A Systematic Review and Meta-Analysis of Preoperative Frailty Instruments Derived From Electronic Health Data : Anesthesia & Analgesia

Secondary Logo

Journal Logo

Featured Articles: Meta-Analysis

A Systematic Review and Meta-Analysis of Preoperative Frailty Instruments Derived From Electronic Health Data

Alkadri, Jamal MD*; Hage, Dima BSc; Nickerson, Leigh H. MD*; Scott, Lia R. MD; Shaw, Julia F. BSc§; Aucoin, Sylvie D. MD, Msc, FRCPC*; McIsaac, Daniel I. MD, MPH, FRCPC*,§,‖

Author Information
doi: 10.1213/ANE.0000000000005595



  • Question: What is the prognostic value of frailty instruments applied to preoperative electronic health data?
  • Findings: In this systematic review and meta-analysis of 90 studies representing 22 unique frailty instruments, electronic frailty assessment was associated with a 3.5-fold increase in the odds of death after adjustment.
  • Meaning: Electronic frailty instruments could improve preoperative risk stratification and guideline recommended care if properly constructed instruments are applied to multidimensional data.

See Article, p 1090

F railty is defined as a biologic state or syndrome that arises due to the accumulation of age- and disease-related deficits across multiple domains.1,2 Having frailty results in increased vulnerability to adverse health outcomes.2,3 As the population ages rapidly, an increasing number of older people are presenting for surgery.4 Over a third of older people having major surgery live with frailty,5 and preoperative frailty is associated with a more than 2-fold increase in rates of morbidity and mortality, as well as increased health care resource utilization.2,6 As the prevalence of frailty increases with age, older people with frailty are likely to present for surgery with increasing frequency.3

Best practice guidelines recommend routine assessment of frailty in older surgical patients, which can result in improved outcomes.7–10 However, routine frailty assessment continues to be underperformed.11,12 Multiple barriers to routine preoperative frailty assessment exist, including time, choice of instrument, and availability of data and instruments required to complete performance-based (eg, physical and cognitive) testing.10,13,14 However, recent advances and development of frailty instruments specifically designed and validated in electronic and administrative data sources suggest that automated frailty assessment could help to address the practice gap related to inadequate preoperative frailty assessment.15–17

Given the increasing availability of electronic medical data and uptake of electronic health records (EHRs), automated preoperative frailty assessment may soon be considered in clinical practice. However, to our knowledge, available frailty instruments that have been applied to perioperative electronic health data have not been synthesized. This is a necessary step to guide decision-making, as choosing an appropriate instrument requires that it be consistent with consensus definitions of frailty while also adding prognostic value. Furthermore, such an instrument would need to be feasible for routine use. Therefore, we performed a systematic review and meta-analysis with a primary objective of describing currently utilized frailty instruments that have been applied to preoperative electronic data and synthesizing their prognostic value in relation to postoperative mortality (primary outcome), discharge disposition after the index surgical hospitalization, length of stay (LOS), and costs (secondary outcomes). Our secondary objective was to evaluate the content validity of available electronic frailty instruments (ie, does the instrument measure the elements recognized to comprise the frailty state or syndrome), while a tertiary objective was to synthesize information provided on the feasibility of using electronically derived preoperative frailty assessments.


Figure 1.:
Guiding research question using the PICOTS format. PICOTS indicates Population, Index prognostic factor, Comparator prognostic factors, Outcomes, Timing, Setting.

Best practice guidelines for systematic reviews were followed, including the meta-analysis of observational studies in epidemiology18 and Cochrane Handbook.19 The protocol was informed by best practice recommendations for systematic review and meta-analysis of prognostic factor studies20 and registered with the Open Science Foundation (https://osf.io/rvc8k/). Our specific research question is provided in Figure 1. Results are reported in keeping with the Preferred Reporting Items for Systematic Reviews and Meta Analyses guidelines.21 All aspects of review and data collection were done using DistillerSR, a cloud based systematic review platform (Evidence Partners, Ottawa, ON, Canada).

Data Sources and Searches

A comprehensive literature search was developed with an information specialist using search terms related to older age, frailty, surgery, and electronic or administrative data. A second, independent information specialist then peer-reviewed the search strategy using the Peer Review of Electronic Search Strategies guideline.22 A copy of the search strategy is included in Supplemental Digital Content, eTable 1, https://links.lww.com/AA/D544. The final search strategy was applied to Medline, Excerpta Medica dataBASE (EMBASE), Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Cochrane databases from inception to December 31, 2019. English- or French-language articles were retrieved. The reference lists of related systematic reviews and included articles were also evaluated to identify relevant articles potentially missed by our initial search.

Study Selection

We included studies that met the following criteria: (1) studied a population of adults ≥18 years presenting for a surgical procedure in an inpatient setting; (2) assessed for the presence of frailty using an explicitly defined frailty instrument applied to any type of health administrative, claims-based, multi-institutional registry or electronic health data; (3) reported relevant outcomes, specifically the association of frailty status with postoperative mortality (primary outcome), discharge disposition (secondary outcome, which is a proxy for functional status and/or independence that is accurately captured in administrative data), and resource use (ie, costs and LOS). Relevant outcome data included absolute rates or differences and relative rates. Measures of predictive accuracy (eg, discrimination, calibration, Brier Scores, and explained variance) were also extracted. Exclusion criteria included studies where >50% of participants were nonsurgical or where surgery-specific data could not be extracted.

Assessment for inclusion involved 2 stages, both of which were conducted in duplicate by independent reviewers. First, titles and abstracts were reviewed; if either reviewer answered yes or unsure, the article proceeded to full-text review, whereas 2 reviewers were required to say no to exclude. All articles that passed stage 1 proceeded to full-text review. In stage 2, each article was reviewed in full; any disagreement about inclusion or exclusion between the 2 reviewers was resolved through discussion with the senior author (D.I.M.).

Data Extraction

Data extraction for included articles was done using a custom-built data-extraction form. Before final use, the form was piloted in 5 articles by the senior author and each reviewer. Data extraction then proceeded independently and in duplicate. Key variables extracted included author; year and country of publication; type and description of the database; surgical specialty and procedure; characteristics of the study population (eg, sample size, age, gender, percent with frailty); type, description, and components of frailty instrument; and association with outcomes (proportions, rates, effect measures and 95% confidence intervals [CIs], and measures of prognostic performance). Feasibility-related data included acceptability, implementation, and practicality.23,24

Characterization and Assessment of Construct Validity

Although no common and agreed-on detailed definition of frailty exists, international consensus states that frailty is multidimensional and reflects contributions from at least 4 domains.13 Furthermore, for accumulating deficits frailty indices, which we expected to be prevalent among electronic instruments, best practice recommends use of at least 30 variables reflecting multiple domains.25 Therefore, our primary approach to construct validity was to assess the subdomain of content validity, specifically that the instrument was based on adequately multidimensional components (including identification of what domains were addressed [eg, physical or physiologic, cognitive, psychosocial, comorbid disease, pharmacology, and disability]) and within the subset of frailty indices, to assess adherence to recommendations from Searle et al.25 We further assessed whether: (1) the instrument was derived de novo, adapted from a previous study, or applied directly based on previous methods and (2) the instrument was based on the frailty phenotype, accumulating deficits model of frailty (the 2 leading conceptual frameworks for frailty),5,26 a separate conceptual model, or no underlying model.


