Author: admin

Digital health information can be pushed or pulled, and while most health information systems are likely to have components of both, it is important to understand that they exist in different paradigms and require different policy and legislation. Pushing of digitalised health data works in ways similar to paper based systems; only electronically. Pushed health information arrives with the recipient only after the sender initiates transfer. Pulling of health data happens when users actively search for and extract the health information from the relevant source.

For example, a patient moving through a healthcare system can either have their medical records pushed by their previous medical practitioner to follow the patient to the new practitioner, or they can have a medical record from which medical practitioners can pull the relevant information when required. Effects of moving towards a pull system of data access need to be considered carefully, as it results in a shift in the burden of responsibility. While previously secondary care institutions may have pushed patient information to the respective general practitioner, the burden shifts to the general practitioner to keep track of the patient’s course through the healthcare system by pulling relevant information from one or more sources and data bases. Pathways through healthcare systems are often complex, and in systems where practitioners are expected to ‘pull’ the relevant health information on patients, the benefits of consolidation of patient health information to a single secure and centralised source become apparent.

The flip side to the shift in burden of responsibility, is that ‘pull’ systems in health information exchange create a shift of ownership over the health information, provided there are effective, mandatory and secure systems in place for reporting (uploading) health information. If information is pulled from a centralised patient health record, there is room for patients to gain more ownership over who can get access to it, resulting in a more patient-driven health information system. Additionally, both practitioners and patients are no longer reliant on finding and chasing medical records, to then have them pushed through by previous practitioners.

The benefits of pulling health information is not limited to health practitioners and personal medical records. If governments and other decision makers can pull public health information when and from where required, they are no longer dependant on reports being pushed to them. This results in more timely and evidence-based health policy, as well as a potential decrease in administrative overhead as health information is pulled only when needed, rather than processing everything which is pushed through.

Understanding the discrepancies between, and ramifications of, pushing and pulling health information is vitally important in moving towards digital health information systems and realizing the mandate of the new Primary Health Networks: more effective and efficient health systems in which there is good coordination of care.

High Blood Cholesterol can be effected by a variety of lifestyle choices including diet, activity level, smoking and weight, all of which can increase the risk of heart disease. It is therefore important for PHNs to focus on intervention strategies given the potential benefit of behaviour modifying education. According to the ABS, Australian Health Survey, 32.8% of the population have high blood cholesterol.

If High Blood Cholesterol data is mapped for the PHN regions in NSW, it clearly highlights a higher than average incidence in a number of PHN Regions, with Northern Sydney PHN standing out as very high. However this first map is a very blunt tool delivering little information for decision makers or information for improving health outcomes. But the ‘Cogent Health Data Decision Tool’ can do better.

Imagine if the PHN or Local Health Network could very easily drill down and understand where the priority suburbs are located, and then allocate the valued Health budget to begin an effective process toward real and targeted community education.

So let’s take the Northern Sydney PHN region as an example. At Cogent we have been working on a big data analysis methodology which allows us to drill down further by reducing big data into manageable population numbers for effective and efficient analysis and consequent allocation of funding against clear strategic health objectives. In fact, we can convert and map ‘big health data’ down to LGA, PHA, PHN and SA3 regions all within seconds and toggle between the health indicators as required.

Here is the Northern Sydney PHN now mapped against smaller Population Heath Areas (PHAs). Gone is the LGA (Local Government Area) based analysis, though this could just as easily be mapped against the SA3 04 SA4 ABS regions or LGAs with just one change of option within the Cogent Health Data Decision Tool.

This map now shows the PHN (red boundary line) with a cluster of PHA regions centred on the Turramurra and Gordon suburbs with an ASR** of over 36.5 per 100: being an incidence of high blood cholesterol against the national average of 32.5%. With our analysis, the PHN can now develop targeted localised responses with a validated argument for additional funding utilising an established benchmark, rather than spending scarce funds across all of its region.

Correlation to Associated Lifestyle Factors
The analysis method as developed at Cogent* can further cross-correlate this localised instance of cholesterol risk against other determinant factors, such as smoking or obesity: utilising the entire PHA set of regions in NSW. This allows for a more efficient and effective use of collaborative strategies to address health issues at a localised level within each PHN. Using the Cogent program and by ruling in or out other instances of health with associated suspicion, the decision maker can instantly focus their intervention strategies. In this example the associated correlation of high cholesterol to Smoking and Obesity in the Northern Sydney PHN is low and could indicate diet as the primary factor driving the regionalised health risk.

The methodology developed by Cogent can further use advanced machine learning to produce optimisation options, thereby allowing for the development of specific strategic responses such as the placing of GP Registrars, or the placing of community education programs in the areas of highest need.

With Cholesterol levels already high across the population, the new ability to target behaviour modifying programs of education is ideal for decision makers. The ‘Cogent Health Data Decision Tool’ allows PHNs and LHNs to use big data analytics with immediacy and localised relevance promising powerful opportunities for program efficiency in delivering improved health outcomes for the Australian Community.

