SEARCH

SEARCH BY CITATION

Keywords:

  • colorectal carcinoma;
  • insurance status;
  • surgical outcomes;
  • complications

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

BACKGROUND

Uninsured and underinsured patients are reported to be at an increased risk for impaired access to healthcare, delayed medical treatment, and the receipt of substandard care. These differences in care may result in disparities in surgical outcomes among patients with different types of insurance. In the current study, the authors examined associations between the insurance provider and short-term surgical outcomes after surgery for colorectal carcinoma and evaluated the extent to which two risk factors (comorbid disease and admission type) might explain any observed association.

METHODS

The authors conducted a nationally representative retrospective cohort study of 13,415 adults ages 40–64 years who were admitted for surgery for colorectal carcinoma to hospitals that participated in the Agency for Healthcare Research and Quality Healthcare Cost and Utilization Project National Inpatient Sample, releases 6 and 7, in 1997 and 1998. Multivariate logistic regression models were developed to describe the correlations between insurance status and the risks of a postoperative complication or postoperative death after adjustment for socioeconomic factors, comorbid conditions, and admission type.

RESULTS

Uninsured and Medicaid patients were found to have more emergent admissions and more comorbid disease compared with patients with private health insurance. Patients without private health insurance had higher rates of postoperative complications and in-hospital death compared with those patients with private insurance. After adjusting for patient and hospital characteristics, patients with Medicaid were found to be 22% more likely to develop a complication during their hospital admission (odds ratio [OR] of 1.22; 95% confidence interval [95%CI], 1.06–1.40) and 57% more likely to die postoperatively (OR of 1.57; 95% CI, 1.01–2.42) compared with patients with private insurance.

CONCLUSIONS

The current study findings suggest that the uninsured and Medicaid populations are at greater risk of developing postoperative complications and dying than the privately insured population due only in part to preexisting medical comorbidities and emergent admission type. Cancer 2004. © 2004 American Cancer Society.

Colorectal carcinoma is the third leading cause of cancer mortality in the U.S. There were an estimated 148,300 incident cases diagnosed in the year 2002.1 Although typically thought of as a disease of the elderly, up to 32% of new cases of colorectal carcinoma are diagnosed in people who are ages 40–64 years.2 Surgical resection is the mainstay of treatment for patients with nonmetastatic colorectal carcinoma.3 As many as 38% of patients with nonmetastatic disease ultimately die despite potentially curative surgery.4 Among patients undergoing colorectal procedures, approximately 20–35% experience complications related to surgery that may contribute to morbidity and mortality.5, 6 Previous data suggest that hospital and surgeon volume can affect surgical outcomes for cancer patients.7–10 Thus, we hypothesized that the perioperative care of colorectal carcinoma patients may play a role in patient outcome.

Health insurance status is alleged to have an important influence on access to cancer care, cancer screening, and long-term cancer outcomes.11–14 Approximately 29% of Americans age < 65 years are not covered by private health care insurance.15 Less than one-third of these people receive their health insurance coverage through the Medicaid program whereas the majority of the others remain uninsured. Research has shown that uninsured patients and Medicaid recipients are more likely to die of colorectal carcinoma than patients with private insurance. This finding has been partially explained by poor access to medical and surgical care,11 but for the uninsured and underinsured patients who do access surgical care for their colorectal carcinoma, data examining the relation between insurance coverage and surgical outcomes are to our knowledge lacking.

Insurance may influence surgical outcomes through its influence on 1) obtaining higher quality surgical care, 2) acuity of presentation for medical care, and 3) a patient's overall state of health.16–18 Alternatively, it is possible that insurance status may just be a marker of demand for medical care, available economic resources, and sociodemographic factors that may influence outcomes from surgical care independent of the influence of insurance status.

We hypothesized that insurance status can influence the outcomes of surgical care delivered to patients with colorectal carcinoma. In the current study, we asked several questions. Are Medicaid patients and the uninsured at greater risk for postoperative complications and postoperative in-hospital mortality after surgery for colorectal carcinoma compared with patients with private insurance? Furthermore, to what extent is the association between insurance status and short-term surgical outcomes explained by comorbid disease and admission type?

