The Home Office is only supposed to lock up pregnant women in immigration detention centres in very exceptional circumstances. But if you try to find out how many of them satisfy this criteria, things get difficult very quickly.
The information started being centrally collected last year, but it wasn't made public. So in February, campaign group Women for Refugee Women sent off a Freedom of Information request. The Home Office waited until the final day of the 20-day time limit and then refused, citing a commercial interests defence.
This was nonsense, but the group kept pressing. By April, the commercial interest argument had disappeared and the Home Office was now refusing on the basis that it was 'management data' - in other words, of an insufficient quality for publication. This is also nonsense. Women for Refugee Women complained to the Information Commissioner's Office. They ruled against the Home Office.
Now, months later, the information has finally been released, although the Home Office decided to send it to an email address no-one had ever heard of. A few more queries and it was sent to the right email address. But now it had the wrong dates on it. The group had asked for stats spanning from when central records began to the date on which they were accessed to answer the inquiry. Instead they were only given stats from August 1st 2015 to December 31st 2015, because that's "in line with the timescale for the publication of the most recent Home Office published statistics". It's ironic, given their refusal to release the information was based on it not being official statistics, that they should now rely on the timetable of official statistics to limit the quantity of information they send.
At the bottom of the Home Office document, in small print, it reveals that the data was extracted on May 17th 2016. This was before the Information Commissioner's Office ruled against them in June. They'd complied with the request ages ago and then just sat on it, refusing to make it public until the information commissioner forced them to.
Why the reluctance? Possibly because when you look at the statistics it's impossible to believe that the Home Office only locks up pregnant women in detention centres in exceptional circumstances. The document shows that in that five-month period, 52 pregnant women were locked up.
That suggests the number is rising slightly. The 2014 Inspector of Prison's report into Yarl's Wood (the only female detention centre) found 99 pregnant women had been held there over the year. Or it could be that the Home Office data includes pregnant women held elsewhere in the very few family detention facilities.
The numbers fluctuate anyway. The Independent Monitoring Board for Yarl’s Wood found 89 pregnant women had been held there in 2015. When MP Caroline Spelman asked Serco about it, she was told 69. It's hard to get a firm idea of the numbers.
But dig down into the data of what happened to these people and it's increasingly obvious the Home Office is breaking its own rules on detaining pregnant women. There should no circumstance in which a pregnant woman is kept for months and released back into the community. They're only supposed to be detained when they are at risk of running away and are about to be deported. If exceptional circumstances mean anything, that's what it is: an urgent need to detain for a very short time. Women being held for a long time or released back into the community - or both - mean the system isn't working.
And that, of course, is exactly what's happening. During that five-month period just 13 were removed from the UK. Thirty-five were released. Four were still in detention as of December 31st 2015.
Most of the women had been detained for less than a month, but 12 were held for longer - four of them for up to two months, six of them for up to three months and two of them for even longer. Of those four women still sat in detention centres on December 31st, two had been held for three months or more.
Maybe the Home Office didn't know they were pregnant. The starting date applies from when they entered the detention centre, not from when they were confirmed as pregnant. Maybe these women hadn’t told staff when they arrived at the centre. Maybe the women themselves didn't know, and only discovered it once they were locked up.
Or maybe, as countless independent reviews and reports have found, the Home Office is failing to live up to its own rules on detaining pregnant women. Maybe, as the inspector of prisons found, "the exceptional circumstances justifying detention of pregnant women were unclear". Maybe, as the independent monitoring board found, there is an "apparent failure" of the Home Office "to follow its own policy".
And maybe that's why they're so reluctant to release information which holds them to account.
Ian Dunt is the editor of Politics.co.uk
The opinions in politics.co.uk's Comment and Analysis section are those of the author and are no reflection of the views of the website or its owners