From: The path from big data analytics capabilities to value in hospitals: a scoping review
Value targets | N | References | Description |
---|---|---|---|
VT1 – Decision making | 62 | ||
Care – Diagnostic | 12 | Hu et al. (2018) [71] | Develop models based on machine learning to aid the diagnosis of hyperlemia at point of care. |
Care – Risk detection | 22 | Genevès et al. (2018) [72] | Use machine learning on prescription data to detect, on the day of hospital admissions, patients at risks of developing complications during their hospital stay. |
Admin. – Assessing hospital activities | 20 | Mahajan et al. (2019) [73] | Develop a data-driven methodology for decision-making supported by the use of quarterly strategic analytics for improvement and learning (SAIL) reports to visualize data, study trends and provide actionable recommendations. |
Admin. – Resource allocation | 14 | McNair (2015) [61] | Use statistical model to forecast the optimal safety level of nurse staffing in intensive care units. |
Research – Hypothesis setting | 2 | Hendricks (2019) [62] | Use process mining to explore available hospital logs and identify areas in clinical operations to further investigate. |
VT2 – Innovation | 54 | ||
Care – Precision medicine | 27 | An et al. (2018) [74] | Develop algorithms using machine learning methods to predict drug-resistant epilepsy in order to ensure these patients receive specific care and interventions following their diagnosis. |
Care – Preventative medicine | 22 | Zolbanin and Delen (2018) [75] | Propose new data processing approaches to predict preventable readmissions for patient with chronic diseases and prescribe the best course of actions for each patient at discharge to prevent readmission. |
Admin. – Adapt strategies | 10 | Navarro et al. (2018) [67] | Develop a machine learning algorithm using perioperative data to predict length of stay and inpatients costs after primary total knee arthroplasty and propose a patient-specific payment model better reflecting patient complexity. |
Research – New research tools | 2 | Johnson et al. (2016) [63] | Develop a dynamic simulation tool suitable for data visualization of both human-designed and data-driven process which can be used for “what if” analysis and used to deep-dive on big data. |
VT3 – Performance | 36 | ||
Care – Patient flow | 25 | Krämer et al. (2019) [76] | Use supervised machine learning techniques to train a model to classify inpatient admissions as either emergency or elective care to reduce the number of hospitals admissions from the emergency department. |
Admin. – Operations management | 7 | Guan et al. (2017) [54] | Use statistical model to investigate platelet usage patterns and better forecast future demand to reduce wastage. |
Research – Research performance | 6 | Karanastasis et al. [68] | Develop a platform with tools and services necessary to explore big data in clinical research to improve the efficiency of clinical trials design and the effectiveness and speed of subject recruitment. |