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Study Design and Sample

After obtaining approval from the Institutional Review Board of the University of Pennsylvania (#705326), we performed a population-based retrospective cohort study of patients who were hospitalized in 1997 and 1998 and enrolled in the Agency for Health Care Research and Quality (AHRQ) Healthcare Cost and Utilization Project (HCUP) National Inpatient Sample (NIS), releases 6 and 7. The NIS is a large national discharge database that records information regarding admissions that occur in > 1000 acute care hospitals in the U.S. The database captures single-episode inpatient hospital stay data and International Classification of Diseases-9th revision (ICD-9) coding for procedures and diagnoses.

Patients enrolled in the HCUP database in 1997 and 1998 were included in the current study if 1) they had a primary or secondary diagnosis of nonrecurrent, nonmetastatic colorectal carcinoma based on the ICD-9, Clinical Modifications (ICD-9-CM) codes, 2) they were ages 40–64 years, and 3) they were admitted for a surgical procedure on the colon or rectum based on the procedural codes of the ICD-9-CM. We restricted the patients to those ages 40–64 years to capture the group of people who were likely to be without Medicare health insurance and would be at risk for colorectal carcinoma. We also excluded those patients ages 40–64 years who were on Medicare because they were an odd group, likely comprised of disabled individuals and patients with end-stage renal disease.

Covariates

The primary variable of interest was insurance status, which was categorized as private, uninsured, or Medicaid. By AHRQ convention, the uninsured represent those admissions listed as either self-pay or no charge within the NIS database. Patient demographic factors and admission type were classified according to the NIS releases 6 and 7. Variables studied included age, race, admission type, median income based on zip code, patient geographic region based on hospital location, and gender. Hospital characteristics investigated included size of the hospital, hospital teaching status, and hospital location. These variables were obtained directly from the NIS database as defined by the American Hospital Association criteria.

Software obtained from the AHRQ was used to generate a comorbid index based on comorbid conditions in the database using the ICD-9-CM codes.19 These HCUP comorbidity indices have been previously shown to capture the majority of the comorbid illnesses experienced among surgical admissions.20 The comorbid conditions included in the current study were congestive heart failure, cardiac arrhythmias, valvular disease, pulmonary circulation disorders, peripheral vascular disorders, hypertension, paralysis, other neurologic disorders, chronic pulmonary disease, diabetes uncomplicated, diabetes complicated, hypothyroidism, renal failure, liver disease, peptic ulcer disease excluding bleeding, acquired immunodeficiency syndrome, lymphoma, rheumatoid arthritis/collagen vascular diseases, coagulopathy, obesity, weight loss, fluid and electrolyte disorders, blood loss anemia, deficiency anemias, alcohol abuse, drug abuse, psychoses, and depression. In addition, we calculated the sum of the number of comorbid conditions listed for each admission. After inspection of the distribution of the sum variable, an additional categoric variable, comorbid burden, was created to assess the burden of comorbid disease carried by each patient. Comorbid burden was defined as 0 for patients without any comorbid disease, 1–2 for patients with one or two comorbid illnesses, and, 3+ for patients with three or more comorbid illnesses.

Outcomes

We evaluated two outcome measures –postoperative complications and in-hospital mortality. Complications of surgery were defined as those known from the literature to be associated with colorectal procedures.5, 6 Complications were identified in the database using the ICD-9CM codes and included respiratory complications, postoperative myocardial infarction, wound disruption, postoperative ileus, sepsis/wound infection, hemorrhage, postoperative stroke, postoperative complication not otherwise specified, and return to the operating room. A binary variable was created to establish whether any complication occurred for each discharge. In-hospital mortality was directly obtained from the NIS releases 6 and 7 databases.

