Association of Fall Timing and Location With Clinical Characteristics in Geriatric Hip Fracture Patients: A Cross-Sectional Study

Abstract

Introduction

Falls can be detrimental for older adults, causing hip fracture that result in disability and increased risk of mortality. This study aimed to investigate the temporal distribution of falls among the geriatric patients with hip fractures and compare the characteristics of falls occurring at different times and locations.

Methods

This cross-sectional study analyzed 801 older adults with hip fractures from falls. We collected data on fall timing (day vs night), location (indoor vs outdoor), and patient characteristics, including demographics, handgrip strength, BMI, and the Charlson Comorbidity Index (CCI). The primary analysis investigated the association between these clinical characteristics and the specific circumstances of the fall.

Results

This study encompassed 801 hip fracture patients, with 560 women (70%) and 241 men (30%), and a median age of 81 years. Among these patients, 546 (68.3%) experienced falls during the daytime, while 255 (31.7%) fell at night. Furthermore, 577 (74.9%) hip fractures occurred indoors, with 193 (25.1%) falls taking place outdoors. The analysis revealed that patients with lower handgrip strength and BMI were significantly more susceptible to nighttime falls compared to daytime falls. Furthermore, patients who fell indoors exhibited notably higher CCI scores, along with lower BMI, handgrip strength, and pre-fracture ADL, in comparison to those who fell outdoors.

Conclusions

Our findings indicate that falls during the daytime were more prevalent among geriatric hip fracture patients in Taiwan compared to nighttime falls. Moreover, we observed that more frail patients were relatively susceptible to falling indoors and at night, emphasizing the potential clinical value for clinicians to take proactive measures in fall prevention.

Introduction

Hip fractures rank among the most severe fall-related injuries, often resulting in challenging recovery and a loss of independence for many individuals. ¹ As life expectancy continues to rise globally, the number of elderly individuals is increasing in every geographical region, consequently elevating the likelihood of experiencing a hip fracture. ² Projections indicate that the global incidence of hip fractures will exceed 6 million by 2050, primarily due to the growing life expectancy and aging population worldwide. ³ Notably, more than 95% of these fractures occur as a result of falls, with sideways falls being the primary cause. As individuals age, clinicians are faced with specific concerns related to educating and preventing elderly falls accidents. ⁴

While over 90% of hip fractures stem from falls, only a small percentage (1-2%) of all falls in the elderly result in hip fractures. ⁵ Previous research indicates that falls result from the interplay of a range of risk factors that can be intrinsic or extrinsic in nature. Intrinsic factors pertain to an individual’s physical, demographic, and health status, while extrinsic factors encompass the physical and socio-economic environment.^6,7 Common extrinsic risks encompass tripping hazards, balance and slipping hazards, and vision hazards.^6,8,9 Increasing age exhibits a strong correlation with nearly all intrinsic factors. These intrinsic factors significantly heighten the risk of falls and fall-related injuries. ¹⁰ However, rectifying these intrinsic factors within the dwelling elderly population necessitates long-term interdisciplinary cooperation. Consequently, fall prevention strategies about extrinsic factors should be considered. Extrinsic factors could have different time distribution depending on the daily activities of the patients. As a result, analyzing the fall time distribution in 1 day could be beneficial to fall prevention.^11,12

In the existing literature, there is a limited discussion of extrinsic factors, specifically the timing and location, concerning falls resulting in hip fractures among the elderly. Within the dwelling elderly population, the risk of falling tends to escalate at specific times during the day due to varying daily activities like getting up, toileting, and performing household tasks. Understanding the intrinsic factors contributing to falls at different times and locations is crucial for enhancing fall prevention strategies. Intrinsic factors, such as muscle strength, balance, vision, and cognitive function, often fluctuate based on an individual’s activity level, fatigue, and daily routine. For instance, early morning falls may be linked to muscle stiffness or dizziness upon waking, while evening falls could be influenced by fatigue or poor lighting. By analyzing how these factors interact with specific environments—such as slippery bathroom floors or stairs—healthcare providers can identify vulnerable periods and locations where individuals are most at risk. This targeted insight allows for more effective interventions, such as tailored exercise programs, environmental modifications, or personalized monitoring systems, ultimately reducing the frequency and severity of falls among the elderly.

