adding, “You do not need to worry about losing any data – we never delete anything!” Those exclamation marks don’t make such reassurances any less creepy.

Perhaps this false sense of intimacy is why it felt more like a betrayal to find myself pregnant than if the pill were at fault. After the abortion, the honeymoon period of my relationship ended abruptly. It felt like we’d begun our romantic race with a false start. I stopped using the thermometer and went back on the pill, but it took me a while longer to part ways with the app. I deleted it from my phone, only to realise the direct debit was rolling and non-refundable. I have just been sent another £60 bill, for a contraceptive app I no longer use, that got me pregnant. But it’s not just the money that bothers me – it’s the reminder that I put so much faith in a technology that in the end relied on something as unreliable as my body. What’s the hashtag for that?

‘I felt colossally naive’: the backlash against the birth control app