Statistical Methods

We used ACCESS XP (Microsoft Corporation, Redmond, WA) for data management and STATA version 6.0 (Stata Corporation, College Station, TX) software for analysis. Differences in population proportions were derived using chi-square analysis. A statistical test was significant if the P value was < 0.05. A Fisher exact test was performed when the sample size was too small for chi-square analysis. Multivariate logistic regression models were developed to describe the relation between insurance status and either complications or in-hospital death with adjustment for socioeconomic factors, comorbid conditions, and admission type. A covariate was included in the multivariate model if it was found to be significantly associated with both insurance status and the outcome of interest in the univariate analyses. Stratified analyses were performed to assess for interactions. Model calibration was assessed with the Hosmer-Lemeshow chi-square statistic.21 In the final analysis, we reported adjusted odds ratios and their 95% confidence intervals (95% CIs). In addition, to address the possibility that our findings might be due to comorbid conditions and admission type, we also ran the regressions without adjustment for these covariates and then reported the odds ratios. The number of procedures performed at the individual hospitals was too small to perform a cluster analysis.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

The final study cohort included 13,415 patients. To arrive at this group of study subjects, we identified 169,206 subjects with admissions for procedures of the colon and rectum; 56,493 of these subjects had a diagnosis of nonmetastatic carcinoma of the colon or rectum. Of these, 15,183 patients were in the 40–64 years age group. Finally, 1776 subjects were excluded from the analysis for the following reasons: unrelated diagnosis-related groups (n = 62; 0.41%), missing information concerning insurance status (n = 63; 0.41%), Medicare health coverage (n = 991; 6.5%), insurance status not specified (n = 647; 4.3%), or missing information regarding in-hospital mortality (n = 13; 0.09%;). Excluded patients were more likely to be Hispanic, to have a median household income < $35,000 or an income that was not specified, and to be admitted emergently compared with patients included in the study. At least 5% of the admissions were missing information regarding median income (n = 756; 5%) or race (n = 3342; 22%); these subjects remained in the study cohort and the missing information in these categories was coded as not specified.

The characteristics of patients included in the final cohort (n = 13,415) are listed in Table 1. Approximately 85% of the patients had private insurance (n = 11,475), 6% were uninsured (n = 758), and 9% were recipients of Medicaid (n = 1182). Approximately 14% of the admissions were classified as emergent and a comorbid illness was identified in 57% of patients. Among the patients in the final study cohort, the median age was 57 years.

Table 1. Characteristics of the Final Study Cohort by Insurance Status (%)
CharacteristicFinal Cohort (n = 13,415)Private (n = 11,475)Medicaid (n = 1182)Uninsured (n = 758)
Insurance status    
 Private85.5
 Medicaid8.8
 Uninsured5.7
Admission type    
 Emergent13.911.626.726.6
 Nonemergent86.188.473.371.4
Comorbid conditions    
 042.844.131.640.5
 1–248.548.052.750.3
 3+8.77.915.79.2
Age (yrs)    
 40–4920.520.123.721.2
 50–5946.346.743.445.0
 60–6433.233.232.933.8
Gender    
 Male54.655.548.050.9
 Female45.444.552.049.1
Race    
 White60.963.543.348.4
 Black9.17.920.49.8
 Hispanic4.74.09.67.1
 Other25.324.526.734.7
Median income ($)    
 0–34,99945.642.766.657.8
 35,000+49.452.330.735.6
 Not specified5.05.12.76.6
Region    
 Northeast20.921.519.514.3
 Midwest22.723.316.922.8
 West38.737.641.250.9
 South17.617.722.312.0
Hospital size    
 Small13.913.615.515.2
 Medium29.328.732.932.9
 Large56.857.751.651.9
Hospital location    
 Rural13.212.514.721.0
 Urban86.887.585.379.0
Hospital teaching status    
 Nonteaching56.857.447.561.8
 Teaching43.242.652.538.2

The characteristics of the study patients based on insurance status also are listed in Table 1. The distribution of comorbid conditions by insurance status can be found in Table 2. A lack of insurance and Medicaid receipt were associated with increased comorbidity and more emergent admissions compared with patients with private insurance (P < 0.0001). Medicaid recipients were found to have the highest burden of comorbid illness followed by the uninsured patients, and patients without private insurance were found to have more comorbid conditions than those with private insurance (P < 0.0001).

Table 2. Comorbid Conditions by Insurance Status (%)
 Private (n = 11,475)Medicaid (n = 1182)Uninsured (n = 758)
  • CHF: congestive heart failure; AIDS: acquired immunodeficiency syndrome.

