Analytics is the process of developing actionable insights through problem definition, application of statistical models or operations research techniques and analysis against existing and/or simulated future data.
The HSR team conducts in-depth analytics into key strategic areas to provide insights to hospital management. The findings form the basis of decision-making in designing interventions to address these areas of concern.
Examples of the work done in analytics include:
CGH Visioning Study
This study determined the collective vision and core values of CGH staff after a collection of vision statements and core values recommended by more than 2,000 respondents. Qualitative and quantitative analyses were used to identify common themes mentioned by the respondents. The classified responses were also profiled with the demographics of the respondents to identify trends. The study allowed the management to condense and gather key themes to craft the future vision statement and core values of CGH staff. This single exercise was also a good platform for management to better understand the needs of different groups of staff, to better align employees' goals with CGH's goals.
Frequent Inpatient Admissions Analysis
Hospitals in Singapore and their Accident and Emergency (A&E) departments are over-utilised as the population is ageing. Inappropriate use of public healthcare services is associated with those who use them frequently. This may have a negative effect on quality of care due to overcrowding. This analysis examines frequently admitted inpatients to identify who they are and how they use hospital services, to propose potential solutions to the issue. We used data from inpatient, A&E and medical social services databases.
Evaluation of Inpatient Pharmacy Workflow
Baseline operating characteristics of CGH's Inpatient Pharmacy were evaluated to identify process gaps (if any) within the pharmacy, that could potentially be addressed to speed up the inpatient discharge process. Historical trends of prescription order times, patient arrival patterns at the Inpatient Pharmacy and waiting times were analysed. The findings were subsequently used to determine whether to deploy pharmacists for discharge rounds.
Strategic Risk Assessment of CGH's Overall Inpatient Bed Demand
A discrete event simulation model was developed to project the daily net bed demand at CGH. The model accounted for the day-dependent effect of daily new admissions via the A&E, Specialist Outpatient Clinics and others, as well as daily new discharges. After the model had been validated by comparing its net bed demand projection with historical data, it was then employed to perform strategic risk assessment under specific scenarios of annual bed demand growth rates and possible schedules of new beds being made available at the Integrated Building.
CGH Acute Bed Demand Projection
This study projected demand for acute beds in 2020 using two different models - negative binomial regression and time series (exponential smoothing and ARIMA) to compare the projected bed demand with that derived using constant annual utilisation rate approach. This analysis included beds in general wards used for inpatient admissions (excluding day surgeries) for acute bed demand projection. Beds in the Intensive Care Unit, High Dependency, isolation and security wards were excluded. About 80% of the data from September 2006 to August 2010 was used as a training data set to develop predictive models. The remaining 20% from September 2010 to August 2011 was used as a validation data set to validate the model. The models' performances were evaluated by mean absolute percentage error.
Profiling Inpatient Falls at CGH
Patient falls are the second most common problem in health systems following medication errors. This study examined risk factors of patients who fell, mechanisms contributing to falls, and patient characteristics by fall aetiologies. This was a retrospective study on falls occurring in 2006 to 2011 across inpatient wards at CGH. Electronic records of patients with falls were reviewed, and qualitative data on free-text descriptions of fall incidents were examined. Patients without fall incidents were used as a comparator to identify factors significantly associated with falls. Based on the content analysis of the narrative descriptions of fall incidents, fall mechanisms were identified. These were further classified into three aetiologies - anticipated physiological fall, unanticipated physiological fall, and accidental fall. Differences in patient characteristics among the three aetiologies were assessed.
Geographic Analysis of A&E Visits
The volume of A&E cases at CGH has increased steadily over the years. This study was a descriptive analysis of A&E visits made by patients from different geographical regions of Singapore in 2011. Postal codes of patients' residences were mapped on digitised Singapore maps with geographical demarcations of development guide plans. These were used to examine volume of A&E cases and distribution of patients by geographical location. Patient acuity status, time of attendance, medical diagnosis and disposition status were also examined in the study.
2 Simei Street 3,
(65) 6788 8833
(65) 6788 0933
Changi General Hospital (CGH)