Using teenage pregnancy data in Kenya to write a powerful story

Introduction

In a previous article, we learned how to extract data from a PDF document using Tabula. Then we proceeded to use extracted data to look closely at how land ownership varies among men and women in Kenya when we consider residence, education, wealth and age. 

Now, we are going to use extracted data to find out more about teenage pregnancy in Kenya. 

Teenage pregnancies in Kenya

Stories about teenage pregnancy are very common in the Kenyan press. In June 2020 while Kenya was in the throes of the Covid-19 pandemic, a news report claimed that nearly 4000 girls had become pregnant in Machakos County over three months. The coverage sparked loud debate and outrage across the country. However, few articles sought to contextualise the figures. Without knowing the size of the county and the national average monthly rate, it’s not immediately apparent the extent of the issue. If 5000 girls tend to fall pregnant monthly, these statistics paint a different picture than if only 100 girls tend to fall pregnant in Machakos in any given month. Avoiding sensationalised reporting means having data at the centre of the story and remembering that context is critical. 

As it turns out, the county had interpreted the data incorrectly. In a press statement released later that month, the ministry explained that the data reflected the number of visits to antenatal clinics by teenage girls who were pregnant. Because each girl is expected to have several visits to antenatal clinics during one pregnancy, the numbers were severely over-inflated and did not show the actual number of girls who were pregnant. This example illustrates the importance of understanding the context in which data is collected, analysed and shared. 

In this article, we are going to deep dive into the latest data on adolescent pregnancy in Kenya and explore potential story ideas. Given that these girls risk health complications and often give birth to underweight children who are born before the full pregnancy period or that adolescent mothers are also more likely to drop out of school affecting their economic advancement, this is a key story. We will also share some tips when reporting on this kind of issue to remain sensitive, humanising the story and at all times keeping your journalism ethics close at hand. 

About the data  

Teenage pregnancy is defined in the Kenya Demographic and Health Survey Key Indicators Report 2022 as the percentage of women aged 15–19 who have ever been pregnant. Respondents were asked whether or not they were or had been pregnant, and whether they had ever had a live birth or pregnancy loss. Their responses are laid out in the Table below. You can access the scraped data here. The respondents were categorised according to their ages, residence, educational attainment and wealth.

Figure 1: The Table above is Table 6 from the 2022 KDHS Key Indicators Report, showing findings on teenage pregnancy. 

‍

Choosing the starting point of your story  

This one dataset (above) already has the potential to support the telling of many different stories. In order to avoid being overwhelmed by a dataset (however complex) remember that when you are embarking on a new story, there are two potential entry points: 

1. Starting with a hypothesis. Do you have an idea of what you expect to see in the data? Do you have an existing question that you hope to answer with the data? You can then use the data to see whether this holds true. 
2. Starting with the dataset. Have you come across a new dataset and are curious about whether there is a story waiting to be told? In this case, exploration and analysis may open up new opportunities for stories that may not be visible at first look. 

Scan the data available and identify from your starting point. Write it down. You can keep coming back to it if you are finding yourself getting confused or uncertain in your analysis. 

For this, we hypothesise that teenage girls with higher rates of education (secondary education and above) are less likely to fall pregnant than those with lower education. We will use this as the base of our story. 

‍

Exploring the data 

Getting the data into a format that shows patterns or outliers is the next step in your data pipeline journey. In this case, we used Flourish to plot the graphs. You can view our training on learning DataWrapper and Flourish here. You can use any tools that you are comfortable or familiar with to start to explore the data. 

To answer our question about the impact of education on teenage pregnancy, we plotted the data in a column graph. Some points to consider: 

• The data on teenage pregnancy involves only one gender (i.e. female), unlike land ownership where both men and women were surveyed. 
• All the data is for a single point in time so line charts are not appropriate because they work best when we are tracking a trend over time. 
• To represent each of the four answers asked of respondents clearly, we decided to use a grouped column chart to illustrate the responses.

What becomes immediately clear from chart is that women who have ever been pregnant occur most frequently among those with no education (37.9%) and least often among those with more than secondary education (4.8%). The graph shows a similar finding for women who have ever had a live birth (30.8% and 4.0% respectively). Women with more than a secondary education are also least likely to have had a pregnancy loss (0%) and to be currently pregnant (0.8%). Those numbers are also highest for women with no education (3.9% and 9.6% respectively). 

Figure 4: Percentage of girls age 15 – 19 who had ever been pregnant, had a live birth, had a pregnancy loss, or were currently pregnant, by education.

