Preoperative assessments for frailty have been shown to predict poor postoperative outcomes [13, 14]. Previously, Owodunni et al. examined the use of the standard EFS and demonstrated that patients classified as frail were at increased risk for LOI and mortality [6]. With the recent changes in healthcare delivery and clinical workflow, our study was aimed at investigating the self-reported domains of the EFS to determine if it would have a similar ability to predict postoperative LOI and mortality as the standard EFS. Our results demonstrate that under extenuating circumstances, where in-person visits are not possible, healthcare providers can utilize the srEFS, with a threshold score of 5 or more indicating HRF, to aid in the prediction of postoperative LOI with similar predictive ability as the previously validated standard EFS.
Additionally, it is important to note that only increased time on the Get Up and Go measured on the standard EFS was independently associated with LOI, while cognitive impairment measured on the Clock Draw was not. This suggests that in ideal practice conditions, the standard EFS would be the preferred tool to capture physical frailty associated with LOI. The clock draw is a screening test used to help diagnose cognitive impairment given its ability to test numerous cognitive domains including memory, visuospatial, concentration, motor function, perception, numerical knowledge etc [15, 16]. It has been demonstrated to be > 85% sensitive for cognitive impairment when used in the MMSE [17]. However, there have been studies that have shown the number of cognitive impairment cases amongst frail patients was not clearly reported [18,19,20]. Similarly, 22% of Alzheimer dementia patients have lacked physical indicators of frailty [21]. The International Association of Gerontology and Geriatrics/International Academy on Nutrition and Aging have proposed the concept of cognitive frailty to address this patient population [22, 23]. In the preoperative setting, studies have found mixed utility in the use of clock draw to identify preoperative patients at risk of developing delirium [24, 25]. Other tools that capture frailty, such as the Reported Edmonton Frail Scale (REFS), also include the clock draw as a component of testing, however a key difference is that the REFS is used in the acute inpatient setting, which may indicate that cognitive frailty may have a larger role in inpatient acute medicine in comparison to the preoperative elective surgery population [4]. In our study, the elimination of the clock draw was not associated with decreasing the predictive ability for LOI. This suggests that physical frailty, which describes a high-risk disposition with increased susceptibility to surgical stressors and low reserve, may be more important for the outcome of LOI [26]. It is important to note that we did not study the outcome of post-operative delirium, which is a vital aspect of surgical care, and thus in ideal practice conditions it may be important to use the standard EFS or REFS if evaluating risk for delirium.
As Nidadavolu et al. describe, there are a variety of preoperative frailty assessment tools that have been validated, however there is no gold standard for preoperative frailty assessments [27]. They also found that the majority of preoperative frailty assessments currently have an in-person component [27]. Although there are numerous self-reported frailty assessments that have been described in the literature, such as the REFS, which focuses on the acute inpatient setting, our goal was to explore tools that could be easily administered in the preoperative setting for elective surgical procedures [28]. Moreover, the REFS includes the clock-draw component of the EFS, which requires either an in-person assessment, or access to a camera and knowledge of how to use secure internet-based video software, which may be challenging for patients [28].
Under the current Age Friendly Health System framework put forth by the John A. Hartford Foundation, there has been a shift in provider focus to “what matters most” to the patient when planning a surgical intervention [29, 30]. The potential for LOI is now considered as deterrent by providers that may prolong quantity over quality of life, particularly in the area of interventions for cancer care [31]. Therefore, identifying risk factors for postoperative LOI have been a focus of recent investigations. Blankenship et al. recently delineated risk factors for postoperative LOI following pelvic organ prolapse in the hopes of improving shared decision making prior to surgery. They found that there was a higher risk of loss of functional independence for patients who were 80 years old or greater, had higher ASA scores and increased LOS [32]. Bonicoli examined outcomes at 2 years following hemiarthroplasty vs. osteosynthesis in patients 80 years or older who suffered hip fracture [33]. Their study demonstrated an increase in mortality risk following hemiarthroplasty but no difference in functional performance measures by the activities of daily living [33]. Finally, Bal et al. examined results of emergent inguinal hernia repair given that current recommendations have encouraged watchful waiting in elderly patients with an inguinal hernia [34]. His study reviewed results following emergent inguinal hernia repair in patients 70 years and older using the National Surgery Quality Improvement Program (NSQIP) database and demonstrated that elective surgery in these patients were associated with improved outcomes, including a lower likelihood for LOI or mortality [34].
Our results demonstrate the novel utilization of the srEFS to identify HRF patients, which can be feasibly performed via telephone interviews or virtual methods such as online patient health portals [35]. The shift in healthcare workflow has led some investigators to create virtual clinics for preoperative assessments. Joughin et al. recently described implementation of their virtual geriatric perioperative medicine clinic [36]. Comprehensive geriatric assessments were converted to questionnaires which could be utilized during a virtual visit. Patient survey results following implementation demonstrated optimism over the virtual methodology, particularly in the area of shared decision making [36]. Hands et al. found that alternative management plans that occurred because of telemedicine proved to be advantageous for both patient and consultants [37]. Similarly, our study highlights that srEFS may be used across a variety of surgical procedures given the diversity of surgical procedures studied.
There are several limitations to this study. The self-reported aspects were taken from our standard EFS evaluations that were already implemented in our preoperative clinic. They were provided as a paper questionnaire to the patient to answer in -person or the patient may have been asked the self-reported questions by a clinic provider. This data was then taken and entered into our institution’s EHR. The use of this data allowed for the timely evaluation using the srEFS, however additional studies examining the use of the srEFS in a virtual setting are warranted. In addition, patients who had an EFS performed up to 6 months prior to surgery were included and their frailty status may have changed in the time prior to surgery. We also included outpatient procedures which are not commonly associated with LOI, however we felt that these procedures are often performed in older patients at risk for LOI. Our data demonstrated that 20% of patients experiencing LOI had an outpatient procedure performed. This study is a single centre study, and confirmation of our results at other institutions is needed. Another limitation is that the EFS was completed by all patients who had undergone surgery including those who had cognitive impairment or other disabilities, which may have impacted their performance on their assessment. However, the EFS was chosen because it integrates many psychosocial aspects into the scoring system and thus is not heavily weighted by a patient’s disability or cognitive impairment. We included patients who required assistance with the clock-draw, or who could not perform an aspect of the test secondary to a disability to enhance its generalizability to all older adults. Another limitation is the heterogeneity of the population of this study, given that both inpatient and outpatients were included. However, we attempted to control for this heterogeneity by including OSS and HCC into our analysis. Finally, this study used the ACS NSQIP® registry database, which records solely 30 day mortality post-surgery [38]. Moreover, it lacks specific information regarding comorbidities, their impact on functioning, surgical outcomes or reason for surgery [38]. Additionally, ACS- NSQIP does not differentiate between unique events that are specific to certain operations, which may have been useful in this study to analyze which specific events had an impact on mortality and LOI. However, ACS-NSQIP is currently creating a procedure-specific form that is aimed at a specific procedure and collecting results and variables [39].
Future directions of this study include exploring the associations between the srEFS being done virtually or using telehealth, and its impacts pre and postoperative planning. We anticipate targeting domains of the srEFS that have been shown to increase risk for LOI. Furthermore, we hope to also apply the srEFS in acute care or emergency care, where it can be completed in emergent/urgent way and allow for early identification of patients at risk for LOI.
