Older adults are at increased risk for adverse events after surgical procedures. Loss of independence (LOI), defined as a decline in function or mobility, increased care needs at home, or discharge to a nonhome destination, is an important patient-centered outcome measure.
To evaluate LOI among older adult patients after surgical procedures and examine the association of LOI with readmission and death after discharge in this population.
Design, Setting, and Participants
This retrospective cohort study examined 9972 patients 65 years and older with known baseline function, mobility, and living situation undergoing inpatient operations from January 2014 to December 2014 at 26 hospitals participating in the American College of Surgeons National Surgical Quality Improvement Program Geriatric Surgery Pilot Project. A total of 4895 patients were excluded because they were totally dependent, classified as class 5 by the American Society of Anesthesiologists, undergoing orthopedic or spinal procedures, or died prior to discharge.
Loss of independence at time of discharge.
Main Outcomes and Measures
Readmission and death after discharge.
Of the 5077 patients included in this study, 2736 (53.9%) were female and 3876 (76.3%) were white, with a mean (SD) age of 75 (7) years. For this cohort, LOI increased with age; LOI occurred in 1386 of 2780 patients (49.9%) aged 65 to 74 years, 1162 of 1726 (67.3%) aged 75 to 84 years, and 479 of 571 (83.9%) 85 years and older (P < .001). Readmission occurred in 517 patients (10.2%). In a risk-adjusted model, LOI was strongly associated with readmission (odds ratio, 1.7; 95% CI, 1.4-2.2) and postoperative complication (odds ratio, 6.7; 95% CI, 4.9-9.0). Death after discharge occurred in 69 patients (1.4%). After risk adjustment, LOI was the strongest factor associated with death after discharge (odds ratio, 6.7; 95% CI, 2.4-19.3). Postoperative complication was not significantly associated with death after discharge.
Conclusions and Relevance
Loss of independence, a patient-centered outcome, was associated with postoperative readmissions and death after discharge. Loss of independence can feasibly be collected across multiple hospitals in a national registry. Clinical initiatives to minimize LOI will be important for improving surgical care for older adults.
With the aging of the population, an increased demand on surgical services can be anticipated. According to the US Census Bureau, people 65 years and older constituted 13.0% of the population in 2010.1 In the same year, people 65 years and older underwent 37.4% of all inpatient procedures, which totaled 19.2 million procedures.2 Older adults undergoing surgical procedures are at increased risk for postoperative morbidity and mortality3 and may prioritize quality over quantity of life years. Among older adults with life-threatening illnesses, one study4 found that most patients would decline life-saving interventions if it resulted in significant functional or cognitive impairment (74.4% and 88.8%, respectively).
Patient-centered care requires that clinicians pay attention to outcomes that matter most to patients. Loss of the ability to live independently is a high-priority patient-centered outcome. In the current study, loss of independence (LOI) is defined as 1 of 3 changes: a decline in function according to activities of daily living (ADLs), a decline in mobility requiring new mobility aid, or an increase in care needs, such as the need for new home-care services or discharge to a nonhome destination. Loss of independence is a novel intermediate outcome (ie, occurring in the early postoperative period). These elements correspond to a conceptual framework known as the Disablement Process, in which multiple factors, such as surgical procedures, the need for special equipment and devices, lifestyle and behavioral changes, and the physical and social environment, all influence a person’s ability to function independently.5 This conceptual model has been used previously in the surgical literature by Lawrence et al6 in a longitudinal evaluation of functional outcomes in older patients after major abdominal surgical procedures.
Currently, quality metrics prioritized by hospitals and medical professionals focus on discrete outcomes, such as readmission or mortality. This study hypothesized that LOI at the time of discharge would be associated with readmission and death after discharge. The objectives of the current study were to evaluate LOI among older adults after surgical procedures and to examine the association of LOI with readmission and death after discharge in this population.
Box Section Ref ID
Question How frequently do older adults experience loss of independence (LOI) after surgical procedures, and is LOI associated with poor outcomes of readmission and death after discharge?
Findings In this cohort study of 5077 older adults after surgical procedures, LOI occurred frequently and increased with age. After risk adjustment, LOI was significantly associated with a 1.7-fold increased risk for readmission and a 6.7-fold increased risk for death after discharge.