Given the limited evidence regarding the timing and location characteristics of falls leading to hip fractures among the elderly and their correlation with patients’ clinical attributes, we initiated this cross-sectional study with two primary objectives: firstly, to establish the diurnal distribution of fall-induced hip fractures, and secondly, to compare the attributes of falls during daytime and nighttime, in order to make a meaningful contribution to the development of fall prevention strategies for the elderly.

Methods

Participants

This prospective cross-sectional study was designed and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. We recruited older adults who underwent surgery for hip fractures at a single medical center in Taiwan from March 1, 2018, to December 31, 2022. The qualifying patients were men and women aged ≥60 years who had hip fractures, including intracapsular femoral neck fractures and extracapsular fractures. Patients were excluded if they underwent hip surgery due to conditions other than primary hip fractures, such as osteoarthritis, trauma, tumor metastasis, infection, or avascular necrosis of the femoral head. The study complied with the code of ethics of the World Medical Association (Declaration of Helsinki). Verbal informed consent was obtained from all participants or their legal proxies before enrollment. This method was approved by the ethics committee due to the advanced age and potential frailty of the participants, which could make signing written forms challenging.

Data Collection and Measurements

Data collection was conducted through interviews with the patients themselves or their families during their admission for hip fracture surgery. Information regarding the self-reported cause of the fall was collected during the patient interview using an open-ended question (‘How did your hip fracture happen?’). Responses were subsequently categorized by two independent researchers into predefined groups.

Fall Time and Location: The fall times and locations leading to the index fracture were obtained via these interviews. Patients’ falls were categorized into two groups: daytime falls and nighttime falls.

Daytime falls were defined as falls occurring from 6:00 a.m. to 6:00 p.m., and nighttime falls as falls occurring from 6:00 p.m. to 6:00 a.m.. This specific 12-hour delineation was chosen based on the average sunrise and sunset times in Taiwan, as provided by the Central Weather Bureau. ¹³ This pragmatic division reflects the distinct natural light cycles and generally aligns with the major active and resting periods for the elderly population, which are influenced by circadian rhythms and can affect activity levels and environmental risks. ¹⁴ Fall locations were also categorized as indoors or outdoors. Due to recall difficulties, fall location data was missing for 31 participants. Consequently, all analyses involving fall location were performed on a complete case basis (n = 770).

Assessment of Fall Causes

Information regarding the self-reported cause of the fall was collected during the admission interview from either the patients themselves or their accompanying family members (proxies). An open-ended question, such as “Can you describe how your fall happened?”, was used to elicit a narrative response rather than using a structured questionnaire with pre-set choices. The narrative responses were subsequently reviewed and categorized into predefined, clinically relevant groups (e.g., rising from bed, bathroom-related, vehicle-related) by two independent researchers to ensure consistency. Any discrepancies in categorization were resolved through discussion and consensus.

Patient Characteristics: Basic demographic data, including age, sex, body mass index (BMI), underlying comorbidities, and information about the type of hip fracture (femoral neck fracture or peritrochanteric fracture), were extracted from medical records for analysis. The underlying comorbidities were analyzed using the Charlson Comorbidity Index (CCI). ¹⁵ Maximal handgrip strength (HGS) was measured preoperatively using a Jamar Hydraulic Dynamometer (Sammons Preston, USA). Following the Asian Working Group for Sarcopenia (AWGS) 2019 guidelines, patients were seated with their elbow flexed at 90° and forearm in a neutral position. Assessments were conducted by a single trained examiner. Patients performed three trials for each hand, with the highest recorded value used for analysis. ¹⁶ Low HGS was defined according to AWGS criteria as <28 kg for males and <18 kg for females. ¹⁷ Additionally, the Barthel Index (BI) was used to assess the patient’s pre-fracture independent living ability, ¹⁸ and the EuroQol-5D (EQ-5D) questionnaire was used to evaluate the patient’s pre-fracture quality of life. ¹⁹

Outcome

The main outcome of the study was to determine the temporal distribution of falls that led to hip fractures among elderly patients. This involved categorizing the falls into daytime and nighttime events and assessing the associated clinical characteristics such as handgrip strength, BMI, Charlson Comorbidity Index (CCI), activities of daily living (ADL) using the Barthel Index (BI), and quality of life using the EuroQol-5D (EQ-5D) questionnaire. The secondary outcomes included examining the location of the falls (indoor vs outdoor) and analyzing the characteristics and clinical parameters associated with each group to identify patterns and risk factors.