  • a

    All P values were < 0.001 based on a chi-square test of category versus insurance status. Conditions that were tested but were not found to be statistically significant and therefore were not listed in the table were valvular disease, pulmonary circulation disorders, diabetes complicated, rheumatoid arthritis/collagen vascular disease, coagulopathy, obesity, hypertension, and depression.

No comorbid conditionsa44.131.640.5
CHF1.54.72.8
Cardiac arrhythmias4.66.43.7
Peripheral vascular disorders0.61.80.4
Paralysis0.20.70.0
Hypertension23.723.620.7
Other neurologic disorders0.62.00.1
Chronic pulmonary disease6.110.85.5
Diabetes uncomplicated9.012.79.6
Hypothyroidism2.91.70.9
Renal failure0.41.50.5
Liver disease1.32.71.7
Peptic ulcer disease, not bleeding0.91.91.1
AIDS0.00.60.1
Weight loss2.15.03.4
Fluid and electrolyte disorders10.416.218.2
Blood loss anemia5.38.17.3
Deficiency anemias9.612.511.5
Alcohol abuse1.13.43.4
Drug abuse0.20.90.4
Depression1.43.11.9
Psychoses0.53.80.4

The overall complication rate was 24% (n = 3190). The overall inhospital mortality rate was 1% (n = 194). The most common complication types were postoperative ileus (13%), respiratory complications (5%), and sepsis/infectious complications (3%).

Univariate Analyses

Univariate analyses, as illustrated in Table 3, demonstrated that insurance status was associated with significant differences in postoperative outcomes. Medicaid and uninsured patients were found to have significantly higher complication rates than privately insured patients (26.7% and 24.5% vs. 21.1%, respectively; P < 0.001). Medicaid and uninsured patients were found to have significantly higher inhospital mortality rates than privately insured patients (2.8% and 2.2% vs. 0.9%, respectively; P < 0.001). Several risk factors that were found to be more prevalent among patients without private health insurance (emergent admission type, comorbid burden, and advanced age) were found to be associated with higher rates of both postoperative complications and mortality. Emergent admission was more closely associated with higher complication rates and in-hospital mortality rates compared with non-emergent admissions (P < 0.001). Patients with three or more comorbid conditions were reported to have higher complication rates and in-hospital mortality rates than patients with none to two comorbid conditions (P < 0.001). And finally, patients ages 60–64 years were found to have significantly higher complication rates and in-hospital mortality rates compared with patients ages 40–59 years.

Table 3. Outcomes after Surgery for Colorectal Carcinoma (n = 13,415) (%)
CategoryNo.ComplicationP valueIn-hospital mortalityP value
Overall13,41521.81.2
Insurance status     
 Private11,47521.1< 0.0010.9< 0.001
 Medicaid118226.7 2.8 
 Uninsured75824.5 2.2 
Admission type     
 Emergent185826.1> 0.0013.7< 0.001
 Nonemergent11,55721.1 0.8 
Comorbid burden     
 0574318.9< 0.0010.5< 0.001
 1–2651223.5 1.4 
 3+116027.0 3.0 
Age (yrs)     
 40–49274520.40.0020.8< 0.001
 50–59621621.2 1.0 
 60–64445423.6 1.7 
Gender     
 Male731723.7< 0.0011.30.202
 Female609519.5 1.0 
Race     
 White816821.20.0921.10.730
 Black122521.5 1.5 
 Hispanic63123.1 1.0 
 Not specified339123.2 1.2 
Median income ($)     
 0–34,999612021.50.3601.40.050
 35,000+663021.9 0.9 
 Not specified66523.9 1.5 
Region     
 Northeast280118.2< 0.0010.90.185
 Midwest304623.3 1.3 
 West518722.0 1.3 
 South238123.8 0.9 
Hospital teaching status     
 Nonteaching758421.70.6811.20.345
 Teaching583122.0 1.1 
Hospital size     
 Small185622.80.1421.00.490
 Medium391322.5 1.3 
 Large764621.2 1.1 
Location     
 Rural13,25923.10.1780.90.191
 Urban15621.6 1.2 

It is interesting to note that median income and race were not found to be associated significantly with differences in complication rates or in-hospital mortality rates. Similarly, hospital characteristics including teaching status, hospital size, and location also were not found to be significantly associated with complication rates or in-hospital mortality rates.

