By creating a consolidated dataset, analysis of different hypothesis could be undertaken in minutes, rather than days.
Sector: Healthcare & Pharmacy
Services: Statistical Analysis
Timescale: 16 weeks
Following ten years of data collection, an NHS department came to us for help with statistical analysis.
Raw data had been stored, but due to poor structuring, it was difficult for the medical researchers to test their various hypotheses.
They wanted to understand what treatment practices they were administering yielded the best results, and use these insights to create a report recommending future best practices.
Initial conversations with the client were important to understand the raw data make-up and agree on the form the outputs should take.
We produced a playback of the anonymised raw data, providing an overview of the information. This highlighted some inconsistencies in the raw data and provided a framework of the complete data set and potential additional analysis that could be undertaken.
The next step was to reconcile differences between the raw datasets as there was no ‘single version of the truth’. To solve this, we visualised the alternate views of how patients had been treated and agreed with the research team how to treat the discrepancies.
With the differences resolved, we were able to build up a dataset to undertake the statistical analysis.
Using Alteryx software, we aggregated multiple datasets into a single, usable source, allowing us to make additional calculations and flag information relevant to the hypotheses. This data source functioned as the core to the data analysis, providing information on demographics, geographies, and specific survival analysis techniques, as required.
Central in this process was the supporting documentation that we worked through with the researchers, to indicate how the analysis would be done, and how further hypotheses could be tested.
By creating a consolidated dataset, analysis of different hypothesis could be undertaken in minutes, rather than days. Multiple iterations could, therefore, be generated and authored to allow all potential hypotheses to be explored.
Speed was achieved by using the right tools for the right problem – Alteryx to handle and consolidate large raw data, Excel and Tableau to playback the data, and R to perform the most complex statistical analysis.
All outputs; datasets, tables, and visualisations were returned to the client, assisting them on the creation of their report on treatment best practices.
This process helped the NHS to save money, staff time, and future patient’s lives.
Using data to understand how well you actually delivered to the customer.
Visualising costs and service of supply chains to set out distribution strategies.
Building a bespoke stock model to balance both resource and costs.