Meaning Hospitals should consider tracking patient-centered outcomes such as LOI, with the goal to intervene and minimize poor outcomes after discharge.
Data were obtained from the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) Geriatric Surgery Pilot Project. The ACS NSQIP has been described previously.7-10 Briefly, it is a prospective, multi-institutional clinical data registry that collects data on more than 200 variables, including preoperative patient characteristics, operative details, and postoperative events. A trained abstractor collects data from medical records using highly standardized data definitions; previous ACS data audits have indicated high reliability.9
The Geriatric Surgery Pilot Project was developed by the ACS Geriatric Surgery Task Force and the ACS NSQIP to capture more granular information valuable to older adults.11 Twenty-six NSQIP hospitals volunteered to collect a set of 14 geriatric-specific variables from 4 important domains: function, cognition, mobility, and decision making. The pilot data were collected between January 2014 and December 2014. Because all data were collected from a deidentified database, informed consent was waived, and the study is exempt from institutional review board oversight according to the Chesapeake Research Review Institutional Review Board.
Patients 65 years and older undergoing an inpatient operation at hospitals in the Geriatric Surgery Pilot Project who had known ADLs, living situation, and mobility were included in the cohort. Because LOI was assessed at the time of discharge, patients who died prior to discharge were excluded.
Risk Factors and Outcomes
The ACS NSQIP collects data on preoperative demographic information and comorbidities, operative details, and postoperative events. Preoperative demographic information and comorbidities used in this analysis includes sex, age, American Society of Anesthesiologists (ASA) classification, and body mass index (calculated as weight in kilograms divided by height in meters squared). Preoperative functional status (independence, partial dependence, or dependence for ADLs) is a standard variable collected for all NSQIP patients. The NSQIP Geriatric Surgery Pilot Project collects 14 geriatric-specific variables, including preoperative living situation (patient living at home without support, living at home with support, or living in a facility), use of mobility aids, previous falls within 1 year, history of cognitive decline or dementia, functional status at time of discharge, fall risk at time of discharge, postoperative use of mobility aid, and hospital discharge destination with or without the need for additional services.
Additional variables used in this analysis include emergency operation and a NSQIP composite measure for serious postoperative complications. This measure calculates serious postoperative complication as 1 or more of the following occurrences: cardiac arrest, myocardial infarction, pneumonia, progressive renal insufficiency, acute renal failure, deep vein thrombosis or pulmonary embolism, a return to the operating room, deep incisional surgical site infection, organ/space surgical site infection, systemic sepsis (sepsis or septic shock), unplanned intubation, urinary tract infection, or wound disruption. Of note, the following were adjusted and not considered complications if identified prior to the operation: surgical site infection, pneumonia, sepsis, progressive renal insufficiency or acute renal failure, and urinary tract infection.
Loss of independence was assessed at the time of discharge and was defined as a decline in functional status, mobility, and/or care needs. Decline in functional status was measured by comparing preoperative with postoperative function according to ADLs, which include bathing, dressing, grooming, eating, toileting, and transferring. Patients who were preoperatively independent or partially dependent had a decline in function if they became partially dependent or fully dependent, respectively. Decline in mobility was estimated using the requirement for a new mobility aid as a proxy. Increased care needs were assessed by discharge destination and the need for additional support or skilled services in the home. Living situation was captured preoperatively (origin from home alone, home with support, or nonhome origin, ie, from facility) and postoperatively (discharge destination to home alone, home with support or skilled services, or nonhome destination). Patients who were discharged to a nonhome destination or who needed new support or skilled services in the home were considered to have increased care needs.
The main outcome measures for the study were readmission and death after discharge, both measured within 30 days postoperatively. The ACS NSQIP collects date of discharge and date of death, allowing for the calculation of mortality after discharge or postdischarge death.
