ABSTRACT
Aim
As the aging population grows, it is crucial to evaluate the quality of surgical care considering geriatric‐specific factors and outcomes. The aim of this study was to develop and validate prediction models for mortality and morbidity in patients aged 65 and older by using a large‐scale, nationwide, real‐world dataset to enable robust and generalizable modeling and to evaluate the contribution of geriatric‐specific risk factors to the model.
Methods
Data from the National Clinical Database in 2021, incorporating 22 geriatric‐specific variables, was used to develop prediction models for 30‐day mortality and major complications using 70% of the dataset (development cohort), with validation on the remaining 30% (validation cohort).
Results
A total of 64 868 cases from gastroenterological surgeries were analyzed. In this geriatric cohort, the 30‐day mortality rate was 1.8%, and the major complication rate was 11.3%. The prediction models demonstrated strong discrimination and calibration (HL = 0.198 and c‐statistics = 0.84 for mortality, HL = 0.0003 and c‐statistics = 0.68 for morbidity). The mortality prediction model identified three significant geriatric‐specific factors as independent predictors: surrogate consent (odds ratio [OR] 1.91; 95% confidence interval [CI] 1.55–2.35), depression (OR 1.68; 95% CI 1.02–2.75), and hospitalization from outside the home (OR 1.24; 95% CI 1.01–1.54), along with age (75–79 years vs. 65–69 years, OR 1.62; 95% CI 1.20–2.19).
Conclusions
Incorporating geriatric‐specific factors along with age identified clinically relevant predictors of mortality and morbidity, reinforcing the importance of geriatric assessment in elderly surgical patients.
Keywords: comprehensive geriatric assessment, geriatric risk model, geriatric surgery
A reliable and effective risk prediction model for mortality and morbidity in elderly patients undergoing gastroenterological surgeries was developed and validated. Geriatric‐specific risk factors, including the newly added variables in the NCD registry, along with age, were identified as significant contributors to the model.
1. Introduction
The global population is aging, with a growing number of individuals over the age of 65 [1]. In Japan, the number of geriatric surgeries is also increasing [2]. This demographic shift heightens the demand for quality surgical care, as older populations generally require more healthcare due to physical decline, cognitive impairments, and social challenges [3, 4]. Given the higher risk of adverse surgical outcomes, there is a growing need for comprehensive geriatric assessments to identify vulnerabilities [5]. These assessments help evaluate a patient's ability to tolerate surgery from various perspectives.
Evaluating the quality of surgical care for elderly patients requires a comprehensive clinical database that includes data on surgical and geriatric outcomes, such as mortality, morbidity, and physical function decline, covering an entire nation or region. In Japan, the National Clinical Database (NCD), established in 2010, includes over 95% of all surgical procedures, collecting preoperative risk factors and tracking adverse outcomes for up to 90 days postoperatively. The accuracy of NCD data has been ensured through systematic audits initiated by the Japanese Society of Gastroenterological Surgery (JSGS) database committee in 2016 [6]. The use of the NCD registry plays a crucial role in assessing the quality of surgical care for elderly patients in Japan, a country experiencing rapid aging. In response to the growing focus on geriatric surgery, the American College of Surgeons (ASC) launched the Geriatric Surgery Pilot Project in 2014. Several studies have shown that incorporating geriatric‐specific variables into clinical databases is crucial for assessing surgical quality in elderly patients [7, 8, 9]. Similarly, the JSGS conducted a geriatric pilot study in 2018 [10], collecting clinical data from over 5000 surgical cases, including geriatric‐specific variables, across 22 academic and community hospitals. This study developed five risk models for geriatric outcomes: (1) postoperative delirium, (2) physical function on postoperative day 30, (3) fall risk at discharge, (4) discharge to a location other than home with social services, and (5) functional decline at discharge. These models were implemented as risk calculators on the NCD web system (http://www.ncd.or.jp). However, the pilot study could not evaluate the relationship between geriatric factors and surgical mortality or morbidity due to the low incidence of these events. To overcome this limitation, 22 geriatric‐specific variables were added to the NCD registry in 2021 (details of these variables are provided in Table S1).
This study used the large‐sale NCD geriatric dataset to develop and validate prediction models for 30‐day mortality and major complications (Clavien–Dindo classification 3b or above). It also examined the influence of geriatric‐specific factors on surgical outcomes through statistical modeling. Additionally, this study is the first to explore the incidence of geriatric‐specific outcomes, such as changes in social environment, functional status, and psychosocial well‐being, across seven major gastroenterological surgeries. The findings are essential for providing patient‐centered care for elderly patients and ensuring the quality of surgical care.
Overall, this study highlights the importance of creating a comprehensive database that includes geriatric‐specific risk factors and outcomes.
2. Methods
The study protocol was approved by the Ethics Committee of Fukushima Medical University in accordance with the Declaration of Helsinki (Approval No. Ippan‐2020‐033).
2.1. Data Source and Patient Selection
Perioperative data for this study were obtained from the NCD for patients who underwent seven major gastroenterological surgeries in 2021: low anterior resection (LAR), pancreaticoduodenectomy (PD), distal gastrectomy (DG), esophagectomy (ESO), hepatectomy (HEP), right hemicolectomy (RC), and total gastrectomy (TG). Clinical data, including at least one entry for geriatric variables and outcomes, were extracted for individuals aged 65 years and older (Figure 1).
FIGURE 1.
Patient flow.
2.2. Preoperative Geriatric‐Specific Variables (Preoperative Risk Factors and Postoperative Outcome Registry in NCD)
The NCD Geriatric Surgery Pilot Project began collecting 22 geriatric‐specific variables for patients aged 65 and older in 2019 among 21 hospitals volunteering to join the project. The NCD Geriatric study based on the project [10] identified 11 of these variables were relevant in the predictive modeling of geriatric outcomes. Consequently, these variables were selected as data items to be widely collected in from all hospitals participating in the NCD. In addition, preoperative depression status, which is increasingly recognized as an important factor in the care of elderly patients [11], was added to the NCD geriatric data collection. In total, 12 geriatric variables are now being collected in the NCD JSGS registry since 2021 (Table S1). Seven preoperative variables are used for comprehensive geriatric assessment: “Origin status from home” and “Preoperative ADL (independent/dependent)” for functional status, “History of dementia” for cognitive status, “Surrogate consent” for decision‐making ability, “Fall history” and “Use of mobility aid” for mobility status, and “Depression” for mental health. These preoperative geriatric‐variables were recorded by the attending surgeon or preoperative nurse based on patient (or family) self‐report and/or a referral letter from the previous physician. Surrogate consent is recorded when a proxy provides consent instead of the patient, regardless of the underlying reason (e.g., cognitive dysfunction, delirium, communication difficulty, or emergency operation). For geriatric‐specific outcomes, six variables were collected in the NCD registry: “Physical function comparing preoperative baseline to 30 days postoperatively”, “Functional status on discharge”, “Fall risk on discharge”, “Postoperative new use of mobility aid”, “Postoperative delirium,” and “hospital destination with or without the need for additional services”. Postoperative delirium is defined as “clinically diagnosed delirium within 30 days postoperatively,” as recorded by the attending physician or nurse based on chart documentation. No standardized assessment tools (e.g., CAM‐ICU, 4AT) were uniformly used.