Multivariate Models

After adjustment for socioeconomic factors, comorbid conditions, and admission type, insurance status remained a significant determinant of postoperative complication and in-hospital mortality (Table 4). Uninsured status and Medicaid insurance were found to be associated with higher complication rates than private insurance. The adjusted odds ratio of having a complication was 1.22 (95% CI, 1.06–1.40) for patients with Medicaid compared with patients with private insurance; however, there was not a statistically significant difference in the odds ratio of a complication between uninsured patients and those with private insurance. The adjusted odds ratio for in-hospital mortality was 1.57 (95% CI, 1.01–2.42) for patients with Medicaid compared with patients with private insurance. Emergent admission type, high comorbid burden (3+), and several specific comorbid conditions were associated with a higher odds ratio of postoperative complications and in-hospital mortality. Female patients and those treated in the Northeast region of the U.S. were less likely to experience a postoperative complication than male patients or those living in other regions of the country (for both variables, P < 0.001). Age > 59 years was found to be associated with higher likelihood of in-hospital mortality compared with patients ages 40–49 years (P = 0.017). No significant interactions were identified.

Table 4. ORs and 95% CIs for Outcomes
Insurance typeUnadjusted OR95% CIAdjusted ORab95% CI
  • OR: odds ratios; 95% CI: 95% confidence interval.

  • Boldface type indicates a P value ≤ 0.05.

  • a

    The adjusted odds ratio in the complication model was adjusted for age, gender, admission type, comorbid burden, congestive heart failure, arrhythmias, chronic lung disease, renal failure, weight loss, fluid and electrolyte disorders, and region. The adjusted odds ratio in the death model was adjusted for age, admission type, comorbid burden, congestive heart failure, arrhythmias, peripheral vascular disorders, paraplegia, other neurologic disorders, renal failure, liver disease, weight loss, fluid and electrolyte disorders, and alcohol abuse.

  • b

    Complication model: Hosmer-Lemeshow chi-square test = 6.75; P = 0.564. In-hospital mortality model: Hosmer-Lemeshow chi-square test = 5.74; P = 0.571.

Complication Model    
Private1.001.00
Medicaid1.36(1.18–1.55)1.22(1.06–1.40)
Uninsured1.21(1.02–1.44)1.12(0.94–1.34)
Death Model    
Private1.001.00
Medicaid3.08(2.07–4.57)1.57(1.01–2.42)
Uninsured2.46(1.47–4.13)1.56(0.91–2.69)

Table 5 illustrates the extent to which the association between insurance status and short-term surgical outcomes can be explained by comorbid disease and admission type. The magnitude of the effect of comorbid disease and admission type was demonstrated by comparing models with and without these variables. When comorbid conditions were removed from the model, the odds ratio for Medicaid jumped from 1.57 to 2.30 in the in-hospital mortality model. Similar findings were shown for uninsured patients in the in-hospital mortality model. The omission of admission type from the in-hospital mortality model resulted in the odds ratio increasing from 1.57 to 1.91 for Medicaid patients, which is a much smaller difference than that observed with the omission of comorbid conditions. However, the relation between in-hospital mortality and uninsured status was even more pronounced when the adjustment for admission type was omitted, with the odds ratio jumping from a statistically insignificant 1.56 to a statistically significant 1.94. It is important to note that the final adjusted results showed a difference in outcomes across all measures between those patients with private health insurance and those without it.

Table 5. Influence of Comorbid Conditions and Admission Type on the Correlation between Insurance Status and Surgical Outcomes (ORs [95% CIs])
Insurance typeUnadjusted ORFinal modelComorbid conditions omittedAdmission type omitted
  1. ORs: odds ratios; 95% CIs: 95% confidence intervals.

