Analyzing Impact of Multimorbidity on Long-Term Outcomes after Emergency General Surgery: A Retrospective Observational Cohort Study : Journal of the American College of Surgeons

Journal Logo

Original Scientific Articles

Analyzing Impact of Multimorbidity on Long-Term Outcomes after Emergency General Surgery: A Retrospective Observational Cohort Study

Rosen, Claire B MD; Roberts, Sanford E MD; Wirtalla, Chris J BA; Ramadan, Omar I MD; Keele, Luke J PhD; Kaufman, Elinore J MD, MSHP, FACS; Halpern, Scott D MD, PhD; Kelz, Rachel R MD, MSCE, MBA, FACS

Author Information
Journal of the American College of Surgeons: November 2022 - Volume 235 - Issue 5 - p 724-735
doi: 10.1097/XCS.0000000000000303
  • Free
  • Infographic

Emergency general surgery (EGS), a major facet of acute care surgery, is a heterogeneous field fraught with complex decision-making regarding high-stakes interventions, and currently lacking the necessary rigorous and innovative outcomes research necessary to understand and improve EGS care.1 EGS patients represent a distinct population and, on average, experience poorer outcomes than patients receiving similar elective operations.1-4 In addition, many EGS patients are older and have many medical problems, which influences outcomes and adds difficulty to preoperative patient counseling and decision-making.5-14

Given the potential for comorbid conditions to interact and influence overall risk of poor outcomes, multimorbidity has been identified as a global health challenge.15,16 In 2018, Silber and colleagues5 identified specific combinations of comorbidities, known as Qualifying Comorbidity Sets (QCSs), that can be used to label patients as multimorbid based on 2-fold increased mortality after general surgery. Our group’s previous work has found that using QCSs to define multimorbidity offers greater selectivity in identifying multimorbid patients and is associated with poorer short-term outcomes after EGS than having 3 or more comorbidities.17 However, the impact of multimorbidity on health outcomes beyond 30 days after EGS currently lacks robust investigation and would be of tremendous use to clinicians in identifying patients at high risk for complications and to aid in preoperative counseling. Patients care about prognosis beyond survival—they care about their function, independence, and about what to expect after surgery.18,19 Outcomes such as the use of durable medical equipment and discharge destination give us insight into patient’s lives that we can evaluate on the population level.

This study aims to examine the impact of multimorbidity, defined by the presence of a QCS, on major health outcomes through 6 months after operative management of EGS conditions. We hypothesize that older, multimorbid, operative EGS patients will have higher rates of mortality persisting through 6 months after discharge from index hospitalization, decreased independence, and increased postoperative resource use than comparable nonmultimorbid patients. This information will allow clinicians and surgeons to discuss postoperative expectations more accurately with older, multimorbid patients with EGS conditions, to improve shared decision-making and goal-concordant care.

METHODS

This is a nationwide, retrospective observational cohort study of Medicare beneficiaries aged 65+ using national Centers for Medicare & Medicaid claims data. Patients were identified using International Classification of Diseases, Ninth/Tenth Revision Clinical Modification billing codes.5,20,21 All included patients underwent operative intervention on hospital day 0, 1, or 2 of admission for an EGS condition. All included patients were admitted through the emergency department to a nonfederal, acute care hospital between July 1, 2015, and June 30, 2018, and enrolled in Medicare fee-for-service. Type of EGS condition (colorectal, general abdominal, hepatopancreatobiliary, intestinal obstruction, and upper gastrointestinal) was identified using International Classification of Diseases, Ninth/Tenth Revision, Clinical Modification codes described by the American Association of the Surgery of Trauma and modified by our research group’s earlier work.20,21 Surgeons were included if they performed 5 or more general surgery operations per year, and general surgery operations were identified using standard Current Procedural Terminology codes.22 Patients were excluded if they did not have continuous Medicare Parts A and B coverage or if they were enrolled in a Health Maintenance Organization at any time between 6 months before and 6 months after the index EGS admission (to ensure that their available claims represented all billable encounters). See Figure 1 for a CONSORT flow diagram of patients screened and included in the final analysis.

F1
Figure 1.:
CONSORT flow diagram of patients screened and included in analysis.

Exposure

First, we identified comorbid conditions using the standard Medicare Hierarchical Condition Category coding system (version 22)23-27 and indicators for disability with Current Procedural Terminology codes for durable medical equipment. Comorbid conditions were defined using a 6-month look-back period of the MEDPAR, Part B, and durable medical equipment files in addition to the codes present on admission at the time of the index hospitalization. We then defined multimorbidity as the presence of a QCS (specific combinations of comorbid conditions) identified by Silber and colleagues5 to be associated with 2-fold increased mortality risk after a general surgery operation (see eTable 1 in the Supplemental Digital Content for a full list of analyzed comorbidities and eTable 2 in the Supplemental Digital Content for QCSs; both at https://links.lww.com/JACS/A118). Patients who did not satisfy a QCS were identified as nonmultimorbid. The purpose of this study was to examine multimorbidity as a binary condition—present for those with a QCS and not present for patients who may still have multiple comorbidities, but whose specific combinations of comorbidities did not satisfy a QCS.

