1097-0142/asset/cover.gif?v=1&s=a7299bc18f075294c232ade468773cd0672bd470 "Cancer")
Fax: (319) 335-8814
1097-0142/asset/olbannerleft.gif?v=1&s=ca681f5719430b26e1bc15e9ea4c9fc0a7110104)
1097-0142/asset/olbannerright.gif?v=1&s=8142566facf7e76aef9be6c51162a2e920b3b9f9)
Original Article
You have full text access to this OnlineOpen article
Characteristics of patients and predictors of in-hospital mortality after hospitalization for head and neck cancers
Article first published online: 25 APR 2006
DOI: 10.1002/cncr.21899
Copyright © 2006 American Cancer Society
Additional Information
How to Cite
Allareddy, V. and Konety, B. R. (2006), Characteristics of patients and predictors of in-hospital mortality after hospitalization for head and neck cancers. Cancer, 106: 2382–2388. doi: 10.1002/cncr.21899
Publication History
- Issue published online: 18 MAY 2006
- Article first published online: 25 APR 2006
- Manuscript Accepted: 11 JAN 2006
- Manuscript Revised: 21 DEC 2005
- Manuscript Received: 11 AUG 2005
- Abstract
- Article
- References
- Cited By
Keywords:
- head and neck cancers;
- in-hospital mortality;
- insurance;
- hospital characteristics
Abstract
BACKGROUND
The objectives of this study were to describe the characteristics of patients who were hospitalized for head and neck cancer (HNC) during the years 2000 through 2003 and to identify predictors of in-hospital mortality.
METHODS
The Nationwide Inpatient Sample for the years 2000 through 2003 was used. All patients who had a primary diagnosis of any of the HNCs were included in the study. Univariate and multivariate logistic regression analyses were used to identify patient and hospital characteristics that were associated with in-hospital mortality.
RESULTS
In total, 24,803 patients were hospitalized for HNCs. The average age of patients was 62 years, the mean length of stay in the hospital was 7.89 days, and the in-hospital mortality rate was 5.18%. Patients who had comorbid conditions and complications and patients who were grouped under the self-pay/no charge/others category had greater odds of in-hospital mortality compared with patients who were covered by private insurance (P<.02). Patients who were treated in large-bed, urban, or teaching hospitals had lower odds of in-hospital mortality compared with patients who were treated in small or medium-bed, rural, or nonteaching hospitals, respectively(P<.03).
CONCLUSIONS
Patients with comorbid conditions and complications and patients without adequate insurance coverage had greater odds of in-hospital mortality. One reason for this may be inadequate access to care because of the absence of insurance or underinsurance. Further studies controlling for disease stage will be required to determine whether insurance status and patient-related factors can influence outcomes from HNC in individual patients independent of their disease stage. Cancer 2006. © 2006 American Cancer Society.
According to the American Cancer Society, close to 30,000 new oropharyngeal cancers and 10,000 laryngeal cancers will be diagnosed in the year 2005 in the United States.1 Approximately 11,000 individuals are expected to die of these cancers.1 To our knowledge, no population-based study had been conducted to date that examined the characteristics of patients who were hospitalized for head and neck cancer (HNC) and identified predictors of in-hospital mortality. The objectives of the current study were to provide demographic characteristics of patients who were hospitalized in the United States for HNC during the years 2000 through 2003 and to identify the predictors of in-hospital mortality. We hypothesized that older patients have high in-hospital mortality rates; whereas patients without comorbid conditions or with private insurance coverage and patients who are treated in large-bed, urban, or teaching hospitals have low in-hospital mortality rates.