  2. Boldface type indicates a P value ≤ 0.05.

  3. The complication final model was adjusted for patient demographics (age, gender, and hospital region), comorbid conditions (comorbid burden, congestive heart failure, arrhythmias, chronic lung disease, renal failure, weight loss, and fluid and electrolyte disorders), and admission type.

  4. The death final model was adjusted for patient demographics (age), comorbid conditions (comorbid burden, congestive heart failure, arrhythmias, peripheral vascular disorders, paraplegia, other neurologic disorders, renal failure, liver disease, weight loss, fluid and electrolyte disorders, and alcohol abuse), and admission type.

Complication Model    
Private1.001.001.001.00
Medicaid1.361.221.32 (1.15–1.52)1.25 (1.08–1.44)
Uninsured1.211.121.16 (0.98–1.38)1.15 (0.97–1.38)
Death Model    
Private1.001.001.001.00
Medicaid3.081.572.30 (1.53–3.46)1.91 (1.25–2.92)
Uninsured2.461.561.74 (1.02–2.95)1.94 (1.13–3.32)

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Utilizing a large population-based cohort of colorectal carcinoma patients receiving surgical treatment, we found that uninsured and Medicaid patients have worse unadjusted and adjusted postoperative outcomes than patients with private health insurance. We found that Medicaid patients had a greater adjusted likelihood of postoperative complications and a higher adjusted risk of in-hospital mortality than patients with private health insurance.

These findings are in keeping with the medical literature that has noted worse outcomes among patients without private health insurance compared with privately insured patients.11, 13, 22, 23 To the best of our knowledge, there has been no report published to date regarding the effects of insurance status on short-term surgical outcomes.

There are several mechanisms that might explain the association between insurance status and surgical outcomes. First, the surgeon may be less skilled. Although private insurance patients with colorectal carcinoma may be referred to the surgeon with the greatest expertise, uninsured and Medicaid populations often are assigned to less experienced surgeons. Second, because of barriers to care, the uninsured and Medicaid populations often receive their care in overcrowded emergency departments,24–26 which previously have been shown to be subject to high rates of medical errors,27 and these patients may receive substandard preoperative resuscitation and care, resulting in poor postoperative outcomes. Finally, poor access to health care for the underinsured and prohibitive costs lead to delayed diagnoses, worse overall health (i.e., more comorbid disease), and presentation at more advanced stages of disease.16, 17, 28

The results of the current study have demonstrated that some of the differences in insurance-related postoperative outcomes could be attributed to comorbid disease and admission type. This finding is in agreement with other studies that have shown differences in surgical outcomes to be associated with these risk factors.18, 29–31 However, the current study findings demonstrate the deleterious effects of a high comorbid disease burden and emergent presentation on surgical patients without private health insurance. To our knowledge, this aspect of impaired healthcare delivery to these populations has not been previously appreciated. Although the other potential mechanisms could not be tested directly in the current study, the fact that the negative association between insurance status and surgical outcomes remained strong after adjustment for comorbid disease and admission type indicates that other possible explanations also may be important.

The remaining negative association between insurance and surgical outcomes may be a signal that insurance status is a marker for other factors that result in worse surgical outcomes. It may be a signal of less medical demand, less health awareness, and reduced compliance, or other socioeconomic factors associated with poor surgical outcomes. The latter appears unlikely in the current study population and is at best controversial in the literature.32 In our cohort, race was not found to be associated significantly with any of our outcome measures. The median income of the patient population was not found to be significantly associated with either outcome. Therefore, at least in the current study population, income and race do not explain the disparate outcomes observed between the uninsured and Medicaid patients and the privately insured patients. Similarly, lifestyle factors common to the poorer patients (such as alcohol abuse, drug use, psychiatric conditions, and obesity) were included in the analyses, and Medicaid patients and those who were uninsured still had worse outcomes than the privately insured patients. Therefore, although insurance may be a marker for these negative risk factors, in our model insurance status functioned as an independent determinant of surgical outcomes, and was not a mere marker of socioeconomic status.