2.3. Outcome Measures and Creating Risk Models
The surgical outcomes for modeling were 30‐day mortality and major postoperative complications, defined as Clavien–Dindo grade 3b or above [12]. Risk models for these two outcomes were developed using the 2021 NCD dataset, which included geriatric variables. In addition, the six geriatric‐specific outcomes were evaluated for the seven gastroenterological surgeries.
2.4. Statistical Analysis
Descriptive statistics were used to summarize the demographic, clinical, and laboratory characteristics of the study participants. Categorical variables were presented as numbers and percentages. A random sample of 70% of the cohort was selected as the development cohort, with the remaining 30% designated as the validation cohort. Multivariable logistic regression with backward stepwise selection was performed to develop risk models in the development cohort based on potential predictors. Cases with missing values were excluded from the analysis. Age categories (65–74, 75–84, > 85 years) were treated as continuous variables. Other variables included in the analysis were those preselected by the investigators as clinically relevant. The discrimination and calibration of the developed models were assessed by applying them to the testing cohort and predicting the outcomes for each patient. C‐statistic values (areas under the receiver operating characteristic [ROC] curve) were reported, and calibration plots along with the Hosmer–Lemeshow (HL) test were used to assess goodness‐of‐fit. All tests were two‐sided, with p < 0.05 considered statistically significant. Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).
3. Results
3.1. Patient Characteristics, Preoperative Risk Profiles, and Laboratory Data of the Study Population
In 2021, the NCD recorded 761 851 gastroenterological surgery cases across 2367 facilities. From this, 64 868 cases were selected, each containing at least one geriatric variable for patients aged 65 and older (Figure S1). These cases came from 1853 facilities, representing 78% of the NCD's gastroenterological surgery facilities. The study cohort characteristics are summarized in Table 1. The majority of patients in this study cohort were in their 70s (52.7%), with those over 80 accounting for 29.6% of the total. In terms of surgical procedures, the cohort had a higher proportion of PDs (11.5%; 7439 cases) and RHCs (22.6%; 14 691 cases) compared to the overall NCD registry, which includes all age groups [2].
TABLE 1.
Profiles and laboratory data of the entire study population and outcome groups.
| Basic demography | Entire study population | Development cohort | Validation cohort | |||
|---|---|---|---|---|---|---|
| n = 64 868 | n = 45 409 | n = 19 459 | ||||
| No. of patients (%) | No. of patients (%) | No. of patients (%) | ||||
| General variables | ||||||
| Age | ||||||
| 65–69 | 11 495 | 17.7% | 8058 | 17.7% | 3437 | 17.7% |
| 70–74 | 18 704 | 28.8% | 13 155 | 29.0% | 5549 | 28.5% |
| 75–79 | 15 489 | 23.9% | 10 786 | 23.8% | 4703 | 24.2% |
| 80–84 | 11 775 | 18.2% | 8187 | 18.0% | 3588 | 18.4% |
| 85–89 | 5752 | 8.9% | 4055 | 8.9% | 1697 | 8.7% |
| 90 | 1653 | 2.5% | 1168 | 2.6% | 485 | 2.5% |
| Sex | ||||||
| Female | 23 845 | 36.8% | 16 676 | 36.7% | 7169 | 36.8% |
| Body mass index | ||||||
| < 18.5 | 8875 | 13.7% | 6211 | 13.7% | 2664 | 13.7% |
| 18.5‐, < 25 | 42 844 | 66.0% | 30 066 | 66.2% | 12 778 | 65.7% |
| 25‐, < 30 | 11 515 | 17.8% | 7999 | 17.6% | 3516 | 18.1% |
| 30— | 1608 | 2.5% | 1119 | 2.5% | 489 | 2.5% |
| Missing | 1026 | 1.6% | 14 | 0.0% | 12 | 0.1% |
| ASA PS classification | ||||||
| 1 and 2 | 52 219 | 80.5% | 36 489 | 80.4% | 15 730 | 80.8% |
| 3, 4, and 5 | 12 649 | 19.5% | 8920 | 19.6% | 3729 | 19.2% |
| Preoperative ADL | ||||||
| Independent | 59 980 | 92.5% | 41 964 | 92.4% | 18 016 | 92.6% |
| Emergent surgery | 1771 | 2.7% | 1247 | 2.7% | 524 | 2.7% |
| Geriatric variables | ||||||
| Origin status from home | ||||||
| From home | 54 544 | 84.1% | 38 133 | 84.0% | 16 411 | 84.3% |
| Not from home | 4548 | 7.0% | 3232 | 7.1% | 1316 | 6.8% |
| Missing | 5776 | 8.9% | 4044 | 8.9% | 1732 | 8.9% |
| Fall history | ||||||
| Yes | 2034 | 3.1% | 1433 | 3.2% | 601 | 3.1% |
| No | 57 494 | 88.6% | 40 235 | 88.6% | 17 259 | 88.7% |
| Missing | 5340 | 8.2% | 3741 | 8.2% | 1599 | 8.2% |
| History of Dementia | ||||||
| Yes | 3270 | 5.0% | 2347 | 5.2% | 923 | 4.7% |
| No | 56 573 | 87.2% | 39 532 | 87.1% | 17 041 | 87.6% |
| Missing | 5025 | 7.7% | 3530 | 7.8% | 1495 | 7.7% |
| Depression | ||||||
| Yes | 758 | 1.2% | 538 | 1.2% | 220 | 1.1% |
| No | 58 843 | 90.7% | 41 177 | 90.7% | 17 666 | 90.8% |
| Missing | 5267 | 8.1% | 1573 | 3.5% | 3694 | 19.0% |
| Surrogate consent | ||||||