MATERIALS AND METHODS
Patient Selection
We performed retrospective analyses of the Nationwide Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project for the years 2000 through 2003. The NIS is a 20% stratified probability sample of all nonfederal, acute-care, general hospitals in the United States and contains from 5 million to 8 million records from approximately 1000 hospitals in 35 states.2 The NIS sampling strata is based on 5 hospital characteristics: geographic region (northeast, midwest, west, and south), control (public, private not-for-profit, and proprietary), location (urban or rural), teaching status (teaching or nonteaching), and bed size (small, medium, and large). All our analyses were performed on the 20% sample provided by the Agency for Healthcare Research and Quality. Patients with a primary diagnosis of malignant neoplasms of lip (International Classification of Diseases-Ninth Revision, Clinical Modification [ICD-9 CM] codes 140.0-140.9); tongue (codes 141.0-141.9); salivary gland (codes 142.0-142.9); gums (codes 143.0-143.9); floor of mouth (codes 144.0-144.9); other parts of the mouth, such as cheek mucosa, vestibule of mouth, hard palate, soft palate, uvula, retromolar area, and unspecified parts of mouth (codes 145.0-145.9); oropharynx (codes 146.0-146.9); nasopharynx (codes 147.0-147.9); hypopharynx (codes 148.0-148.9); other and ill-defined sites within the lip, oral cavity, and pharynx (codes 149.0-149.9); nasal cavities, middle ear, or accessory sinuses (codes 160.0-160.9); larynx (codes 161.0-161.9); and ill-defined sites in the head, neck, and face region (code 195.0) were selected for analyses.
Statistical Analyses
Simple descriptive statistics were used to describe the characteristics of patients who were hospitalized for HNC. The impact of patient characteristics (age, gender, comorbid conditions, complications from medical or surgical care, and insurance status) and hospital characteristics (hospital bed size, hospital location, and hospital teaching status) on in-hospital mortality were examined by univariate logistic regression analyses.
A multivariate logistic regression analysis model was built to assess the impact of patient and hospital characteristics on in-hospital mortality. Age was not adjusted for in the multivariate model because of collinearity with Medicare insurance status. The effects of gender (with males as the reference category), comorbid conditions, complications from medical or surgical care (with no complications as the reference category), health insurance status (with private insurance as the reference category), hospital bed size (with large bed size as the reference category), hospital location (with rural as the reference category), and hospital teaching status (with nonteaching as the reference category) were examined. The effect of year of hospitalization (with 2003 as the reference category) and the type of cancer (with cancer of ill-defined sites in the head, neck, and face region as the reference category) were adjusted for in the analysis. Insurance status was characterized as private, Medicare, Medicaid, and self-pay/no charge/others (included Workers Compensation, the Civilian Health and Medical Program of the Uniformed Services [CHAMPUS], the Civilian Health and Medical Program of the Veterans Administration [CHAMPVA], Title V, and other government programs). We used the 29 comorbid conditions identified by Elixhauser et al. to adjust for the confounding effects of comorbid disease severity on outcomes.3 All comorbid conditions that were associated with significantly higher mortality rates at P<.05 were adjusted for in the multivariate models. The presence of complications from surgical or medical care were identified by using the clinical classification software codes “238” and “2” in the data set. Hospitals with small and medium beds were collapsed into 1 category. We also examined the impact of patient and hospital characteristics on patients being admitted to the hospitals on an elective basis versus a nonelective basis (emergent/urgent).
We used the Generalized Estimating Equations method to the fit the univariate and multivariate regression models to correct for possible clustering of similar outcomes within hospitals. An exchangeable correlation matrix was specified for the analyses to adjust for clustering. The empirical standard errors were used to compute the confidence intervals of the odds ratios. Two-sided P values were calculated for all analyses, and a P value <.05 was considered significant. All analyses were performed using SAS software (version 9.1; SAS Institute, Cary, NC).
