Delivering patient outcomes with data

Health organisations that collect, engineer and harness data effectively can elevate the quality of care they provide, building trust and improving the experience for clinicians and patients alike. 

The more data that is captured, aggregated and analysed, the better patient conditions can be monitored and managed. Ultimately, this can enable faster, better, and earlier clinical decisions that benefit patients, clinicians and system managers. 

However, many health institutions currently lack sufficient processes and functionality to manage this data. This exposes organisations, clinicians and patients to unnecessary risks and missed opportunities. The good news? With many virtual health channels still in their infancy – now is the ideal time to get the right data foundations in place ahead of scaling up and scaling out.

Having a clear view of patient journeys (across multiple health providers as well as in-person and virtual models of care) can help identify opportunities to capture the right data, store it securely yet accessibly, share it with providers and patients and, critically, translate it into meaningful insights and actions for better patient outcomes and improved health services.

To do this, however, providers need a comprehensive strategy that covers three key aspects of data use.

1. Standards, governance, and interoperability

In its rawest form, a lot of virtual health data is unstructured, stored in different places and formats. Standardisation allows health organisations to process unstructured data and combine it with structured data for analysis. For example, natural language processing, information retrieval and machine-learning techniques can convert unstructured free-text data (such as information from free-text contents of pathology reports) into computable data for use with AI. 

To convert unstructured data into meaningful information, healthcare providers also require governance practices to underpin data recording, optimisation, and analysis. Governance should extend across the organisation itself and also encompass data-sharing agreements with third parties. Therefore, governance needs to be flexible enough to adapt data and analytics to different healthcare scenarios. 

One essential pillar of good data governance is privacy and security. It’s important to embed this early in the design phase of new health models, rather than making it an afterthought that is (expensively) retrofitted later on. Mapping entire patient and clinician journeys can help identify where data needs to be exchanged or combined with different parties, and in what format. At each point, security and privacy requirements should be specified. 

Another essential pillar of good data governance is interoperability; establishing connections between systems to allow the secure, timely, accurate exchange of data. When this is managed well, health data and health relevant data can be combined from various sources (such as virtual health interactions, electronic health records, patient case histories) and activity and lifestyle data (like real-time digital data from sensors, wearable devices and trackers).

To enable interoperability, health organisations should identify silos and barriers to sharing data, and then seek ways to remove these. This can pave the way to adopt digital tools that empower patients to be more proactive about their own health.*

2. Privacy and security

Activating virtual health, in particular the use of remote patient monitoring and asynchronous communication, increases the surface area an organisation must defend from a cybersecurity perspective. It’s therefore important to assess whether the organisation’s existing technologies and procedures can handle the increased and varied scope of devices and threats.

Assessments should review current data protection policies and security program controls to establish whether they are sufficient to cover the more complex network boundary and architecture brought on by virtual health (not to mention the full ‘reach’ of data interactions with third parties). 

It is also important to look ahead and consider the organisation’s long-term goals for virtual health expansion. This helps clarify what security and privacy capabilities will be required in future, and therefore what measures should be established now in readiness for this (eg. data interoperability, tracking and interventions, device integration, cyber threat mitigation).

Broadly, good security and privacy processes for virtual health include:

• Security assessments: Testing the efficacy of the security controls protecting health information (using a variety of tests such as vulnerability assessment, penetration testing, security configuration review, etc).

• Digital identity: Establishing the security discipline to enable the right individuals to access the right resources at the right times for the right reasons. 

• Monitoring: Clarifying the required capabilities to construct, collect, aggregate, and correlate cyber risks and events.

3. Monitoring and evaluating outcomes

The beauty of virtual health models is that – with the right preparation – they can evolve and flex in line with changes in consumer behaviour and clinical results. In this way, service delivery and patient outcomes can continually improve.

To enable this, it’s essential to understand how results will be measured. While virtual health is in its (relative) infancy, now is the time to collect baseline outcomes to compare against future outcomes. Longitudinal data can reveal patterns and trends that can inform further improvements.

To determine what data matters most, consider this from the clinicians’ perspective: What data could best support their decision-making? And how might findings be presented to clinicians to support their decisions on ‘best practice’ given a patient’s characteristics (for example, visualisations or dashboards)? It’s also important not to overlook the possibility that – if clinicians know certain metrics are being calculated – this might affect their behaviour when recording certain data.

In this design phase, it’s vital to closely consult with clinicians. Health professionals understand the basic logic that strong evidence generally delivers stronger outcomes for patients. Similarly, the stronger that data (i.e. evidence) is managed, the better that health leaders can assess best practice and optimise delivery models to achieve higher quality clinical (value-based) care. 

Of course, the outcomes of health service delivery extend beyond the immediate experiences of patients and clinicians too. So, in clarifying what data to measure, it’s also worth considering how to track economic and social participation, and improved quality of life.

Having clarified what outcomes will be measured, health organisations can establish the methodologies to collect data, calculate metrics, and present outcomes.

Getting started

Introducing a comprehensive data management strategy across an entire organisation, service or sector can be a daunting prospect. But it doesn’t have to happen overnight. Often, the best chance of lasting, large-scale success is actually to start small. 

Many health leaders begin with a pilot project for data management, where lessons can be learned to inform wider rollouts later on. (For example, a pilot might involve an outpatient virtual service that is data rich and doesn’t involve care for critically ill patients.) The guiding philosophy for such pilots should be ‘build, test and learn’, so that the concepts can be gradually shaped and improved by experience and evidence.

For every organisation, good data management is a journey, not a destination. When it comes to data-enabled healthcare, the important thing is to make a start.