Univariate and bivariate analyses were conducted with descriptive statistics, χ2 tests, and t tests as appropriate. Multivariable regression models were developed using forward selection. Missing data were evaluated and found to be missing at random. Two variables were missing more than 5% of data (fall history missing 12% and preoperative cognitive impairment missing 9%). Missing data were handled with multiple imputation, using the SAS fully conditional specification imputation algorithm (preferred because of predominantly categorical data). To exceed the highest percentage of missing data, 15 imputed data sets were created. Variables included for imputation were those included in the final model: age, sex, body mass index classification, ASA classification, race/ethnicity, smoking status, preoperative living situation (origin), preoperative cognitive impairment, surrogate consent, fall history, emergency status, serious postoperative complications, postoperative loss of function or mobility or increased care needs, readmission, and death after discharge. All statistical analyses were performed using SAS version 9.4 (SAS Institute).
A total of 9972 eligible patients 65 years and older underwent inpatient operations in 26 hospitals taking part in the ACS NSQIP Geriatric Surgery Pilot Project between January 2014 and December 2014. The 100 patients who were totally dependent in ADLs and the 16 classified as ASA class 5 (moribund) were excluded. Because of the expected loss of mobility following orthopedic or spinal procedures, the 4680 patients undergoing these operations were also excluded. Because LOI was assessed at the time of discharge, the 99 patients who died prior to the time of discharge were excluded. The final cohort for analysis consisted of 5077 older adults.
The cohort included 2736 female patients (53.9%) with a mean (SD) age of 75 (7) years who underwent a range of inpatient operations. Most operations were general surgical procedures (2934 [57.8%]), followed by vascular surgical procedures (965 [19.0%]), and urologic or gynecologic surgical procedures (622 [12.3%]), and the remainder were surgical subspecialties. The group of patients experiencing LOI was significantly older with more comorbid conditions and geriatric risk factors (falls, weight loss, and cognitive impairment) (Table 1). Patients with LOI experienced higher rates of serious complications after the procedure than patients without LOI (5.2% vs 2.5%; P < .001) and stayed in the hospital longer (mean [SD] length of stay, 7.3 [7.5] vs 3.3 [3.3] days; P < .001).
Loss of independence was defined as a decline in functional status or mobility or increased care needs at the time of discharge. A decline in functional status occurred in 1348 patients (26.6%), and a decline in mobility occurred in 1626 (32.0%) (Figure, A). Increased care needs were observed in 2339 (46.0%), with 1414 (27.8%) requiring additional skilled or supportive services at home and 925 (18.2%) requiring discharge to a nonhome destination (Figure, A). Overall, 3027 (59.6%) of the cohort experienced at least some degree of LOI (loss in at least 1 domain), with 666 (13.1%) experiencing losses in all 3 domains (Figure, B). Loss of independence increased significantly with age, occurring in 1386 of 2780 patients (49.9%) aged 65 to 74 years, 1162 of 1726 (67.3%) aged 75 to 84 years, and 479 of 571 (83.9%) 85 years and older (P < .001).
Factors associated with LOI were identified through multivariable regression with forward selection, controlling for the preoperative patient characteristics identified in Table 1, as well as postoperative complication and length of stay. The strongest factor associated with LOI was age of 85 years and older, with an odds ratio (OR) of 4.4 (95% CI, 3.4-5.7) (Table 2). Having experienced a fall in the year prior to the operation was also strongly associated with LOI, with an OR of 2.4 (95% CI, 1.9-3.0). Although serious postoperative complications were not significant in the model, postoperative length of stay demonstrated a significant association, with each day increasing risk by 30% (OR, 1.3; 95% CI, 1.2-1.3).
Readmission occurred in 517 patients (10.2%) from the cohort. A multivariable model with forward selection controlled for age, sex, ASA classification, body mass index category, emergency operation, preoperative living situation, unintentional weight loss (loss of ≥10% body weight in prior 6 months), smoking status, surrogate consent, preoperative cognitive impairment, fall history (within 1 year), serious postoperative complication, and LOI. The model identified 4 risk factors demonstrating a strong association with readmission. Loss of independence exhibited the second strongest association with readmission (OR, 1.7; 95% CI, 1.4-2.2) after postoperative surgical complication (OR, 6.7; 95% CI, 4.9-9.0). Additional significant factors included an ASA class of 3 or 4 (OR, 1.4; 95% CI, 1.1-1.7) and preoperative living situation with support in the home (OR, 1.4; 95% CI, 1.1-1.7) (Table 3).