Our primary outcome was all-cause postoperative mortality at any point after surgery (as previous research demonstrates that relative effect measures for mortality are consistent across the first postoperative year27). The secondary outcomes were discharge disposition after the index surgical hospitalization, LOS, and costs. These outcomes were selected as all are accurately captured in electronic health data and are relevant to patients, clinicians, and health care systems. We did not collect complications, functional, or cognitive outcomes, as they are not accurately captured in most electronic data sources.28 We extracted unadjusted and adjusted outcomes; however, in keeping with best practice recommendations20 and previous perioperative reviews,29 we specifically identified whether adjustment had been made for a minimum confounder set (at least 4/5 of: illness severity [comorbidity or American Society of Anesthesiologists’ score], sex or gender, age, type of surgery, and urgency of surgery), which we postulated would be routinely available in electronic data sources and would represent a minimum set of variables that would be associated with the presence of frailty and the risk of adverse outcomes (recognizing that this set represents a possibly adequate but not necessarily sufficient set of confounders29–31).


Descriptive statistics were used to summarize key methodological and population characteristics across included studies. We also described the number of domains captured by each instrument, as well as the proportion that adhered to our prespecified metrics for construct validity (ie, multidimensionality).

As new prognostic factors should provide additional value above and beyond factors already used in assessment,30–32 our primary analyses focused on studies where prognostic value was estimated after adjustment for a minimum confounder set. Pooled unadjusted estimates were estimated to ensure complete reporting of these observational data but were not of primary interest. Prognostic value analyses addressed strength of association (ie, effect estimates). Effect estimates were meta-analyzed overall as a primary approach (which addressed the general question of whether use of any preoperative electronic frailty assessment added prognostic value) and by instrument (where >2 studies were available that provided required adjusted data, which addressed the question of how much prognostic value–specific frailty instruments may provide); otherwise, results were narratively synthesized. A 5% level of significance was specified for all analyses. We prespecified that our main pooled analyses would assess the association between frailty expressed as a binary exposure (as this is the most common manner in which frailty is represented in epidemiologic studies).10 Where frailty was categorized in >2 levels, the level without frailty (or the lowest frailty category) was defined as the reference, with the moderate frailty category as the comparator.

All meta-analyses were performed using the statistical package “metafor” (R statistical programming language, R Foundation, Vienna, Austria) and were based on random effects models (as content knowledge suggested that this area of study would not meet the assumptions of fixed-effects meta-analysis). Outcomes were pooled using the Sidik-Jonkman estimator, which provides better error rates than the more frequently used DerSimonian and Laird estimator when small study numbers or variable sized study samples are pooled.33 We measured heterogeneity using the I2 statistic but did not base any analytic decisions on this value (as we were already using random effects models and evidence suggests that heterogeneity statistics like I2 tend to be artificially high in studies with large n34). We also prespecified meta-regression analyses to test whether there was an evidence of effect modification by frailty instrument, urgency, or surgery type when predicting mortality. These analyses also helped to explore potential sources of heterogeneity.

During the peer-review process, we were asked to perform a formal meta-analysis of measures of predictive accuracy. Therefore, using the “metafor” package, we pooled c-statistics for studies that reported such values for logistic regression models where the frailty instrument was the only predictor variable (ie, unadjusted). Where CIs for the c-statistic were not reported, we estimated variances using the number of events and used the Sidik-Jonkman estimator to derive CIs for pooled estimates, as recommended.35

Recognizing that some sources of electronic data (eg, National Surgical Quality Improvement Program [NSQIP] database) are widely studied, overlapping cohorts were possible. Therefore, we had to exclude some studies from adjusted pooled analyses to avoid “double counting” participants. Where ≥50% overlap was identified, we excluded the study that had the smaller number of total participants, or if cohort sizes were equal, we excluded the older study. If studies had overlapping study cohorts but used differing frailty instruments, the study was excluded from the overall analysis but included in subgroup analysis of frailty instruments.

Risk of Bias Assessment

Risk of bias was assessed using the Cochrane Quality in Prognostic Studies (QUIPS) tool. For each included study, a risk of bias assessment was performed in duplicate, with at least 1 performed by the senior author. Disagreements were resolved through consensus with a third team member.


Figure 2.:
Preferred reporting items for systematic reviews and meta-analyses flow diagram documenting process of including and excluding studies.

We identified 1596 titles and abstracts, and after removal of duplicates, we reviewed 1130 articles. We ultimately assessed 406 full-text articles and included 90 studies (Figure 2; 7 studies reported predictive accuracy only but not effect sizes—our primary focus). Together, the included studies involved 4,698,454 participants and were published between 2013 and 2019. Full details of studies are provided in Supplemental Digital Content, eTable 2, https://links.lww.com/AA/D544.

Study Characteristics

General surgical procedures were the most studied (30 studies [33%]), followed by orthopedics (19 studies [21%]), neurosurgery (10 studies [11%]), cardiac (7 studies [8%]), urology and mixed surgical procedures (6 studies each [7%]), vascular (4 studies [4%]), gynecology and otolaryngology (3 studies each [3%]), and 2 studies (2%) examining thoracic surgical patients. Average study population age ranged from 46 to 82. Proportion of female patients ranged from 3.6% to 100%.

Frailty Instruments

Frailty was defined using 22 different instruments. The most prevalent was the 11-item NSQIP modified frailty index (42 studies),36–77 followed by the 5-item NSQIP modified frailty index (16 studies),78–93 the John Hopkins Adjusted Clinical Groups (ACG) frailty-defining diagnosis indicator (6 studies),94–99 and the Risk Analysis Index (4 studies).100–103 Other instruments used were reported in ≤3 studies.17,104–124 Only 2 studies compared different frailty instruments head-to-head.101,117 Seventy-six studies (84%) used a frailty index as their frailty instrument. Of the instruments identified, 11 (50%) were developed de novo, and 4 (18%) were modified from preexisting instruments (Supplemental Digital Content, eTable 3, https://links.lww.com/AA/D544, for a summary of instruments used).

Prognostic Value—Mortality

Fifty-five studies (n = 1,925,619)36,37,40–46,48,49,52–59,62–65,70,72,74–76,81–83,87,89,91,94–96,98–110,112,115,116,120,122 reported unadjusted mortality data (Supplemental Digital Content, eTable 4, https://links.lww.com/AA/D544). The pooled unadjusted odds ratio (OR) for frailty and mortality across instruments was 3.55 (95% CI, 3.03–4.17; P < .001; I2 = 98%; see Supplemental Digital Content, eFigure 1, https://links.lww.com/AA/D544). Forty studies (n = 3001,537)17,36,38,39,44–48,50,52,55,57,60,62,64,70,73,74,77–80,82,84–86,92,96,98,101,104,107,109,111,115,116,120–122 reported adjusted mortality data; 9 studies36,44,45,48,52,62,64,84,92 were excluded from the pooled adjusted analysis as they provided overlapping data with newer or larger studies. Twenty-three studies reported adequately adjusted ORs and 4 adequately adjusted hazard ratios (HRs; 3 studies reported mortality data separately by procedure).55,78,86 The pooled adjusted OR was 3.57 (95% CI, 2.68–4.75; P < .001; I2 = 97%; Figure 3A); the pooled adjusted HR was 1.31 (95% CI, 1.03–1.65; P < .001; I2 99%; Figure 3B). Instrument-specific adjusted pooling was possible for the 11-item NSQIP modified frailty index and the 5-item NSQIP modified frailty index (Supplemental Digital Content, eFigure 2, https://links.lww.com/AA/D544).