RESULTS
Overall, 24,803 patients were hospitalized during the years 2000 through 2003 for 1 of the HNCs. The average patient age (± standard deviation) was 62 years (± 14.13 years), and the mean length of stay in the hospital was 7.89 days (± 9.78 days). The in-hospital mortality rate was 5.18% (n = 1284 patients). Data regarding patient disposition were not available for 39 patients. Among all HNCs, the cancers that required hospitalization most frequently were malignant neoplasms of the larynx (n = 6470 patients) followed by malignant neoplasms of the tongue (n = 4520 patients) and the oropharynx (n = 2263 patients). Patient and hospital characteristics of all hospitalized HNCs are provided in Table 1. Malignant neoplasms of ill-defined sites in the head, neck, and face region had the highest in-hospital mortality rates (15.61%). Close to 90% of all hospitalizations for all cancer types occurred in urban hospitals. Teaching hospitals and large-bed hospitals also accounted for a large proportion of hospitalizations. Patients who had ill-defined cancers of the mouth, head, neck, and face were admitted more frequently on a nonelective basis compared with patients who had other cancer types. Close to 70% of patients were males. Medicare patients accounted for 43.65% (n = 10,800 patients) of admissions, whereas patients with Medicaid and with private insurance accounted for 13.25% (n = 3279 patients) and 35.11% (n = 8686 patients) of admissions, respectively. One thousand forty patients were self-pay (4.20%), 188 patients were not charged (0.76%), and 749 patients (3.03%) had other insurance coverage, such as Workers Compensation, CHAMPUS, CHAMPVA, Title V, and other government programs.
| Characteristic/ Response | No. of Patients (%) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Lip | Tongue | Salivary | Gums | Floor | Oropharynx | Unspecified* | Nasopharynx | Hypopharynx | Ill defined† | Nasal | Larynx | Ill defined: HNF‡ | |
| |||||||||||||
| Mean age, y | 67.8 | 60.5 | 64.7 | 67.6 | 62.1 | 58.4 | 64.6 | 54.5 | 63.6 | 63.5 | 60.4 | 62.8 | 61.9 |
| Gender | |||||||||||||
| Male | 192 (72.73) | 2960 (65.54) | 1190 (60.96) | 315 (49.53) | 883 (66.04) | 1719 (76.03) | 925 (54.99) | 646 (70.76) | 1002 (77.92) | 666 (70.85) | 704 (60.22) | 5015 (77.56) | 1005 (74.28) |
| Female | 72 (27.27) | 1556 (34.46) | 762 (39.04) | 321 (50.47) | 454 (33.96) | 542 (23.97) | 757 (45.01) | 267 (29.24) | 284 (22.08) | 274 (29.15) | 465 (39.78) | 1451 (22.44) | 348 (25.72) |
| Missing | 1 | 4 | 10 | 1 | 0 | 2 | 1 | 3 | 0 | 0 | 1 | 4 | 1 |
| Disposition of patient | |||||||||||||
| Routine | 205 (77.36) | 2857 (63.24) | 1642 (83.73) | 445 (69.86) | 812 (60.87) | 1450 (64.19) | 1085 (64.47) | 555 (60.66) | 572 (44.62) | 419 (44.72) | 854 (73.05) | 3038 (47.09) | 750 (55.47) |
