Our ability to harness and interpret data has revolutionized animal health, safeguarding human well-being, and advanced food security, sustainability, and economies. Animal health companies can unlock transformative improvements by leveraging data about medical treatments, vaccinations, and yield.

Based on my experience working with data solutions in this sector, I have compiled a list of best practices around effective data collection, analysis, interpretation, and communication. In this article, I share these practices, which I hope can be used as guidelines for data utilization in animal pharma as well as other industries.

Humans and animals are deeply connected

We keep pets as companions, breed animals for sustenance, and we try to protect wild animals.

Our fates are also closely linked by disease. Animal health is vital for humanity and in his remarkably prescient 2011 book The Viral Storm, biologist Nathan Wolfe explains why:

“Pandemics almost always begin with the transmission of an animal microbe to a human.”

Wolfe explains that many diseases and pandemics originate from animals and explores how technology can fight these diseases.

We believe that a data-first approach to improving animal health can also help us improve human health, food security, the economy, and reduce environmental impact.

The value of animal health data*

Outbreaks of animal diseases such as BSE, CJD, bird flu, SARS, swine flu and Ebola are hazardous to human health and can cause huge loss of animal life. For example, just one bird flu variant – H5N1 led to the culling of over 300 million geese in China.

Since animals are so crucial to our food supply, we must rely on data to understand how animal diseases impact food security, economy, and environment.

Animal health and food security

Data about the effect of vaccines on livestock productivity allow us to design interventions that measurably improve food security, for example:

• If 60% of the world’s beef cattle were vaccinated, it could increase beef production by 52.6%. This would be enough to feed 3.1 billion people – almost 40% of the world’s population.
• Vaccinating 40% of cattle in Nigeria in a given year is associated with lifting 2.4 million people out of severe food insecurity. Every 1% increase in beef cattle vaccination in Brazil correlates to a 0.7% increase in production.
• Poultry disease was associated with a 2% increase in global hunger (13.6 million people) in 2018 and 5% (34.39 million people) in 2019.

Financial and Environmental impact

Data reveals the cost of animal disease. In 2018, disease appears to have reduced global egg production by 3 million tonnes, costing producers $5.6 billion – this is the equivalent of around four times the UK’s egg market. This data can help us develop practices to reduce these costs.

We can also make connections between vaccination rates and land use to make appropriate, sustainable production choices:

• A global cattle vaccination rate of 40% could reduce cattle land use by 5.2%.
• When 20% of global poultry are affected by disease annually, approximately 8.6% more land is needed to maintain production levels.

The environmental and economic case is clear that if we improve animal health, we will use fewer financial and environmental resources (such as land and feed), and increase productivity.

Monitoring animal health and disease

Data, the Internet of Things, and smart devices allow us to remotely diagnose, track, and protect animal health.

• Bird flu has caused staggering costs – over 500 million birds have been lost to bird flu since 2005. The USDA identifies where the disease is present, tracks bird movement and helps farmers protect their flocks.
• Cowlar uses smart collars to track dairy cow vital signs including movement, temperature, and heart rate. If the data indicates signs of disease, the farmer is notified and can act appropriately. Other companies such as Digitanimal use similar approaches to monitor the health of other animals.

Controlling disease at scale

Data can also protect animal health on a large scale by improving the supply chain. At The World Organisation for Animal Health, Dr David Swayne talks about strategies to deal with bird flu:

“Vaccinating is not the end, it is just the beginning. Vaccination application needs to be managed along the supply chain including a surveillance programme which is able to detect active infection in vaccinated animals.”

A transparent and traceable supply chain allows animal health companies to monitor vaccine efficacy and invest in improving inoculations to keep up with changing seasonal strains.

Once animals are vaccinated, they can be protected through the supply chain. By identifying if birds carry antibodies from vaccines (not disease), healthy birds can be tracked, and disease surveillance data, biosecurity, and hygiene measures can keep them healthy.

My cheat sheet for great data analysis practice

The statistics above clearly show that we can significantly improve our well-being by understanding data and taking appropriate action.

Animal health companies do a fantastic job of analyzing data. We are lucky that so many automation, machine learning and AI tools are emerging that means we can process more data faster and with more precision.

I like to share my cheat sheet of principles and good practices. Just checking in with the fundamentals helps me harness and steer the data processing power we have at our disposal. I hope you find them helpful.

1. Define the problem. The first step of any data analysis project is to define the problem area you want to explore.
2. Collect data. You might find relevant data internally, such as sales records or customer surveys, external data sources, such as social media or government databases, or aggregated data such as industry bodies.
3. Clean and prepare data. Check for errors, inconsistencies, and omissions. Reformat the data to make it suitable for analysis.
4. Analyze. Use statistical methods to identify patterns and trends. Data visualization tools can help you explore the data and gain insights.
5. Interpret the results. To interpret the results, you need to step back from processing the data and think about: what do the results of your analysis mean? What insights can you draw from the data? Interpreting results can be the most challenging part of data analysis.
6. Communicate the results. The final step is to communicate the results to stakeholders using visualizations, presentations, and clear reports. Stakeholders need to understand the results of your analysis so they can make informed decisions.

And data is never as clean and straightforward as I would like it to be, so I try to be explicit about how I process my data to generate clear, meaningful results. Perhaps some of the considerations I keep in mind here may help you too:

1. Data limitations. It is important to be explicit about the quality of your data and factor any potential data bias, inaccuracies, or errors in when interpreting.
2. Missing data. You are likely to encounter missing data, and it is important to handle it explicitly.
3. Regression modelling.  It is important to use multiple independent and control variables to identify the relationships between the independent and dependent variables you are analyzing.
4. Multicollinearity. Each regression model should be tested for multicollinearity to accurately understand the dependent variables.
5. Lagging variables. Be mindful of the delay between inputs, such as disease treatments, and outcomes, such as animal health, disease outbreaks, productivity, and financial and economic returns. Delays will vary depending on species and product. For instance, cattle meat production cycles are longer than egg production cycles.
6. Clustering.  Clustering using unsupervised learning algorithms can help explore large datasets if you are not sure what patterns exist. Clustering may render patterns visible that would be difficult to see using traditional statistical methods.
7. Scenario analysis.  Once you have key variables, scenarios allow you to understand how various circumstances may impact your dataset. They give you the opportunity to think of the consequences of each scenario, understand risks and opportunities, make informed decisions, and adapt when conditions change.

Data analysis can be a time-consuming process. I try not to expect to get all the answers overnight and instead remain patient.

Creativity is also important – there is no one right way to analyze data, so it pays to be open to new ideas and approaches, and interrogate data with no expectations. I find automations helpful to free up my time for creative thinking and interpreting.

We have seen how data insights can transform populations, and we are only scratching the surface of its potential. As we continue to refine data collection, analysis, interpretation, and communication practices, we unlock the doors to further innovation and progress. This holds immense importance, particularly in sectors that are vital for humanity. Let us embrace the almost endless possibilities that data presents and collaborate to build a healthier future for both animals and humans. Together, we can make a remarkable difference.

About C&F

C&F is a technology firm with deep experience in animal health and data analysis. Despite the abundance of data, companies don’t always leverage data as well as they could. This may be down to siloed teamwork, conflicting reports based on different sources, and ineffective communication between IT and business areas.

As digital transformation specialists, we support companies in leveraging data to become data driven. We work with IT teams and leadership teams during their digitalization journey. See an overview of our capabilities, or talk to us to find out more.

* all data sourced from Animal health and Sustainability: A Global Data Analysis