| No; the patients signed his/her own consent | 54 344 | 83.8% | 38 047 | 83.8% | 16 297 | 83.8% |
| Yes; consent is signed by a surrogate | 5386 | 8.3% | 3769 | 8.3% | 1617 | 8.3% |
| Missing | 5138 | 7.9% | 3593 | 7.9% | 1545 | 7.9% |
| Use of mobility aid | ||||||
| Yes | 7629 | 11.8% | 5369 | 11.8% | 2260 | 11.6% |
| No | 54 949 | 84.7% | 38 456 | 84.7% | 16 493 | 84.8% |
| Missing | 2290 | 3.5% | 1584 | 3.5% | 706 | 3.6% |
| Preoperative comorbidity variables | ||||||
| Dyspnea | 1160 | 1.8% | 793 | 1.7% | 367 | 1.9% |
| Respirator | 222 | 0.3% | 152 | 0.3% | 70 | 0.4% |
| COPD | 2976 | 4.6% | 2064 | 4.5% | 912 | 4.7% |
| Hypertension | 33 793 | 52.1% | 23 646 | 52.1% | 10 147 | 52.1% |
| Diabetes mellitus | 16 100 | 24.8% | 11 290 | 24.9% | 4810 | 24.7% |
| Congestive heart failure | 634 | 1.0% | 449 | 1.0% | 185 | 1.0% |
| PVD with a symptom | 265 | 0.4% | 176 | 0.4% | 89 | 0.5% |
| Acute renal failure | 98 | 0.2% | 71 | 0.2% | 27 | 0.1% |
| Hemodialysis | 553 | 0.9% | 402 | 0.9% | 151 | 0.8% |
| Cerebrovascular disease | 4151 | 6.4% | 2928 | 6.4% | 1223 | 6.3% |
| Steroid use | 906 | 1.4% | 617 | 1.4% | 289 | 1.5% |
| Weight loss | 2981 | 4.6% | 2116 | 4.7% | 865 | 4.4% |
| Disseminated cancer | 1301 | 2.0% | 939 | 2.1% | 362 | 1.9% |
| Laboratory values | ||||||
| Hemoglobin < 13.5 mg/dL for males, < 12.5 mg/dl for females | 36 609 | 56.4% | 25 688 | 56.6% | 10 921 | 56.1% |
| Platelet < 15 × 104/ul | 4926 | 7.6% | 3414 | 7.5% | 1512 | 7.8% |
| Albumin < 3.5 g/dL | 17 024 | 26.2% | 11 938 | 26.3% | 5086 | 26.1% |
| AST > 35 IU/L | 6892 | 10.6% | 4812 | 10.6% | 2080 | 10.7% |
| Serum Na < 137 mEq/L | 5656 | 8.7% | 3981 | 8.8% | 1675 | 8.6% |
| BUN > 25 mg/dL | 4926 | 7.6% | 3440 | 7.6% | 1486 | 7.6% |
| PT INR > 1.25 | 2029 | 3.1% | 1449 | 3.2% | 580 | 3.0% |
| Procedure | ||||||
| Distal gastrectomy | 18 483 | 28.5% | 12 902 | 28.4% | 5581 | 28.7% |
| Low anterior resection | 10 602 | 16.3% | 7437 | 16.4% | 3165 | 16.3% |
| Right hemicolectomy | 14 691 | 22.6% | 10 383 | 22.9% | 4308 | 22.1% |
| Total gastrectomy | 6528 | 10.1% | 4594 | 10.1% | 1934 | 9.9% |
| Pancreaticoduodenectomy | 7439 | 11.5% | 5102 | 11.2% | 2337 | 12.0% |
| Esophagectomy | 3279 | 5.1% | 2302 | 5.1% | 977 | 5.0% |
| Hepatectomy | 3846 | 5.9% | 2689 | 5.9% | 1157 | 5.9% |
Abbreviations: ADL, activity of daily living; ASA, American Society of Anesthesia; AST, aspartate aminotransferase; BUN, Blood urea nitrogen; COPD, chronic obstructive pulmonary disease; INR, international normalized ratio; PS, Performance Status; PT, prothrombin time; PVD, peripheral arterial disease.
Regarding geriatric risk factors, 84.1% of patients were admitted to the hospital from home, and 92.5% were functionally independent in activities of daily living (ADL). There were 2034 cases (3.1%) with a history of falls, 7629 cases (11.8%) using mobility aids, 3270 cases (5.0%) with dementia, and 758 cases (1.2%) with depression. Detailed information on these factors for each surgical procedure is provided in Table S2.
3.2. Mortality, Morbidity, and Geriatric‐Specific Outcomes in Seven Major Gastroenterological Surgeries
The 30‐day mortality rate and major complication rate for all procedures were 1.8% and 11.3%, respectively. The incidence of these outcomes and geriatric‐specific outcomes for each surgical procedure is summarized in Table 2. Notably, both PD and ESO had significantly higher rates of major complications (PD 23.7%, ESO 23.6%), while RHC and TG had higher postoperative mortality rates (RHC 2.4%, TG 2.6%) compared to other procedures. The 30‐day mortality rates for these seven major surgeries in elderly patients were higher than those in the NCD registry cohort, which includes all age groups [2]. For the six geriatric‐specific outcomes, physical functional decline on postoperative day 30 was more common in ESO (28.6%) and PD (19.8%). Regarding ADL at discharge, 11.9% of patients across all surgical procedures were either partially independent or totally dependent, with the highest rate observed in RHC (17.5%). The fall risk at discharge was classified as high in 16% of patients across all procedures. New use of a mobility aid at discharge, discharge to a destination other than home, and postoperative delirium occurred in 4.2%, 7.1%, and 19.5% of patients, respectively, across all surgical procedures. Among discharge destinations, ESO (9.0%) had the highest rate of patients being transferred to another hospital or clinic. Patients who underwent RHC showed a higher risk of falls (20.0%), new use of mobility aids (4.9%), and being transferred to a long‐term care facility (1.4%) compared to other procedures.
TABLE 2.
Incidence of surgical and geriatric outcomes according to surgical procedure.