Death after discharge occurred in 69 patients (1.4%). A multivariable model with forward selection controlled for age, emergency operation, unintentional weight loss (loss of ≥10% body weight in prior 6 months), surrogate consent, preoperative cognitive impairment, fall history within 1 year, serious postoperative complication, and LOI. There were 4 significant risk factors selected by the model. Loss of independence demonstrated the strongest association with death after discharge (OR, 6.7; 95% CI, 2.4-19.3). A surrogate decision maker having signed the consent form was also strongly associated with death after discharge (OR, 6.0; 95% CI, 3.4-10.5). Additional risk factors associated with death after discharge included emergency operation (OR, 2.0; 95% CI, 1.2-3.6) as well as age of 75 to 84 years (OR, 1.9; 95% CI, 1.1-3.6) and age of 85 years and older (OR, 2.4; 95% CI, 1.2-4.8) compared with patients aged 65 to 74 years. Serious postoperative complication was not significantly associated with death after discharge and was not selected in the model (Table 4).
The current study aimed to evaluate the association between LOI, a novel and patient-centered outcome, and the discrete end points of readmission and postdischarge death. Loss of independence encompasses several elements that influence independent living, including function in performing tasks necessary for survival and personal care (ie, ADLs, such as eating and toileting), use of aids to ambulate (mobility devices), and physical and social environmental factors (home with support or skilled care vs nonhome environment). In this retrospective cohort study, 3027 older patients (59.6%) undergoing surgical procedures experienced a decline in at least 1 of 3 domains—function, mobility, or home care needs—at the time of discharge. Loss of independence increased significantly with age, with those 85 years and older at greatest risk.
After serious postoperative complication, LOI was the second most important factor associated with readmission, increasing the risk by 70%. Serious postoperative complications were most significantly associated with readmission, increasing the risk by 6.7-fold. Complications may presumably be directly associated with the indication for readmission. However, the significant association of readmission with LOI and preoperative support in the home may suggest the critical role of environment and patient resources in prompting readmission. When examining death after discharge, LOI was associated with a 6.7-fold increased risk. Additional significant factors included surrogate-signed consent and emergency operations, as well as advancing age. Postoperative complications were not significantly associated with death after discharge. The need for a surrogate decision maker or an emergency operation suggests that patients had cognitive or psychological limitations or severe illness prior to an operation. Patients experiencing LOI were older and sicker; patients 85 years and older experienced a 4.4-fold increased risk for LOI compared with patients aged 65 to 74 years, while those with a fall in the year prior to an operation had a 2.4-fold increased risk for LOI. Certainly, older, sicker patients often have a poorer prognosis, and this may help explain the increased risk for death after discharge.
These findings may be best interpreted by considering the literature on frailty, a phenotypic state of decreased physiologic reserve not confined to chronological age.12 Patients experiencing LOI in this study may have had underlying frailty predisposing them to worse outcomes. This subset of vulnerable patients are likely the same group that experience high rates of LOI. Although rates of frailty vary widely when studied among surgical populations (10% to more than 50%), studies consistently demonstrate increased risks for poor postoperative outcomes (eg, infectious complications, cardiac events, respiratory compromise, or death) among frail patients.13-15 Studies examining discharge to nonhome destinations have identified that frailty significantly increases this risk, ranging from a 1.5- to a 20-fold increase.13,16,17 Although frailty represents a critical prognostic factor, it is imperative to identify patient-centered outcomes, such as LOI, that may serve as targets for intervention and relevant metrics for high-quality care.