| Transfer to short-term facility | DS | 51 (1.13) | 10 (0.51) | DS | 13 (0.97) | 31 (1.37) | 16 (0.95) | 16 (1.75) | 37 (2.89) | 23 (2.45) | 13 (1.11) | 139 (2.15) | 31 (2.29) |
| Other transfers | 26 (9.81) | 497 (11) | 109 (5.56) | 65 (10.20) | 172 (12.89) | 229 (10.14) | 203 (12.06) | 99 (10.82) | 244 (19.03) | 165 (17.61) | 119 (10.18) | 964 (14.94) | 183 (13.54) |
| Home health care | 24 (9.06) | 915 (20.25) | 148 (7.55) | 98 (15.38) | 299 (22.41) | 398 (17.62) | 310 (18.42) | 166 (18.14) | 366 (28.55) | 207 (22.09) | 134 (11.46) | 1986 (30.78) | 167 (12.35) |
| Against medical advice | DS | 17 (0.38) | DS | DS | DS | 10 (0.44) | DS | DS | DS | 10 (1.07) | DS | 29 (0.45) | DS |
| Died | DS | 178 (3.94) | 50 (2.55) | 21 (3.30) | 34 (2.55) | 139 (6.15) | 66 (3.92) | 72 (7.87) | 59 (4.60) | 111 (11.85) | 44 (3.76) | 292 (4.53) | 211 (15.61) |
| Discharged alive, destination unknown | DS | DS | DS | DS | DS | DS | DS | DS | DS | DS | DS | DS | DS |
| Missing | 0 | 2 | 1 | 0 | 3 | 4 | 0 | 1 | 4 | 3 | 1 | 18 | 2 |
| Admission type | |||||||||||||
| Elective | 218 (82.58) | 3163 (70.26) | 1565 (79.93) | 508 (79.87) | 1012 (76.03) | 1407 (62.45) | 1197 (71.46) | 391 (42.87) | 709 (55.48) | 446 (47.70) | 763 (65.61) | 3690 (57.31) | 651 (48.37) |
| Nonelective | 46 (17.42) | 1339 (29.74) | 393 (20.07) | 128 (20.13) | 319 (23.97) | 846 (37.55) | 478 (28.54) | 521 (57.13) | 569 (44.52) | 489 (52.30) | 400 (34.39) | 2749 (42.69) | 695 (51.63) |
| Missing | 1 | 4 | 4 | 1 | 6 | 10 | 8 | 1 | 8 | 5 | 7 | 31 | 8 |
| Insurance | |||||||||||||
| Medicare | 163 (61.74) | 1800 (39.95) | 1026 (52.51) | 388 (60.91) | 588 (44.18) | 717 (31.82) | 824 (49.02) | 267 (29.24) | 603 (46.96) | 444 (47.39) | 507 (43.37) | 2921 (45.20) | 552 (40.86) |
| Medicaid | 18 (6.82) | 533 (11.83) | 100 (5.12) | 34 (5.34) | 186 (13.97) | 337 (14.96) | 181 (10.77) | 191 (20.92) | 225 (17.52) | 142 (15.15) | 105 (8.98) | 1002 (15.51) | 225 (16.65) |
| Private | 57 (21.59) | 1831 (40.63) | 729 (37.31) | 185 (29.04) | 457 (34.34) | 994 (44.12) | 548 (32.60) | 342 (37.46) | 360 (28.04) | 283 (30.20) | 490 (41.92) | 1969 (30.47) | 441 (32.64) |
| Self-pay/no charge/other | 26 (9.85) | 342 (7.59) | 99 (5.07) | 30 (4.71) | 100 (7.51) | 205 (9.10) | 128 (7.61) | 113 (12.38) | 96 (7.48) | 68 (7.26) | 67 (5.73) | 570 (8.82) | 133 (9.84) |
| Missing | 1 | 14 | 8 | 0 | 6 | 10 | 2 | 3 | 2 | 3 | 1 | 8 | 3 |
| Hospital bed size | |||||||||||||
| Small | 25 (9.43) | 459 (10.16) | 239 (12.18) | 70 (10.99) | 134 (10.04) | 230 (10.16) | 157 (9.34) | 67 (7.32) | 144 (11.21) | 91 (9.70) | 133 (11.37) | 663 (10.25) | 147 (10.86) |
| Medium | 56 (21.13) | 878 (19.44) | 407 (20.74) | 96 (15.07) | 271 (20.30) | 501 (22.14) | 361 (21.48) | 228 (24.92) | 272 (21.18) | 222 (23.67) | 217 (18.55) | 1445 (22.34) | 323 (23.87) |