Outcomes

The primary outcome was mortality (from any cause) during index admission and at 1 month (30 calendar days), 3 months (90 calendar days), and 6 months (180 calendar days) after discharge from index hospitalization. Secondary outcomes included complications,20 readmission rates at 1, 3, and 6 months after discharge from index hospitalization (excluding patients who died by each time interval), hospital length of stay of index admission (differentiated by mortality), discharge status (to home, to home with home services, to hospice, to rehabilitation/nursing facility), new durable medical equipment use among patients discharged to home (home oxygen, walker, cane, wheelchair/scooter, hospital bed, patient lift, enteral or parenteral supplemental nutrition/hydration; see eTable 3 in the Supplemental Digital Content at https://links.lww.com/JACS/A118 for list of Current Procedural Terminology and Healthcare Common Procedure Coding System codes used to identify durable medical equipment use), and cost (of index admission and of subsequent care related to EGS diagnosis at 1, 3, and 6 months after discharge from index hospitalization).

Covariates

Covariates included age, sex, race category (White, Black, Asian, Hispanic, Native American, Other, Unknown), number of comorbidities at admission and/or within the 6-month look-back period (0, 1, 2, or 3+ comorbid conditions), surgical condition bucket (colorectal, general abdominal, hepatopancreatobiliary, intestinal obstruction, and upper gastrointestinal),20,21 and dual-eligibility status (having both Medicare and Medicaid coverage, as a proxy for socioeconomic status). The presence of sepsis/septic shock was calculated using the Angus-Sepsis Score.28 Frailty was determined according to the method of Kim and Schneeweiss.29 Covariates were assessed for collinearity and multicollinearity with variance inflation factor of <10 required for inclusion in our regression models.30,31

The multimorbid population was compared with nonmultimorbid patients. Outcomes were reported as numbers of patients and percentage of patients applicable to each outcome. Length of stay was reported as the number of in-patient calendar days, and cost was reported in 2018 US dollars using the medical component of the consumer price index for all urban consumers.32 Continuous variables were reported as mean ± standard deviation. Univariate analyses of continuous variables were performed using 1-way ANOVAs, and categorical variables were analyzed using Pearson chi-square tests. A Kaplan-Meier curve of all-cause mortality was constructed for multimorbid patients and for nonmultimorbid patients with survival curve comparison using the log-rank test.33 Risk-adjusted outcomes for binary variables were generated using multivariable logistic regression models for categorical variables and linear regression for continuous outcomes. We clustered the data by hospital to account for within-hospital correlations that might understate the standard errors. Univariate statistical significance was based on p < 0.05, and the significance of adjusted outcomes was assessed after Bonferroni correction at p < 0.00238. In the adjusted models, the beta coefficients for length of stay and costs represent the increased or decreased days in the hospital or dollars spent for multimorbid compared with nonmultimorbid patients.

This study was deemed exempt from continuing review by the Institutional Review Board of the University of Pennsylvania (Protocol number 832059). Data analysis and statistics were performed using Stata version 17.0 (StataCorp LLC, College Station, TX). This article was prepared in compliance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines for observational studies.34

RESULTS

Demographics

In this retrospective observational study of 174,891 patients who underwent operative intervention for an EGS condition within 48 hours of their initial presentation, 45.5% (n = 79,607) were identified as multimorbid (satisfaction of a QCS; see Table 1). Although most patients overall were female (54.2%, n = 94,732), there were slightly more men in the multimorbid group (49.2%, n = 39,189) than the nonmultimorbid group (43.0%, n = 40,970; p < 0.001).

Table 1. - Multimorbid vs Nonmultimorbid Patient Demographics
Variable Nonmultimorbid Multimorbid Total p Value
Total, n (%) 95,284 (54.5) 79,607 (45.5) 174,891 (100)
Age, y, mean ± SD 71.8 ± 11.9 75.1 ± 11.5 73.3 ± 11.9 <0.001
Sex, n (%) <0.001
 Male 40,970 (43.0) 39,189 (49.2) 80,159(45.8)
Race, n (%) <0.001
 White 81,668 (85.7) 67,710 (85.1) 149,378 (85.4)
 Black 6,976 (7.3) 6,885 (8.6) 13,861 (7.9)
 Asian 1,337 (1.4) 1,074 (1.3) 2,411 (1.4)
 Hispanic 2,157 (2.3) 1,713 (2.2) 3,870 (2.2)
 Native American 497 (0.5) 535 (0.7) 1,032 (0.6)
 Other 1,353 (1.4) 1,003 (1.3) 2,356 (1.3)
 Unknown 1,296 (1.4) 687 (0.9) 1,983 (1.1)
Number category of comorbid conditions,* n (%) <0.001
 0 6,513 (6.8) 0 (0) 6,513 (3.7)
 1 12,772 (13.4) 539 (0.7) 13,301 (7.6)
 2 21,788 (19.0) 1,994 (2.5) 20,109 (11.5)
 3+ 57,884 (60.7) 77,084 (96.8) 134,968 (77.2)
Number of comorbid conditions,* mean ± SD 3.2 ± 2.0 7.6 ± 3.5 5.2 ± 3.5 <0.001
Emergency general surgery condition, n (%) <0.001
 Colorectal 4,903 (5.1) 5,370 (6.7) 10,273 (5.9)
 General abdominal 3,252 (3.4) 11,301 (14.2) 14,553 (8.3)
 Hepatopancreatobiliary 44,744 (47.0) 25,687 (32.3) 70,431 (40.3)
 Intestinal obstruction 21,788 (22.9) 21,381 (26.9) 43,169 (24.7)
 Upper gastrointestinal 20,597 (21.6) 15,868 (19.9) 36,465 (20.9)
Claims-Based Frailty Index, 26 mean ± SD 0.129 ± 0.032 0.171 ± 0.061 0.148 ± 0.052 <0.001
Preoperative angus sepsis 25 yes, n (%) 2,196 (2.3) 17,215 (21.6) 19,411 (11.1) <0.001
Medicaid dual-eligibility, yes, n (%) 15,350 (16.5) 15,258 (20.1) 30,608 (18.1) <0.001
Multimorbidity defined by the satisfaction of Qualifying Comorbidity Set; nonmultimorbid patients’ comorbidities do not satisfy a Qualifying Comorbidity Set.
*See eTable 1 for comorbid conditions, https://links.lww.com/JACS/A118.