| Surgical outcomes and geriatric outcomes | Surgical procedure | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| LAR (n = 10 602) | % | PD (n = 7439) | % | DG (n = 18 483) | % | ESO (n = 3279) | % | HEP (n = 3846) | % | RHC (n = 14 691) | % | TG (n = 6528) | % | Total (n = 64 868) | % | |
| Postoperative complication of Grade 3b or above on POD30 | 1065 | 10.0 | 1765 | 23.7 | 1393 | 7.5 | 773 | 23.6 | 566 | 14.7 | 996 | 6.8 | 775 | 11.9 | 7333 | 11.3 |
| Postoperative mortality on POD30 | 103 | 1.0 | 146 | 2.0 | 229 | 1.2 | 63 | 1.9 | 97 | 2.5 | 353 | 2.4 | 172 | 2.6 | 1163 | 1.8 |
| Physical function comparing the preoperative baseline to 30 days postoperatively | ||||||||||||||||
| Death at discharge | 63 | 0.6 | 82 | 1.2 | 127 | 0.7 | 29 | 0.9 | 61 | 1.6 | 242 | 1.7 | 89 | 1.4 | 693 | 1.1 |
| Improved physical function | 848 | 8.3 | 325 | 4.6 | 1394 | 7.8 | 125 | 4.0 | 201 | 5.4 | 1442 | 10.2 | 439 | 7.0 | 4774 | 7.7 |
| Diminished physical function | 1190 | 11.7 | 1410 | 19.8 | 2270 | 12.8 | 905 | 28.6 | 474 | 12.7 | 1808 | 12.8 | 1075 | 17.1 | 9132 | 14.6 |
| Similar physical function | 7864 | 77.3 | 5199 | 73.0 | 13 643 | 76.7 | 2046 | 64.7 | 2902 | 78.1 | 10 417 | 73.6 | 4536 | 72.3 | 46 607 | 74.7 |
| Unknown details | 211 | 2.1 | 109 | 1.5 | 349 | 2.0 | 55 | 1.7 | 80 | 2.2 | 246 | 1.7 | 139 | 2.2 | 1189 | 1.9 |
| Functional status on discharge | ||||||||||||||||
| Independent | 9168 | 88.8 | 6392 | 88.7 | 15 540 | 86.5 | 2807 | 88.3 | 3353 | 89.5 | 11 393 | 79.7 | 5453 | 85.9 | 54 106 | 85.8 |
| Partial dependent | 890 | 8.6 | 574 | 8.0 | 1790 | 10.0 | 257 | 8.1 | 243 | 6.5 | 2048 | 14.3 | 589 | 9.3 | 6391 | 10.1 |
| Total dependent | 123 | 1.2 | 57 | 0.8 | 308 | 1.7 | 28 | 0.9 | 25 | 0.7 | 451 | 3.2 | 110 | 1.7 | 1102 | 1.7 |
| Death at discharge | 88 | 0.9 | 137 | 1.9 | 208 | 1.2 | 58 | 1.8 | 94 | 2.5 | 331 | 2.3 | 161 | 2.5 | 1077 | 1.7 |
| Unknown details | 61 | 0.6 | 47 | 0.7 | 111 | 0.6 | 29 | 0.9 | 33 | 0.9 | 79 | 0.6 | 32 | 0.5 | 392 | 0.6 |
| Fall risk on discharge | ||||||||||||||||
| High risk | 1381 | 13.5 | 1063 | 14.9 | 2807 | 15.7 | 503 | 15.9 | 477 | 12.9 | 2828 | 20.0 | 968 | 15.4 | 10 027 | 16.1 |
| Low risk | 6302 | 61.6 | 4422 | 62.1 | 10 665 | 59.8 | 2011 | 63.4 | 2304 | 62.2 | 7905 | 55.9 | 3660 | 58.3 | 37 269 | 59.7 |
| No fall risk evaluation performed | 997 | 9.7 | 678 | 9.5 | 1816 | 10.2 | 277 | 8.7 | 407 | 11.0 | 1426 | 10.1 | 676 | 10.8 | 6277 | 10.0 |
| Not applicable | 1547 | 15.1 | 957 | 13.4 | 2541 | 14.3 | 379 | 12.0 | 518 | 14.0 | 1982 | 14.0 | 976 | 15.5 | 8900 | 14.2 |
| Postoperative new use of mobility aid | ||||||||||||||||
| No | 9897 | 96.2 | 6862 | 95.7 | 17 132 | 96.1 | 3057 | 96.4 | 3583 | 96.2 | 13 513 | 95.1 | 6002 | 95.3 | 60 046 | 95.8 |
| Yes | 390 | 3.8 | 305 | 4.3 | 696 | 3.9 | 113 | 3.6 | 143 | 3.8 | 697 | 4.9 | 296 | 4.7 | 2640 | 4.2 |
| Discharge destination | ||||||||||||||||
| Discharge to home | 9594 | 92.4 | 6627 | 91.6 | 16 500 | 91.4 | 2800 | 87.6 | 3468 | 92.2 | 12 427 | 86.6 | 5719 | 89.9 | 57 135 | 90.2 |
| Transferring to a different hospital or clinic | 394 | 3.8 | 354 | 4.9 | 821 | 4.5 | 287 | 9.0 | 151 | 4.0 | 861 | 6.0 | 317 | 5.0 | 3185 | 5.0 |
| Transferring to a long‐term care health facility | 221 | 2.1 | 43 | 0.6 | 352 | 1.9 | 12 | 0.4 | 23 | 0.6 | 530 | 3.7 | 103 | 1.6 | 1284 | 2.0 |
| Transferring to another ward in the same hospital | 71 | 0.7 | 56 | 0.8 | 122 | 0.7 | 22 | 0.7 | 20 | 0.5 | 167 | 1.2 | 51 | 0.8 | 509 | 0.8 |
| Death at discharge | 91 | 0.9 | 139 | 1.9 | 217 | 1.2 | 60 | 1.9 | 94 | 2.5 | 332 | 2.3 | 162 | 2.5 | 1095 | 1.7 |
| Unknown details | 17 | 0.2 | 16 | 0.2 | 42 | 0.2 | 14 | 0.4 | 7 | 0.2 | 41 | 0.3 | 10 | 0.2 | 147 | 0.2 |
| Postoperative delirium | ||||||||||||||||
| No | 8474 | 81.7 | 6047 | 82.7 | 14 619 | 80.9 | 2640 | 81.9 | 3089 | 81.8 | 11 347 | 78.9 | 5159 | 80.4 | 51 375 | 80.8 |
| Yes | 1899 | 18.4 | 1261 | 18.6 | 3451 | 19.2 | 585 | 18.4 | 685 | 18.9 | 3040 | 21.1 | 1257 | 19.8 | 12 178 | 19.5 |
Note: The number of cases classified as “death at discharge” differs across the listed outcome categories. These discrepancies may result from differences in the timing of outcome measurement (in‐hospital death vs. 30‐day postoperative death) and may also be partially influenced by incomplete registry documentation or missing data.
Abbreviations: LAR, low anterior resection; PD, pancreaticoduodenectomy; DG, distal gastrectomy; ESO, esophagectomy; HEP, hepatectomy; RHC, right‐hemi colectomy; and TG, total gastrectomy.
3.3. Importance of Geriatric‐Specific Risk Factors in Predicting 30‐Day Mortality or Major Postoperative Complications
The impact of geriatric factors on the prediction model for 30‐day mortality or major complications is shown in Tables 3, S3 and Figure 2. Figure 2 highlights that all six factors negatively impacted postoperative mortality across all age groups. Notably, a significant difference in mortality rates was observed between those with or without “Surrogate consent” and “Hospitalization from outside the home.”
TABLE 3.
Risk models of mortality and postoperative complications.
