Relationship Between Asking an Older Adult About Falls and Surgical Outcomes

Importance  More than one-third of all US inpatient operations are performed on patients aged 65 years and older. Existing preoperative risk assessment strategies are not adequate to meet the needs of the aging population.

Objectives  To evaluate the relationship of a history of falls (a geriatric syndrome) to postoperative outcomes in older adults undergoing major elective operations.

Design, Setting, and Participants  This prospective, cohort study was conducted at a referral medical center. Persons aged 65 years and older undergoing elective colorectal and cardiac operations were enrolled. The predictor variable was having fallen in the 6 months prior to the operation.

Main Outcomes and Measures  Postoperative outcomes measured included 30-day complications, the need for discharge institutionalization, and 30-day readmission.

Results  There were 235 subjects with a mean (SD) age of 74 (6) years. Preoperative falls occurred in 33%. One or more postoperative complications occurred more frequently in the group with prior falls compared with the nonfallers following both colorectal (59% vs 25%; P = .004) and cardiac (39% vs 15%; P = .002) operations. These findings were independent of advancing chronologic age. The need for discharge to an institutional care facility occurred more frequently in the group that had fallen in comparison with the nonfallers in both the colorectal (52% vs 6%; P < .001) and cardiac (62% vs 32%; P = .001) groups. Similarly, 30-day readmission was higher in the group with prior falls following both colorectal (P = .04) and cardiac (P = .02) operations.

Conclusions and Relevance  A history of 1 or more falls in the 6 months prior to an operation forecasts increased postoperative complications, the need for discharge institutionalization, and 30-day readmission across surgical specialties. Using a history of prior falls in preoperative risk assessment for an older adult represents a shift from current preoperative assessment strategies.

More than one-third of all inpatient operations in the United States are performed on patients aged 65 years and older,¹ a proportion which will increase during the next several decades.² Existing preoperative risk assessment strategies are not adequate to meet the needs of the aging population. Current tactics either quantify risk of a single organ system (eg, the American Heart Association cardiovascular risk assessment³) instead of the whole patient, or they sum chronic disease burden (eg, cumulative illness rating scale⁴) as the measure of risk rather than quantifying global reduced physiologic reserve of the older adult termed frailty.

Falling represents 1 of the 5 core geriatric syndromes, which reflect reduced physiologic reserve unique to the older adult.^5,6 A geriatric syndrome is a “multifactorial health condition that occur[s] when the accumulated impairments in multiple systems render [older] persons vulnerable to situational challenges.”⁵ In short, geriatric syndromes are clinical symptoms that represent the frail older adult.⁶ In community-dwelling older adults, the presence of a geriatric syndrome is closely linked to the development of functional dependence.⁵ While there is some data evaluating the relationship of these symptoms to outcomes in surgical^7,8 and medical⁹ hospitalized older adults, to our knowledge, no literature directly addresses the relationship of a history of prior falls to postoperative outcomes.

The purpose of this study was to evaluate the relationship of a history of falls to surgical outcomes in older adults undergoing major elective colorectal and cardiac operations. The specific aims were to compare outcomes of patients with and without a fall within the 6 months prior to their operation including 30-day morbidity, the need for discharge to an institutional care facility, and 30-day readmission.

This prospective cohort study was performed at the Denver Veterans Affairs Medical Center. Regulatory approval was obtained through the Colorado Multiple Institutional Review Board (08-1071) and written informed consent was obtained from patients. Participants were enrolled between January 6, 2006, and October 20, 2010. Inclusion criteria were age 65 years and older and having underwent an elective colorectal or cardiac operation. Exclusion criteria for both groups were emergent (defined as an operation within 12 hours of admission) and urgent (defined as an operation between 12 and 72 hours following admission) operations. Additional exclusion criterion for the colorectal group was the performance of an additional procedure in combination with the segmental colectomy (eg, liver resection and exenteration).

A fall was defined as unintentionally coming to rest on the ground, floor, or other lower level.¹⁰ Patients were considered to have had a fall if they had a history of 1 or more falls in the 6 months preceding surgery. A history of falls was recorded preoperatively. In addition to the fall history, other routine preoperative and intraoperative variables were recorded.

