Comorbid Illness and the Early Detection of Cancer.
Background Comorbidity may be associated with later detection of cancer.
Methods. Incident cases of colorectal, breast, and prostate cancer and melanoma were determined from the 1994 Florida state tumor registry (N = 32,074). The relationship between comorbidity and late stage at diagnosis was examined using multiple logistic regression.
Results. Patients with comorbid conditions had greater odds of late stage diagnosis for each of the four cancers (colorectal, melanoma, breast, and prostate). Higher mortality rates were observed among patients with comorbid illness, not as a result of later stage at diagnosis, but rather due to their underlying disease.
Conclusions. Comorbidity was associated with later stage diagnosis. Further research is needed to determine mechanisms by which comorbidity might influence stage at diagnosis.
COMORBIDHY is defined as the presence of concurrent chronic illnesses.  Comorbidity generally increases with advancing age and may be the reason behind age-related differences in cancer diagnosis, treatment, and outcome. [2-7] Because the incidence of most cancers increases with advancing age, it is not surprising that comorbidity has been frequently found among patients with cancer.  Comorbidity has often been associated with less aggressive treatment and poor cancer outcomes. [9-15] It is less clear, however, whether comorbidity influences the early detection of cancer.
Comorbidity could influence stage at diagnosis of cancer in a number of ways. Some have argued that comorbidity might mask early symptoms of cancer and lead to later stage diagnosis. [16,17] Comorbidity could also serve as a competing demand for primary care physicians, decreasing the likelihood of cancer screening recommendations.  Finally, patients and physicians may place less value on cancer screening in the face of competing causes of morbidity and mortality.  A number of clinic based studies have in fact found lower cancer screening rates among patients having comorbid conditions. [20-22]
On the other hand, some aspects of comorbidity could facilitate cancer screening. Patients having comorbid conditions generally have increased contact with the health care system and thus have more opportunities for preventive care. Stange et al  found that family physicians delivered some preventive service in 39% of visits for chronic disease. Some population-based studies have shown higher cancer screening rates among patients having comorbidity. [24-26]
Few studies have assessed the impact of comorbidity on cancer stage at diagnosis. Studies by Satariano and Ragland  and West et al  found trends toward earlier stage at diagnosis of breast cancer for patients having comorbid conditions. These studies were limited by not using multivariate analysis to examine the relationship between comorbidity and stage at diagnosis. A separate study by Satariano  with the use of multivariate analysis showed a statistically nonsignificant trend toward earlier diagnosis for breast cancer patients having comorbidity.
It is unclear, therefore, whether comorbidity affects stage at diagnosis for patients with breast cancer. Whether comorbidity affects the early diagnosis of other cancers amenable to screening also remains unknown. We used administrative data from the state of Florida to determine whether patients having comorbid conditions were more likely to have diagnosis at late stage and if so, whether resultant later stage at diagnosis had an impact on their survival. We hypothesized that cancer in patients having comorbidity would more likely be diagnosed at late stage and that these patients would have poorer survival than those without comorbid conditions.
We studied 1994 Florida incident cases (the most recent year for which all relevant data were available) of four cancers for which screening is associated with detection of early stage disease: colorectal, breast (female only), prostate, and melanoma (N = 34,616). [28-36] Cervical cancer was not included because of different reporting requirements for this site (in situ cervical cancers are not reportable). Incident cases were identified from the Florida Cancer Data System (FCDS), Florida's population-based statewide cancer registry. The FCDS has well-established methods to ensure complete case-finding, including cooperative arrangements with other state tumor registries, linkage with other databases, and ad hoc audits of reporting facilities. The FCDS is a member of the North American Association of Central Cancer Registries, whose audits have estimated the completeness of case ascertainment for the period 1990 to 1994 to be 97%.
To include information that is not routinely available from the FCDS (insurance payer, comorbidity, socioeconomic status, urban/ nonurban residence), cases were linked with state discharge abstracts and the 1990 US Census. The State of Florida Agency for Health Care Administration (AHCA) maintains discharge abstracts for admissions to all nonfederal acute care hospitals, and patient visits to ambulatory surgical centers, freestanding radiation therapy centers, and diagnostic imaging centers. Data abstracted include Social Security number, date of birth, sex, race-ethnicity, discharge diagnoses (up to 10), procedures done (up to 10), and insurance payer. The methods of linking FCDS and AHCA records have been previously described  and resulted in a match rate of 82.8%, a rate similar to that achieved in a comparable study. 