Data Analysis

Data were analyzed using IBM SPSS Statistics version 25.0. Descriptive statistics were presented as frequencies (n), proportions (%), means, and standard deviations. The normality of distribution for all continuous variables was assessed by visual inspection of histograms and Q-Q plots. As the variables were found to be approximately normally distributed, they are presented as mean ± standard deviation (SD). Independent sample t-tests were used for continuous variables, and Chi-square tests were used for categorical variables. Categorical variables were expressed as frequencies and percentages. To identify independent predictors of fall location, variables with a P-value <0.1 in the univariate analysis were entered into a multivariate logistic regression model using a backward stepwise selection method. The final model was adjusted for age, BMI, handgrip strength, Barthel Index, and CCI. While a clear recall of the fall timing was required for study inclusion, a subset of participants was unable to provide a detailed account of the specific cause of their fall. Therefore, for the analyses pertaining to the causes of falls, only participants with complete data for this variable were included. Cases with missing information on the cause of the fall were excluded from this specific part of the analysis.

Results

Eight hundred one eligible patients with hip fracture were enrolled, including 241 (30%) men and 560 (70%) women. The baseline demographic and clinical characteristics stratified by sex are presented in Table 1. The mean age of the cohort was 81 ± 9.5 years, with no significant difference between sexes (P = 0.96). While most baseline characteristics such as BMI, pre-fracture functional status (Barthel Index), and comorbidity scores were similar between men and women, two key differences were observed.

Table 1.

Baseline Demographic Characteristics of the Study Population

Clinical characteristics	Mean ± SD/Number (Percentage) (Male, n = 241)	Mean ± SD/Number (Percentage) (Female, n = 560)	P-value
Age	81 ± 10.0	81 ± 9.2	0.96
Body mass index	22.3 ± 3.6	22.4 ± 4.0	0.6
Hand grip strength	18.0 ± 10.1	9.6 ± 5.6	<0.001
Barthel index	88.6 ± 19.8	86.5 ± 20.9	0.18
EQ5D	0.87 ± 0.19	0.87 ± 0.18	0.95
Charlson comorbidity index	1.76 ± 1.73	1.58 ± 1.6	0.16
Type of fracture			0.009
Femoral neck fracture	121 (50%)	330 (59%)
Peritrochanteric fracture	120 (50%)	230 (41%)
Time of fall			0.83
Fall in day	164 (68%)	381 (68%)
Fall at night	77 (32%)	179 (32%)
Place of fall(n = 232)			0.074
Indoors	165 (71%)	414 (67%)
Outdoors	67 (29%)	124 (23%)

Abbreviations: SD, standard deviation. P-values were derived from two-sided tests. The independent sample t-test was used for continuous variables (Age, BMI, Handgrip strength, etc.), and the Chi-square test was used for categorical variables (Sex, Place of fall, etc.).

First, a profound and statistically significant difference was found in handgrip strength, with males having a substantially higher mean handgrip strength compared to females (18.0 ± 10.1 kg vs 9.6 ± 5.6 kg, P < 0.001). Second, the distribution of fracture types also differed significantly between sexes (P = 0.009), with a higher proportion of femoral neck fractures observed in females compared to males (59% vs 50%).

Overall, the majority of patients fell during the daytime (68.3%) and indoors (72.1%), and there was no significant difference in the distribution of fall timing or location between sexes (Table 1).

Fall Time Distribution

Figure 1 shows the distribution of fall events across the entire day. The highest peak of fall events occurred between 9:00 and 10:00. The second highest peak was observed between 8:00 and 9:00, followed by a third peak between 12:00 and 13:00. All three of these peaks occurred during the daytime. Additionally, the highest peak during the night was observed between 18:00 and 19:00.

Figure 1.