Figure 3.:
Forest plots for pooled strengths of association between frailty and mortality, subdivided by frailty instrument. A, Adjusted ORs. B, Adjusted HRs. CI indicates confidence interval; HR, hazard ratio; OR, odds ratio; TAVR, transarterial valve replacement.

Meta-regression did not identify evidence of effect modification on the OR scale (as we had >10 studies to support meta-regression)19 by frailty instrument (P = .202), surgery type (P = .077), or urgency (P = .301). Heterogeneity also remained in each meta-regression (based on studies with adequate adjusted data listed in Figure 3A), with I2 values consistently >90%.

Predictive accuracy for mortality was reported in 19 studies, including 9 for the 11-item NSQIP modified frailty index (c-statistic range, 0.54–0.94), 2 for the 5-item NSQIP modified frailty index (c-statistic, 0.89–0.91), 2 for the Preoperative Frailty Index (c-statistic, 0.80–0.82), and 2 for the Risk Analysis Index (c-statistic, 0.68–0.78). Seven studies reported unadjusted c-statistics, leading to a pooled estimate for 0.69 (95% CI, 0.63–0.75). No studies reported expected number of deaths, precluding a meta-analysis of calibration. Full descriptions of predictive accuracy data can be found in Supplemental Digital Content, eTable 5, https://links.lww.com/AA/D544.

Prognostic Value—Nonhome Discharge

Twenty-two studies (n = 853,701)36,52,61,62,64,72,75,81,87–89,91,93–98,103,115,120,123 reported unadjusted nonhome discharge (defined as new admission to a nursing home, long-term care center, rehabilitation, palliative care, or short-term hospital; Supplemental Digital Content, eTable 6, https://links.lww.com/AA/D544) data. The pooled unadjusted OR for frailty and nonhome discharge was 3.16 (95% CI, 2.48–4.03, P < .001; I2 = 99%; Supplemental Digital Content, eFigure 3, https://links.lww.com/AA/D544). Twenty studies (n = 1177,905)17,39,52,61,62,64,78,79,84,85,87–89,92,93,96,97,115,120,123 reported adjusted data for nonhome discharge. The pooled adjusted OR was 2.40 (95% CI, 1.99–2.89, P < .001; I2 = 99%; Supplemental Digital Content, eFigure 4, https://links.lww.com/AA/D544). The most frequently used frailty instruments were the 5-item NSQIP, which was associated with nonhome discharge (adjusted OR, 2.34; 95% CI, 1.75–3.13, P < .001; I2 = 98%; Supplemental Digital Content, eFigure 5a, https://links.lww.com/AA/D544), the 11-item NSQIP modified frailty index (adjusted OR, 1.84; 95% CI,, 1.53–2.21, P < .001; I2 = 76%; Supplemental Digital Content, eFigure 5b, https://links.lww.com/AA/D544), and the ACG frailty defining diagnoses indicator (adjusted OR, 3.39; 95% CI, 2.19–5.26, P < .001; I2 = 99%; Supplemental Digital Content, eFigure 5c, https://links.lww.com/AA/D544).

Predictive accuracy was reported in 9 studies for nonhome discharge, including 4 for the 11-item NSQIP modified frailty index (c-statistic range, 0.59–0.63) and 3 for 5-item NSQIP modified frailty index (c-statistic range, 0.60–0.85). Seven studies reported unadjusted c-statistics, leading to a pooled estimate of 0.65 (95% CI, 0.58–0.71). No studies reported expected number of nonhome discharges, precluding a meta-analysis of calibration. See full predictive accuracy data in Supplemental Digital Content, eTable 5, https://links.lww.com/AA/D544.

Prognostic Value—Costs

Six studies (n = 342,726)94–98,102 reported unadjusted cost (eg, hospital, physicians/clinician billings, laboratory tests, diagnostic procedures, medications, and medical equipment; Supplemental Digital Content, eTable 7, https://links.lww.com/AA/D544) data. The pooled unadjusted standardized mean difference (SMD) for frailty and cost was 4.56 with a 95% CI of −4.78 to 13.91, P = .334; I2 = 100% (Supplemental Digital Content, eFigure 6, https://links.lww.com/AA/D544). Four studies (n = 217,152)94–97 reported an adjusted ratio of means for costs of care, all of which used the ACG frailty-defining diagnosis indicator (adjusted ratio of means, 1.54; 95% CI, 1.46–1.63, P < .001; I2=47%; Supplemental Digital Content, eFigure 7, https://links.lww.com/AA/D544).

Prognostic Value—LOS

Eighteen studies (n = 468,018)44,46,74,82,83,90,94–98,102,113,118–120,123,124 reported unadjusted LOS data, defined as the number of days from surgery to hospital discharge (Supplemental Digital Content, eTable 8, https://links.lww.com/AA/D544). The pooled unadjusted SMD for frailty and LOS was 2.44 with a 95% CI of −0.17 to 5.05 (P = .589; I2 = 100%; Supplemental Digital Content, eFigure 8, https://links.lww.com/AA/D544). Four studies reported predictive data for frailty and prolonged LOS, all using the 11-item NSQIP modified frailty index (c-statistic range, 0.53–0.62). Five studies reported unadjusted c-statistics, leading to a pooled estimate of 0.66 (95% CI, 0.57–0.74; see Supplemental Digital Content, eFigure 7, https://links.lww.com/AA/D544). No studies reported expected number of deaths, precluding a meta-analysis of calibration.

Construct Validity

All instruments identified (n = 22) used comorbidities as a domain to define frailty, with markers of function being the next most prevalent domain. The median number of domains captured was 3 (interquartile range, 2–4), with only 8 of 22 instruments (36%) capturing 4 or more domains contributing to frailty (Supplemental Digital Content, eTable 3, https://links.lww.com/AA/D544). Among the subset of 16 frailty indices identified, the median number of deficits included was 10 (interquartile range, 7–11), and only 4 (25%) included 30 or more deficits. The median number of domains captured was 3 (interquartile range, 2–3), with only 6 (38%) capturing at least 4 domains.


No studies examined feasibility of instruments used or their application in a clinical setting.

Risk of Bias

Risk of bias results according to the Quality in Prognostic Studies tool are reported in Supplemental Digital Content, eTable 9, https://links.lww.com/AA/D544. Overall, forty-eight (53%) studies had low risk, thirty-four (38%) studies had moderate risk, 6 (7%) studies had high risk, and for 2 (2%) studies, risk of bias was unclear. There was low risk of bias across all studies with regard to participation, attrition, prognostic factor measurement, and outcome measurement. The main factors for bias were issues of confounding (combining surgical procedures or urgency category without adjustment) and statistical reporting.


