5.3 Data modelling categories

Data analytics has various modelling categories:

  1. Descriptive/prescriptive data (also called ‘small data’)
  2. Predictive modelling
  3. Exploratory analytics
  4. Real-time analytics
    (Blackett, 2013; Wills, 2014)

These methods allow health organisations to use raw patient and related data to support improvement in the quality of care of their patients, provide predictive modelling of progression of disease states as well as improve allocation of available resources in a proactive approach, conducting gap analysis to address organisations goals (Wills, 2014). Organisations can also adapt one or two types of data analytics to assess various aspects of performance and set goals (Chong & Hui, 2015).

Descriptive/prescriptive data

This concept includes examining a specific dataset that is collected for a solution or an improved outcome to an existing identified challenge in health service (Blackett, 2013; Wills 2014). For example, you could focus on a targeted approach in an area where high medication errors have been identified. Prescriptive analytics usually focuses on recommendations and help for decision-making. The concept helps end users to identify issues and find optimal solutions, direction and strategy (Wills, 2014).

Small data is easily retrievable in a relatively cost-effective method from electronic systems and various software and other IT databases; for example, descriptive data such as patient demographics for a particular cohort or historical data used to summarise a situation (Blackett, 2013).

This data offers invaluable information to leaders and managers across health services to easily analyse and identify patterns, trends and provide insight on patient cohorts and demographics. This information can be a source for workforce planning, strategic planning, goal setting and resource allocation to address specific demands (Kibbe & Kuraitis, 2012; Terry, 2012).

Predictive modelling

Predictive modelling includes extrapolation of current data to predict future outcomes (Ingenix, 2006). For example, in healthcare modelling can be used to target a specific group of patients while they are inpatients to review a specific disease state and ideally prevent readmission and hospitalisation (Wills, 2014).

Exploratory analytics

Exploratory analytics includes examining a set or sets of data from various sources to yield unpredictable observations or identify unexpected correlations between various parameters. For example, in healthcare you could analyse patient feedback about their hospital stay and management during an episode of care. This type of data analytics also includes examining patient cohorts admitted for management of a specific disease.

Real-time analytics

This concept uses real-time data at the point of care to help clinicians make timely decisions for patient care, often at the bedside (Murphy, 2013). Examples of real-time analytics include identification of medication interactions through prescribing alerts at the time of medication order entry, availability of pathology and vital signs data, as well as availability of clinical decision tools such as various treatment algorithms at the point of care (Taylor, 2010).

The uptake of various categories of data analytics by healthcare organisations will likely have a profound impact on the quality and efficiency of models of care for the short- and long-term care of patients. (Chen et al., 2020a)

Licence

Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Value-based Healthcare Copyright © 2024 by La Trobe University is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book