The 1990 United States Census was used to obtain aggregate measures of socioeconomic status by either Census tract or, if unavailable, by ZIP code. Each individual was assigned the median income and education level of either the Census tract (87% of cases) or ZIP code (13% of cases) of their residence. The use of Census-derived measures of socioeconomic status have been validated in previous studies. [39-12] Patients were defined as having an urban residence if they lived in a ZIP code that was classified as 100% urban by the US Census. Patients were defined as having nonurban residence if they lived in a ZIP code that contained outside urban or rural components.
Stage at diagnosis was defined using the SEER Site-Specific Summary Staging Guide.  Stage at diagnosis is based on a combination of pathologic, operative, and clinical assessments available within 2 months of diagnosis. Stage categories included in situ, local (invasive disease confined to the organ of origin), regional (direct spread to adjacent structures or regional lymph nodes), and distant (distant metastases). For these analyses, stage at diagnosis was reclassified as either early stage (insitu, local) or late stage (regional, distant). Stage at diagnosis was available for 32,074 FCDS cases (92.7% of all cases: colon 93.5%, melanoma 93.6%, breast 95.5%, prostate 88.9%). Vital status was assessed through December 31, 1997, using FCDS-derived mortality files. The length of time from diagnosis to death, or until the last follow-up contact, was measured in months.
Comorbidity was determined using methods described by Deyo et al  and Charison et al.  The Charlson comorbidity index was chosen because it has been validated specifically in studies of cancer patients.  The Charlson comorbidity index is not an exhaustive list of all possible comorbid conditions but is rather a weighted index of 19 selected categories of disease that were found to be associated with mortality and other important health outcomes. Charlson comorbid conditions (and their corresponding weightings) include myocardial infarction (1), congestive heart failure (1), peripheral vascular disease (1), cerebrovascular disease (1), dementia (1), chronic pulmonary disease (1), connective tissue disease (1), peptic ulcer disease (1), mild liver disease (1), moderate/severe liver disease (3), diabetes without complications (1), diabetes with complications (2), hemiplegia (2), renal disease (2), and acquired immunodeficiency syndrome (6). Increasing scores on the Charlson comorbidity index reflect an increasing burden of comorbid conditions. [15,45,46]
We identified Charlson comorbid conditions using all inpatient and ambulatory discharge abstracts for the calendar year 1994. We used methods described by Deyo et al  that were specifically developed to measure Charlson comorbidity from administrative databases. Patients having no Charlson comorbid condition identified in discharge abstracts were assigned a comorbidity score of zero, as were patients who had no record of inpatient or outpatient admissions during the calendar year. We created two variables related to comorbidity: a dichotomous variable indicating the presence or absence of comorbid conditions and a second categorical variable defined by three levels of comorbidity (0, 1, 2+) based on the patient's index score.
All analyses were conducted separately by site. The proportion of cases diagnosed at a late stage (regional or distant) was first compared for patients having any comorbid condition using the chi-square test. We then used the Mantel-Haenszel [[chi].sup.2] test for trend to examine whether the likelihood of late stage diagnosis increased with increasing severity of overall comorbidity.
The multivariate relationship between late stage at diagnosis and comorbidity was then examined using multiple logistic regression. Indicator variables were created for the presence or absence of comorbid conditions, and for the three comorbid index categories (0, 1, 2+). Based on previous research, the following variables were included in all logistic models to control for potential confounding: age, sex (if appropriate), race-ethnicity, marital status, educational level, income level, urban residence, and insurance payer. [47-49] The statistical significance of individual indicator variables was assessed using the chi-square likelihood ratio test. 
To determine whether findings might differ, we also repeated logistic models with cases restricted to invasive cancers only. For colorectal, breast, and prostate cancers, we also repeated analyses with cases restricted to ages for which screening is most often recommended and for which physician agreement is high (ages 50 to 75 years) We also repeated analyses after excluding those FCDS cases that did not match with inpatient and ambulatory discharge abstracts.
Survival was examined for patients having Charlson comorbid conditions compared with those who did not. Survival curves were constructed using the Kaplan-Meier product-limit method.  The duration of potential follow-up varied from 36 months to a maximum of 48 months; depending on the patient's date of diagnosis. Survival curves were compared using the Mantel-Cox log-rank test.  We examined the adjusted risk of death from all-cause mortality for patients with and without comorbidity using Cox proportional-hazards regression analysis. Hazard rates were adjusted for other factors that might be associated with mortality, including age, sex, marital status, smoking status, cancer stage at diagnosis, and community measures of socioeconomic status. To determine the degree to which greater mortality among patients with comorbidity was the result of later stage at diagnosis, we repeated models both with and without variables for stage at diagnosis. We did not have data on the cause of death to allow analysis o f cancer specific mortality.