In this systematic review and meta-analysis of 90 studies of frailty instruments operationalized in electronic health data, we found that frailty assessments that were performed using electronic data provided prognostic value for postoperative mortality, discharge location, and resource use, even after adjustment for baseline illness, sex, age, procedure type, and urgency. This suggests that electronic frailty assessment could improve preoperative risk stratification and help to meet best practice guidelines for preoperative assessment of older patients. However, the composition of most electronic frailty instruments was not consistent with the multidimensional nature of frailty. Finally, most instruments have not been widely evaluated, minimal data exist comparing electronic frailty instruments head-to-head, and no data were available to describe feasibility. These knowledge gaps should urgently be addressed to support clinicians and administrators in making informed decisions when choosing a frailty instrument for use in preoperative electronic data.

Sources of electronic health data are growing exponentially; over 80% of US hospitals now have some type of EHR system.125 These systems and the data that they contain could be leveraged to increase preoperative frailty assessment, a best practice recommended since at least 2012126 but which has been rarely implemented.11 However, barriers exist in using electronic data for frailty assessment. First, new prognostic factors deployed in clinical practice should provide additional information above factors already typically considered (such as age, sex, baseline illness and procedural complexity, and urgency).32,127 While no single “typical” set of assessed preoperative risk factors exist, our meta-analyses suggest that frailty, even after adjustment for a set of baseline covariates consistent with other perioperative prognostic studies,30,31 is significantly associated with increased odds of postoperative mortality (3.5-fold), institutional discharge (2.4-fold), and greater resource use (50% relative increase in costs). This means that performing frailty assessment using electronic data should help clinicians to identify patients at increased risk of adverse events independent of their other common risk factors, in addition to helping to provide guideline-based care.

Although electronic frailty assessment appears to add prognostic value, limited data were available to assess the predictive accuracy (eg, discrimination [how well frailty instruments assigned higher risk to people who actually experience bad outcomes] or calibration [how closely predicted risks aligned with actual outcome rates]) of electronic frailty instruments. Studies reported a range of c-statistics, and pooled estimates suggest that frailty instruments provide weak-to-moderate discrimination for predicting death, nonhome discharge, and long LOS. Calibration that may be more directly relevant to preoperative risk quantification was almost never assessed or reported. Future research that follows best practice recommendations for predictive modeling (including adequate internal and external validations)128,129 is urgently needed to determine the predictive accuracy of electronic frailty instruments, the extent to which they improve predictive accuracy beyond typical risk factors, and head-to-head comparisons between different frailty instruments.

While our data support using electronic frailty assessment before surgery, there is far greater uncertainty in the choice of a best or optimal instrument. A multitude of frailty instruments exist, and although some data now suggest that tools like the Clinical Frailty Scale or Edmonton Frail Scale may be the preferred tools when performing prospective, clinical frailty assessment, both of these tools require prospective evaluation and, therefore, cannot be automated in electronic data.10 Of instruments that can be used electronically, an ideal choice should both add prognostic value and reflect the multidimensional constructs accepted to underlie frailty.13 Unfortunately, across included studies, only 36% of identified instruments included variables from at least 4 different domains, meaning that most instruments currently used represent proxies to actual frailty assessment or reflect related concepts (such as multimorbidity). Adherence to multidimensionality matters for at least 3 reasons. First, an assessment labeled frailty should directly reflect the presence of frailty, not a related entity. For example, the 2 most studied instruments, the 11- and 5-item NSQIP modified frailty indices, are more consistent with modified comorbidity indices than multidimensional frailty instruments. Second, recent studies demonstrate that by failing to capture the multidimensional nature of frailty, the NSQIP 5-item instrument provides inferior predictive accuracy compared to a multidimensional instrument (the Risk Analysis Index) and worsened predictive accuracy compared to baseline risk factors alone.130 Third, frailty likely represents an important first step in risk assessment for older people, and recent recommendations suggest that further assessment of deficits within different domains could help to optimize older people with frailty before surgery.5 Finally, while a lack of multidimensionality is a limitation in most instruments, it likely also reflects the fact that many electronic data sources do not routinely capture aspects of cognition, function, or nutrition. This highlights the need to also continue to focus on improving perioperative data sources along with instruments.

Moving forward, future evaluation of electronic frailty instruments should focus on multidimensional tools. Frailty indices derived based on best practice recommendations25 and other multidimensional tools like the Risk Analysis Index are both promising and should be applicable to many electronic data sources. Required evaluations to help select optimal instruments should compare multidimensional frailty instruments head-to-head in different data ecosystems and consider both added prognostic value (ie, effect sizes) measures of added predictive accuracy (eg, discrimination and calibration). Feasibility data and barriers and facilitators to implementing an electronic preoperative frailty assessment are also required. Finally, as emerging evidence suggests that frailty assessment could result in improved outcomes,131 appropriate prospective evaluation of electronic frailty assessment as an intervention to improve outcomes should also be conducted.

Strengths and Limitations

Our study features several strengths. First, we followed a prespecified protocol and adhered to best practices in conducting and reporting our review. Our outcomes were both relevant to patients and the health system and validly ascertained in electronic data. We were inclusive of tools used to identify frailty but evaluated them using preestablished criteria based on consensus and best practices for frailty measurement.13,25 We used random effects models to pool outcomes and focused on adjusted estimates, as recommended. Limitations also exist. The most commonly cited instrument identified (11-item NSQIP index) is no longer supported by NSQIP data collection. As electronic data do not accurately capture functional or cognitive outcomes that patients value highly, our review provides limited insights into patient-oriented prognosis. Heterogeneity was high in all analyses and was not explained by meta-regression. This was likely due to large sample sizes in included studies,34 but true heterogeneity cannot be ruled out. As little agreement is available on how to define validity of frailty instruments, we were limited to assessing content validity, which is only one aspect of the larger concept of construct validity.


Frailty identified using electronic data could improve preoperative risk stratification and increase adherence to best practices in preoperative assessment of older people if multidimensional frailty instruments can be applied to multidimensional data. However, clinicians and policy makers should be cautious of instruments labeled as frailty but that are not consistent with consensus definitions. Future research is needed to evaluate the feasibility of routine electronic frailty assessment before surgery and should compare different multidimensional instruments head-to-head.


All authors acknowledge the support of Ms Sasha Davis (The Ottawa Hospital Information Services) for her assistance in developing the search strategy and The Ottawa Hospital Department of Anesthesiology for support of Distiller SR licenses.


Name: Jamal Alkadri, MD.

Contribution: This author helped contribute to conception; design; data acquisition; analysis; interpretation; and drafting, revising, and approving the final manuscript.

Name: Dima Hage, BSc.

Contribution: This author helped contribute to data acquisition; analysis; interpretation; and drafting, revising, and approving the final manuscript.

Name: Leigh H. Nickerson, MD.

Contribution: This author helped contribute to design; analysis; interpretation; and drafting, revising, and approving the final manuscript.

