top of page

Case Studies

Te Waipounamu Data Platform

Partnering with South Island PHOs to Deliver Region-Wide Insights

South Island PHOs partnered with DataCraft Analytics to create a single, shared platform for population health data across Te Waipounamu. The goal was simple but ambitious: to provide timely, secure, and consistent insights that would support PHOs, Iwi Māori Partnership Boards (IMPBs), and the wider health system.

​

The Challenge

​

Before this project, each PHO worked with data mostly in silos. This made it difficult to deliver insights at a regional level and limited the ability to plan at a system level across the South Island. PHOs needed a trusted partner to enable stakeholders to agree on common standards, and deliver a solution quickly.

​

Our Approach

​

Working collaboratively with all South Island PHOs, we worked with each to establish a project structure, including a governance supported by both a Technical Advisory Group and a Clinical Advisory Group. DataCraft Analytics was primarily responsible for:

​

  • Developing and documenting agreed data specifications

  • Designing the secure data-sharing protocol to hold the data

  • Built the platform in Microsoft Azure, leveraging the proven DataCraft Analytics stack already in use by PHOs

  • Building flexible data pipelines to allow PHOs to easily supply data to the platform

  • Building Power BI reports within the secure Thalamus environment

  • Administering access to the reports and platform

 

Rapid Results

 

Delivery was achieved at pace:

​

  • 2 weeks – first data capture from PHOs

  • 4 weeks – live Power BI insights covering the entire South Island population

 

From the start, the platform produced valuable reporting on:

​

  • Population insights

  • PHO project expenditure

  • Immunisation coverage

  • Secondary care events

  • Realised enrolment and practice access

  • Diabetes outcomes

 

Collaboration with Iwi Māori Partnership Boards

​

The project was designed from the outset to support wider stakeholders and took a collaborative approach with Māori, through partnership with Iwi Māori Partnership Boards (IMPBs). The IMPBs contributed along-side PHOs at governance, technical and clinical levels to help to shape the solutions to be fit for purpose for all in Te Waipounamu.

​

The Impact

​

The Te Waipounamu Primary Data Platform has transformed the way South Island PHOs work with collaborative data in strategic planning and convesations. It provides:

​

  • A single, trusted view of primary care data across the South Island

  • Faster insights to guide planning and service delivery

  • A scalable foundation for adding new datasets and reports as regional needs grow

  • A model of collaboration between PHOs, Iwi Māori Partnership Boards, stakeholders and technology partners such as DataCraft Analytics

SHIVERS V Primary Care Study

Partnering with the University of Auckland and University of Otago
 

DataCraft Analytics partnered with researchers from the University of Auckland and the University of Otago on the SHIVERS V Primary Care Study, a major international research initiative funded by FluLab (USA). The project set out to answer a critical question: what impact did New Zealand’s border closures during the COVID-19 pandemic have on respiratory illness presentations to general practice?

 

Our Role

 

Using our Analytique platform, we applied Sherlock, our novel clinical text classifier, to analyse over 10 million patient encounters across a five-year period, spanning both pre-COVID years and the pandemic restrictions.

 

Unlike traditional approaches, Sherlock does not rely on clinicians manually coding presentations. Instead, it unlocks the rich clinical narrative recorded in free-text notes to classify presentations at scale. Critically, this is done within each practice’s own system: Sherlock processes the raw notes locally and only transmits the resulting classification. This ensures that patient privacy and the most sensitive health information remain securely inside the practice’s environment, while still enabling population-level insights.

 

Clinical Expert Classifier Development

 

Working closely with the SHIVERS V research team and practicing GPs, we built and validated a clinical expert classifier capable of reliably identifying presentations of:

 

  • Influenza-like Illness (ILI)

  • Upper Respiratory Tract Infections (URTI)

  • Lower Respiratory Tract Infections (LRTI)

  • Asthma

  • COPD

  • Confirmed COVID-19 cases

  • COVID-19 related presentations (e.g. negative results, suspected cases, patient discussions)

  • Otitis Media

  • Sore Throat

 

This classifier was rigorously tested against expert GP opinion to ensure accuracy, reliability, and clinical relevance.

 

Impact

 

By combining Sherlock’s novel, privacy-preserving approach with expert clinical validation, the project overcame the typical shortfalls of studies that depend on clinical coding. The result was robust, population-level insights into how respiratory illness presentations changed under border closures. These findings help inform public health policy, pandemic preparedness, and reinforce the vital role of general practice in monitoring community respiratory health.

bottom of page