The current study has several limitations. By using the HCUP database, we limited the scope of the current study to events that occurred during the surgical admission. Many Medicaid patients receive coverage during hospitalization and therefore the Medicaid population in the current study may contain some uninsured patients. Similarly, many patients classified as being insured actually may have inadequate coverage33 and should not have been classified as insured if they lack the proposed service benefits of having insurance. In addition, information regarding income was available only according to zip code. Although we included a comprehensive adjustment for comorbid conditions, we could not account for severity or stage of disease. It is possible that our adjustments were incomplete; however, because the uninsured and Medicaid patients have been shown to have worse overall health than the privately insured patients,16 a better adjustment might weaken the association between insurance status and outcome while strengthening the finding that access to care affects health status and therefore impacts surgical outcomes as well. To our knowledge, national databases that control for stage of disease are not available in this age group. The uninsured and Medicaid populations are reported to present with more advanced disease than patients with private health insurance.13 However, although advanced stage of disease is a contributing factor to the poor long-term outcomes reported in the Medicaid and uninsured patients,11, 14 stage of disease does not generally alter the colorectal procedures performed for nonmetastatic colorectal carcinoma. With rare exceptions, a colon resection is always performed for colorectal carcinoma without systemic metastases. Therefore, short-term surgical outcomes are unlikely to be affected by stage of disease at the time of presentation for nonmetastatic disease. Similarly, we did not adjust for the anatomic resection performed; however, we have no reason to suspect that the distribution of tumors along the colon would differ by insurance status.

