Your browser doesn't support javascript.
A big-data analysis dashboard for auditing, evaluation and AI-based prediction of acute neurosurgical referrals
British Journal of Neurosurgery ; 36(1):162, 2022.
Article in English | EMBASE | ID: covidwho-1937541
ABSTRACT

Objectives:

Healthcare dashboards provide a visual, interactive presentation of clinical and service data that facilitates interpretation and decision-making. We present an interactive dashboard app for neurosurgical referrals that not only offers important audit insights into but can also make robust machine learning predictions regarding future surgical demand. As a test case, we evaluated referrals from the start of the Covid-19 lockdown to present.

Design:

Single-centre retrospective study with mixed-methods design and app feasibility testing.

Subjects:

10,033 Acute referrals were made via referapatient to our neurosurgical centre between March, 2020 to October, 2021 (female=4938, mean age [SD]=61.1 years [18.8]) from 116 hospital sites.

Methods:

Data were anonymised then analysed in Python before transfer to Plotly Dash / Heroku. Forecasting was performed with state-of-the-art artificial intelligence methods (Prophet). User experience was tested using semi-structured interviews and scales of feasibility, acceptability and usability.

Results:

96% were stated as 'emergency' or 'urgent' by the referrer and 79% of referrals were made by a junior/FY-doctor. 9.5% of referrals were accepted for immediate transfer in this time period. Weekly referral timing was concentrated at 2-6 PM on weekdays with weekend effect/reduction for certain diagnoses (p<0.0001). There was a significant increase between early and late lockdown periods, mainly driven by an increase in spinal referrals (early median-wkly-vol=34, late median-wkly-vol=44, p<0.05). Referrals were forecasted to significantly increase beyond this with low algorithmic error (mean percentage error=2-8%). 20 participants were recruited for feasibility testing 5 consultants, 12 registrars and 3 management-staff. All user groups gave high usability, feasibility and acceptability scores.

Conclusions:

We showcase a big-data analytics dashboard that can audit and rapidly predict acute neurosurgical referral volume that is highly generalisable to other neurosurgical centres. This type of software is critical in enabling a dynamic, flexible surgical service.
Keywords

Full text: Available Collection: Databases of international organizations Database: EMBASE Type of study: Experimental Studies / Prognostic study Language: English Journal: British Journal of Neurosurgery Year: 2022 Document Type: Article

Similar

MEDLINE

...
LILACS

LIS


Full text: Available Collection: Databases of international organizations Database: EMBASE Type of study: Experimental Studies / Prognostic study Language: English Journal: British Journal of Neurosurgery Year: 2022 Document Type: Article