Introduction. Patient falls occur frequently in the acute hospital setting and are one of the most common adverse events experienced by hospitalised patients. In-hospital falls have negative outcomes for patients, causing injuries in up to half of those who fall. Falls in hospital create additional costs for health services due to increased length of stay (LOS), and greater health resource use. In contrast to much research focused on in-hospital falls worldwide, little is known about the rate, contributing factors and outcomes of inpatient falls in the state sector in South African hospitals. At the research hospital, a Falls Policy has been in place since 2013. The chosen falls risk screening tool, the Morse Falls Scale (MFS), had not been locally validated, and therefore its ability to accurately discriminate between patients who fall and patients who do not fall was unknown. A focused analysis of local falls incident reporting, and a description of contributory factors and consequences of falls, could better inform and target falls and fall injury prevention. Furthermore, this research may assist in service development and refining the Falls Policy. Methodology. The aim of this study was to obtain broad-based data on the magnitude of patient falls, and to identify factors contributing to falls. The aim was achieved in two parts, the first was a retrospective record review design. Predictive risk factors for falls were explored by comparing two patient groups, a Fall-Group and a Non-fall Group. In the FallGroup, further objectives related to describing circumstances surrounding fall events, including activities patients were performing at the time of the fall, the time of day and day of week the fall occurred, locations of fall events, and the clinical consequences sustained as a result of the fall. The use of the existing falls risk screening tool, the MFS, as well as its predictive accuracy to correctly identify patients at increased risk of falling was investigated. Second, a survey of nurses at the research hospital was undertaken to examine nurses' knowledge, attitudes and beliefs around the Falls Policy and current falls prevention practices. Results. There were 171 reported fall events during the ten-month period, representing 11.77% of adverse events and a falls rate of 0.73 per 1000 patient occupied bed days (POBD) during this time. Significant predictive risk factors for falling were a longer LOS and having a greater number of comorbid conditions. While the mean age of the sample was 50.0 years (SD=17.3 years), the Fall Group was significantly older than the Non-fall Group (p = .004). There were significantly more deaths in the Fall Group (p = .001), and this group had a longer average LOS (p < .001) compared to the Non-fall Group. The only sub-scale from the MFS that was significantly associated with falls was walking status. Minor-moderate clinical consequences were experienced as a result of the fall in 97% of cases (n=124). This study demonstrated that the MFS in use in the hospital has a low predictive accuracy of 55% at the current cut-off score of 50. At this score, the MFS has a sensitivity of 35.9% and a specificity of 75.4%. While an initial MFS was found in each of the cases, there was only evidence of a repeat MFS in 13 participants (9.7%) in the Fall Group. The nursing survey showed 70% of respondents had not had training on the Falls Policy (n=93) and only 37% (n=49) reported receiving regular feedback on fall rates. Receptiveness of most (66%, n=91) nurses to more training in falls prevention is encouraging. Discussion. The fall rate of 0.73 falls per POBD was lower than expected when compared to international studies. At the research hospital, when the Falls Policy was introduced in 2013, a fall was not defined in the policy and as highlighted in the nursing survey, there still appears to be lack of clarity on the fall definition. The MFS had a low predictive accuracy at the current cut-off score. The low sensitivity and specificity of the MFS in this setting may be due to the MFS not being updated regularly as per the Falls Policy. A further reason for the MFS poor predictive value may be the younger age group found in this sample when compared to international studies where the scale has performed better. Recommendations. The poor predictive value of the current risk screening tool found in this study is concerning. Therefore, further investigation into whether the MFS performs better if it is updated more frequently, and if completed in full, as per the Falls Policy, is recommended. Alternatively, the hospital should consider all patients with multiple comorbidities and those with longer length of stays at high risk, and provide interventions to minimise risk as per the Falls Policy. Future research into factors contributing to fall events and falls prevention should follow a prospective design and be supported at management as well as ward level. Further investigation into the most appropriate way to reduce harm from falls is recommended at the research site. Conclusion. This descriptive study provides a starting point for the hospital to examine the Falls Policy and falls prevention strategies currently in use. It is hoped that the study will contribute to local awareness-raising and capacity-building and help the hospital evaluate current practice and set a baseline for improvement.