Previous studies evaluating the role of insurance status on cancer outcomes have focused on screening practices and long-term outcomes after multiple therapeutic interventions. The current study focused on short-term surgical outcomes that are crucial to the long-term survival of these patients. Similar to other studies, our findings indicate that future efforts need to be made to improve healthcare delivery to vulnerable populations. In addition, further studies are needed to evaluate the processes of perioperative care provided to the at-risk populations to help eliminate disparities in care.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES
  • 1
    American Cancer Society. Cancer facts and figures 2002. Atlanta: American Cancer Society, Inc., 2002. Available from URL: http://www.cancer.org/eprise/main/docroot/STT_1x_Cancer_Facts_Figures_2002 [accessed October 9, 2002]
  • 2
    National Cancer Institute. SEER cancer statistics review 1973-1997. Table VI-3. Available from URL: http://www.nci.nih.gov [accessed May 9, 2001].
  • 3
    Physician Data Query. Colon cancer treatment for health professionals. Bethesda, MD: National Cancer Institute, 2004. Available from URL; http://www.cancer.gov/cancerinfo/pdq/treatment/colon/healthprofessional/ [accessed April 21, 2004].
  • 4
    Kaufman HS, Efron J, Wexner SD. Colon cancer and rectal cancer. In: CameronJL, editor. Current surgical therapy. 7th edition. St Louis: Mosby, Inc., 2001: 229245.
  • 5
    Schrock TR. Colorectal procedures. In: WilmoreDW, CheungLY, HarkenAH, et al., editors. ACS surgery: principles and practice. New York: WebMD Professional Publishing, 2002: 825836.
  • 6
    Kodner IJ, Fry RD, Fleshman JW, Birnbaum EH. Colon, rectum and anus. In: SchwartzSI, editor. Principles of surgery. 6th edition. New York: McGraw-Hill Inc., 1996: 11911306.
  • 7
    Schrag D, Cramer LD, Bach PB, Cohen AM, Warren JL, Begg CB. Influence of hospital procedure volume on outcomes following surgery for colon cancer. JAMA. 2000; 284(23): 30283035.
  • 8
    Begg CB, Cramer LD, Hoskins WJ, Brennan MF. Impact of hospital volume on operative mortality for major cancer surgery. JAMA. 1988; 280: 17471751.
  • 9
    Bach PB, Cramer LD, Schrag D, Downey RJ, Gelfand SE, Begg CB. The influence of hospital volume on survival after resection for lung cancer. N Engl J Med. 2001; 345(3): 181188.
  • 10
    Gordon TA, Bowman HM, Bass EB, et al. Complex gastrointestinal surgery: impact of provider experience on clinical and economic outcomes. J Am Coll Surg. 1999; 189(1): 4656.
  • 11
    Roetzheim RG, Pal N, Gonzalez EC, Ferrante JM, Van Durme DJ, Krischer JP. Effects of health insurance and race on colorectal cancer treatments and outcomes. Am J Public Health. 2000; 90(11): 17461754.
  • 12
    Judith H, Elizabeth K, Caterina K, et al. The importance of health insurance as a determinant of cancer screening: evidence from the Women's Health Initiative. Prev Med. 2000; 31: 261270.
  • 13
    Roetzheim RG, Pal N, Tennant C, et al. Effects of health insurance and race on early detection of cancer. J Natl Cancer Inst. 1999; 91: 14091415.
  • 14
    Ayanian JZ, Kohler BA, Abe T, Epstein AM. The relation between health insurance coverage and clinical outcomes among women with breast cancer. N Engl J Med. 1993; 329: 326331.
  • 15
    Centers for Disease Control and Prevention, National Center for Health Statistics. National Health Interview Survey. Health, United States, 2000. Atlanta: Centers for Disease Control and Prevention, 2000. Available from URL: http://www.cdc.gov [accessed May 9, 2001].
  • 16
    Franks P, Clancy CM, Gold MR. Health insurance and mortality: evidence from a national cohort. JAMA. 1993; 270: 737741.
  • 17
    Weissman JS, Stern R, Fielding SL, Epstein AM. Delayed access to healthcare: risk factors, reasons and consequences. Ann Intern Med. 1991; 114: 325331.
  • 18
    MacArthur DC, Nixon SJ, Aitken RJ. Avoidable deaths still occur after large bowel surgery. Br J Surg. 1998; 85: 8083.
  • 19
    Comorbidity software. Rockville, MD: Agency for Healthcare Research and Quality, 2001. Available from URL: http://www.ahrq.gov/data/hcup/comorbid.htm [accessed October 12, 2001].
  • 20
    Elixhauser A, Steiner C, Harris DR, Coffey RM. Comorbidity measures for use with administrative data. Med Care. 1998; 36: 827.
  • 21
    Hosmer DW, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons Inc., 1989.
  • 22
    Sada M, French WJ, Goldberg SL, et al. Comparison of HMO and fee-for-service utilization of angiography and in-hospital mortality in patients with acute myocardial infarction. J Am Coll Cardiol. 1998; 31: 14741480.
  • 23
    McDavid K, Tucker TC, Sloggeett A, Coleman MP. Cancer survival in Kentucky and health insurance coverage. Arch Intern Med. 2003; 163: 21352144.
  • 24
    Cornelius L, Beauregard K, Cohen J. Usual sources of medical care and their characteristics. AHCPR Publication No. 91-0042. Rockville, MD.: Agency for Health Care Policy and Research, 1991.
  • 25
    Cohen JW. Medicaid policy and the substitution of hospital outpatient care for physician care. Health Serv Res. 1989; 24: 3366.
  • 26
    Yudkowsky BK, Cartland JDC, Flint SS. Pediatrician participation in Medicaid: 1978 to 1989. Pediatrics. 1990; 85: 567577.
  • 27
    Leape LL, Brennan TA, Laird NM, et al. The nature of adverse events in hospitalized patients: results of the Harvard Medical Practice Study II. N Engl J Med. 1991; 324: 377384.
  • 28
    Shea S, Misra D, Ehrlich MH, Field L, Francis CK. Predisposing factors for severe, uncontrolled hypertension in an inner-city minority population. N Engl J Med. 1992; 327: 776781.
  • 29
    Liu LL, Leung J. Predicting adverse postoperative outcomes in patients aged 80 years or older. J Am Geriatr Soc. 2000; 48: 405412.
  • 30
    Longnecker DE, Murphy FL. Introduction to anesthesia. Philadelphia: W.B. Saunders Company, 1997: 18.
  • 31
    Kee F, Wilson RH, Harper C, et al. Influence of hospital and clinician workload on survival from colorectal cancer: cohort study. BMJ. 1999; 318: 13811385.
  • 32
    Laantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality: results from a nationally representative study of US adults. JAMA. 1998; 279: 17031708.
  • 33
    Farley PJ. Who are the underinsured? Milbank Q. 1985; 63: 476503.