Comorbid conditions

Multimorbid patients had a mean of 7.86 ± 3.5 comorbid conditions, compared with a mean of 3.2 ± 2.0 comorbid conditions among nonmultimorbid patients (p < 0.001). Most multimorbid patients had 3 or more comorbid conditions (96.8%, n = 77,084), and almost all had 2 or more comorbid conditions (99.3%, n =7 9,078). Among nonmultimorbid patients, 60.7% (n = 57,884) had 3 or more and 79.7% (n = 79,762) had 2 or more comorbid conditions. The most common comorbid condition was hypertension among both multimorbid patients (81.8%, n = 65,098) and nonmultimorbid patients (71%, n = 67,687). See eTable 4 in the Supplemental Digital Content (https://links.lww.com/JACS/A118) for the prevalence of specific comorbid conditions among multimorbid and nonmultimorbid patients.

Surgical Condition

The most common EGS condition category for both multimorbid and nonmultimorbid patients was hepatopancreatobiliary (32.3% multimorbid, 47% nonmultimorbid), followed by intestinal obstruction (26.9% multimorbid, 22.9% nonmultimorbid), and upper gastrointestinal (19.9% multimorbid, 21.6% nonmultimorbid) conditions (see Table 1). The frequency of EGS subconditions among multimorbid and nonmultimorbid patients can be seen in eTable 5 in the Supplemental Digital Content (https://links.lww.com/JACS/A118). Gallstones and related diseases constituted the most common indication for surgery in our population: 30% of multimorbid patients and 43.4% of nonmultimorbid patients. There was a statistically significant difference in the distribution of types of operations received by multimorbid and nonmultimorbid patients (p < 0.001). The most common operation performed was cholecystectomy (33.9% multimorbid, 45.8% nonmultimorbid) followed by open partial colectomy (15.3% multimorbid, 8.8% nonmultimorbid), laparoscopic appendectomy (7.4% multimorbid, 12.9% nonmultimorbid), and small-bowel resection (10.4% multimorbid, 6.4% nonmultimorbid; see Table 2).

Table 2. - Procedure Type by Multimorbidity
Category, procedure type Nonmultimorbid, n (%) Multimorbid, n (%) Total, n (%)
Appendix
 Appendectomy 2,219 (2.3) 1,433 (1.8) 3,652 (2.1)
 Appendectomy – laparoscopic 12,297 (12.9) 5,890 (7.4) 18,187 (10.4)
Hepatopancreatobiliary
 Biliary common duct 658 (0.7) 497 (0.6) 1,155 (0.7)
 Biliary other 34 (<0.1) 54 (0.1) 88 (0.1)
 Cholecystectomy 43,639 (45.8) 27,015 (33.9) 70,654 (40.4)
 Liver other 45 (<0.1) 38 (<0.1) 83 (<0.1)
 Liver partial hepatectomy 115 (0.1) 164 (0.2) 279 (0.2)
 Pancreatectomy 25 (<0.1) 59 (0.1) 84 (<0.1)
Enterostomy
 Closure of enterostomy 23 (<0.1) 49 (0.1) 72 (<0.1)
 Closure of enterostomy with resection 43 (<0.1) 101 (0.1) 144 (0.1)
Colon and rectum
 Colectomy partial: laparoscopic 1,417 (1.5) 1,217 (1.5) 2,634 (1.5)
 Colectomy partial: open 8,404 (8.8) 12,151 (15.3) 20,555 (11.8)
 Colectomy total: laparoscopic 17 (<0.1) 30 (<0.1) 47 (<0.1)
 Colectomy total: open 247 (0.3) 916 (1.2) 1,163 (0.7)
 Large-bowel other 567 (0.6) 795 (1.0) 1,362 (0.8)
 Proctectomy 690 (0.7) 746 (0.9) 1,436 (0.8)
 Proctectomy: laparoscopic 200 (0.2) 107 (0.1) 307 (0.2)
 Proctopexy 66 (0.1) 40 (0.1) 106 (0.1)
Hernia
 Hernia abdominal: laparoscopic 1,340 (1.4) 1,020 (1.3) 2,360 (1.3)
 Hernia abdominal: open 5,179 (5.4) 4,901 (6.2) 10,080 (5.8)
 Hernia groin: laparoscopic 580 (0.6) 388 (0.5) 968 (0.6)
 Hernia groin: open 4,448 (4.7) 3,904 (4.9) 8,352 (4.8)
Small intestine
 Small-bowel other 971 (1.0) 1,367 (1.7) 2,338 (1.3)
 Small-bowel resection 6,125 (6.4) 8,304 (10.4) 14,429 (8.3)
Adhesions
 Lysis of adhesions 2,874 (3.0) 2,998 (3.8) 5,872 (3.4)
Stomach and duodenum
 Pyloroplasty 30 (<0.1) 59 (0.1) 89 (0.1)
 Stomach antireflux 39 (<0.1) 37 (<0.1) 76 (<0.1)
 Stomach gastric bypass (nonbariatric) 106 (0.1) 184 (0.2) 290 (0.2)
 Stomach other 160 (0.2) 302 (0.4) 462 (0.3)
 Stomach partial gastrectomy 279 (0.3) 556 (0.7) 835 (0.5)
 Stomach total gastrectomy 13 (<0.1) 44 (0.1) 57 (<0.1)
 Ulcer 2,365 (2.5) 4,033 (5.1) 6,389 (3.7)
Spleen
 Splenectomy 53 (0.1) 139 (0.2) 192 (0.1)
Multimorbidity defined by the satisfaction of Qualifying Comorbidity Set; nonmultimorbid patients’ comorbidities do not satisfy a Qualifying Comorbidity Set.