The total fall induced hip fracture time distribution

Figure 2 shows the time distribution of indoor and outdoor fall events. The peaks of indoor fall events are the same as the total fall events, occurring between 9:00 and 10:00, 8:00 and 9:00, and 12:00 and 13:00. The peaks of outdoor fall events are observed between 9:00 and 10:00, 12:00 and 13:00, and 10:00 and 11:00.

Figure 2.

The indoor and outdoor fall induced hip fracture time distribution

Differences in Characteristics of Daytime and Nighttime Hip Fractures

Table 2 shows the characteristics of the included patients grouped by falls during the day or at night. Among the hip fracture patients who fell in day, only the hand grip strength was significant higher (12.52 vs 11.22, P value = 0.036) compared to hip fracture patients who fell at night. There was no statistically significant difference in BMI between daytime and nighttime fallers (22.6 vs 22.0, P = 0.051). The ratio of falls at day in the outdoor group is significantly higher than the ratio in the indoor group.

Table 2.

Comparison of Patient Characteristics by Fall Timing (Day vs Night)

Variable	Falls in day (n = 546)	Falls at night (n = 255)	P Value
Age, mean (SD)	80.91 ± 9.517	81.34 ± 9.321	0.594
Sex, no. (%)
Male	163 (30.0)	78 (30.6)	0.833
Female	383 (70.0)	177 (69.4)	0.833
BMI(SD)	22.6 ± 3.91	22.0 ± 3.86	0.051
Hand grip strength (SD)	12.52 ± 8.33	11.22 ± 7.81	0.036
Barthel index (SD)	87.26 ± 20.73	87.08 ± 20.2	0.901
EQ5D (SD)	0.878 ± 0.174	0.85 ± 0.2	0.159
Charlson comorbidity index (SD)	1.63 ± 1.57	1.6 ± 1.77	0.776
Type of fracture, no. (%)
Femoral neck fracture	316 (57.9)	135 (53.1)	0.158
Peritrochanteric fracture	230 (42.1)	120 (46.9)	0.158
Place of fall, no. (%)
indoors	361 (68.9)	216 (87.8)	<0.001
outdoors	163 (31.1)	30 (12.2)

Abbreviations: SD, standard deviation. P-values were derived from two-sided tests. The independent sample t-test was used for continuous variables (Age, BMI, Handgrip strength, etc.), and the Chi-square test was used for categorical variables (Sex, Place of fall, etc.).

Differences in Characteristics of Hip Fractures Indoors vs Outdoors

Table 3 shows the characteristics of the included patients grouped by falls indoors or outdoors. Among the hip fracture patients who fell indoors, there were significant differences in age (82.3 vs 77.5), BMI (22.2 vs 23.0), handgrip strength (7.4 vs 9.5), BI (84.1 vs 96.7), as well as higher scores CCI (1.75 vs 1.28) compared to hip fracture patients who fell outdoors. However, in multiple variance analysis using logistic regression (Table4.), only the BMI (odds ratio(OR) = 1.057; 95%CI = 1.003, 1.114; P = 0.038) and hand grip strength (odds ratio(OR) = 1.033; 95%CI = 1.006, 1.06; P = 0.014) were significant predictors of hip fractures indoors or outdoors. The ratio of falls at day in the outdoor group is significantly higher than the ratio in the indoor group.

Table 3.

Characteristics of Patients Fall Indoors and Outdoors

Variable	Fall indoors (n = 577)	Fall outdoors (n = 193)	P Value
Age, mean (SD)	82.29 ± 8.984	77.53 ± 10.038	<0.001
Sex, no. (%)
Male	164 (28.4)	68 (35.2)	0.074
Female	413 (71.6)	125(64.8)	0.074
BMI(SD)	22.2 ± 3.92	23.0 ± 3.88	0.012
Handgrip strength, mean (SD)	10.83 ± 7.36	16.01 ± 9.54	<0.001
Barthel index, mean (SD)	84.12 ± 22.18	96.74 ± 10.1	<0.001
EQ5D (SD)	0.868 ± 0.18	0.872 ± 0.19	0.717
Charlson comorbidity index (SD)	1.74 ± 1.683	1.3 ± 1.366	<0.001
Type of fracture, no. (%)
Femoral neck fracture	321 (55.6)	110 (57)	0.731
Peritrochanteric fracture	256 (44.4)	83 (43)	0.731
Time of fall, no. (%)
Day	361 (62.5)	163 (84.5)	<0.001
Night	216 (37.5)	30 (15.5)

Abbreviations: SD, standard deviation. P-values were derived from two-sided tests. The independent sample t-test was used for continuous variables (Age, BMI, Handgrip strength, etc.), and the Chi-square test was used for categorical variables (Sex, Place of fall, etc.).