Postoperative complications were defined using the following Veterans Affairs Surgery Quality Improvement Program definitions that were recorded prospectively by the research team: cardiac (cardiac arrest requiring cardiopulmonary resuscitation or myocardial infarction); respiratory (pneumonia, pulmonary embolism or reintubation for respiratory/cardiac failure); renal insufficiency; neurologic (cerebral vascular accident/stroke or coma >24 hours); postoperative infection (deep wound surgical site infection, superficial surgical site infection, or urinary tract infection); sepsis; deep vein thrombosis; and reoperation (return to operating room). Institutionalization was defined as discharge to an institutional care facility (eg, not home). If a patient resided in an institutional care facility preoperatively, they were not considered newly institutionalized postoperatively.

Statistical analysis was performed using bivariate comparisons for the presence or absence of a fall within 6 months of the operation (the independent variable). Preoperative, intraoperative, and postoperative variables were analyzed using χ² tests for categorical variables and the nonparametric Wilcoxon rank-sum test for continuous variables. Logistic regression was performed, with the dependent variable of the occurrence of 1 or more complications and both prior falls and age as predictor variables to determine whether falls were related to the occurrence of 1 or more complications independent of advancing age. The variable age was tested separately as a continuous variable and a categorical variable (65-69 years, 70-74 years, 75-79 years, and ≥80 years) in bivariable and multivariable logistic regression models for both the colorectal and cardiac groups to look at age as a single predictor and to adjust for age in the relationship of having fallen in relation to the occurrence of a postoperative complication. The nonparametric Spearman correlation test was used to determine how the number of falls related to the number of complications. Finally, the ability of a fall history to forecast 1 or more postoperative complications was compared with established variables used to predict risk (Charlson Comorbidity Index score, American Society of Anesthesiologists [ASA] score, and chronologic age). To accomplish this comparison, a multivariable logistic regression was used in which the dependent variable was the occurrence of 1 or more complications and the independent variables were fall history and dichotomized Charlson and ASA scores for which 3 or greater indicated the high-risk category and age (used as a continuous variable). Separate analyses were performed for the colorectal and cardiac groups.

A total of 235 patients were included in the study (81 colorectal and 154 cardiac) (Figure 1). The mean (SD) age of patients was 74 (6) years and 98% (231 of 235) were male. One or more falls occurred in 33% of patients (78 of 235) in the 6 months prior to the operation. One or more complications occurred in 28% (65 of 235). Postoperative inpatient mortality occurred in 2% (5 of 235). Baseline characteristics of fallers and nonfallers were compared. Among patients who underwent colorectal operations, baseline characteristics different in the faller’s group were older age, higher ASA class, lower albumin, higher creatinine, lower hematocrit, and higher Charlson score (Table 1). Among patients who underwent cardiac operations, baseline characteristics different in the faller’s group were older age, lower albumin, lower hematocrit, and higher Charlson score (Table 2).

Operative characteristics for patients who underwent colorectal and cardiac operations were compared. The type of operation, operative time, blood loss, and transfusion requirements were similar for fallers and nonfallers in both the colorectal (Table 1) and cardiac (Table 2) groups. In the colorectal group, disease state was similar in the faller and nonfaller groups (Table 1).

Postoperative characteristics were compared between the fallers and nonfallers. Preoperative falls were associated with higher incidence of the occurrence of 1 or more postoperative complications and higher rates of discharge institutionalization in both the colorectal and cardiac groups (Table 3). Thirty-day readmission rates were higher in patients with a history of falls following both colorectal (P = .04) and cardiac (P = .02) operations (Table 3).