Unadjusted Outcomes: Multimorbid vs Nonmultimorbid Patients

Multimorbid patients had higher rates of mortality during the index hospitalization (5.9% vs 0.7%, p < 0.001), and also within 1 month (9.4% vs 1.4%, p < 0.001), 2 months (12.2% vs 2%, p < 0.001), 3 months (13.8% vs 2.4%, p < 0.001), and 6 months (17.1% vs 3.4%, p < 0.001) of discharge from index hospitalization. Kaplan-Meier survival curves between multimorbid and nonmultimorbid patients were significantly different (p < 0.001; see Fig. 2). Multimorbid patients were also less likely than nonmultimorbid patients to be discharged to home (42.4% vs 74.2%, p < 0.001), and more likely to be discharged to a rehabilitation center or skilled nursing facility (28.3% vs 11.3%, p < 0.001) or to hospice (2.1% vs 0.3%, p < 0.001). Multimorbid patients were more likely to be readmitted to the hospital within 1 month (22.9% vs 11.4%, p < 0.001), 2 months (27.2% vs 13.8%, p < 0.001), 3 months (30.5% vs 16%, p < 0.001), and 6 months (38.2% vs 21.2%, p < 0.001) of discharge from their index hospitalization (see Table 3).

Table 3. - Comparing Unadjusted Outcomes for Multimorbid vs Nonmultimorbid Patients
Variable Nonmultimorbid Multimorbid Total p Value
Mortality, n (% of total patients)
 During index hospitalization 657 (0.7) 4,706 (5.9) 5,363 (3.1) <0.001
 Within 30 days of index hospitalization discharge 1,337 (1.4) 7,518 (9.4) 8,855 (5.1) <0.001
 Within 60 days of index hospitalization discharge 1,911 (2.0) 9,677 (12.2) 11,588 (6.6) <0.001
 Within 90 days of index hospitalization discharge 2,315 (2.4) 10,977 (13.8) 13,292 (7.6) <0.001
 Within 180 days of index hospitalization discharge 3,273 (3.4) 13,614 (17.1) 16,887 (9.7) <0.001
Discharge status, n (% of patients who survived index hospitalization)
 Discharged to home 70,726 (74.2) 33,871 (42.5) 104,597 (59.8) <0.001
 Discharged to home with home health services 21,177 (22.2) 34,508 (43.3) 55,685 (31.8) <0.001
 Discharged to hospice 310 (0.3) 1,697 (2.1) 2,007 (1.1) <0.001
 Discharged to rehabilitation center or nursing facility 10,798 (11.3) 22,506 (28.3) 33,304 (19.0) <0.001
Readmission, n (% of patients who survived to that time point)
 Within 30 days of index hospitalization discharge 10,703 (11.4) 16,507 (22.9) 27,210 (16.4) <0.001
 Within 60 days of index hospitalization discharge 12,867 (13.8) 19,029 (27.2) 31,896 (19.5) <0.001
 Within 90 days of index hospitalization discharge 14,900 (16.0) 20,948 (30.5) 35,848 (22.2) <0.001
 Within 180 days of index hospitalization discharge 19,516 (21.2) 25,212 (38.2) 44,728 (28.3) <0.001
New durable medical equipment use within 30 days of index hospitalization, n (% of patients discharged to home) n=68,130 n=36467 n=104597
 Home oxygen 777 (1.0) 1,274 (3.7) 2,051 (2.0) <0.001
 Cane/crutch 141 (0.2) 147 (0.4) 288 (0.3) 0.060
 Walker 3,173 (4.5) 3,896 (11.5) 7,069 (6.8) <0.001
 Wheelchair or scooter 280 (0.4) 681 (2.0) 961 (0.9) <0.001
 Bedside commode 832 (1.2) 1,244 (3.7) 2,076 (2.0) <0.001
 Hospital bed 175 (0.2) 441 (1.3) 616 (0.6) <0.001
 Patient lift 21 (<0.1) 74 (0.1) 95 (0.1) <0.001
 Home infusion therapy supplies 91 (0.1) 194 (0.6) 285 (0.3) <0.001
Length of stay (calendar days), mean ± SD
 Index hospital admission 4.5 ± 3.9 7.5 ± 6.7 5.9 ± 5.6 <0.001
 Index hospital admission (excluding in-hospital death) 4.5 ± 3.9 7.5 ± 6.5 5.8 ± 5.4 <0.001
 ICU 0.7 ± 2.4 2.7 ± 5.1 1.6 ± 4.0 <0.001
 ICU (excluding in-hospital death) 0.7 ± 2.3 2.5 ± 4.9 1.5 ± 3.8 <0.001
Emergency general surgery–related cost, $, mean ± SD
 In-hospital 13,764 ± 9,737 21,258 ± 17,047 17,175 ± 14,066 <0.001
 30-day 14,395 ± 10,828 23,189 ± 19,532 18,398 ± 16,022 <0.001
 90-day 14,957 ± 12,265 24,951 ± 23,139 19,506 ± 18,720 <0.001
 180-day 15,565 ± 13,778 26,452 ± 25,794 20,521 ± 20,873 <0.001
Multimorbidity defined by the satisfaction of Qualifying Comorbidity Set; nonmultimorbid patients’ comorbidities do not satisfy a Qualifying Comorbidity Set.