Table 4.

Logistic Regression Model on Patients Fall Indoors and Outdoors

	OR	LLCI	ULCI	P-value
Age	0.974	0.955	0.994	0.009
BMI	1.029	0.982	1.078	0.227
Hand grip strength	1.036	1.014	1.059	0.002
Barthel index	1.049	1.028	1.069	<0.001
CCI	0.919	0.812	1.040	0.180

Abbreviations: OR, Odds Ratio; LLCI, Lower Limit Confidence Interval; ULCI, Upper Limit Confidence Interval.

Independent Predictors of Fall Location

The multivariate logistic regression analysis identified two significant independent predictors of fall location.

The multivariate logistic regression analysis identified three significant independent predictors of fall location after adjusting for potential confounders (Table 4). Age was a significant predictor, with each increasing year associated with a 2.6% lower odds of falling indoors (OR = 0.974; 95% CI = 0.955-0.994; P = 0.009), indicating that older patients were more likely to fall outdoors.

Conversely, both handgrip strength (OR = 1.036; 95% CI = 1.014-1.059; P = 0.002) and pre-fracture Barthel Index (OR = 1.049; 95% CI = 1.028-1.069; P < 0.001) were significant predictors of indoor falls. For every one-unit increase in handgrip strength and Barthel Index score, the odds of an indoor fall increased by 3.6% and 4.9%, respectively. This suggests that patients who were physically stronger and more functionally independent were more likely to sustain their fractures indoors. Body Mass Index (BMI) was not a significant predictor in the final adjusted model (P = 0.227).

Causes of Falls in Hip Fracture Patients

Among the patients who fell indoors, 394 patients can clearly documented the cause of their falls. The most common cause of indoor falls during the daytime was moving (18.4%). The most common cause of indoor falls at night was related to the bathroom (34.3%). The next common cause in both groups was getting up in bed (Table 5).

Table 5.

Indoors Fall Causes

Indoors, No (%)	Day	Night
Rising from bed	39 (16.3)	32 (20.3)
Bathroom related	35 (14.6)	54 (34.3)
Moving on flat ground	44 (18.4)	22 (14)
Rising from a chair	17 (7.1)	9 (5.7)
Housework	16 (6.6)	0 (0)
Stairs related	8 (3.3)	5 (3.1)
Stumbling	20 (8.4)	9 (5.7)
Slippery floor	15 (6.2)	3 (1.9)
Willing to get something	12 (5)	12 (7.6)
Dizziness	17 (7.1)	4 (2.5)
Other reasons	16 (6.7)	7 (4.5)

Among the patients who fell outdoors, 219 patients can clearly documented the cause of their falls. The most common cause of falls in both the day and night groups was related to vehicles. The next common cause in both groups was falling while walking (Table 6).

Table 6.

Outdoors Fall Causes

Outdoors, No (%)	Day	Night
Vehicles related	54 (37)	10 (34.5)
Fall during walking	22 (15.1)	8 (27.6)
Stumbling	9 (6.2)	3 (10.3)
Physical activity	7 (4.8)	2 (6.9)
Bus related	11 (7.5)	0 (0)
Slippery floor	11 (7.5)	1 (3.4)
Stair related	9 (6.2)	0
Work	7 (4.8)	0
Other reasons	16 (11)	5 (17.2)

Discussion

This study reveals critical insights into the circumstances surrounding hip fractures in older adults, highlighting distinct patterns based on the time and location of the fall. We identified that the majority of fractures resulted from indoor, daytime falls, with a significant peak in the morning between 8:00 and 10:00 a.m.. Our findings delineate two divergent patient profiles: healthier individuals who fell during the day, and more frail individuals who fell indoors or at night. These results challenge a one-size-fits-all approach to fall prevention, advocating instead for stratified strategies tailored to specific risk profiles and environmental contexts.

