Abortion Rates in Aotearoa

This is my first experiment doing data analysis in R.

This year the Abortion Legislation Act 2020 has been approved by the New Zealand Parliament, finally decriminalising abortion and streamlining the process (at least for women who are no more than 20 weeks pregnant).

I decided to analyse abortion statistics in Aotearoa focusing on differences throughout the years and among age groups. To do so I used the data provided by Stats NZ.

Analysis

About the data

There are two main groups of data:

  1. general abortion rates: these values represents the number of abortions both induced and spontaneous. For general data the rates are expressed by number of abortions per 1000 women aged between 15 and 44. For data grouped by age groups, the rates are expressed by number of abortions per 1000 women in the corresponding age group.
  2. Induced abortion numbers: these values represents the raw number of abortions for each year. Even in this case the data is provided both summarised and grouped by age.

Rates throughout the years

The following graph displays the trend of abortions from 1980 to 2019.

As it can be seen from this graph, the abortion rate increased steadily between 1980 and 2003, reaching a peak that year. More in the details, while at the beginning of the considered period the rate was as low as 8.5 abortions per 1000 women (aged 15-44), in 2003 it was recorded a rate of 20.8 abortions per 1000 women (aged 15-44). However, in the last 8 years, this rate declined markedly, reaching approximately 13 abortions per 1000 women (aged 15-44).

I then focused on the differences between age groups. Following a gif I created with gganimate and ggplot showing the figures for 8 age groups from the year 2000 to 2019.

While women younger than 15 years and older than 29 years maintained stabel rates, important fluctuations can be observed for the three groups in between: 15-19, 20-24, 25-29. New Zealand women between 20 and 24 year old have been more likely to abort than any other group throughout the two decades. However, starting from 2008 their abortion rate plummeted significantly. In 2019 the rate of abortions for this group was lower than 20 abortions per 1000 women (in the same age group), less than a half than the maximum rate reached in 2003 (over 40 abortions per 1000 women (in the same age group). Furthermore, in 2019 there was almost no difference between the group 20-24 and 25-29. Clearly evident is also the drop in abortions in teenagers (15-19). Last year it was lower than women aged between 30 and 39. I suspect this is driven by two factors:

  1. More availability of contraceptives
  2. Increase age at which women start to have children which could increase the possibility of having a spontaneous abortion (these figure do not distinguish between induced and spontaneous abortions).

A focused gif on these three groups:

Induced abortions

If we focus only on induced abortions we can see that the trend is basically the same as when we consider all kind of abortions together.

There was a steadily rise until 2003, and then a significant drop in the following years. The numbers divided by age groups are shown in the following plot:

I do not have population data to adjust these plots in order to compare induced abortions vs spontaneous, however, something interesting happened between 2014 and 2019. The number of abortions in the age group 25-29 raised between 2017 and 2018, outreaching the number for the age group 20-24. Given that the rates of general abortions for the former group remained the higher over the years, this means that, probably, in the last few years there has been:

  1. More spontaneous abortions and less induced abortions for women between the age of 20-24
  2. More induced abortions and less spontaneous abortions for women aged 25-29.

To be sure of this, however, I would need to adjust for the population numbers.

Final Disclaimer

This was just a project I started to test and challenge myself in order to learn R and its packages (in particular tidyverse, ggplot2 and gganimate) better. The analysis and the conclusions are not verified.