In the bivariable analysis, increasing age (where age was used as a continuous variable) was found to not be associated with the occurrence of 1 or more postoperative complications in both the colorectal (P = .51) and cardiac (P = .06) groups. Logistic regression was performed to adjust for the effect of advancing age on the relationship of the occurrence of 1 or more postoperative complications and having fallen in the 6 months prior to the operation. In the colorectal group, when age was used as a continuous variable, a history of falls remained associated with the occurrence of 1 or more complications (odds ratio, 7.380; 95% CI, 1.994-27.311; P = .003) and advancing age remained nonsignificant (P = .18). In the cardiac group, when age was used as a continuous variable, a history of falls remained associated with the occurrence of 1 or more complications (odds ratio, 3.095; 95% CI, 1.361-7.042; P = .007) and advancing age remained nonsignificant (P = .33). Because the age of those who had fallen was an average of 9 years older than the nonfallers in the colorectal group, additional statistical analysis was performed where age was used as a categorical variable (65-69 years, 70-74 years, 75-79 years, and the reference group was ≥80 years) in the logistic regression model rather than as a continuous variable (results reported here). In the bivariable analysis, increasing age was again not found to be associated with the occurrence of 1 or more postoperative complications in both the colorectal (P = .98) and cardiac (P = .23) groups. When adjusting for age, a history of falls remained associated with the outcome (odds ratio, 10.214; 95% CI, 2.401-43.455; P = .002) and advancing age remained nonsignificant in all groups (P = .32). Adjusting for age in the cardiac group, a history of falls remained associated with the occurrence of 1 or more complications (odds ratio, 3.372; 95% CI, 1.481-7.678; P = .004) and advancing age remained nonsignificant in all groups. The correlation between the number of fall events in the 6 months prior to the operation and the number of complications was determined. There was a positive correlation between the number of prior falls and the number of complications in both the colorectal (Spearman rho = 0.411; P < .001) and cardiac (Spearman rho = 0.260; P = .001) groups (Figure 2).

The ability of a fall history to forecast 1 or more postoperative complications was compared with currently used methods to assess a patient’s risk including the Charlson score, the ASA score, and chronologic age (Table 4).

The current study examined the association of a history of falling to postoperative adverse outcomes in individuals aged 65 years and older who underwent elective colorectal and cardiac operations. The main result was that having fallen in the 6 months prior to an operation was related to the occurrence of 1 or more postoperative complications, regardless of what procedure was performed. This finding is independent of advancing age in both groups. Having fallen was also associated with increased 30-day readmission and the need for discharge to an institutional care facility. In addition, a positive correlation between an increased number of falls and an increased number of complications existed for both the colorectal and cardiac groups. Finally, having fallen in the 6 months prior to an operation was compared with established methods to define a patient’s risk. A history of falls was comparable with the Charlson score, and it was favorable to both the ASA score and chronologic age at forecasting 1 or more postoperative complications.

The concept of a common group of symptoms reflecting the breakdown of frail older adults was first described in Isaacs’ The Giants of Geriatrics¹¹: (1) incontinence, (2) immobility, (3) instability (falls), and (4) intellectual impairment. Gerontologists now recognize 5 geriatric syndromes (pressure ulcers, incontinence, falls, functional decline, and delirium) to represent symptoms that clinically unify a core geriatric concept.⁶ While these 5 clinical presentations are disparate, they all result from an accumulation of subtle impairments across multiple systems. For example, the cause of a fall in an older adult is not owing to a single cause; instead, it results from the interaction of subclinical deficiencies in multiple systems (eg, decreased proprioception, reduced neuromuscular response, slowed mobility, and weakened skeletal muscles). From the standpoint of the clinician, the presence of a geriatric syndrome reflects a frail individual with reduced physiologic reserves and its expression may be instigated by acute illness. An example of a geriatric syndrome encountered in the perioperative setting is when an older adult with appendicitis presents with delirium and not right lower quadrant pain.

Geriatric syndromes have been directly linked to adverse outcomes in hospitalized medical patients. Campbell and colleagues¹² performed a multicenter trial studying patients aged 65 years and older who were admitted to the hospital nonelectively for medical treatment and found that the presence of the geriatric giants (problems with falling, mobility, continence, or cognition) prior to admission was at least as important as the presenting diagnosis at forecasting the need for discharge institutionalization. Subsequently, Anpalahan and Gibson⁹ found in 110 medical patients aged 75 years and older that the presence of a geriatric syndrome (impaired cognition, having fallen, impaired mobility, dependence in an activity of daily living, or urinary incontinence) was associated with the occurrence of adverse outcomes (longer length of stay, institutionalization, unplanned 3-month readmission, and increased 3-month mortality). In addition to the geriatric giants, the geriatric syndrome of pressure ulcers has also been related to adverse hospitalized outcomes.¹³

There is a paucity of data relating preexisting geriatric syndromes to surgical outcomes. To our knowledge, no study in the literature directly relates geriatric syndromes to surgical outcomes. However, multiple single-center studies have examined frailty and adverse postoperative outcomes. To quantify frailty, these studies sum the number of frailty characteristics (some of which are geriatric syndromes) present in an older adult prior to an operation. In these studies, the geriatric syndromes of impaired cognition,^7,8,14-16 poor mobility,^7,8,16-18 incontinence,¹⁴ functional decline,^7,8,14,15 and having fallen^7,8 were abnormal characteristics used in sum to define frailty, which has been closely related to adverse postoperative outcomes.