| Large | 184 (69.43) | 3179 (70.39) | 1316 (67.07) | 471 (73.94) | 930 (69.66) | 1532 (67.70) | 1163 (69.19) | 620 (67.76) | 868 (67.60) | 625 (66.63) | 820 (70.09) | 4359 (67.40) | 883 (65.26) |
| Missing | 0 | 4 | 0 | 0 | 2 | 0 | 2 | 1 | 1 | 2 | 0 | 3 | 1 |
| Hospital location | |||||||||||||
| Urban | 239 (90.19) | 4225 (93.56) | 1803 (91.90) | 596 (93.56) | 1262 (94.53) | 2095 (92.58) | 1564 (93.04) | 849 (92.79) | 1201 (93.54) | 827 (88.17) | 1103 (94.27) | 5883 (90.97) | 1206 (89.14) |
| Rural | 26 (9.81) | 291 (6.44) | 159 (8.10) | 41 (6.44) | 73 (5.47) | 168 (7.42) | 117 (6.96) | 66 (7.21) | 83 (6.46) | 111 (11.83) | 67 (5.73) | 584 (9.03) | 147 (10.86) |
| Missing | 0 | 4 | 0 | 0 | 2 | 0 | 2 | 1 | 2 | 2 | 0 | 3 | 1 |
| Hospital teaching status | |||||||||||||
| Teaching | 183 (69.05) | 3053 (67.60) | 1234 (62.90) | 460 (72.21) | 975 (73.03) | 1425 (62.97) | 1171 (69.66) | 579 (63.28) | 805 (62.69) | 535 (57.04) | 851 (72.74) | 4145 (64.09) | 708 (52.33) |
| Nonteaching | 82 (30.94) | 1463 (32.40) | 728 (37.10) | 177 (27.79) | 360 (26.97) | 838 (37.03) | 510 (30.34) | 336 (36.72) | 479 (37.31) | 403 (42.96) | 319 (27.26) | 2322 (35.91) | 645 (47.67) |
| Missing | 0 | 4 | 0 | 0 | 2 | 0 | 2 | 1 | 2 | 2 | 0 | 3 | 1 |
The results of the univariate and multivariate models for predictors of in-hospital mortality are shown in Table 2. For each year increase in age, the odds of in-hospital mortality increased by 1.3% (P<.0001). Females were associated with 12% lower odds of in-hospital mortality compared with males in the univariate analysis (P = .005). However, after adjusting for other confounding factors, gender was not a significant predictor of in-hospital mortality. Patients with comorbid conditions, such as congestive heart failure, pulmonary circulation disorders, paralysis, neurologic disorders, chronic pulmonary disease, renal failure, liver disease, coagulopathy, weight loss, fluid/electrolyte disorders, chronic blood loss anemia, or deficiency anemia, were associated with greater odds of in-hospital mortality (P<.05) in the univariate analysis. After adjusting for confounding factors in the multivariate analysis, only congestive heart failure, neurologic disorders, renal failure, liver disease, coagulopathy, and fluid/electrolyte disorders were associated significantly with increased in-hospital mortality (P<.04). Patients who developed complications from surgical or medical care during hospitalizations were associated with increased odds of in-hospital mortality in both the univariate and multivariate analyses (P<.001). Among the different complications, septicemia was associated with the greatest odds of in-hospital mortality. Even after adjusting for all other confounding factors, septicemia was associated with the greatest odds of in-hospital mortality (adjusted odds ratio, 5.03; 95% confidence interval, 3.44-7.36; P<.0001).