Name: Lia R. Scott, MD.

Contribution: This author helped contribute to design; data acquisition; interpretation; and drafting, revising, and approving the final manuscript.

Name: Julia F. Shaw, BSc.

Contribution: This author helped contribute to design; data acquisition; analysis; interpretation; and drafting, revising, and approving the final manuscript.

Name: Sylvie D. Aucoin, MD, Msc, FRCPC.

Contribution: This author helped contribute to design; analysis; interpretation; and drafting, revising, and approving the final manuscript.

Name: Daniel I. McIsaac, MD, MPH, FRCPC.

Contribution: This author helped contribute to conception; design; data acquisition; analysis; interpretation; and drafting, revising, and approving the final manuscript and is the guarantor.

This manuscript was handled by: Robert Whittington, MD.


    1. Rockwood K, Theou O, Mitnitski A. What are frailty instruments for? Age Ageing. 2015;44:1–3.
    2. Fried LP, Tangen CM, Walston J, 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:M146–M156.
    3. Rockwood K, Song X, MacKnight C, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ. 2005;173:489–495.
    4. McIsaac DI, Wijeysundera DN, Huang A, Bryson GL, van Walraven C. Association of the hospital volume of frail surgical patients cared for with outcomes after elective, major noncardiac surgery: a retrospective population-based cohort study. Anesthesiology. 2017;126:602–613.
    5. McIsaac DI, MacDonald DB, Aucoin SD. Frailty for perioperative clinicians: a narrative review. Anesth Analg. 2020;130:1450–1460.
    6. Fried LP, Ferrucci L, Darer J, Williamson JD, Anderson G. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. J Gerontol A Biol Sci Med Sci. 2004;59:255–263.
    7. Lin HS, Watts JN, Peel NM, Hubbard RE. Frailty and post-operative outcomes in older surgical patients: a systematic review. BMC Geriatr. 2016;16:157.
    8. Kim DH, Kim CA, Placide S, Lipsitz LA, Marcantonio ER. Preoperative frailty assessment and outcomes at 6 months or later in older adults undergoing cardiac surgical procedures: a systematic review. Ann Intern Med. 2016;165:650–660.
    9. Watt J, Tricco AC, Talbot-Hamon C, et al. Identifying older adults at risk of harm following elective surgery: a systematic review and meta-analysis. BMC Med. 2018;16:2.
    10. Aucoin SD, Hao M, Sohi R, et al. Accuracy and feasibility of clinically applied frailty instruments before surgery: a systematic review and meta-analysis. Anesthesiology. 2020;133:78–95.
    11. Eamer G, Gibson JA, Gillis C, et al. Surgical frailty assessment: a missed opportunity. BMC Anesthesiol. 2017;17:99.
    12. Deiner S, Fleisher LA, Leung JM, Peden C, Miller T, Neuman MD; ASA Committee on Geriatric Anesthesia and the ASA Perioperative Brain Health Initiative. Adherence to recommended practices for perioperative anesthesia care for older adults among US anesthesiologists: results from the ASA Committee on Geriatric Anesthesia-Perioperative Brain Health Initiative ASA member survey. Perioper Med (Lond). 2020;9:6.
    13. Rodríguez-Mañas L, Féart C, Mann G, et al. Searching for an operational definition of frailty: a Delphi method based consensus statement. the frailty operative definition-consensus conference project. J Gerontol A Biol Sci Med Sci. 2013;68:62–67.
    14. Tang VL, Covinsky K. Frailty and surgical decision making. Ann Intern Med. 2016;165:669–670.
    15. Clegg A, Bates C, Young J, et al. Development and validation of an electronic frailty index using routine primary care electronic health record data. Age Ageing. 2016;45:353–360.
    16. Gilbert T, Neuburger J, Kraindler J, et al. Development and validation of a hospital frailty risk score focusing on older people in acute care settings using electronic hospital records: an observational study. Lancet. 2018;391:1775–1782.
    17. McIsaac DI, Wong CA, Huang A, Moloo H, van Walraven C. Derivation and validation of a generalizable preoperative frailty index using population-based health administrative data. Ann Surg. 2019;270:102–108.
    18. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283:2008–2012.
    19. Higgins J, Thomas J, Chandler J, et al. Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (Updated September 2020). 2020.Cochrane;
    20. Riley RD, Moons KGM, Snell KIE, et al. A guide to systematic review and meta-analysis of prognostic factor studies. BMJ. 2019;364:k4597.
    21. Moher D, Liberati A, Tetzlaff J, Altman DG; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097.
    22. McGowan J, Sampson M, Salzwedel DM, Cogo E, Foerster V, Lefebvre C. PRESS peer review of electronic search strategies: 2015 guideline statement. J Clin Epidemiol. 2016;75:40–46.
    23. Bowen DJ, Kreuter M, Spring B, et al. How we design feasibility studies. Am J Prev Med. 2009;36:452–457.
    24. Aucoin S, Hao M, Sohi R, et al. Accuracy and feasibility of clinically applied frailty instruments before surgery: a systematic review and meta-analysis. Anesthesiology. 2020;133:78–95.
    25. Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC Geriatr. 2008;8:24.
    26. Robinson TN, Walston JD, Brummel NE, et al. Frailty for surgeons: review of a National Institute on Aging Conference on Frailty for Specialists. J Am Coll Surg. 2015;221:1083–1092.
    27. Smith T, Li X, Nylander W, Gunnar W. Thirty-day postoperative mortality risk estimates and 1-year survival in veterans health administration surgery patients. JAMA Surg. 2016;151:417–422.
    28. McIsaac DI, Hamilton GM, Abdulla K, et al. Validation of new ICD-10-based patient safety indicators for identification of in-hospital complications in surgical patients: a study of diagnostic accuracy. BMJ Qual Saf. 2020;29:209–216.
    29. Hamilton GM, Wheeler K, Di Michele J, Lalu MM, McIsaac DI. A systematic review and meta-analysis examining the impact of incident postoperative delirium on mortality. Anesthesiology. 2017;127:78–88.
    30. McIsaac DI, Harris EP, Hladkowicz E, et al. Prospective comparison of preoperative predictive performance between 3 leading frailty instruments. Anesth Analg. 2020;131:263–272.
    31. Wijeysundera DN, Pearse RM, Shulman MA, et al.; METS study investigators. Assessment of functional capacity before major non-cardiac surgery: an international, prospective cohort study. Lancet. 2018;391:2631–2640.
    32. Cook NR. Quantifying the added value of new biomarkers: how and how not. Diagn Progn Res. 2018;2:14.