F2
Figure 2.:
Survival rates by multimorbid status over time.

During index hospitalization, multimorbid patients stayed on average 7.5 ± 6.7 days in the hospital, but nonmultimorbid patients stayed 4.5 ± 3.9 days (p < 0.001). Excluding those who died during the index hospitalization, multimorbid patients stayed 7.5 ± 6.5 days compared with 4.5 ± 3.9 days for nonmultimorbid patients (p < 0.001). In addition, multimorbid patients stayed an average of 2.7 ± 5.1 days in the intensive care unit during index hospitalization compared with 0.7 ± 2.4 days for nonmultimorbid patients (p < 0.001), and for those who survived index hospitalization, multimorbid patients stayed 2.5 ± 4.9 days in the intensive care unit during index hospitalization compared with 0.7 ± 2.3 days for nonmultimorbid patients.

New use of durable medical equipment within 30 days of discharge from index hospitalization was more common among multimorbid patients than nonmultimorbid patients, although there was low prevalence overall. Multimorbid patients were more than twice as likely to require a walker (11.5% vs 4.5%, p < 0.001), home oxygen (3.7% vs 1%, p < 0.001), a bedside commode (3.7% vs 1.2%, p < 0.001), a wheelchair/scooter (2% vs 0.4%, p < 0.001), or a hospital bed (1.3% vs 0.2%, p < 0.001).

Multimorbid patients cost significantly more to treat than nonmultimorbid patients during the index hospitalization ($21,258 ± $17,047 vs $13,764 ± $9,737 per patient, p < 0.001), and at 1 month ($23,189 ± $19,532 vs $14,395 ± $10,828 per patient, p < 0.001), 3 months ($24,951 ± $23,139 vs $14,957 ± $12,265 per patient, p < 0.001), and 6 months ($26,452 ± $25,794 vs $15,565 ± $13,778 per patient, p < 0.001) after discharge from index hospitalization.

Adjusted Outcomes: Multimorbid vs Nonmultimorbid Patients

Multimorbid patients had 3 times the odds of death during index hospitalization compared with nonmultimorbid patients (odds ratio [OR] 3.05, p < 0.001), and at 6 months after index hospitalization still had more than double the odds of death (OR 2.33, p < 0.001; see Table 4). In addition, multimorbid patients had 72% higher odds of a complication during the index admission (OR 1.72, p < 0.001). Multimorbid patients had half as likely odds of being discharged to home (OR 0.52, p < 0.001) and 2.51 times the odds of being discharged to hospice than nonmultimorbid patients (p < 0.001). Multimorbid patients had 50% higher odds of readmission to the hospital at 1, 3, and 6 months compared with nonmultimorbid patients (p < 0.001) and stayed, on average, 1.3 calendar days longer in the hospital than nonmultimorbid patients during the index admission, even when stratified by in-hospital mortality (p < 0.001). The cost discrepancy between multimorbid and nonmultimorbid patients increased over time, from $4,051 extra dollars spent per multimorbid patient in-hospital (p < 0.001) to $5,162 extra dollars spent per multimorbid patient at 6 months after discharge (p < 0.001). For our population of 79,607 multimorbid patients, this is 410 million extra dollars.

Table 4. - Adjusted Outcomes
Variable Measure, comparison p Value*
Mortality
 During index hospitalization 3.05 <0.001
 Within 30 days of index hospitalization discharge 2.68 <0.001
 Within 90 days of index hospitalization discharge 2.45 <0.001
 Within 180 days of index hospitalization discharge 2.33 <0.001
Discharge status
 Discharged to home 0.52 <0.001
 Discharged to home with home health services 1.54 <0.001
 Discharged to hospice 2.51 <0.001
 Discharged to rehabilitation center or nursing facility 1.62 <0.001
Complication
 Complication (during index admission) 1.72 <0.001
Readmission
 Within 30 days of index hospitalization discharge 1.48 <0.001
 (Excluding patients who died within 30 days) 1.50 <0.001
 Within 90 days of index hospitalization discharge 1.49 <0.001
 (Excluding patients who died within 90 days) 1.49 <0.001
 Within 180 days of index hospitalization discharge 1.48 <0.001
 (Excluding patients who died within 180 days) 1.49 <0.001
Length of stay
 Index hospital admission +1.33 <0.001
 Index hospital admission (excluding in-hospital death) +1.32 <0.001
Emergency general surgery–related cost§
 In-hospital 4,051.1 <0.001
 30-day 4,538.9 <0.001
 90-day 4,914.7 <0.001
 180-day 5,162.1 <0.001
Multimorbidity defined by the satisfaction of Qualifying Comorbidity Set; nonmultimorbid patients’ comorbidities do not satisfy a Qualifying Comorbidity Set.
*Bonferroni correction yields significance threshold of <0.00238.
Odds ratio, multimorbid vs non-multimorbid patients.
Additional inpatient calendar days compared with nonmultimorbid patients.
§ Additional US dollars spent compared with nonmultimorbid patients.