All analysis was conducted using SAS statistical software (LOGISTIC, LIFETEST, PHREG procedures).  We present 95% confidence intervals for adjusted odds and risk ratios and unless specified, all P values are two-tailed. Statistical significance was determined using an [alpha] level of .05.
The study population consisted of the 32,074 Florida residents who had colorectal, breast, or prostate cancer or melanoma diagnosed in 1994 and for whom information on stage was available (Table 1). Reflecting the demographics of the state, most patients were over age 65, and Medicare was the most common type of insurance. The majority of the breast and prostate cancers and melanomas were diagnosed at an early stage (either in situ or local). The majority of colorectal cancers, however, were diagnosed at a late stage (either regional or distant).
The percentage of patients having Charlson comorbid conditions varied from 7% for melanoma to 30% for colorectal cancer (Table 1). Patients having any comorbid condition were more likely to have diagnosis at late stage for each of the four cancer sites examined (Table 2). For breast and prostate cancer, the likelihood of late stage diagnosis increased in a dose-response fashion with increasing levels of comorbidity. This was not the case for patients with colorectal cancer or melanoma. In multivariate analysis, the presence of any comorbid condition was a significant predictor of late stage diagnosis for all four cancer sites that were examined (Table 3). The magnitude of the effect ranged from 17% greater odds of late stage diagnosis for colorectal cancer to a 62% increased odds of late stage diagnosis for patients with melanoma. Again, only for breast and prostate cancers did the effects of comorbidity demonstrate a dose-response, patients with comorbidity scores of 2 or greater having greater likelihood o f late stage diagnosis than patients with a score of 1.
Results were similar to those described when logistic models were repeated with cases restricted to invasive cancers only. In addition, results did not vary when cases were restricted to ages for which screening is most often recommended (ages 50 to 75 years) or when unmatched FCDS cases were excluded from the analysis (data not presented). We also did not find evidence of statistical interaction between the effects of comorbidity and other patient characteristics (age, sex, race-ethnicity).
Survival for patients with and without Charlson comorbidity is presented in Figures 1 through 4. For each cancer type examined, the proportion of patients surviving through the 4-year follow-up period was lower for patients having Charlson comorbidity than for those who did not (colorectal 58.8% versus 79.3%; breast 71.8% versus 85.4%; prostate 75.8% versus 82.8%; melanoma 58.8% versus 79.3%). Table 3 presents the results of proportional hazards regression analysis. Hazard rates were controlled for other potential predictors of mortality (age, sex, race-ethnicity, marital status, stage at diagnosis, smoking status, socioeconomic status, urban versus rural residence). The presence of comorbidity was associated with higher mortality rates for each cancer type examined. Adjusting for stage at diagnosis, however, did not alter this finding, suggesting that the higher mortality observed among patients with comorbid illness is primarily the result of their underlying illness, rather than later stage at diagnosis.
We found that the presence of comorbidity among patients with cancer was associated with later stage at diagnosis and greater overall mortality. The magnitude of the effects of comorbidity varied considerably by cancer type, however. Among patients with melanoma, those having comorbidity had 62% greater odds of late stage diagnosis and more than double the mortality rate of patients lacking comorbidity. The effects of comorbidity were least among patients with colorectal cancer, for which comorbidity was associated with 17% greater odds of late stage diagnosis and a 27% higher mortality rate.
One possible mechanism by which comorbidity could affect stage at diagnosis is by influencing cancer screening. Comorbidity would seemingly have two opposing influences on cancer screening. Comorbidity should increase the number of patient encounters with the health care system, increasing opportunities for cancer screening recommendations.  On the other hand, comorbidity may decrease the likelihood of cancer screening discussions during encounters by serving as a competing demand for primary care physicians, and by decreasing the perceived importance of cancer screening. [18,19] If comorbidity does have separate and opposing influences on cancer screening, our results suggest than in balance it reduces the likelihood of early detection. Further clinical studies examining screening behavior among patients with comorbid illness would be helpful in understanding these issues.