| The 30‐day mortality | Major complication (Clavien–Dindo 3b or above) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Term | B. coefficient | p | OR | 95% CI | B. coefficient | p | OR | 95% CI | |
| Age | 65–69 | Reference | — | — | — | ||||
| 70–74 | 0.181 | 0.2513 | 1.20 | 0.88–1.63 | |||||
| 75–79 | 0.481 | 0.0019 | 1.62 | 1.20–2.19 | |||||
| 80–84 | 0.744 | < 0.0001 | 2.10 | 1.55–2.86 | |||||
| 85–89 | 0.957 | < 0.0001 | 2.60 | 1.86–3.64 | |||||
| 90— | 1.329 | < 0.0001 | 3.78 | 2.53–5.63 | |||||
| Sex | Male | Reference | — | — | — | Reference | — | — | — |
| Female | −0.4948 | < 0.0001 | 0.61 | 0.51–0.73 | −0.470 | < 0.001 | 0.625 | 0.56–0.70 | |
| ASA PS classification: 1 and 2 | 1 and 2 | Reference | — | — | Reference | — | — | — | |
| 3, 4, and 5 | 3, 4, and 5 | 0.515 | < 0.001 | 1.67 | 1.41–1.99 | 0.370 | < 0.001 | 1.447 | 1.29–1.62 |
| Preoperative ADL | Independent | Reference | — | — | Reference | — | — | — | |
| Dependent | 0.656 | < 0.001 | 1.93 | 1.56–2.38 | 0.328 | < 0.001 | 1.389 | 1.18–1.63 | |
| Dyspnea | No symptom | Reference | — | — | |||||
| Resting dyspnea | 0.975 | 0.0014 | 2.65 | 1.46–4.83 | |||||
| Effort dyspnea | 0.546 | 0.0069 | 1.73 | 1.16–2.57 | |||||
| Ventilator dependent within 48 h | No | Reference | — | — | Reference | — | — | — | |
| Yes | 1.166 | < 0.0001 | 3.21 | 1.86–5.54 | 0.979 | < 0.001 | 2.663 | 1.70–4.17 | |
| COPD | No | Reference | — | — | — | ||||
| Yes | 0.346 | < 0.001 | 1.409 | 1.17–1.70 | |||||
| Congestive heart failure within 30 days | No | Reference | — | — | |||||
| Yes | 0.587 | 0.004 | 1.80 | 1.21–2.68 | |||||
| Hypertension with no treatment | No | Reference | — | — | — | ||||
| Yes | 0.013 | 0.9347 | 1.013 | 0.74–1.39 | |||||
| Hypertension with medication | No | Reference | — | — | — | ||||
| Yes | 0.170 | 0.0007 | 1.185 | 1.07–1.31 | |||||
| Previous PVD with symptom | No | Reference | — | — | |||||
| Yes | 0.626 | 0.0832 | 1.87 | 0.92–3.80 | |||||
| Hemodialysis within 14 days | no | Reference | — | — | Reference | — | — | — | |
| Yes | 1.266 | < 0.0001 | 3.55 | 2.40–5.25 | 0.717 | < 0.001 | 2.048 | 1.48–2.82 | |
| Cerebrovascular disease | No | Reference | — | — | |||||
| COMA lasting more than 24 h | 1.362 | 0.162 | 3.90 | 0.58–26.3 | |||||
| CVA lasting more than 72 h | −0.462 | 0.1401 | 0.63 | 0.34–1.16 | |||||
| RIND | 0.795 | 0.0411 | 2.214 | 1.03–4.75 | |||||
| TIA | 0.583 | 0.1045 | 1.792 | 0.89–3.62 | |||||
| Unknown details | 0.010 | 0.9509 | 1.01 | 0.74–1.38 | |||||
| Depression | No | Reference | — | — | |||||
| Yes | 0.517 | 0.0401 | 1.68 | 1.02–2.75 | |||||
| Surrogate consent | No | Reference | — | — | Reference | — | — | — | |
| Yes | 0.647 | < 0.0001 | 1.91 | 1.55–2.35 | 0.415 | < 0.001 | 1.52 | 1.30–1.76 | |
| Origin status | Hospitalization from home | Reference | — | — | Reference | — | — | — | |
| Hospitalization from outside the home | 0.217 | 0.0447 | 1.24 | 1.01–1.54 | 0.243 | 0.0022 | 1.28 | 1.09–1.49 | |
| Emergent surgery | No | Reference | — | — | Reference | — | — | — | |
| Yes | 1.419 | < 0.0001 | 4.13 | 3.22–5.31 | 1.114 | < 0.001 | 3.047 | 2.50–3.71 | |
| Advanced cancer with dissemination | No | Reference | — | — | Reference | — | — | — | |
| Yes | 1.236 | < 0.0001 | 3.44 | 2.52–4.71 | 0.499 | 0.0002 | 1.647 | 1.27–2.14 | |
| Weight loss | No | Reference | — | — | |||||
| Yes | 0.567 | < 0.0001 | 1.76 | 1.36–2.28 | |||||
| Platelet < 15 × 104/ul | No | Reference | — | — | Reference | — | — | — | |
| Yes | 0.408 | 0.0003 | 1.50 | 1.20–1.88 | 0.241 | 0.0024 | 1.272 | 1.09–1.49 | |
| Albumin < 3.5 g/dl | No | Reference | — | — | Reference | — | — | — | |
| Yes | 0.584 | < 0.0001 | 1.79 | 1.51–2.14 | 0.341 | < 0.001 | 1.407 | 1.26–1.57 | |
| AST > 35 IU/L | No | Reference | — | — | Reference | — | — | — | |
| Yes | 0.464 | < 0.0001 | 1.59 | 1.31–1.94 | 0.205 | 0.0041 | 1.228 | 1.07–1.41 | |
| BUN > 25 mg/dL | No | Reference | — | — | Reference | — | — | — | |
| Yes | 0.385 | 0.0003 | 1.47 | 1.19–1.81 | 0.301 | < 0.001 | 1.351 | 1.16–1.57 | |
| PT INR > 1.25 | No | Reference | — | — | Reference | — | — | — | |
| Yes | 0.571 | < 0.0001 | 1.77 | 1.37–2.30 | 0.458 | < 0.001 | 1.581 | 1.30–1.92 | |
| Type of surgery | Distal gastrectomy | Reference | — | — | Reference | — | — | — | |
| Low anterior resection | −0.027 | 0.8693 | 0.97 | 0.71–1.34 | 0.834 | < 0.001 | 2.302 | 1.98–2.67 | |
| Pancreaticoduodenectomy | 1.030 | < 0.0001 | 2.80 | 2.12–3.71 | 0.744 | < 0.001 | 2.105 | 1.77–2.50 | |
| Esophagectomy | 1.050 | < 0.0001 | 2.86 | 1.96–4.15 | 1.277 | < 0.001 | 3.585 | 2.97–4.32 | |
| Hepatectomy | 1.314 | < 0.0001 | 3.72 | 2.72–5.09 | 0.407 | 0.0006 | 1.502 | 1.19–1.89 | |
| Right hemicolectomy | −0.079 | 0.5201 | 0.92 | 0.73–1.18 | −0.076 | 0.36 | 0.927 | 0.79–1.09 | |
| Total gastrectomy | 0.839 | < 0.0001 | 2.31 | 1.77–3.02 | 0.511 | < 0.001 | 1.666 | 1.39–1.99 | |
| Intercept | −5.92 | < 0.001 | — | — | −3.92 | < 0.001 | — | — | |
| H‐L score | 0.198 | 0.000277 | |||||||
| C‐index | 0.84 (0.82–0.86) | 0.68 (0.66–0.70) | |||||||
Abbreviations: ADL, activity of daily living; ASA, American Society of Anesthesia; AST, aspartate aminotransferase; BUN, Blood urea nitrogen; CI, confidence interval; COPD, chronic obstructive pulmonary disease; INR, International normalized ratio; OR, odds ratio; PS, Performance Status; PT, prothrombin time.