DISCUSSION

In this national Centers for Medicare & Medicaid claims-based study, we found that the presence of a QCS, when used to designate a patient as multimorbid, confers substantially increased odds of mortality through 6 months after discharge from index hospitalization for surgical management of an EGS condition. Our multimorbid EGS population experienced increased readmission rates through 6 months, increased rates of complications, decreased rates of discharge to home, increased rates of discharge to a skilled nursing facility or to hospice, increased use of durable medical equipment, and increased cost of care through 6 months compared with nonmultimorbid patients. Several of these outcomes, especially readmission risks, discharge status, and dependency on medical equipment, are important for expectation planning with older patients in the preoperative setting. The considerable risk associated with multimorbidity suggests that the use of QCSs to identify multimorbid patients is a useful tool in the EGS setting.

Multimorbidity, recognized as a global health challenge16,35 and highlighted by the National Institutes of Health through an initiative to expand research on measurement, causes, and consequences of multimorbidity,36 has various definitions. This variation, and the complexity of multimorbid patients, limits evidence-based clinical decision support for multimorbid patients.37,38 Suls and colleagues39 recommend consideration of the purpose of multimorbidity identification when selecting the appropriate measurement; for example, a claims-based approach when attempting to predict morbidity, mortality, and healthcare use of patients, as in our study. The most straightforward definition of multimorbidity is the presence of 2 or more comorbidities.11,40,41 Although this definition of multimorbidity may allow providers to quickly recognize a patient as multimorbid, this count-based definition fails to value the interactions between specific medical conditions that may contribute to overall health and risks. In response, Silber and colleagues5 developed QCSs to define multimorbidity using International Classification of Diseases, Ninth/Tenth Revision, Clinical Modification claims to identify specific patient populations at high risk for adverse outcomes after surgery. Our current investigation highlights the selectivity of using QCSs to define multimorbidity as opposed to simply counting comorbid conditions, reaffirming and expanding on our earlier work17; in our total cohort, 45.5% of patients were found to have a QCS and were thus identified as multimorbid, but 79.7% of patients without a QCS had 2 or more comorbidities and 60.7% had 3 or more comorbidities. Our cohort with a QCS had worse outcomes, through 6 months, than patients without a QCS, of whom 79.7% may be labeled as multimorbid with a simple count-based method. Ho and colleagues42 recently found that older patients with complex multimorbidity (ie functional limitations, geriatric syndrome, and chronic medical conditions) have higher risk of mortality within 3 years after an EGS admission. QCSs incorporate chronic and acute diagnoses, include concordant and discordant conditions, and consider function-related concepts (like home oxygen use and wheelchair/hospital bed use) when defining multimorbidity, thus representing a holistic and patient-centered approach to multimorbid patient identification, in line with the National Institutes of Health aims of multimorbidity research.36 Our focus on outcomes beyond mortality is an important extension of earlier multimorbidity research.

After emergency abdominal surgery, older patients not only have higher rates of mortality, but are also less likely than their younger counterparts to be discharged home.43-46 Patients care about more than just survival, they care about what their life looks like in recovery and they want honest preparation for what to expect.18,19 As such, it is important for surgeons to consider and communicate outcomes regarding what patients want to know: outcomes related to their independence and anticipated functional status. Our study’s findings of increased complications, increased use of durable medical equipment, decreased rates of discharge to home, and increased rates of discharge to nursing facilities or rehabilitation among multimorbid, older patients provide valuable quality-of-life metrics. In 2015, Faurot and colleagues47 analyzed the use of durable medical equipment as a predictor for dependency in activities of daily living and found charges for a home hospital bed or wheelchair to be strong predictors of activities of daily living dependency, but, overall, the use durable medical equipment information in health services research, especially in nonorthopaedic surgery, is extremely limited. As such, the inclusion of durable medical equipment as an outcome metric in this analysis is a novel approach to using claims data to paint a picture of life quality that could be important to patients. This information, along with risk of readmission and likelihood of discharge to home or to another level of care, can help guide conversations that surgeons can have with multimorbid patients presenting with EGS conditions regarding postoperative expectations for recovery and subsequent independence/functional status.

It is important to note that, in this analysis, which only considered patients who received an operation, a substantial proportion of EGS diagnoses were those that could have been managed operatively or nonoperatively: gallstones, diverticular disease, and appendicitis, for example.48-52 Given that this investigation uses claims-based data, there are other important limitations. The first is that the accuracy of our results depends on the accuracy of coding, which can vary between institutions and individual coders. This was controlled for, in part, through clustering our analysis by hospital. Because this study was aimed to understand the impact of multimorbidity as a whole, it lacks granular information for differences between QCSs that is beyond the scope of this study. Finally, this study only included Medicare beneficiaries aged 65 and older, which could limit the generalizability of our findings. Our research team is currently developing a mobile App to easily identify patients as multimorbid, based on the presence of a QCS, to prospectively study this population and improve the generalizability and validity of these findings.

CONCLUSIONS

In this national, retrospective observational study of Medicare beneficiaries who underwent EGS, we found that the established poor short-term outcomes of multimorbid patients5,17 extend through 6 months after the index hospitalization. These poor outcomes include higher mortality, higher rates of readmission, and higher cost of care. Placing value on and understanding quality-of-life metrics, such as discharge status and durable medical equipment use, gives patients the information that they want to know, especially regarding independence. Multimorbidity is a powerful prognostic indicator, and further work to provide granular details would improve shared decision-making and goal-concordant care.