The Dichotomy of Fall Timing: Activity Exposure versus Intrinsic Frailty

Contrary to the common assumption that darkness makes nighttime the highest-risk period, our data show that over two-thirds of hip fractures occurred during the daytime.^20-22 This finding is consistent with literature that identifies “activity exposure” as a dominant risk factor. The morning peak corresponds directly with a period of increased activity after waking, including transfers from bed and toileting. ²³ Such movements are well-known triggers for orthostatic hypotension, a major contributor to falls in the elderly. The fact that daytime fallers in our cohort had significantly higher handgrip strength suggests that while they were physically more robust, their increased activity levels amplified their exposure to fall opportunities.

Conversely, nighttime fallers presented as a more vulnerable group, characterized by lower handgrip strength and BMI. This profile strongly suggests underlying frailty and sarcopenia, which are established independent risk factors for falls and fractures.^24,25 The finding that the most common cause of indoor nighttime falls was bathroom-related aligns with extensive research on nocturia as a key precipitant of falls. ²⁶ This creates a high-risk scenario where a frail individual with sarcopenia and nocturia navigates a poorly lit path to the bathroom, leading to a devastating fall. It is crucial to interpret the findings from our multivariate model with nuance. While factors like age, handgrip strength, and functional independence were statistically significant predictors, their odds ratios indicate modest effect sizes. For instance, a higher Barthel Index was associated with increased odds of an indoor fall. This may seem counterintuitive, but it could suggest that more independent individuals are also more active within their homes, increasing their exposure to indoor hazards. These factors should not be viewed as powerful standalone predictors, but rather as components of a larger, multifactorial risk profile. This underscores that a comprehensive fall risk assessment is essential for effective prevention strategies.

The Environmental Context: the Frail-at-Home Versus the Healthy-In-Hazard

Our analysis draws a sharp distinction between patients who fell indoors vs outdoors. While the univariate analysis showed that indoor fallers were significantly older and more frail across multiple domains, our multivariate model provides a more refined and nuanced picture. After adjusting for other factors, age was found to be an independent predictor for outdoor falls (OR = 0.974), whereas greater handgrip strength and higher pre-fracture functional independence (Barthel Index) were paradoxically associated with indoor falls.

This seemingly counterintuitive finding for age highlights the powerful confounding effect of frailty. Once the model accounts for the stronger predictors of indoor falls—namely, lower muscle strength and poorer functional status—the independent effect of age suggests that among individuals with a similar level of function, the older ones may be those who remain active enough to venture outdoors, thereby increasing their exposure to external hazards. This reinforces our concept of the “Healthy-in-Hazard” faller, who is comparatively robust but encounters significant environmental risks in the community. This profile supports the concept of a “fear of falling” cycle, where frail individuals may limit their outdoor activities, leading to further deconditioning and an increased risk of falling even in a familiar home environment. For this population, multifactorial interventions, including comprehensive home safety assessments and modifications, are paramount, as recommended by major clinical guidelines like the CDC’s STEADI (Stopping Elderly Accidents, Deaths & Injuries) initiative. ²⁷ In contrast, patients who sustained fractures outdoors were comparatively healthier but faced more significant external threats. The leading cause of outdoor falls was vehicle-related incidents. This is a critical public health issue, particularly in Taiwan where high motorcycle density creates a uniquely hazardous environment for older pedestrians. ²⁸ This finding underscores that fall prevention must extend beyond individual clinical care into the realm of public health and urban planning to create age-friendly, safe communities.

Based on the distinct patient profiles identified, a multi-layered approach to fall prevention is warranted, moving beyond generic advice to targeted, evidence-based strategies. The pronounced morning peak in falls, particularly between 8:00 and 10:00 a.m., suggests a universal intervention focused on mitigating risks associated with morning routines. Educating all older adults on behavioral techniques to minimize orthostatic hypotension, such as rising slowly from a lying or sitting position, is a crucial first step.