At first glance, an increase in education appears to reduce the chances of teenage pregnancy. Girls with little to no education appear to be especially vulnerable to teenage pregnancy. 

This type of basic analysis can be done for any of the attributes. We’ve shown some of the other options for exploration below. 

Figure 6: Comparison of graphs showing percentage of pregnancy among girls aged 15-19 by wealth, education age and residence. Notice the comparatively higher share of pregnant girls who have had no education.

‍

You have your data, now what? 

Time to start understanding the context and building your story. 

After a basic analysis of data, you’ve noticed that relatively few teenage girls go through pregnancy loss. However, its relatively infrequent occurrence does not mean it is unimportant. Given that it's a painful event for those who go through it, it’s necessary to find out more about it.

Once you’ve sense-checked the data, there are a few different steps to take to start shaping your narrative. ‍

1. Compare to other countries or to global statistics 

How widespread is pregnancy loss in Kenya? In East Africa? In Africa? Worldwide? This is where comparative statistics by credible organisations come in. Leveraging organisations such the World Health Organisation or national statistics agencies, you can start to get a sense if similar patterns are seen in other countries. You can use the analysis methods above to do similar types of explorations. Importantly, remember to use footnotes or endnotes to track figures in global databases to their original sources, which are often surveys by national statistical authorities. It is important to follow up and report any discrepancy. If the numbers underestimate the situation, badly needed attention might be focused elsewhere. If they exaggerate the situation, resources could be wrongly directed from urgent health emergencies.

Global figures can also provide a basis to hold government officials such as ministers accountable, and demand that the government prioritise reducing pregnancy loss, particularly if it is clear that authorities are either ill-equipped or unwilling to handle the problem.

2. Identifying and reaching out to experts 

Your previous work building up contacts will pay off handsomely now but if you don’t have any, it’s never too late to start. Work with hospitals, health research institutions and organisations who support people who are living with pregnancy loss. Does your understanding of the data tally with their experience and knowledge of the sector? In their expert opinion, why have these types of trends emerged? What are key issues to be considered in the write-up of the story? ‍

3. Is there a solutions angle to the story? 

Sometimes stories in the news about unhappy events, such as pregnancy loss, may be avoided by readers. How can we construct stories that showcase positive outcomes? Given the severity of the issue, are there programmes or initiatives that exist locally that seek to overcome the deleterious effects of teenage pregnancy and/or pregnancy loss that are worth showcasing to the public? Are there specific clinics that are doing something different that are resulting in positive change in their districts on pregnancy outcomes? 

4. Humanising the story ‍

Remembering that each value in this dataset represents a person, not a random value. Showing the personal impacts of teenage pregnancy loss, outside of the sensationalised victim-blaming reporting that we’re used to, could go a long way to shifting the narrative around pregnancy loss. 

Our story can look into how families who live with pregnancy loss cope, and perhaps thrive. It can provide insights into the physical, emotional and financial effects of pregnancy loss, in order to better understand the condition, reduce stigma, combat unfounded conspiracy theories and myths, reinforce acceptance into the community, and protect the human rights of people living with pregnancy loss. 

We can also interview experts who work with people who have experienced pregnancy loss for advice, and provide a list of organisations that provide this help to the public. Because professional help is often expensive, it’s a good idea to discuss the cost of such care and any possible assistance, such as insurance or charity, that might reduce the cost.

In our story we can interview health experts for advice on the steps families can take to minimise the chances of pregnancy loss. This could include what to do and what to avoid. 

5. Reporting with sensitivity and maintaining your ethical credibility 

Given this as a particularly sensitive subject, keep ethical considerations in mind at all times. Ensure you state clearly you are a journalist who wants to write a story on pregnancy loss for publication and obtain the informed consent of the family. If they do not approve, accept this decision. Remember that if your interview subject is a minor, you need to ensure their identity and that of their relatives is withheld. Consider whether photographs or videos are appropriate in the context of their story. Use your discretion wisely. 

6. Always follow up

Stories with sad beginnings can have happy endings. A woman who suffered a pregnancy loss could later have a successful birth. A community that did not have antenatal services could start receiving them. Statistics could show an improvement in live births. 

Or nothing could change, leaving us to write even more stories.

Writing your own story! 

In this piece, we’ve learnt how a powerful story can spring from statistics. By asking questions from our data, we came up with building blocks for a story around pregnancy loss. It’s your turn now. Go on, and write stories that point out important problems, suggest compelling solutions and bring real change.

‍