Author Contributions

Conceptualization: Rosen, Roberts, Wirtalla, Ramadan, Keele, Kaufman, Halpern, Kelz

Formal analysis: Rosen, Roberts, Wirtalla, Ramadan, Keele, Kaufman, Halpern, Kelz

Investigation: Rosen, Roberts, Wirtalla, Ramadan, Keele, Kaufman, Kelz

Methodology: Rosen, Roberts, Wirtalla, Ramadan, Keele, Kaufman, Halpern, Kelz

Validation: Rosen, Roberts, Wirtalla, Ramadan, Keele, Kaufman, Halpern, Kelz

Funding acquisition: Rosen, Wirtalla, Keele, Kelz

Resources: Rosen, Wirtalla, Ramadan, Keele, Kelz

Project administration: Rosen, Kelz

Supervision: Keele, Kelz

Data curation: Wirtalla, Kelz

Visualization: Rosen, Wirtalla, Kelz

Writing - original draft: Rosen, Roberts, Wirtalla, Ramadan, Keele, Kaufman, Kelz

Writing - review and editing: Rosen, Roberts, Wirtalla, Ramadan, Keele, Kaufman, Kelz

Acknowledgment:

With instrumental support from and appreciation for the Center for Outcomes Research at the Children’s Hospital of Philadelphia Research Institute, Philadelphia, PA, led by Jeffrey H Silber, MD, PhD, for identification of comorbid conditions with International Classification of Diseases, Current Procedural Terminology, and Healthcare Common Procedure Coding System Codes and for identification of Qualifying Comorbidity Sets.