    33. IntHout J, Ioannidis JP, Borm GF. The Hartung-Knapp-Sidik-Jonkman method for random effects meta-analysis is straightforward and considerably outperforms the standard DerSimonian-Laird method. BMC Med Res Methodol. 2014;14:25.
    34. Li S, Jiang H, Yang H, et al. The dilemma of heterogeneity tests in meta-analysis: a challenge from a simulation study. PLoS One. 2015;10:e0127538.
    35. Debray TPA, Damen JAAG, Snell KIE, et al. A guide to systematic review and meta-analysis of prediction model performance. BMJ. 2017;356:i6460.
    36. Vu CCL, Runner RP, Reisman WM, Schenker ML. The frail fail: increased mortality and post-operative complications in orthopaedic trauma patients. Injury. 2017;48:2443–2450.
    37. Shin JI, Kothari P, Phan K, et al. Frailty index as a predictor of adverse postoperative outcomes in patients undergoing cervical spinal fusion. Spine (Phila Pa 1976). 2017;42:304–310.
    38. Augustin T, Schneider E, Alaedeen D, et al. Emergent surgery does not independently predict 30-day mortality after paraesophageal hernia repair: results from the ACS NSQIP database. J Gastrointest Surg. 2015;19:2097–2104.
    39. Flexman AM, Charest-Morin R, Stobart L, Street J, Ryerson CJ. Frailty and postoperative outcomes in patients undergoing surgery for degenerative spine disease. Spine J. 2016;16:1315–1323.
    40. Adams P, Ghanem T, Stachler R, Hall F, Velanovich V, Rubinfeld I. Frailty as a predictor of morbidity and mortality in inpatient head and neck surgery. JAMA Otolaryngol Head Neck Surg. 2013;139:783–789.
    41. Hodari A, Hammoud ZT, Borgi JF, Tsiouris A, Rubinfeld IS. Assessment of morbidity and mortality after esophagectomy using a modified frailty index. Ann Thorac Surg. 2013;96:1240–1245.
    42. Mosquera C, Spaniolas K, Fitzgerald TL. Impact of frailty on surgical outcomes: the right patient for the right procedure. Surgery. 2016;160:272–280.
    43. Mosquera C, Spaniolas K, Fitzgerald TL. Impact of frailty on approach to colonic resection: laparoscopy vs open surgery. World J Gastroenterol. 2016;22:9544–9553.
    44. Augustin T, Burstein MD, Schneider EB, et al. Frailty predicts risk of life-threatening complications and mortality after pancreatic resections. Surgery. 2016;160:987–996.
    45. Congiusta DV, Palvannan P, Merchant AM. The impact of frailty on morbidity and mortality following open emergent colectomies. Biomed Res Int. 2017;2017:5126452.
    46. Palumbo C, Knipper S, Dzyuba-Negrean C, et al. Complication rates, failure to rescue and in-hospital mortality after cytoreductive nephrectomy in the older patients. J Geriatr Oncol. 2020;11:718–723.
    47. Lakomkin N, Zuckerman SL, Stannard B, et al. Preoperative risk stratification in spine tumor surgery: a comparison of the modified Charlson index, frailty index, and ASA score. Spine (Phila Pa 1976). 2019;44:E782–E787.
    48. Ali TZ, Lehman EB, Aziz F. Modified frailty index can be used to predict adverse outcomes and mortality after lower extremity bypass surgery. Ann Vasc Surg. 2018;46:168–177.
    49. Kothari P, Congiusta DV, Merchant AM. Laparoscopic versus open colectomy: the impact of frailty on outcomes. Updates Surg. 2019;71:89–96.
    50. Joseph WJ, Cuccolo NG, Baron ME, Chow I, Beers EH. Frailty predicts morbidity, complications, and mortality in patients undergoing complex abdominal wall reconstruction. Hernia. 2020;24:235–243.
    51. Barbey SM, Scali ST, Kubilis P, et al. Interaction between frailty and sex on mortality after elective abdominal aortic aneurysm repair. J Vasc Surg. 2019;70:1831–1843.
    52. Youngerman BE, Neugut AI, Yang J, Hershman DL, Wright JD, Bruce JN. The modified frailty index and 30-day adverse events in oncologic neurosurgery. J Neurooncol. 2018;136:197–206.
    53. Venkat R, Pandit V, Telemi E, Trofymenko O, Pandian TK, Nfonsam VN. Frailty predicts morbidity and mortality after colectomy for clostridium difficile colitis. Am Surg. 2018;84:628–632.
    54. Telemi E, Trofymenko O, Venkat R, Pandit V, Pandian TK, Nfonsam VN. Frailty predicts morbidity after colectomy for ulcerative colitis. Am Surg. 2018;84:225–229.
    55. Rege RM, Runner RP, Staley CA, Vu CCL, Arora SS, Schenker ML. Frailty predicts mortality and complications in chronologically young patients with traumatic orthopaedic injuries. Injury. 2018;49:2234–2238.
    56. Kessler RA, De la Garza Ramos R, Purvis TE, et al. Impact of frailty on complications in patients with thoracic and thoracolumbar spinal fracture. Clin Neurol Neurosurg. 2018;169:161–165.
    57. Guyton RL Jr, Mosquera C, Spaniolas K, Fitzgerald TL. Association of increasing frailty with detrimental outcomes after pancreatic resection. Am Surg. 2018;84:512–519.
    58. Crozier-Shaw G, Joyce WP. Too frail for surgery? A frailty index in major colorectal surgery. ANZ J Surg. 2018;88:1302–1305.
    59. Leven DM, Lee NJ, Kim JS, et al. Frailty is predictive of adverse postoperative events in patients undergoing lumbar fusion. Global Spine J. 2017;7:529–535.
    60. Akyar S, Armenia S, Ratnani P, Merchant A. The impact of frailty on postoperative cardiopulmonary complications in the emergency general surgery population. Surg J. 2018;04:e66–e77.
    61. Adedayo P, Resnick K, Singh S. Preoperative frailty is a risk factor for non-home discharge in patients undergoing surgery for endometrial cancer. J Geriatr Oncol. 2018;9:513–515.
    62. Runner RP, Bellamy JL, Vu CCL, Erens GA, Schenker ML, Guild GN 3rd. Modified frailty index is an effective risk assessment tool in primary total knee arthroplasty. J Arthroplasty. 2017;32:S177–S182.
    63. Konstantinidis IT, Lewis A, Lee B, et al. Minimally invasive distal pancreatectomy: greatest benefit for the frail. Surg Endosc. 2017;31:5234–5240.
    64. Bellamy JL, Runner RP, Vu CCL, Schenker ML, Bradbury TL, Roberson JR. Modified frailty index is an effective risk assessment tool in primary total hip arthroplasty. J Arthroplasty. 2017;32:2963–2968.
    65. Abt NB, Richmon JD, Koch WM, Eisele DW, Agrawal N. Assessment of the predictive value of the modified frailty index for Clavien-Dindo grade IV critical care complications in major head and neck cancer operations. JAMA Otolaryngol Head Neck Surg. 2016;142:658–664.
    66. Ehlert BA, Najafian A, Orion KC, Malas MB, Black JH 3rd, Abularrage CJ. Validation of a modified frailty index to predict mortality in vascular surgery patients. J Vasc Surg. 2016;63:1595–1601.e2.
    67. Fu MC, Ondeck NT, Nwachukwu BU, et al. What associations exist between comorbidity indices and postoperative adverse events after total shoulder arthroplasty? Clin Orthop Relat Res. 2019;477:881–890.