*Cogent Business Solutions developed the ‘My Hospitals’ portal for the Commonwealth
** Age Standardised Rate

If you’re interested in developments in this area. Just provide your email by clicking here.
or contact Ian Hook
Senior Consultant – Cogent Health and NGO Division
0455165508

The Australian Health Policy Collaboration has just published the Australian Health Tracker Report 2016.*  This first national health report card is a welcome addition to the debate on health service needs.

As an example, the report puts the national average for Diabetes at 4.7% for the 25-65 age group and rising. To plot this against the Primary Health Network regions of NSW (PHN) delivers an interesting and concerning result.

If this data is mapped for the PHNs, it clearly highlights the same elevated incidence in a number of PHN Regions, but it is a very blunt tool delivering no further information for decision makers or information for improved health outcomes within communities.

Imagine if the PHN or Local Health Network could drill down and understand where the priority suburbs are located, and then allocate the valued Health budget to begin an effective process toward real community outcomes.

So let’s take the South Western Sydney PHN region as an example, where in the graph above diabetes is identified at almost twice the national average.

At Cogent we have been working on a big data analysis methodology which allows us to drill down further, reducing big data into manageable population numbers for effective and efficient analysis and consequent allocation of funding against clear strategic health objectives.

Here is the same South Sydney PHN now mapped against smaller Population Heath Areas (PHAs). Gone is the LGA (Local Government Area) based analysis, though this could just as easily be mapped against the SA3 ABS regions or LGAs.

Our analysis can be translated to a map of the PHN (red boundary line) showing a cluster of PHA regions centred on Canley Vale and Fairfield with an ASR** of over 15 per 100 incidence of diabetes against the national average of 4.7. With our analysis, the PHN can now develop targeted localised responses with a validated argument for additional funding utilising an established benchmark, rather than spending scarce funds across all of its region.

The analysis method we have developed can further cross correlate this localised instance of diabetes against other health indicators, such as obesity or alcohol-use: utilising the entire PHA set of regions in NSW. This allows for a more efficient and effective use of collaborative strategies to address health issues at a localised level within each PHN:  by ruling in or out other instances of health with associated suspicion, the decision maker can instantly focus their programs. The methodology developed to date can further use advanced machine learning and artificial intelligence tools to develop optimisation, thereby allowing for the development of specific strategic responses such as the placing of GP Clinics or GP Registrars, or the placing of mobile Cancer Screening facilities in the areas of highest need.

With a National Diabetes Target of 4.1% by 2025, the use of big data analytics with immediacy and localised relevance promises powerful opportunities for Health decision makers to target programs and focus health strategies so they achieve the most effective health outcomes for the Australian Community.

Contact Cogent if you’re interested in developments in this area or utilising this capability in your organisation.

Reference

* https://www.vu.edu.au/sites/default/files/AHPC/pdfs/australias-health-tracker-adult.pdf

** Age Standardised Rate

 

We have the numbers… now we can use them.

It is the question of the decade: how do we not just collect but effectively use data to plan for the future of health? Increasing concerns surrounding rising healthcare costs, population densities, chronic diseases and ageing populations only add fuel to this discussion.

Data collection itself is omnipresent in health, ranging from increasingly digitalised medical records, Australia’s well established Census and National Health Surveys, documenting and reporting of hospital data, Cause of Death reporting, specific disease reporting, and data recorded under the Pharmaceutical Benefit Scheme, to just name a few. In addition to this, there is a number of other more creative data sources which have powerful potential, such as tobacco tax collection per location, data from ‘healthy living’ phone applications, or Google behavioural search trends.

However, large-scale and well-executed data collection often comes at a cost (Australia’s 2011 Census alone cost 440 million dollars) and is defendable only if we can use the data to enable effective and efficient healthcare monitoring, evaluation and planning leading to better, more cost-effective health systems in the long run.

‘Big data’ in health has a lot to live up to. It is credited with potential benefits ranging from improving effectiveness of current health systems, decreasing health costs, spotting trends and epidemics, increasing accountability in healthcare spending, to improving planning in health through predictive modelling and simulation.

At Cogent Data Analytics we have been exploring innovative ways of managing big data in health to maximise its potential. While a lot of health information is being collected, there are barriers in the accessibility of this data to the public and to key decision makers.

How can we improve the access to data from a range of sources to decision makers at all levels, allowing them to make evidence based decisions on health budgets and improve targeted spending in health? Can we portray the data in a way that enables good data literacy by the users, both those statistically trained and not? What can we do to make health data less siloed and more integrated, to reflect the reality of health? How and to what extent can we best exploit the predictive powers of our health data, using correlation and optimised solutions? What is the best way to ‘dig deeper’ and enable informed analyses of the data by key decision makers, which goes further than simply ‘having the numbers’? Finally, can we give local decision makers the tools needed to become health leaders?

Here at Cogent Data Analytics, we have been pulling apart health data and reworking it for our clients: to give the data meaning and practical application. In a recent project, we worked on integrating a number of indicators to provide a multi-variable analysis of health discrepancies across a Primary Health Network. If you have an application for local data management and analysis, please contact us.