PREDICTIVE ANALYTICS FOR EARLY DETECTION AND INTERVENTION
The Selangor International Healthcare Conference 2022 at the Kuala Lumpur Convention Centre from 7 to 9 October 2022 featured speakers on three main themes – namely Future of Healthcare, Population Health and Healthcare Financing.
Enterprise IT News covered speakers on use of data analytics, artificial intelligence and the digital transformation of healthcare.
“As time progresses, while the aims and visions of the physician have remained consistent since time immemorial, the healthcare landscape and vehicles for care delivery have been transformed by rapid advancements in new technologies”, said Dr Paul Tan Sze Hiang, Medical Director of Singapore-based Mesh Bio Pte Ltd, a technology company which applies predictive analytics to medical data, with resultant enhancements in medical outcomes.
Dr. Tan has also been in private practice for 15 years, where his clinic deals mainly with chronic diseases associated wit the elderly.
As a subset of artificial intelligence, predictive analytics is gaining traction throughout the medical industry, amongst healthcare stakeholders, and for many reasons, relying on vast improvements in computational power and speed, the technology has the stability to synthesize insights, and trends from masses of data, achieved mainly from a two step process – namely data harvesting and modelling the use of statistics.
Predictive analytics can also be applied and used as additional support in the healthcare industry and advances have enable the analysis of the abundance of underutilised data obtained from several sources such as diagnostic tests, physical examinations, imaging, genome analysis, and even patient and family histories.
By structuring and ingesting these medical parameters, algorithms which guide the process are able to evaluate how these biomarkers affect individuals health trajectories and enhance of gnostic value.
So the more information available, the stronger the algorithms become at generating higher quality data which guides doctors to provide higher quality care.
For example, Mayo Clinic in the United States utilises analytics to enhance patient care and they have attempted to develop an algorithm that can mask heart abnormalities long before patients begin experiencing symptoms.
This insight will make it possible to better prevent the onset of stroke and predict the likelihood of future arrhythmias.
“Instead of reactively, diagnosing individuals that disease onset, we can prevent disease progression, giving individuals the opportunity to adjust their lifestyle or in fact start prophylactic treatment”, said Dr. Tan.
Whilst the Mayo Clinic’s algorithm is still under trial, however it does illustrate the attitude to uncover potentially novel means of practising preventative medicine.
According to the World Health Organisation (WHO), there are 18 million deaths annually due to cardiovascular disease, whilst The Lancet, the U.K.-based, peer-reviewed general medical journal has observed that US$825 billion are spent on diabetes care each year.
Chronic metabolic diseases typically progress silently and present themselves later as an acute event such as a heart attack or stroke.
Since these diseases can be attributed to both genetic predisposition and lifestyle habits, analytics offers an opportunity for individuals to understand the extent to which each risk factor categorically impacts their health futures, so patients with high risk can be identified so they benefit from early intervention.
“So harnessing the potential of analytics provides us with near limitless possibilities and decisive doctoring aspects”, said Dr. Tan.
Common issues faced by patients through long waiting times, often due to the inefficiencies of clinical workflows, poor IT infrastructure, or perhaps just simply inadequate staffing.
This highlights glaring inefficiencies, and therefore one of the key uses of analytics is the reduction of administrative work by hospital staff at a ward clinic and managerial levels, so time and energy can be diverted to meaningful increments in patient care, while minimising medical error and service failure.
Used to its full potential, there is no doubt that analytics will yield valuable information that provides guidance to national strategies for forward planning, and reshaping the glue of the respective healthcare sectors.
For example, it can be a useful tool for monitoring various aspects of population health demographics, ensuring that future needs are adequately catered for.
In Australia, the health ministry has been working to improve services to the Victorian community by using such technologies to study data trends.
The obtained valuable insights on supply and demand needs of the healthcare workforce, hospital inpatient capacity requirements and future projections on emergency service requirements are geared towards supporting service development capital planning overall strategy.
Why now?
Firstly, interest in the utility of analytics and healthcare has been picking up lately, secondly, the world’s population of aged people is rapidly increasing, and there’s an exponential rise in healthcare costs which puts a colossal strain and cost pressures on healthcare workforces, so increased staffing is one solution.
“However, we must keep looking for novel ways of easing this burden, so the answer lies in technology, and that’s the case for the use of powerful analytics tools to optimise chronic disease control and workflow efficiency”, said Dr. Tan.
Healthcare standards have advanced tremendously and possibly developed nations physical medical records have been increasingly replaced by some form of electronic medical record system. Automation and natural language processing technologies have greatly increased the efficiency of medical note taking and transcribing, whilst cloud based storage services allow interoperability.
As digitalisation of healthcare systems gain popularity, analytics will be expected to spearhead another big role in bridging the gap between patient and doctor.
However, despite its promise, the utility of AI is not without its inherited limitations. Data privacy and security can be a huge issue when it comes to digitalising medical information and trust. Cybersecurity should be every stakeholders priority to protect the interests of patients.
“By implementation of an adherence to strict protocols, you can ensure that patient confidentiality is upheld and some are concerned with the use have AI due to the potential bias that stems from inaccurate data sets, or inadequate representation”, said Dr. Tan.
Since algorithms take in numerical data and trends, it’s possible that social factors, genetic and even geographical influences might be overlooked during the design of datasets. However, this can be mostly overcome if clinicians, data scientists, and creators of algorithms are mindful of these and collaborate to design software that accounts for such bias.
Effective data sets must be based on a diverse range of race, gender and geography to in order to avoid bias algorithms. On top of that, users of AI must adopt impartial practices and effectuate guidelines to avoid this.