    68. Seib CD, Rochefort H, Chomsky-Higgins K, et al. Association of patient frailty with increased morbidity after common ambulatory general surgery operations. JAMA Surg. 2018;153:160–168.
    69. Meng X, Press B, Renson A, et al. Discriminative ability of commonly used indexes to predict adverse outcomes after radical cystectomy: comparison of demographic data, American Society of Anesthesiologists, modified Charlson comorbidity index, and modified frailty index. Clin Genitourin Cancer. 2018;16:e843–e850.
    70. Leven DM, Lee NJ, Kothari P, et al. Frailty index is a significant predictor of complications and mortality after surgery for adult spinal deformity. Spine (Phila Pa 1976). 2016;41:E1394–E1401.
    71. Ondeck NT, Bohl DD, Bovonratwet P, McLynn RP, Cui JJ, Grauer JN. Discriminative ability of Elixhauser’s comorbidity measure is superior to other comorbidity scores for inpatient adverse outcomes after total hip arthroplasty. J Arthroplasty. 2018;33:250–257.
    72. Murphy PB, Savage SA, Zarzaur BL. Impact of patient frailty on morbidity and mortality after common emergency general surgery operations. J Surg Res. 2020;247:95–102.
    73. Louwers L, Schnickel G, Rubinfeld I. Use of a simplified frailty index to predict Clavien 4 complications and mortality after hepatectomy: analysis of the National Surgical Quality Improvement Project database. Am J Surg. 2016;211:1071–1076.
    74. Brahmbhatt R, Brewster LP, Shafii S, et al. Gender and frailty predict poor outcomes in infrainguinal vascular surgery. J Surg Res. 2016;201:156–165.
    75. Chimukangara M, Frelich MJ, Bosler ME, Rein LE, Szabo A, Gould JC. The impact of frailty on outcomes of paraesophageal hernia repair. J Surg Res. 2016;202:259–266.
    76. Shin JI, Keswani A, Lovy AJ, Moucha CS. Simplified frailty index as a predictor of adverse outcomes in total hip and knee arthroplasty. J Arthroplasty. 2016;31:2389–2394.
    77. Arya S, Kim SI, Duwayri Y, et al. Frailty increases the risk of 30-day mortality, morbidity, and failure to rescue after elective abdominal aortic aneurysm repair independent of age and comorbidities. J Vasc Surg. 2015;61:324–331.
    78. Traven SA, Reeves RA, Slone HS, Walton ZJ. Frailty predicts medical complications, length of stay, readmission, and mortality in revision hip and knee arthroplasty. J Arthroplasty. 2019;34:1412–1416.
    79. Traven SA, McGurk KM, Reeves RA, Walton ZJ, Woolf SK, Slone HS. Modified frailty index predicts medical complications, length of stay, readmission, and mortality following total shoulder arthroplasty. J Shoulder Elbow Surg. 2019;28:1854–1860.
    80. Subramaniam S, Aalberg JJ, Soriano RP, Divino CM. New 5-factor modified frailty index using American College of Surgeons NSQIP data. J Am Coll Surg. 2018;226:173–181.e8.
    81. Segal DN, Wilson JM, Staley C, Michael KW. The 5-item modified frailty index is predictive of 30-day postoperative complications in patients undergoing kyphoplasty vertebral augmentation. World Neurosurg. 2018;116:e225–e231.
    82. Chen SY, Stem M, Gearhart SL, Safar B, Fang SH, Efron JE. Functional dependence versus frailty in gastrointestinal surgery: are they comparable in predicting short-term outcomes? Surgery. 2018;164:1316–1324.
    83. Chen SY, Stem M, Cerullo M, et al. The effect of frailty index on early outcomes after combined colorectal and liver resections. J Gastrointest Surg. 2018;22:640–649.
    84. Chimukangara M, Helm MC, Frelich MJ, et al. A 5-item frailty index based on NSQIP data correlates with outcomes following paraesophageal hernia repair. Surg Endosc. 2017;31:2509–2519.
    85. Traven SA, Horn RW, Reeves RA, Walton ZJ, Woolf SK, Slone HS. The 5-factor modified frailty index predicts complications, hospital admission, and mortality following arthroscopic rotator cuff repair. Arthroscopy. 2020;36:383–388.
    86. Traven SA, Reeves RA, Sekar MG, Slone HS, Walton ZJ. New 5-factor modified frailty index predicts morbidity and mortality in primary hip and knee arthroplasty. J Arthroplasty. 2019;34:140–144.
    87. Taylor BL, Xia L, Guzzo TJ, Scherr DS, Hu JC. Frailty and greater health care resource utilization following major urologic oncology surgery. Eur Urol Oncol. 2019;2:21–27.
    88. Sathianathen NJ, Jarosek S, Lawrentschuk N, Bolton D, Konety BR. A simplified frailty index to predict outcomes after radical cystectomy. Eur Urol Focus. 2019;5:658–663.
    89. Holzgrefe RE, Wilson JM, Staley CA, Anderson TL, Wagner ER, Gottschalk MB. Modified frailty index is an effective risk-stratification tool for patients undergoing total shoulder arthroplasty. J Shoulder Elbow Surg. 2019;28:1232–1240.
    90. Hampton JP, Owodunni OP, Bettick D, et al. Compliance to an enhanced recovery pathway among patients with a high frailty index after major gastrointestinal surgery results in improved 30-day outcomes. Surgery. 2019;166:75–81.
    91. Balla FM, Yheulon CG, Stetler JL, Patel AD, Lin E, Davis SS. Ventral hernia repair outcomes predicted by a 5-item modified frailty index using NSQIP variables. Hernia. 2019;23:891–898.
    92. AL-Khamis A, Warner C, Park J, et al. Modified frailty index predicts early outcomes after colorectal surgery: an ACS-NSQIP study. Color Dis. 2019;21:1192–1205.
    93. Wilson JM, Holzgrefe RE, Staley CA, Schenker ML, Meals CG. Use of a 5-item modified frailty index for risk stratification in patients undergoing surgical management of distal radius fractures. J Hand Surg Am. 2018;43:701–709.
    94. McIsaac DI, Moloo H, Bryson GL, van Walraven C. The association of frailty with outcomes and resource use after emergency general surgery: a population-based cohort study. Anesth Analg. 2017;124:1653–1661.
    95. McIsaac DI, Beaulé PE, Bryson GL, Van Walraven C. The impact of frailty on outcomes and healthcare resource usage after total joint arthroplasty: a population-based cohort study. Bone Joint J. 2016;98-B:799–805.
    96. Michel J, Goel AN, Golla V, et al. Predicting short-term outcomes after radical cystectomy based on frailty. Urology. 2019;133:25–33.
    97. Goel AN, Lee JT, Gurrola JG 2nd, Wang MB, Suh JD. The impact of frailty on perioperative outcomes and resource utilization in sinonasal cancer surgery. Laryngoscope. 2020;130:290–296.
    98. Asemota AO, Gallia GL. Impact of frailty on short-term outcomes in patients undergoing transsphenoidal pituitary surgery. J Neurosurg. 2020;132:360–370.
    99. Tran DTT, Tu JV, Dupuis JY, Bader Eddeen A, Sun LY. Association of frailty and long-term survival in patients undergoing coronary artery bypass grafting. J Am Heart Assoc. 2018;7:e009882.
