Our study was aimed at identifying fall predictors in hospitalized patients with cancer and developing an assessment tool for screening patients at risk for fall at admission to the hospital. Even though several studies conducted in different countries have investigated the issue before, ours is the first study conducted in an Italian population in the cancer setting [3–18]. Risk for fall varies according to type and specificity of oncology wards . The need to implement an assessment tool based on our patients and organizational characteristics justified the study. Hospitalized patients with different types of advanced cancer diseases were included in this study. The median age of the study population was in line with Stone et al, while the prevalence of falls with minor or moderate injuries (44%) was in line with two previous studies [9, 23], but lower if compared with Tsai (84.8%) . Forty per cent of study participants were women, 86.8% had solid tumors and 13.2% were affected from hematologic malignancies.
Because the fall rates and injuries and the environment and organizational characteristics of included centres are not constantly described in the previous studies focusing on fall predictors, a full comparison of our findings with them is not possible. The fall rate per 1000 patient days in this study resulted lower than in Weed-Pfaff, Fischer and Pautex et al: respectively 2.07 vs. 3.41, 3.83 and 6.89. This also applies to the fall rate per 100 patients compared to Lorca, Weed-Pfaff and Capone’s studies (i.e., respectively 1.19% vs.1.98%, 2.4% and 3%) [3, 8, 11, 15, 23]. Clinical characteristics of patients and organizational variables may partially explain such differences in fall rates between studies.
In line with previous studies, we found for the most of the candidate fall predictors higher prevalences in the cases group as compared to controls. This was applies to the need for walking aid, having a neurologic disease, impaired vision/hearing, delirium, anxiety/depression, gate imbalance disorders, fatigue, hyposthenia of lower limbs, use of ambulatory devices, diuretics, hypnotics and opioids [3, 8- 11, 13]. Conversely, in contrast with previous studies, we found no statistically significant differences in prevalences of having comorbidities such as cardiovascular, endocrine and respiratory diseases, treatment induced neuropathy, the presence of plegies, and the use of antihypertensive, hypoglicemic agents and having received blood products transfusions in the last 24 hours. History of previous fall prevalence in the study population resulted low (2,6%) with no statistically significant difference between fallers and not fallers.
Having a hematologic malignancy (tumor site), neurologic disease (including brain metastases, or other neurologic diseases), gait imbalance disorders, fatigue and the use of diuretics, hypnotics and opioids resulted as independent predictors for fall.
Patients with hematologic malignancies at our centre are hospitalized mainly for receiving high intensity treatments (e.g., high dose chemotherapies, bone-marrow transplantation) or for the management of serious toxicities that can’t be managed in outpatient settings. Treatment regimens of such patients could explain having hematologic malignancies as a fall predictor. This result is in line with the studies of Capone et al, Weed-Pfaff et al and Jun et al, who reported that hematologic patients receiving bone marrow transplantation present higher risk for falls [3, 11, 13].
Having a neurologic disease resulted as an independent fall predictor. Indirectly, this finding confirms the study conducted by Stone et al in 2012, in which having primary brain cancer or brain metastases resulted as a hospital fall predictor .
Gait imbalance disorders, asthenia of lower limbs and fatigue resulted as independent fall predictors as well. These factors are mutually influenced and are related with disease itself as well as cancer treatment side effects. Besides, patients experiencing fatigue likely have asthenia of lower limbs which may result in gait imbalance. These fall predictors were frequently found in most if not all of previous studies [3, 6, 7, 11–13, 25–28].
The use of diuretics resulted as an independent predictor of fall in the study population. It is known that the need for elimination as result of the effect of this medication category and the hypotensive effect of diuretics, which can explain such predictor. In fact, having elimination issues have resulted as fall predictor in other studies before [7, 24]. Among studied medications, the use of hypnotics and opioids resulted as independent fall predictors in line with findings reported from previous authors [3, 8–10, 28–30]. Both these two medication categories have effects on patient’s cognitive status by influencing their space and time orientation and motor capabilities, which may explain this finding [7–9, 11].
The presence of medical devices in the study population was analysed singularly for each medical device limiting ambulation and by considering having at least one of them (i.e., nose-gastric tube, having infusion lines, drainage tubes, urinary catheter). In our study population having an IV line resulted protective for fall since the prevalence of patients with IV lines was higher in the controls group (OR 0.54; 95% CI 0.34-0.85; P-value = 0.008). This finding may be explained by the characteristics of the drip stands used in our hospital for infusions and with the fact that patients with an IV line probably use them as an ambulatory aid while walking.
Among fall predictors, tumor site, gait imbalance disorders, fatigue and sensor-cognitive disorders such as anxiety/depression resulted related with falls with injury, while previous falls, use of diuretics, hypnotics and opioids were associated with falls without harms.
It is worth noting that despite the retrospective study design, the quality of the information contained in the EMRs helped minimize the known risk of recall bias.
In conclusion, the definition of a screening tool able to identify patients at higher risk for fall based on patients, organizational and facility characteristics represents the first step for implementing tailored evidence-based strategies for fall prevention. We believe that having described the characteristics of the study center, including facility, staffing and nursing organization could help other researchers to evaluate whether the results of our study apply or not to their own context. Thus, future study is needed to externally validate our results and investigate the clinical utility of our screening tool.