The demand for surgical therapy from older adults will only increase in the coming decades. The association of LOI with readmission and death after discharge should prompt increased attention from physicians, hospital quality personnel, and administrators. Efforts to decrease LOI at the time of discharge may potentially prevent readmission at a later time. Multidisciplinary care with an emphasis on geriatric medicine may provide benefits for surgical patients. Established models of care (such as the Hospital Elder Life Program18) and geriatric comanagement strategies focus on simple interventions, such as frequent reorientation, hydration, and early ambulation to prevent delirium, deconditioning, and malnutrition.19,20 Length of stay was a significant risk factor for LOI, with each additional hospital day adding 30% risk. Certainly, deconditioning over the hospital stay is a logical mechanism that may explain LOI. The current study identified functional decline among 1348 surgical inpatients (26.6%) 65 years and older. This is consistent with literature that indicates most functional loss occurs within the first week of surgical procedures and may even persist up to 6 months later.6
Therapies to promote preservation of function and mobility may minimize the detrimental effects of a hospitalization in vulnerable elderly patients. Recognition that LOI carries an increased risk for readmission should prompt a more detailed, personalized discharge plan. A 2013 Cochrane database systematic review21 found that personalized, detailed discharge planning improved transitions of care, with a significant decrease in readmission rates and length of stay. Attention to postoperative care needs may allow early recognition of limited home resources or an overreliance on family and friends for support. Caregiver burnout can occur in chronic conditions and contribute to hospital readmission and poor outcomes.22,23 The significant association of LOI and support in the home preoperatively with risk of readmission in our study underscores the importance of evaluating the patient’s social support system. Further investigation should examine mechanisms to prevent LOI and promote care processes that facilitate successful transitions from the hospital to the community.
This study has several limitations. First, the ACS NSQIP Geriatric Surgery Pilot Project includes a subset of 26 NSQIP hospitals, with voluntary participation. These hospitals and their patients may represent a homogenous sample, potentially limiting the generalizability of these findings. However, it remains possible that LOI in a more heterogeneous sample, with less awareness of geriatric-specific issues, may demonstrate an even stronger association with readmission and postdischarge death. Second, LOI at discharge, death after discharge, and readmission within 30 days remain short-term outcomes. Function, cognition, and mobility are dynamic, long-term processes. The current study presents an assessment of 1 patient-centered outcome at 1 time point; however, future studies should include longitudinal assessments of patient-centered postoperative outcomes. Third, the current study is retrospective in nature and cannot establish the cause of LOI nor solutions to prevent readmission and death after discharge. These interventions will require a concerted prospective effort.
Data on function, mobility, and care needs are more difficult to collect than discrete outcomes like readmission and death. However, there is growing support for the use of patient-centered outcomes in defining high-quality care for older adults.24 The association of LOI with readmission and death after discharge should highlight the potential value of these patient-centered outcomes as targets for clinical initiatives. This study demonstrates that data on function, mobility, and care needs can be feasibly collected in a multi-institutional, national data registry.
Loss of independence, as defined by decreases in function or mobility or an increase in care needs, is associated with readmission and death after discharge. Patient-centered outcomes such as LOI can, and should, be collected in multi-institutional data registries. Loss of independence is a potential target for intervention, and future work should move beyond its use as a factor for prognostication. To best serve the aging population, clinical initiatives must focus on efforts to minimize LOI and better understand its association with discrete outcomes like readmission and death after discharge.
Corresponding Author: Julia R. Berian, MD, Division of Research and Optimal Patient Care, American College of Surgeons, 633 N St Clair St, 22nd Floor, Chicago, IL 60611 (firstname.lastname@example.org).
Accepted for Publication: April 12, 2016.
Published Online: July 13, 2016. doi:10.1001/jamasurg.2016.1689.
Author Contributions: Dr Berian had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Mohanty, Ko, Robinson.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Berian, Ko.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Berian, Mohanty.
Administrative, technical, or material support: Berian, Rosenthal.
Study supervision: Ko, Rosenthal, Robinson.
Conflict of Interest Disclosures: Dr Berian’s position as the James C. Thompson Geriatric Surgery Clinical Scholar in Residence at the American College of Surgeons for the 2015-2016 academic year is supported by a grant from the John A. Hartford Foundation (coprincipal investigators, Drs Rosenthal and Ko). Additional support for Dr Berian’s position as the American College of Surgeons Clinical Scholar in Residence for the 2014-2015 academic year was provided by the University of Chicago and the American College of Surgeons. Dr Mohanty’s position as the James C. Thompson Geriatric Surgery Clinical Scholar in Residence at the American College of Surgeons for the 2013-2015 academic years was supported by the American Geriatric Society, the John A. Hartford Foundation, and the American College of Surgeons. No other disclosures were reported.
Previous Presentation: This work was presented at the 87th Annual Pacific Coast Surgical Association Meeting; February 14, 2016; Kohala Coast, Hawaii.