    100. Hall DE, Arya S, Schmid KK, et al. Development and initial validation of the risk analysis index for measuring frailty in surgical populations. JAMA Surg. 2017;15213:175–182.
    101. Esses G, Andreopoulos E, Lin HM, Arya S, Deiner S. A comparison of three frailty indices in predicting morbidity and mortality after on-pump aortic valve replacement. Anesth Analg. 2018;126:39–45.
    102. Wilkes JG, Evans JL, Prato BS, Hess SA, MacGillivray DC, Fitzgerald TL. Frailty cost: economic impact of frailty in the elective surgical patient. J Am Coll Surg. 2019;228:861–870.
    103. George EL, Chen R, Trickey AW, et al. Variation in center-level frailty burden and the impact of frailty on long-term survival in patients undergoing elective repair for abdominal aortic aneurysms. J Vasc Surg. 2020;71:46–55.e4.
    104. Lascano D, Pak JS, Kates M, et al. Validation of a frailty index in patients undergoing curative surgery for urologic malignancy and comparison with other risk stratification tools. Urol Oncol. 2015;33:426.e1–426.12.
    105. Kundi H, Popma JJ, Reynolds MR, et al. Frailty and related outcomes in patients undergoing transcatheter valve therapies in a nationwide cohort. Eur Heart J. 2019;40:2231–2239.
    106. Kundi H, Valsdottir LR, Popma JJ, et al. Impact of a claims-based frailty indicator on the prediction of long-term mortality after transcatheter aortic valve replacement in Medicare beneficiaries. Circ Cardiovasc Qual Outcomes. 2018;11:e005048.
    107. De la Garza Ramos R, Goodwin CR, Jain A, et al. Development of a Metastatic Spinal Tumor Frailty Index (MSTFI) using a nationwide database and its association with inpatient morbidity, mortality, and length of stay after spine surgery. World Neurosurg. 2016;95:548–555.e4.
    108. Johnson RL, Abdel MP, Frank RD, Chamberlain AM, Habermann EB, Mantilla CB. Impact of frailty on outcomes after primary and revision total hip arthroplasty. J Arthroplasty. 2019;34:56–64.e5.
    109. Ho B, Lewis A, Paz IB. Laparoscopy can safely be performed in frail patients undergoing colon resection for cancer. Am Surg. 2017;83:1179–1183.
    110. Sacco R, Condoluci A, Curto LS, et al. A new frailty index as a risk predictor of morbidity and mortality: its application in a surgery unit. Eur J Oncol. 2018;23:41–46.
    111. Tomlinson SB, Piper K, Kimmell KT, Vates GE. Preoperative frailty score for 30-day morbidity and mortality after cranial neurosurgery. World Neurosurg. 2017;107:959–965.
    112. Hui Y, van Walraven C, McIsaac DI. How do common comorbidities modify the association of frailty with survival after elective noncardiac surgery? A Population-Based Cohort Study. Anesth Analg. 2019;129:1699–1706.
    113. Ahmed AK, Goodwin CR, De la Garza-Ramos R, et al. Predicting short-term outcome after surgery for primary spinal tumors based on patient frailty. World Neurosurg. 2017;108:393–398.
    114. Gani F, Cerullo M, Amini N, et al. Frailty as a risk predictor of morbidity and mortality following liver surgery. J Gastrointest Surg. 2017;21:822–830.
    115. Ganapathi AM, Englum BR, Hanna JM, et al. Frailty and risk in proximal aortic surgery. J Thorac Cardiovasc Surg. 2014;147:186–191.e1.
    116. Johnson R, Abdel M, Frank R, Chamberlain A, Habermann E, Mantilla C. Frailty index is associated with periprosthetic fracture and mortality after total knee arthroplasty. Orthopedics. 2019;42:335–343.
    117. Ramanathan R, Rieser C, Kurtom S, et al. Simplified preoperative tool predicting discharge destination after major oncologic gastrointestinal surgery. J Surg Oncol. 2020;121:249–257.
    118. Miller EK, Vila-Casademunt A, Neuman BJ, et al.; European Spine Study Group; International Spine Study Group. External validation of the adult spinal deformity (ASD) frailty index (ASD-FI). Eur Spine J. 2018;27:2331–2338.
    119. Miller EK, Neuman BJ, Jain A, et al.; International Spine Study Group. An assessment of frailty as a tool for risk stratification in adult spinal deformity surgery. Neurosurg Focus. 2017;43:E3.
    120. Pandit V, Khan M, Martinez C, et al. A modified frailty index predicts adverse outcomes among patients with colon cancer undergoing surgical intervention. Am J Surg. 2018;216:1090–1094.
    121. Martin GP, Sperrin M, Ludman PF, et al. Do frailty measures improve prediction of mortality and morbidity following transcatheter aortic valve implantation? An analysis of the UK TAVI registry. BMJ Open. 2018;8:e022543.
    122. Kumar A, Langstraat CL, DeJong SR, et al. Functional not chronologic age: frailty index predicts outcomes in advanced ovarian cancer. Gynecol Oncol. 2017;147:104–109.
    123. Yao T, DeJong SR, McGree ME, Weaver AL, Cliby WA, Kumar A. Frailty in ovarian cancer identified the need for increased postoperative care requirements following cytoreductive surgery. Gynecol Oncol. 2019;153:68–73.
    124. Cohan JN, Bacchetti P, Varma MG, Finlayson E. Outcomes after ileoanal pouch surgery in frail and older adults. J Surg Res. 2015;198:327–333.
    125. Adler-Milstein J, Holmgren AJ, Kralovec P, Worzala C, Searcy T, Patel V. Electronic health record adoption in US hospitals: the emergence of a digital “advanced use” divide. J Am Med Inform Assoc. 2017;24:1142–1148.
    126. Chow WB, Rosenthal RA, Merkow RP, Ko CY, Esnaola NF; American College of Surgeons National Surgical Quality Improvement Program; American Geriatrics Society. Optimal preoperative assessment of the geriatric surgical patient: a best practices guideline from the American College of Surgeons National Surgical Quality Improvement Program and the American Geriatrics Society. J Am Coll Surg. 2012;215:453–466.
    127. Walker EMK, Bell M, Cook TM, Grocott MPW, Moonesinghe SR; Central SNAP-1 Organisation; National Study Groups. Patient reported outcome of adult perioperative anaesthesia in the United Kingdom: a cross-sectional observational study. Br J Anaesth. 2016;117:758–766.
    128. Steyerberg EW, Vickers AJ, Cook NR, et al. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010;21:128–138.
    129. Harrell FE. Regression Modeling Strategies. 2015.Springer International Publishing;
    130. McIsaac DI, Aucoin S, van Walraven C. A Bayesian comparison of frailty instruments in non-cardiac surgery: a cohort study. Anesth Analg. 2021;133:366–373
    131. Hall DE, Arya S, Schmid KK, et al. Association of a frailty screening initiative with postoperative survival at 30, 180, and 365 days. JAMA Surg. 2017;152:233–240.

    Supplemental Digital Content

    Copyright © 2021 International Anesthesia Research Society