| Characteristics | Univariate Model*† | Multivariate Model† | ||
|---|---|---|---|---|
| OR (95%CI) | P value | OR (95%CI) | P value | |
| ||||
| Age (for each year increase) | 1.013 (1.01-1.016) | <.0001 | NA | |
| Gender | ||||
| Female | 0.88 (0.81-0.96) | .0049 | 0.91 (0.82-1.01) | .09 |
| Male‡ | 1.00 | 1.00 | ||
| Comorbid conditions | ||||
| Congestive heart failure | 2.17 (1.85-2.56) | <.0001 | 2.17 (1.79-2.63) | <.0001 |
| Pulmonary circulation disorders | 2.48 (1.39-4.42) | .002 | 1.90 (0.99-3.64) | .05 |
| Paralysis | 1.55 (1.08-2.23) | .017 | 1.33 (0.81-2.16) | .24 |
| Neurologic disorders | 1.76 (1.46-2.12) | <.0001 | 1.71 (1.38-2.12) | <.0001 |
| Chronic pulmonary disease | 1.12 (1.02-1.23) | .017 | 1.04 (0.92-1.17) | .45 |
| Renal failure | 1.80 (1.32-2.47) | .0002 | 1.49 (1.02-2.17) | .03 |
| Liver disease | 1.56 (1.23-1.98) | .0003 | 1.37 (1.01-1.86) | .03 |
| Coagulopathy | 2.04 (1.54-2.69) | <.0001 | 1.54 (1.11-2.15) | .009 |
| Weight loss | 1.50 (1.30-1.73) | <.0001 | 1.16 (0.98-1.38) | .08 |
| Fluid and electrolyte disorders | 1.95 (1.75-2.18) | <.0001 | 1.73 (1.51-1.98) | <.0001 |
| Chronic blood loss anemia | 1.54 (1.001-2.39) | .049 | 1.26 (0.878-2.05) | .33 |
| Deficiency anemias | 1.40 (1.21-1.63) | <.0001 | 1.13 (0.95-1.36) | .15 |
| Complications of surgical procedures or medical care | ||||
| Yes | 1.65 (1.50-1.83) | <.0001 | 1.72 (1.49-1.99) | <.0001 |
| No | 1.00 | 1.00 | ||
| Insurance | ||||
| Medicare | 1.12 (1.02-1.22) | .011 | 1.01 (0.90-1.14) | .75 |
| Medicaid | 1.13 (0.98-1.29) | .077 | 0.99 (0.82-1.18) | .92 |
| Self-pay/no charge/others | 1.41 (1.15-1.74) | .0009 | 1.42 (1.10-1.83) | .006 |
| Private‡ | 1.00 | 1.00 | ||
| Hospital bed size | ||||
| Small/medium | 1.19 (1.001-1.41) | .047 | 1.22 (1.03-1.46) | .018 |
| Large‡ | 1.00 | 1.00 | ||
| Hospital location | ||||
| Urban | 0.46 (0.37-0.57) | <.0001 | 0.54 (0.43-0.67) | <.0001 |
| Rural‡ | 1.00 | 1.00 | ||
| Hospital teaching status | ||||
| Teaching | 0.51 (0.42-0.60) | <.0001 | 0.72 (0.58-0.89) | .002 |
| Nonteaching‡ | 1.00 | 1.00 | ||
Patients who were covered by Medicare or who were grouped under the self-pay/no charge/others category were associated with higher odds of in-hospital mortality compared with patients who were covered by private insurance (P<.02). Patients who were grouped under the self-pay/no charge/others category were associated with higher odds of mortality compared with patients who were covered by private insurance even after adjusting for the effects of other confounding factors (P = .006). The results of the multivariate analysis showed that patients who were treated in small or medium-bed hospitals were associated with higher odds of in-hospital mortality compared with patients who were treated in large-bed hospitals (P = .018), patients who were treated in urban hospitals were associated with lower odds of in-hospital mortality compared with patients who were treated in rural hospitals (P<.0001), and patients who were treated in teaching hospitals were associated with lower odds of in-hospital mortality compared with patients who were treated in nonteaching hospitals (P = .002).
The results of the multivariate model examining the predictors of whether patients would be admitted to the hospital on an elective basis revealed that patients who were covered by Medicare were associated with lower odds of being admitted on an elective basis compared with patients who had private insurance (P<.0001) (Table 3). Patients who were covered by Medicaid and who were grouped under the self-pay/no charge/others category were associated with lower odds of being treated on an elective basis compared with patients who were covered by private insurance (P<.0001) (Table 3).