    REFERENCES

    1. Becher RD, Davis KA, Rotondo MF, et al. Ongoing evolution of emergency general surgery as a surgical subspecialty. J Am Coll Surg. 2018;226:194–200.
    2. Havens JM, Peetz AB, Do WS, et al. The excess morbidity and mortality of emergency general surgery. J Trauma Acute Care Surg. 2015;78:306–311.
    3. Patel SS, Patel MS, Goldfarb M, et al. Elective versus emergency surgery for ulcerative colitis: a National Surgical Quality Improvement Program analysis. Am J Surg. 2013;205:333–337; discussion 337.
    4. Lilley EJ, Kelley A, Bollens-Lund E, et al. Pre-existing illness burden is associated with emergency department visits and mortality after emergency laparotomy. J Am Coll Surg. 2017;225:S95–S96.
    5. Silber JH, Reiter JG, Rosenbaum PR, et al. Defining multimorbidity in older surgical patients. Med Care. 2018;56:701–710.
    6. Chi MJ, Lee CY, Wu SC. Multiple morbidity combinations impact on medical expenditures among older adults. Arch Gerontol Geriatr. 2011;52:e210–e214.
    7. Gross CP, Guo Z, McAvay GJ, et al. Multimorbidity and survival in older persons with colorectal cancer. J Am Geriatr Soc. 2006;54:1898–1904.
    8. Shah AA, Haider AH, Zogg CK, et al. National estimates of predictors of outcomes for emergency general surgery. J Trauma Acute Care Surg. 2015;78:482–490; discussion 490.
    9. Kassin MT, Owen RM, Perez SD, et al. Risk factors for 30-day hospital readmission among general surgery patients. J Am Coll Surg. 2012;215:322–330.
    10. Tooth L, Hockey R, Byles J, et al. Weighted multimorbidity indexes predicted mortality, health service use, and health-related quality of life in older women. J Clin Epidemiol. 2008;61:151–159.
    11. Condelius A, Edberg AK, Jakobsson U, et al. Hospital admissions among people 65+ related to multimorbidity, municipal and outpatient care. Arch Gerontol Geriatr. 2008;46:41–55.
    12. Columbus AB, Morris MA, Lilley EJ, et al. Critical differences between elective and emergency surgery: identifying domains for quality improvement in emergency general surgery. Surgery. 2018;163:832–838.
    13. Lee TA, Shields AE, Vogeli C, et al. Mortality rate in veterans with multiple chronic conditions. J Gen Intern Med. 2007;22(Suppl 3):403–407.
    14. Lee KC, Streid J, Sturgeon D, et al. The impact of frailty on long-term patient-oriented outcomes after emergency general surgery: a retrospective cohort study. J Am Geriatr Soc. 2020;68:1037–1043.
    15. Koroukian SM, Warner DF, Owusu C, et al. Multimorbidity redefined: prospective health outcomes and the cumulative effect of co-occurring conditions. Prev Chronic Dis. 2015;12:E55.
    16. Nanna M, Alexander KP. A call to ACTION in caring for older patients with multimorbidity. July 13, 2019. American College of Cardiology. Available at: https://www.acc.org/latest-in-cardiology/articles/2019/07/03/08/22/a-call-to-action-in-caring-for-older-patients-with-multimorbidity. Accessed November 9, 2020.
    17. Rosen CB, Wirtalla C, Keele LJ, et al. Multimorbidity confers greater risk for older patients in emergency general surgery than the presence of multiple comorbidities: a retrospective observational study. Med Care Published online May 30, 2022.
    18. Doering LV, McGuire AW, Rourke D. Recovering from cardiac surgery: what patients want you to know. Am J Crit Care. 2002;11:333–343.
    19. McNair AGK, MacKichan F, Donovan JL, et al. What surgeons tell patients and what patients want to know before major cancer surgery: a qualitative study. BMC Cancer. 2016;16:258.
    20. Sharoky CE, Sellers MM, Keele LJ, et al. Does surgeon sex Matter?: Practice patterns and outcomes of female and male surgeons. Ann Surg. 2018;267:1069–1076.
    21. Shafi S, Aboutanos MB, Agarwal S Jr, et al.; AAST Committee on Severity Assessment and Patient Outcomes. Emergency general surgery: definition and estimated burden of disease. J Trauma Acute Care Surg. 2013;74:1092–1097.
    22. Kelz RR, Niknam BA, Sellers MM, et al. Duty hour reform and the outcomes of patients treated by new surgeons. Ann Surg. 2020;271:599–605.
    23. Pope GC, Kautter J, Ellis RP, et al. Risk adjustment of Medicare capitation payments using the CMS-HCC model. Health Care Financ Rev. 2004;25:119–141.
    24. Chronic Conditions Warehouse. Chronic condition warehouse user guide: medicare risk score files. version 1.1.
    25. Medicode (firm). ICD-9-CM: International classification of diseases, 9th revision, clinical modification. Salt Lake City, UT: Medicode1996.
    26. American Medical Association. ICD-10-CM 2017: The Complete Official Code Book. Chicago, IL: American Medical Association2016.
    27. Health Alliance. RxHCC model. Available at: https://provider.healthalliance.org/coding-counts-post/rxhcc-model/. Accessed April 11, 2022.
    28. Iwashyna TJ, Odden A, Rohde J, et al. Identifying patients with severe sepsis using administrative claims: patient-level validation of the angus implementation of the international consensus conference definition of severe sepsis. Med Care. 2014;52:e39–e43.
    29. Kim DH, Schneeweiss S, Glynn RJ, et al. Measuring frailty in Medicare data: development and validation of a claims-based frailty index. J Gerontol A Biol Sci Med Sci. 2018;73:980–987.
    30. Vatcheva KP, Lee M, McCormick JB, et al. Multicollinearity in regression analyses conducted in epidemiologic studies. Epidemiology (Sunnyvale). 2016;6:227.
    31. Kutner M, Nachtsheim C, Neter J. Applied Linear Statistical Models. 4th ed. Irwin: McGraw-Hill2004.
    32. US Bureau of Labor Statistics. CPI for all urban consumers. Available at: https://data.bls.gov/timeseries/CUUR0000SAM?output_view=data. Accessed April 21, 2022.
    33. Bland JM, Altman DG. Survival probabilities (the Kaplan-Meier method). BMJ. 1998;317:1572–1580. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1114388/. Accessed April 11, 2022.
    34. STROBE. Checklists – STROBE. Available at: https://www.strobe-statement.org/checklists/ Web site. https://www.strobe-statement.org/checklists/. Accessed September 8, 2021.
    35. The Academy of Medical Sciences Working Group. Multimorbidity: a priority for global health research. The Academy of Medical Sciences; 2018:1–127.
    36. Salive ME, Suls J, Farhat T, et al. National Institutes of Health advancing multimorbidity research. Med Care. 2021;59:622–624.
    37. Muth C, Beyer M, Fortin M, et al. Multimorbidity’s research challenges and priorities from a clinical perspective: the case of ‘Mr Curran’. Eur J Gen Pract. 2014;20:139–147.
    38. Campbell-Scherer D. Multimorbidity: a challenge for evidence-based medicine. Evid Based Med. 2010;15:165–166.
    39. Suls J, Bayliss EA, Berry J, et al. Measuring multimorbidity: selecting the right instrument for the purpose and the data source. Med Care. 2021;59:743–756.
    40. Salive ME. Multimorbidity in older adults. Epidemiol Rev. 2013;35:75–83.
    41. King DE, Xiang J, Pilkerton CS. Multimorbidity trends in United States adults, 1988–2014. J Am Board Fam Med. 2018;31:503–513.
    42. Ho VP, Bensken WP, Warner DF, et al. Association of complex multimorbidity and long-term survival after emergency general surgery in older patients with Medicare. JAMA Surg. 2022;157:499–506.
    43. Massarweh NN, Legner VJ, Symons RG, et al. Impact of advancing age on abdominal surgical outcomes. Arch Surg. 2009;144:1108–1114.
    44. Cooper Z, Mitchell SL, Gorges RJ, et al. Predictors of mortality up to 1 year after emergency major abdominal surgery in older adults. J Am Geriatr Soc. 2015;63:2572–2579.
    45. Duron JJ, Duron E, Dugue T, et al. Risk factors for mortality in major digestive surgery in the elderly: a multicenter prospective study. Ann Surg. 2011;254:375–382.
    46. Turrentine FE, Wang H, Simpson VB, et al. Surgical risk factors, morbidity, and mortality in elderly patients. J Am Coll Surg. 2006;203:865–877.
    47. Faurot KR, Jonsson Funk M, Pate V, et al. Using claims data to predict dependency in activities of daily living as a proxy for frailty. Pharmacoepidemiol Drug Saf. 2015;24:59–66.
    48. Salminen P, Paajanen H, Rautio T, et al. Antibiotic therapy vs appendectomy for treatment of uncomplicated acute appendicitis: the APPAC randomized clinical trial. JAMA. 2015;313:2340–2348.
    49. Sceats LA, Trickey AW, Morris AM, et al. Nonoperative management of uncomplicated appendicitis among privately insured patients. JAMA Surg. 2019;154:141–149.
    50. Garfinkle R, Almalki T, Pelsser V, et al. Conditional risk of diverticulitis after non-operative management. Br J Surg. 2020;107:1838–1845.
    51. Long B, Robertson J, Koyfman A. Emergency medicine evaluation and management of small bowel obstruction: evidence-based recommendations. J Emerg Med. 2019;56:166–176.
    52. McGillicuddy EA, Schuster KM, Barre K, et al. Non-operative management of acute cholecystitis in the elderly. Br J Surg. 2012;99:1254–1261.

    Supplemental Digital Content

    © 2022 by the American College of Surgeons. Published by Wolters Kluwer Health, Inc. All rights reserved.