Additional Contributions: We acknowledge the surgical clinical reviewers and surgeon champions at the hospitals participating in the American College of Surgeons National Surgical Quality Improvement Program Geriatric Surgery Pilot Project for their efforts in data collection.
KY. Identification of specific quality improvement opportunities for the elderly undergoing gastrointestinal surgery [published correction appears in Arch Surg
. 2010;145(3):225]. Arch Surg
. 2009;144(11):1013-1020.PubMedGoogle ScholarCrossref
H. Understanding the treatment preferences of seriously ill patients. N Engl J Med
. 2002;346(14):1061-1066.PubMedGoogle ScholarCrossref
et al. Functional independence after major abdominal surgery in the elderly. J Am Coll Surg
. 2004;199(5):762-772.PubMedGoogle ScholarCrossref
et al; National VA Surgical Quality Improvement Program. The Department of Veterans Affairs’ NSQIP: the first national, validated, outcome-based, risk-adjusted, and peer-controlled program for the measurement and enhancement of the quality of surgical care. Ann Surg
. 1998;228(4):491-507.PubMedGoogle ScholarCrossref
et al; Principal Investigators of the Patient Safety in Surgery Study. Successful implementation of the Department of Veterans Affairs’ National Surgical Quality Improvement Program in the private sector: the Patient Safety in Surgery Study. Ann Surg
. 2008;248(2):329-336.PubMedGoogle ScholarCrossref
et al. Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg
. 2010;210(1):6-16.PubMedGoogle ScholarCrossref
et al. Optimizing ACS NSQIP modeling for evaluation of surgical quality and risk: patient risk adjustment, procedure mix adjustment, shrinkage adjustment, and surgical focus. J Am Coll Surg
. 2013;217(2):336-346.e1.PubMedGoogle ScholarCrossref
RA. The ACS NSQIP Geriatric Surgery Pilot Project: improving care for older surgical patients. Bull Am Coll Surg
. 2014;99(10):21-23.PubMedGoogle Scholar
et al; Cardiovascular Health Study Collaborative Research Group. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci
. 2001;56(3):M146-M156.PubMedGoogle ScholarCrossref
et al. Frailty as a predictor of surgical outcomes in older patients. J Am Coll Surg
. 2010;210(6):901-908.PubMedGoogle ScholarCrossref
et al. Too frail for surgery? initial results of a large multidisciplinary prospective study examining preoperative variables predictive of poor surgical outcomes. J Am Coll Surg
. 2013;217(4):665-670.e1.PubMedGoogle ScholarCrossref
et al. Redefining geriatric preoperative assessment using frailty, disability and co-morbidity. Ann Surg
. 2009;250(3):449-455.PubMedGoogle Scholar
et al. Accumulated frailty characteristics predict postoperative discharge institutionalization in the geriatric patient. J Am Coll Surg
. 2011;213(1):37-42, 42-44.PubMedGoogle ScholarCrossref
et al. Predicting hospital discharge disposition in geriatric trauma patients: is frailty the answer? J Trauma Acute Care Surg
. 2014;76(1):196-200.PubMedGoogle ScholarCrossref
SK. Effects of a modified Hospital Elder Life Program on frailty in individuals undergoing major elective abdominal surgery. J Am Geriatr Soc
. 2014;62(2):261-268.PubMedGoogle ScholarCrossref
et al. Geriatric outcomes are improved by a geriatric trauma consultation service. J Trauma
. 2006;61(5):1040-1046.PubMedGoogle ScholarCrossref
A. Quality of care delivered before vs after a quality-improvement intervention for acute geriatric trauma. J Am Coll Surg
. 2015;220(5):820-830.PubMedGoogle ScholarCrossref
SL. Discharge planning from hospital to home. Cochrane Database Syst Rev
. 2013;1(1):CD000313.PubMedGoogle Scholar
EM. Caregiving burden, stress, and health effects among family caregivers of adult cancer patients. JAMA
. 2012;307(4):398-403.PubMedGoogle ScholarCrossref
L. Caregiver status: a simple marker to identify cardiac surgery patients at risk for longer postoperative length of stay, rehospitalization, or death. J Cardiovasc Nurs
. 2014;29(1):12-19.PubMedGoogle ScholarCrossref