It is a capital mistake to theorise before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts.” - Sherlock Holmes.

It can be easy to fall into the trap of confirmation bias, and alter data finding to fit previously agreed theories. In the health sector, this approach can and does result in death, which is why collecting health data and conducting unbiased health data analysis is vital.

Health data analysis is the extraction of meaning from the health data, which is then used to make better-informed decisions concerning the management of an individual’s or population’s health. At Fjelltopp, we’re passionate about improving quality of life for all - but especially those in low and middle income countries - through the implementation and promotion of software solutions and research.

Better health data, better health data analysis

At the time of writing, the UK is conducting an ongoing examination of the decisions made by the government and beyond in response to the COVID-19 pandemic. This COVID-19 inquiry has put the spotlight on public health decision making, questioning whether the decisions taken by individuals were reasonable given the evidence available at the time.

On a national level, decisions made concerning health systems have a far-reaching impact and directly affect matters of life and death for a vast number of people. The global pandemic made that evident to people outside the public health sector. It’s because of these life or death decisions that choices concerning public health must be grounded in clear, objective evidence - including health data and health data analysis.

At Fjelltopp, we are passionate about collecting, managing and understanding health-related data, especially public health data. Better data leads to better analysis, which leads to better decisions and better health for people worldwide. Good data analysis starts with great data collection.

The types of data we collect for health data analysis

The quality of the health data analysis depends significantly on the quality of the underlying health data. Here are the different types of public health data we work with at Fjelltopp:

• Survey data - This data comes from large-scale surveys that take place periodically and offer highly reliable snapshots of a population. Due to the size and cost of these surveys, other data sources are required, alongside statistical estimation methods, to answer many research questions.
• Medical records data - Medical records should store a detailed account of a patient’s interactions with a health system. The primary aim of medical records is to facilitate smooth clinical patient pathways through the health system, but case-based data can also be effectively extracted from them.
• Case-based data - Health data collected on a case-by-case basis (e.g. per consultation), including detailed information about an individual’s demographics, symptoms, diagnosis and prescribed treatments. Unlike medical records, the data is often anonymised and sent to a central institution for analysis. In many ways, this is the gold standard of public health surveillance because it captures the highest level of data detail but is also the most challenging to implement effectively.
• Aggregate data - Where the resources don’t exist to collect data on a case-by-case basis, the alternative approach is to collect pre-aggregated data from health facilities. Health facilities count and submit numbers of interest, e.g. number of consultations, diagnoses, and prescriptions that week. Whilst this is much cheaper and simpler to implement, you can’t disaggregate to understand more detailed trends in the data.
• Programmatic data - This is typically aggregated data collected or organised by “health programme” where patients with a specific disease are treated and monitored with data collected routinely. The data often contributes to international programmes responding to specific diseases such as malaria or HIV. Read our UNAIDS case study to put this into context.
• Laboratory data - Diagnoses may often be confirmed by a test that requires laboratory equipment. This testing process can produce important public health data, but laboratory test results often take time to process.
• Civil registration data - Accurate registration of births, deaths and vital statistics is essential in understanding a country’s population. Many low and middle-income countries struggle to collect and manage this data.
• Research data - Data may be collected to address specific public health research questions that can’t be answered with pre-existing data. Such data is typically associated with a journal or published report.
• Other reference data - Many other sources of data may prove crucial in public health data analysis. Examples include population, geographic, climate and weather data, animal/veterinary data, and economic data. More recently, we have become increasingly interested in unstructured data such as social media posts, or news articles scraped from the web, which may provide further insight into particular types of analyses. Epidemiological Intelligence From Open Sources (EIOS) is a WHO tool to collect such data.
• Surveillance data - Public health surveillance is a broad umbrella term that includes much of the above. On their website, CDC quotes the book “Field Epidemiology” to define public health surveillance as the ongoing, systematic collection, analysis, and interpretation of health-related data essential to planning, implementing, and evaluating public health practice.

At Fjelltopp, we help you collect the best public health data possible using our technical know-how and experience with large-scale public health research. Just take a look at our previous projects. We’ll be writing some more blog posts about health data and health data analysis. Stay tuned for part two of this blog!

In the meantime, if you have any questions or need a sounding board for your own health data problem, please get in touch.