For the frail population, who are more susceptible to indoor and nighttime falls, a more intensive, multifactorial approach is essential. Our findings show these individuals often have lower handgrip strength and poorer functional status, underscoring the need for programs aimed at improving physical resilience. Key components of this targeted strategy should include strength and balance training to combat sarcopenia, a condition linked to nighttime falls, and professional assessment for home environment modifications. Installing assistive devices like grab bars, improving ambient lighting with night lights, and ensuring non-slip surfaces can directly address the hazards associated with nighttime bathroom use, a primary cause of falls in this group.^29,30 Concurrently, a thorough review of medications and specific management of conditions like nocturia are vital parts of this comprehensive plan.

Finally, for the comparatively healthier and more active older adults who primarily fall outdoors, prevention efforts must extend into the community. Given that vehicle-related incidents were a leading cause of outdoor falls in our study, individual resilience is often insufficient against significant external threats. Therefore, advocating for public health policies that improve pedestrian safety, such as creating safer walking paths and implementing traffic-calming measures, is a necessary third layer of a truly comprehensive fall prevention strategy.

Limitation

This study has several limitations that should be considered in the interpretation of the findings. First and foremost, the study is subject to significant recall bias. The causes and precise circumstances of falls were determined through retrospective interviews with patients or their families after a traumatic event. The recall of patients, who may also have had cognitive impairments, could be imprecise, potentially impacting the accuracy of the categorized fall etiologies.

Second, daytime was operationally defined as 06:00-18:00 rather than using precise daily sunrise and sunset times. While this simplification does not fully capture seasonal variation, it was adopted as a pragmatic definition for analytic consistency. Taiwan (latitude 22-25°N) has relatively small seasonal variation in daylight. Based on data from the Central Weather Administration and astronomical databases, sunrise ranges from about 05:00 to 06:45 and sunset from 17:10 to 19:05. On average, this approximates to ∼12 hours of daylight, with sunrise and sunset occurring close to 06:00 and 18:00, respectively. ¹³

Third, as a single-center study, the findings may not be fully generalizable to other populations or healthcare systems. The study design also presented challenges in establishing a control group; it was not feasible to recruit a comparative group of elderly individuals from the community who had experienced falls without resulting in a fracture. Consequently, this study was unable to compare the characteristics of falls that lead to fractures vs those that do not.

Fourth, including hip fractures resulting from traffic accidents introduces heterogeneity into the mechanism of injury. Although most of these were low-energy incidents, the study could not definitively distinguish whether the fracture was caused by the fall itself or by a direct crush injury. Our inclusion criterion was the presence of a hip fracture rather than a strict medical definition of a fall, which may have resulted in the inclusion of injury mechanisms that differ from the commonly understood concept of a fall.

Finally, our analysis did not account for several important unmeasured confounders, such as specific medication use (e.g., psychotropics, antihypertensives), visual acuity, or detailed environmental hazards within the home. Furthermore, a formal a priori sample size calculation was not performed, as the study was based on a consecutive cohort of all eligible patients over a defined period.

Conclusion

In conclusion, this study demonstrates that fall-induced hip fractures in the geriatric population are more frequent during the daytime, with a distinct peak between 8:00 a.m. and 10:00 a.m. Our findings reveal divergent clinical profiles based on the circumstances of the fall. Patients who fell during the day exhibited stronger handgrip strength and a lower susceptibility to indoor falls compared to those who fell at night. Furthermore, when distinguishing by location, a multivariate analysis revealed that older age was a significant predictor of outdoor falls, while greater handgrip strength and higher pre-fracture functional independence were significant predictors of indoor falls.. The etiology of falls also varied by context; indoor falls were most often related to movement during the day and bathroom use at night, while outdoor falls were predominantly attributed to vehicle-related incidents. These insights underscore the importance of developing targeted fall prevention strategies that account for specific high-risk times, locations, and the distinct characteristics of different patient subgroups within the community-dwelling elderly population.

ORCID iD

Yu-Pin Chen https://orcid.org/0000-0002-9729-6375

Ethical Considerations

This study complied with the Declaration of Helsinki and ethical guidelines for medical research involving human participants. The study was reviewed and approved by Ethics Committee of Taipei Medical University (TMU-JIRB N201709053).

Consent to Participate

Informed consent was obtained from all participants.

Consent for Publication

Informed consent was obtained from all individual participants included in the study.

Data Availability Statement

The datasets used and analyzed in the current study are available from the corresponding authors upon reasonable request. *