FIGURE 2.
Impact of geriatric factors on 30‐day mortality. The 30‐day mortality rates for patients with and without a geriatric factor are shown by age category.
Risk models for 30‐day mortality and major complications were developed using multivariate logistic analysis (Table 3). The model for 30‐day mortality included 22 predictors, with three geriatric‐specific variables identified as significant: surrogate consent (odds ratio [OR] 1.91; 95% confidence interval [CI] 1.55–2.35), depression (OR 1.68; 95% CI 1.02–2.75), and hospitalization from outside the home (OR 1.24; 95% CI 1.01–1.54). The model for major complications retained 16 predictors, with two geriatric‐specific variables as significant predictors: surrogate consent (OR 1.52; 95% CI 1.30–1.76) and hospitalization from outside the home (OR 1.28; 95% CI 1.09–1.49).
The C‐index, representing the area under the ROC curve, and the 95% CI were calculated using the validation cohort. The C‐indices of 30‐day mortality and major complications were 0.84 (95% CI 0.82–0.86) and 0.68 (95% CI 0.66–0.70), respectively (Figure 3), which were similar to those of established risk models developed using NCD registry data without geriatric‐specific variables [13, 14, 15, 16, 17, 18, 19, 20].
FIGURE 3.
ROC curve and calibration plots of the developed risk models.
Calibration plots for the risk model predicting each outcome are shown in Figure 3. The predicted surgical outcomes closely matched the actual probabilities across both low‐ and high‐risk groups, demonstrating consistent calibration in both models.
4. Discussion
This study developed and validated risk models for 30‐day mortality and major complications in elderly patients undergoing seven major gastroenterological surgeries, using a nationwide surgical registry that covers almost all of Japan. The models showed good predictive performance and reliability. The discrimination performance for predicting 30‐day mortality was high, with a C‐index of 0.84 (95% CI 0.82–0.86), comparable to established models based on general variables rather than geriatric‐specific ones [13, 14, 15, 16, 17, 18, 19, 20]. In the mortality prediction model, three geriatric‐specific factors—surrogate consent (OR 1.91; 95% CI 1.55–2.35), depression (OR 1.68; 95% CI 1.02–2.75), and hospitalization from outside the home (OR 1.24; 95% CI 1.01–1.54)—along with age, were identified as independent predictors. The impact of the three factors on prediction was similar to that of patient age, with those aged 75–79 years having a higher OR (OR 1.62; 95% CI 1.20–2.19) compared to those aged 65–69 years. In addition, this study provides real‐world evidence on geriatric‐specific outcomes, highlighting potential changes in patients' physical functional and social environment following each procedure in Japan.
A key finding of this study is that geriatric‐specific risk factors, such as preoperative depression (mental state), surrogate consent (cognitive status), and hospitalization from outside the home (origin status), negatively impacted 30‐day mortality in major gastroenterological surgeries. While other geriatric risk factors, including fall history, dementia, and use of mobility aids, were not identified as independent prognosticators, a correlation with postoperative mortality or complications was observed in univariate analysis (Table S3). The NCD Geriatric Surgery Pilot Project [10] previously indicated a close relationship between geriatric‐specific factors and outcomes, but due to its small sample size, the association with 30‐day mortality or major complications was not fully explored. This study helps address the limitations of the initial NCD Geriatric Surgery Pilot Project.
Patient age is widely recognized as a strong predictor of postoperative mortality and is included in the established NCD risk calculators for all seven major gastroenterological surgeries. In this study, as expected, age was identified as an independent predictor of 30‐day mortality but not of major postoperative complications. When comparing the contribution of age and geriatric‐specific risk factors to the prediction using ORs in the developed risk model, the ORs for the three geriatric‐specific risk factors were similar to that of age for patients aged 75–79 years (Table 3). In addition, the 30‐day mortality rate was higher for those with a geriatric risk factor across all age groups, as shown in Figure 2, indicating that geriatric‐specific factors add to the risk stratification for surgical outcomes in elderly patients.
Studies have demonstrated that frailty associated with aging is an independent risk factor for surgical outcomes [21, 22]. However, there is limited evidence on which geriatric‐specific factors are significant for postoperative mortality or morbidity. Brian et al. investigated the impact of geriatric‐specific risk factors on adjusting risk for 30‐day mortality and serious morbidity or mortality in general–vascular and orthopedic surgery, finding that surrogate consent, fall history, and preoperative use of mobility aids were key factors [9]. The value of this study is in identifying which geriatric factors are most important for predicting mortality or morbidity in gastroenterological surgeries. Among the three risk factors, surrogate consent and hospitalization from outside the home were also found to be independent predictors for major complications (Table 3). As shown in Figure 2, the difference in mortality rates between patients with and without “Surrogate consent” or “Hospitalization form outside the home” appeared more pronounced than other factors. Surrogate consent may indicate significant cognitive decline beyond a history of dementia and could be linked to declines in physical function, making it a potentially important indicator of vulnerability in elderly patients. The discrepancy between the prevalence of surrogate consent (8.3%) and dementia (5.0%), as shown in Table 1, suggests that surrogate consent may capture not only severe cognitive impairment but also other conditions, such as delirium, emergency surgery, or cultural factors in Japan where family involvement in medical decision‐making is common. Regarding preoperative depression, which was newly added to the NCD registry's geriatric dataset, this study is the first to show that it is an independent predictor of postoperative surgical outcomes. It is well‐known that postoperative depression is a common cause of morbidity or mortality following surgery [23]. Preoperative depression is recognized as a risk factor for postoperative depression [11], which can negatively affect healthcare outcomes after surgery. The results emphasize the importance of assessing preoperative depressive status and support its inclusion in the NCD registry. These findings highlight that the addition of geriatric‐specific variables to the NCD registry is crucial for a thorough geriatric assessment.
In this study, the prevalence of preoperative geriatric factors was lower than that reported in community‐based studies. For example, the prevalence of depression in our cohort was 1.2%, whereas community‐based studies of older adults using standardized assessment tools such as the Geriatric Depression Scale have reported rates of 23.4%–34.2% [24]. Because the NCD data collection relies on medical record abstraction or attending surgeon assessment rather than standardized questionnaires, underreporting is possible. Similarly, the prevalence of fall history in our cohort (3.1%) was markedly lower than that reported in a community‐based study (18.3%) [25] as well as a prior surgical geriatric cohort (20.0%) [9]. Taken together, the lower prevalence of geriatric factors in our cohort likely reflects both selection bias toward healthier surgical candidates and limited ascertainment sensitivity. This study is the first to provide real‐world data on geriatric outcomes for each of the seven major gastroenterological surgeries in Japan (Table 2). For physical function changes after surgery, ESO (28.64%) and PD (19.79%) showed higher rates of diminished physical function compared to the other procedures. ESO also had the highest rate of patients transferred to a different hospital or clinic (8.98%). Unexpectedly, postoperative outcomes for RHC were worse than for other procedures, with higher mortality (2.4%) and fall risk (20.0%). It is important to consider that RHC patients were more frail, with higher rates of total dependence before surgery or surrogate consent (Table S2). This evidence underscores the potential changes in physical function or the need for social support after surgery, which is important for patient‐centered care, particularly in shared decision‐making before surgery [26, 27]. Since elderly patients are at a higher risk of functional decline or loss of independence postoperatively, healthcare providers must focus on the outcomes most important to these patients. Some elderly individuals may prioritize overall well‐being and quality of life over the length of life after surgery. Traditionally, surgical quality has been evaluated based on prognosis and perioperative morbidity and mortality [13, 14, 15, 16, 17, 18, 19, 20]. To ensure high‐quality surgery for elderly patients, it is essential to evaluate outcomes using geriatric metrics, such as postoperative functional status (ADL, need for new mobility aids), psychosocial well‐being (postoperative depression), and social environment (discharge destination). During shared decision‐making, the medical team must provide sufficient information to support collaborative decisions, including not only the potential benefits of surgery, complications, and mortality but also the risks of functional decline and the need for life support.