| Insurance | Multivariate Model* | |
|---|---|---|
| OR (95%CI) | P value | |
| ||
| Medicare | 0.86 (0.81-0.92) | <.0001 |
| Medicaid | 0.48 (0.43-0.53) | <.0001 |
| Self-pay/no charge/others | 0.64 (0.57-0.73) | <.0001 |
| Private† | 1.00 | |
DISCUSSION
The current findings provide nationally representative estimates of demographic characteristics and predictors of in-hospital mortality for patients who are hospitalized with HNCs in the United States. The factors that were identified as related to mortality can serve as targets for future quality-improvement efforts. Based on our retrospective data analysis, insurance payer status and hospital characteristics, along with other disease-related factors (such as stage), may play a role in influencing clinical outcomes for patients who are hospitalized for HNC. The findings of this study are similar to those reported in earlier studies, which assessed the influence of health insurance status on clinical outcomes for patients with other cancers.4–6 Patients who were grouped under the self-pay/no charge/others category had higher in-hospital mortality rates compared with patients who were covered by private insurance. One potential explanation is that patients with other than private insurance may not have health care services that are as accessible or of the same quality as those provided to individuals with private insurance. When we examined the factors associated with patients being admitted on an elective basis to the hospital, it was clear that patients in the self-pay/no charge/others category were admitted more often on an emergent or urgent basis compared with patients who were covered by private insurance. This also suggests that patients who do not have private insurance may seek care on an emergent basis or with advanced disease stage that requires an emergent admission because of their restricted access to care.
Patients with comorbid conditions were associated with higher odds of in-hospital mortality compared with patients without any comorbid conditions even after adjusting for all other confounding factors. Previous studies have demonstrated that comorbid conditions exert substantial influence on clinical outcomes.7, 8 Older age was a strong, independent predictor of in-hospital mortality in the current study. It is likely that older patients have higher in-hospital mortality rates because of severe comorbid conditions, advanced stages of cancer, or a decreased functional status that precludes aggressive cancer therapy.9–11
The results of this study indicated that patients who developed complications from surgical or medical care were associated with higher odds of in-hospital mortality. Although this is not surprising, previously published data are lacking to compare population-based estimates of complications in hospitalized patients with HNC. Patients who had septicemia were associated with the highest odds of in-hospital mortality. This further confirms the results of previous studies, which showed that severe sepsis was a significant cause of death in patients with cancer.12, 13
Even after adjusting for the effects of insurance status, comorbid conditions, the presence of complications, and cancer type, hospital characteristics (such as large-bed hospitals, urban hospitals, and teaching hospitals) were associated with lower in-hospital mortality rates compared with small or medium-bed hospitals, rural hospitals, and nonteaching hospitals, respectively. These results are not surprising, because large-bed hospitals, urban hospitals, and teaching hospitals treat larger numbers of patients. Better outcomes in terms of lower in-hospital mortality rates at these hospitals may be because of improved structural components of the system, such as nursing support, anesthesia, intensive care unit care, and laboratory and radiology support, which may be the focus of future studies directed at seeking explanations for the differences in observed outcomes. Previous studies that investigated the correlations between hospital volumes and clinical outcomes after complex surgical procedures demonstrated that hospitals treating a high volume of patients had better outcomes in terms of lower in-hospital mortality rates, lower lengths of stays, and lower total charges.14–18
The current study had several limitations. We used administrative discharge data, and the data set did not have information regarding disease stage or tumor size. Thus, cancer severity is not captured. It is well known that patients who have advanced-stage cancers have poor outcomes. However, limiting the data set to patients who received inpatient care may have limited some of the variation in disease severity, because all of those patients had disease at least severe enough to warrant inpatient treatment. We did not adjust for race in the multivariate model that assessed in-hospital mortality, because nearly 30% of patients did not provide information about race. Previous studies have shown that race exerts significant influence on clinical outcomes,6 but this may indicate that race is only a proxy for other socioeconomic variables, such as income levels, educational levels, or insurance status. Finally, in the current study, we used in-hospital mortality as a proxy for assessing clinical outcomes for patients with HNC. A better approach would have been to examine 30-day, 60-day, or 90 day mortality after hospitalization for cancer. The NIS does not capture long-term mortality or follow-up clinical information. It is likely that uninsured or inadequately insured patients will have even worse long-term mortality rates, because they may not be able to afford postdischarge care. Therefore, the estimated, risk-adjusted, in-hospital mortality rates in the current study may have underestimated the true mortality rates for uninsured patients. Finally, these retrospective data identify predictors of outcome in a large sample set. The predictors may not apply uniformly to all patients at the individual level. These data can serve as guides to factors that may need to be included in more robust models that can be evaluated prospectively for their ability to predict disease outcomes in patients with HNCs.