The importance of the current study is 2-fold. First, this study directly implicated the presence of a geriatric syndrome, falling, to adverse postoperative outcomes; a finding not altered by the operation performed. Using the presence of geriatric syndromes to forecast adverse postoperative outcomes instead of chronic disease burden or single end-organ dysfunction is a departure from current strategies. Second, a history of falls has the potential to be incorporated into surgical risk calculators (the method which most likely represents the future of preoperative risk assessment). The key difference between a fall history and other characteristics that define the frail older adult (eg, gait speed) is that a history of falls, routinely documented by nursing assessments for inpatient fall risk, can be collected retrospectively from the medical record. This allows such a variable to be recorded in retrospective surgical data sets (eg, the National Surgical Quality Improvement Program database) from which preoperative risk calculators are developed. The addition of variables specific to geriatric physiologic vulnerability would allow these risk calculators to move beyond quantifying surgical risk in older adults using chronic diseases (eg, hypertension) and single end-organ dysfunction (eg, end-stage renal disease) to using frailty characteristics, which more appropriately quantify physiologic vulnerability of the older adult. A schematic of additional geriatric-specific variables that can be recorded retrospectively, thereby being amenable to inclusion in retrospective surgical outcome databases, can be found in Figure 3.

There were 2 main limitations of this study. First, these data show efficacy and not effectiveness of preoperative falls at forecasting adverse postoperative outcomes. In other words, data on falls collected by a research team focused on quantifying the frail older adult (efficacy) may or may not be similar to falls data retrospectively collected out of the nursing notes in the clinical medical record (effectiveness). Second, most patients in this study were male; a fact that does not allow a sex bias of the relationship of falls to postoperative outcomes to be detected. The sex distribution of our study reflects the sex distribution of a Veterans Affairs medical center and not selection bias.

Given the high volume of surgical care provided for the elderly population, improving preoperative risk assessment for the older adult is becoming increasingly important. Incorporating geriatric-specific variables that reflect physiologic vulnerability of the older adult into large surgical outcomes data sets used to construct preoperative risk calculators has real potential to improve the accuracy of these tools at forecasting risk in older adults. A history of falls has the unique property of both being a marker of physiologic frailty and being amenable to retrospective data collection; facts that make a history of falls an ideal candidate for incorporation into large surgical outcomes data sets. Future directions include assessing the relationship of other geriatric-specific variables amenable to retrospective medical record review to surgical outcomes (Figure 3) and to begin a geriatric specialty collaborative within the National Surgical Quality Improvement Program to test the effectiveness of geriatric-specific variables at improving the accuracy of surgical risk calculators for older adults.

Corresponding Author: Thomas N. Robinson, MD, MS, University of Colorado School of Medicine, MS C313, 12631 E 17th Ave, Aurora, CO 80045 (thomas.robinson@ucdenver.edu).

Accepted for Publication: March 12, 2013.

Published Online: October 9, 2013. doi:10.1001/jamasurg.2013.2741.

Author Contributions: Ms Dunn and Dr Robinson had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Jones, Wu, Robinson.

Acquisition of data: Jones, Dunn, Wu, Robinson.

Analysis and interpretation of data: Jones, Dunn, Kile, Cleveland, Robinson.

Drafting of the manuscript: Jones, Dunn, Robinson.

Critical revision of the manuscript for important intellectual content: Jones, Wu, Kile, Cleveland, Robinson.

Statistical analysis: Jones, Wu, Kile, Robinson.

Obtained funding: Robinson.

Administrative, technical, or material support: Dunn, Robinson.

Study supervision: Cleveland, Robinson.

Conflict of Interest Disclosures: None reported.

Previous Presentation: This paper was presented in part at the Annual Clinical Congress of the American College of Surgeons; October 1, 2012; Chicago, IL.