This study has several limitations. First, while this study conducted internal validation of the newly developed risk models, external validation using an independent dataset is crucial. Validation with a database outside Japan is also essential to evaluate the model's generalizability. Second, restricting inclusion to patients with at least one geriatric variable recorded may have introduced information bias. Specifically, cases with more comprehensive recording might reflect institutions with more attention to frailty or complexity, which could influence outcome estimates. Regarding missing data, Table S4 summarizing the proportion of missing data for each variable showed missingness was not concentrated in specific surgical or demographic groups. Third, while it examined geriatric‐specific outcomes at discharge, it is likely that these outcomes are related to long‐term results after discharge. Brian et al. identified a correlation between loss of independence at discharge and increased risks of readmission and mortality following surgical procedures in older patients [28]. Further research is needed to assess long‐term outcomes after discharge in elderly patients who have undergone surgery in Japan.
5. Conclusion
Risk models for 30‐day mortality and major complications in elderly patients undergoing gastroenterological surgery were developed and validated, demonstrating strong predictive performance and reliability. In the mortality model, three geriatric‐specific risk factors—“Hospitalization from outside the home”, “Surrogate consent,” and “Depression”—along with age, were identified as significant contributors. Incorporating geriatric‐specific factors along with age identified clinically relevant predictors of mortality and morbidity, reinforcing the importance of geriatric assessment in elderly surgical patients. Overall, this study highlights the importance of creating a comprehensive database that incorporates geriatric‐specific risk factors and outcomes.
Author Contributions
Naoya Sato: investigation, validation, writing – original draft, writing – review and editing, methodology, data curation. Hiraku Kumamaru: conceptualization, methodology, software, data curation, validation, formal analysis, supervision, writing – review and editing. Mitsukazu Gotoh: conceptualization, methodology, data curation, funding acquisition, writing – review and editing, validation. Yoshihiro Kakeji: conceptualization, investigation, writing – review and editing, resources, validation, funding acquisition. Yuko Kitagawa: conceptualization, validation, investigation, writing – review and editing, resources, funding acquisition. Yasuyuki Seto: conceptualization, investigation, funding acquisition, validation, resources, writing – review and editing. Hiromi Rakugi: writing – review and editing, conceptualization, investigation, funding acquisition, validation, resources. Masahiro Akishita: writing – review and editing, validation, formal analysis, investigation, conceptualization. Kazue Nakajima: investigation, conceptualization, writing – review and editing, data curation, validation. Arata Takahashi: methodology, software, data curation, validation, visualization, formal analysis. Hiroaki Miyata: methodology, software, data curation, formal analysis, validation, supervision. Shigeru Marubashi: supervision, project administration, conceptualization, investigation, visualization, funding acquisition, writing – review and editing, methodology, data curation, formal analysis, validation.
Funding
This work was funded by the Health and Labor Sciences Research Grants (21EA0401).
Ethics Statement
The study protocol was approved by the Japanese Society of Gastroenterological Surgery and the Ethics Committee of Kobe University (approval number 20190128).
Conflicts of Interest
Yuko Kitagawa is Editor in Chief of Annals of Gastroenterological Surgery. Yoshihiko Kakeji is an editorial board member of the Annals of Gastroenterological Surgery. Hiraku Kumamaru, Arata Takahashi, and Hiroaki Miyata are affiliated with the Department of Healthcare Quality Assessment at the University of Tokyo. The department is a social collaboration department supported by National Clinical Database, Johnson & Johnson K.K., Nipro Corporation, and Intuitive Surgical Sàrl.
Supporting information
Table S1: Geriatric surgery variables and definitions.
Table S2: Detailed information on preoperative geriatric factors by types of surgical procedure.
Table S3: Relationship between geriatric factors and postoperative outcomes.
Table S4: Number of missing data by age and surgical procedure.
Acknowledgments
We appreciate the data managers and hospitals involved in the NCD project for their ongoing efforts in registering surgical and other medical data.
Data Availability Statement
Research data are not shared. We used data from the National Clinical Database (NCD) registry in Japan. The NCD committee prohibits us from sharing the data with anyone outside this research group.