In conclusion, the current study has provided insight into the influence of patient and hospital characteristics on in-hospital mortality after hospitalization for head and neck cancers. Patients who had comorbid conditions or who developed complications had high mortality rates. Patients who were grouped under the self-pay/no charge/others category were associated with higher odds of in-hospital mortality after admission and treatment for head and neck cancers. One potential explanation for this may be the delay in seeking treatment and later stage at presentation. Large beds or urban or teaching hospitals were associated with lower odds of in-hospital mortality.
REFERENCES
- 1American Cancer Society. Cancer reference information—oral cavity and oropharyngeal cancer; laryngeal cancer 2005. Available from URL: http://www.cancer.org/docroot/home/index.asp [accessed July 25, 2005].
- 2Healthcare Cost and Utilization Project (HCUP). Nationwide Inpatient Sample. Rockville, MD: Agency for Health Care Policy and Research, 2005.
- 3, , , . Comorbidity measures for use with administrative data. Med Care. 1998; 36: 8–27.
- 4, , , . The relation between health insurance coverage and clinical outcomes among women with breast cancer. N Engl J Med. 1993; 329: 326–331.
- 5, , , , , . Morbidity and mortality of colorectal carcinoma surgery differs by insurance status. Cancer. 2004; 101: 2187–2194.Direct Link:
- 6, , , , , . Effects of health insurance and race on colorectal cancer treatments and outcomes. Am J Public Health. 2000; 90: 1746–1754.
- 7, , , et al. Comorbidity and survival of elderly head and neck carcinoma patients. Cancer. 2001; 92: 2109–2116.Direct Link:
- 8, , . Mortality and morbidity in gastro-oesophageal cancer surgery: initial results of ASCOT multicentre prospective cohort study. BMJ. 2003; 327: 1192–1197.
- 9, , , et al. Do age and comorbidity impact treatment allocation and outcomes in limited stage small-cell lung cancer? A community-based population analysis. Int J Radiat Oncol Biol Phys. 2003; 55: 1321–1330.
- 10, , , . Age and the risk of in-hospital death: insights from a multihospital study of intensive care patients. J Am Geriatr Soc. 2002; 50: 1205–1212.Direct Link:
- 11, , . Functional status, health problems, age and comorbidity in primary care patients. Qual Life Res. 2001; 10: 141–148.
- 12, , , , , . Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Crit Care Med. 2001; 29: 1303–1310.
- 13, , , et al. Hospitalized cancer patients with severe sepsis: analysis of incidence, mortality, and associated costs of care. Crit Care. 2004; 8: R291–R298.
- 14, . Population-based study of relationships between hospital volume of prostatectomies, patient outcomes, and length of hospital stay. J Natl Cancer Inst. 1999; 91: 1950–1956.
- 15, , , . Association between volume and charges for most frequently performed ambulatory and nonambulatory surgery for bladder cancer. Is more cheaper? J Urol. 2004; 172: 1056–1061.
- 16, , , . Impact of hospital and surgeon volume on in-hospital mortality from radical cystectomy: data from the Health Care Utilization Project. J Urol. 2005; 173: 1695–1700.
- 17, , , et al. Hospital volume and surgical mortality in the United States. N Engl J Med. 2002; 346: 1128–1137.
- 18, , , . Impact of hospital volume on operative mortality for major cancer surgery. JAMA. 1998; 280: 1747–1751.