References
- 1. Beard J. R., Officer A., de Carvalho I. A., et al., “The World Report on Ageing and Health: A Policy Framework for Healthy Ageing,” Lancet 387, no. 10033 (2016): 2145–2154. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Marubashi S., Takahashi A., Kakeji Y., et al., “Surgical Outcomes in Gastroenterological Surgery in Japan: Report of the National Clinical Database 2011‐2019,” Annals of Gastroenterological Surgery 5, no. 5 (2021): 639–658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Zietlow K. E., Wong S., Heflin M. T., et al., “Geriatric Preoperative Optimization: A Review,” American Journal of Medicine 135, no. 1 (2022): 39–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Berian J. R., Rosenthal R. A., Baker T. L., et al., “Hospital Standards to Promote Optimal Surgical Care of the Older Adult: A Report From the Coalition for Quality in Geriatric Surgery,” Annals of Surgery 267, no. 2 (2018): 280–290. [DOI] [PubMed] [Google Scholar]
- 5. Mohanty S., Rosenthal R. A., Russell M. M., Neuman M. D., Ko C. Y., and Esnaola N. F., “Optimal Perioperative Management of the Geriatric Patient: A Best Practices Guideline From the American College of Surgeons NSQIP and the American Geriatrics Society,” Journal of the American College of Surgeons 222, no. 5 (2016): 930–947. [DOI] [PubMed] [Google Scholar]
- 6. Kanaji S., Takahashi A., Miyata H., et al., “Initial Verification of Data From a Clinical Database of Gastroenterological Surgery in Japan,” Surgery Today 49, no. 4 (2019): 328–333. [DOI] [PubMed] [Google Scholar]
- 7. Hornor M. A., Ma M., Zhou L., et al., “Enhancing the American College of Surgeons NSQIP Surgical Risk Calculator to Predict Geriatric Outcomes,” Journal of the American College of Surgeons 230, no. 1 (2020): 88–100. [DOI] [PubMed] [Google Scholar]
- 8. Robinson T. N. and Rosenthal R. A., “The ACS NSQIP Geriatric Surgery Pilot Project: Improving Care for Older Surgical Patients,” Bulletin of the American College of Surgeons 99, no. 10 (2014): 21–23. [PubMed] [Google Scholar]
- 9. Berian J. R., Zhou L., Hornor M. A., et al., “Optimizing Surgical Quality Datasets to Care for Older Adults: Lessons From the American College of Surgeons NSQIP Geriatric Surgery Pilot,” Journal of the American College of Surgeons 225, no. 6 (2017): 702–712. [DOI] [PubMed] [Google Scholar]
- 10. Kofunato Y., Takahashi A., Gotoh M., et al., “Geriatric Risk Prediction Models for Major Gastroenterological Surgery Using the National Clinical Database in Japan: A Multicenter Prospective Cohort Study,” Annals of Surgery 275 (2020): 1112–1120. [DOI] [PubMed] [Google Scholar]
- 11. Geoffrion R., Koenig N. A., Zheng M., et al., “Preoperative Depression and Anxiety Impact on Inpatient Surgery Outcomes: A Prospective Cohort Study,” Annals of Surgury Open 2, no. 1 (2021): e049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Clavien P. A., Barkun J., de Oliveira M. L., et al., “The Clavien‐Dindo Classification of Surgical Complications: Five‐Year Experience,” Annals of Surgery 250, no. 2 (2009): 187–196. [DOI] [PubMed] [Google Scholar]
- 13. Nakagoe T., Miyata H., Gotoh M., et al., “Surgical Risk Model for Acute Diffuse Peritonitis Based on a Japanese Nationwide Database: An Initial Report on the Surgical and 30‐Day Mortality,” Surgery Today 45, no. 10 (2015): 1233–1243. [DOI] [PubMed] [Google Scholar]
- 14. Takeuchi H., Miyata H., Gotoh M., et al., “A Risk Model for Esophagectomy Using Data of 5354 Patients Included in a Japanese Nationwide Web‐Based Database,” Annals of Surgery 260, no. 2 (2014): 259–266. [DOI] [PubMed] [Google Scholar]
- 15. Watanabe M., Miyata H., Gotoh M., et al., “Total Gastrectomy Risk Model: Data From 20,011 Japanese Patients in a Nationwide Internet‐Based Database,” Annals of Surgery 260, no. 6 (2014): 1034–1039. [DOI] [PubMed] [Google Scholar]
- 16. Kurita N., Miyata H., Gotoh M., et al., “Risk Model for Distal Gastrectomy When Treating Gastric Cancer on the Basis of Data From 33,917 Japanese Patients Collected Using a Nationwide Web‐Based Data Entry System,” Annals of Surgery 262, no. 2 (2015): 295–303. [DOI] [PubMed] [Google Scholar]
- 17. Kobayashi H., Miyata H., Gotoh M., et al., “Risk Model for Right Hemicolectomy Based on 19,070 Japanese Patients in the National Clinical Database,” Journal of Gastroenterology 49, no. 6 (2014): 1047–1055. [DOI] [PubMed] [Google Scholar]
- 18. Matsubara N., Miyata H., Gotoh M., et al., “Mortality After Common Rectal Surgery in Japan: A Study on Low Anterior Resection From a Newly Established Nationwide Large‐Scale Clinical Database,” Diseases of the Colon and Rectum 57, no. 9 (2014): 1075–1081. [DOI] [PubMed] [Google Scholar]
- 19. Kenjo A., Miyata H., Gotoh M., et al., “Risk Stratification of 7,732 Hepatectomy Cases in 2011 From the National Clinical Database for Japan,” Journal of the American College of Surgeons 218, no. 3 (2014): 412–422. [DOI] [PubMed] [Google Scholar]
- 20. Kimura W., Miyata H., Gotoh M., et al., “A Pancreaticoduodenectomy Risk Model Derived From 8575 Cases From a National Single‐Race Population (Japanese) Using a Web‐Based Data Entry System: The 30‐Day and In‐Hospital Mortality Rates for Pancreaticoduodenectomy,” Annals of Surgery 259, no. 4 (2014): 773–780. [DOI] [PubMed] [Google Scholar]
- 21. Ketelaers S. H. J., Jacobs A., van der Linden C. M. J., et al., “An Evaluation of Postoperative Outcomes and Treatment Changes After Frailty Screening and Geriatric Assessment and Management in a Cohort of Older Patients With Colorectal Cancer,” Journal of Geriatric Oncology 14, no. 8 (2023): 101647. [DOI] [PubMed] [Google Scholar]
- 22. Sioutas G. and Tsoulfas G., “Frailty Assessment and Postoperative Outcomes Among Patients Undergoing General Surgery,” Surgeon 18, no. 6 (2020): e55–e66. [DOI] [PubMed] [Google Scholar]
- 23. Ghoneim M. M. and O'Hara M. W., “Depression and Postoperative Complications: An Overview,” BMC Surgery 16 (2016): 5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Haseda M., Kondo N., Takagi D., and Kondo K., “Community Social Capital and Inequality in Depressive Symptoms Among Older Japanese Adults: A Multilevel Study,” Health and Place 52 (2018): 8–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Shiratsuchi D., Makizako H., Nakai Y., et al., “Associations of Fall History and Fear of Falling With Multidimensional Cognitive Function in Independent Community‐Dwelling Older Adults: Findings From ORANGE Study,” Aging Clinical and Experimental Research 34, no. 12 (2022): 2985–2992. [DOI] [PubMed] [Google Scholar]
- 26. Kunneman M., Henselmans I., Gartner F. R., Bomhof‐Roordink H., and Pieterse A. H., “Do Shared Decision‐Making Measures Reflect Key Elements of Shared Decision Making? A Content Review of Coding Schemes,” Medical Decision Making 39, no. 7 (2019): 886–893. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Legare F. and Witteman H. O., “Shared Decision Making: Examining Key Elements and Barriers to Adoption Into Routine Clinical Practice,” Health Affairs (Millwood) 32, no. 2 (2013): 276–284. [DOI] [PubMed] [Google Scholar]
- 28. Berian J. R., Mohanty S., Ko C. Y., Rosenthal R. A., and Robinson T. N., “Association of Loss of Independence With Readmission and Death After Discharge in Older Patients After Surgical Procedures,” JAMA Surgery 151, no. 9 (2016): e161689. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1: Geriatric surgery variables and definitions.
Table S2: Detailed information on preoperative geriatric factors by types of surgical procedure.
Table S3: Relationship between geriatric factors and postoperative outcomes.
Table S4: Number of missing data by age and surgical procedure.
Data Availability Statement
Research data are not shared. We used data from the National Clinical Database (NCD) registry in Japan. The NCD committee prohibits us from sharing the data with anyone outside this research group.