Many would argue that because of competing causes of mortality, cancer screening has less value in patients of advanced age and comorbidity. [14,56] If so, one might expect that comorbidity would have greater effects on stage at diagnosis among patients of advanced age. We found no evidence, however, that the effects of comorbidity on stage at diagnosis increased with advancing age.
Comorbidity showed a dose-response effect only for breast and prostate cancers. Charlson comorbidity index scores are based on the number of comorbid conditions and their assigned weights. Weights were chosen, however, to predict outcomes related to inpatient care, such as in-hospital mortality, length of stay, and health care resource use. It is possible that a different assignment of weights would be required to predict early cancer detection activities that largely occur outside of the hospital.
This study has a number of limitations that should be considered. First, we relied on administrative data only, the accuracy of which could not be independently verified. The inherent limitations in assessing comorbidity using administrative data sources have been well described. [57,58] We also did not have information on the severity of individual comorbidities, which may have affected our results. [59-61] We used the Charlson method to assess comorbidity, and it is possible that another method of assessment would have yielded different results. Most current measures of comorbidity have proven to be highly correlated and have shown similar effects, however. [62,63] We also did not have information on the cause of death to allow differentiation of cancer mortality from all-cause mortality. Information about whether patients had cancer screening tests was also not available and would have helped in understanding our findings. Finally, our study was restricted to incident cases in Florida, and our findings mi ght not be generalizable to other parts of the country.
In conclusion, comorbidity was associated with later stage at diagnosis and higher overall mortality rates among patients with colorectal cancer, female breast cancer, prostate cancer, and melanoma. Further research is needed to confirm these findings and if confirmed, to determine the mechanisms by which comorbidity might influence stage at diagnosis. A better understanding of the degree to which the benefits of cancer screening are attenuated with comorbidity and how comorbidity influences physicians' and patients' decisions to pursue cancer screening is also needed.
From the Department of Family Medicine, University of South Florida, and the H. Lee Moffitt Cancer Center and Research Institute, Tampa, Fla.
Supported by a Robert Wood Johnson Foundation Generalist Physician Faculty Scholars Award (Dr. Roetzheim).
(1.) Guralnik J, LaCroix AZ, Everett DF, et al: Aging in the Eighties: The Prevalence of Camorbidity and Its Association With Disability. Hyattsville, Md, National Center for Health Statistics, 1989
(2.) Bennett C: Patterns of care related to age of men with prostate cancer. Cancer 1991; 67:2633-2641
(3.) Greenberg E, Chute CG, Stukel T, et al: Social and economic factors in the choice of lung cancer treatment: a population based study in two stages. N Engl J Med 1988; 313:612-617
(4.) Newcomb P, Carbone P: Cancer treatment and age: patient perspectives. J Natl Cancer Inst 1993; 85:1580-1584
(5.) Lazovich D, White E, Thomas DB, et al: Underutilization of breast-conserving surgery and radiation therapy among women with stage I or II breast cancer. JAMA 1991; 266:3433-3438
(6.) Greenfield S, Blanco, DM, Elashoff RM, et al: Patterns of care related to age of breast cancer patients. JAMA 1987; 257:2766-2770
(7.) Silliman R, Guadagnoli E, Weitberg AB, et al: Age as a predictor of diagnostic and initial treatment intensity in newly diagnosed breast cancer patients. J Cerontol 1989; 44:46-50
(8.) Yancik R, Ries LG, Yates JW: Cancer and comorbidity in older patients: a descriptive profile. Ann Epidemiol 1996;6:399-412
(9.) Desch CE, Penberthy L, Newschaffer CJ, et al: Factors that determine the treatment for local and regional prostate cancer. Med Care 1996; 34:152-162
(10.) Newschaffer CJ, Penberthy L, Desch CE, et al: The effect of age and comorbidity in the treatment of elderly women with nonmetastatic breast cancer. Arch Intern Med 1996; 156:85-90
(11.) Newschaffer CJ, Bush TL, Penberthy LT: Comorbidity measurement in elderly female breast cancer patients with administrative and medical records data. J Clin Epidemiol 1997; 50:725-733
(12.) Payne JE, Meyer HJ: The influence of other diseases upon the outcome of colorectal cancer patients. Aust N Z J Surg 1995; 65:398-402
(13.) Satariano WA: Aging, comorbidity, and breast cancer survival: an epidemiologic view. Adv Exp Med Biol 1993; 330:1-11
(14.) Satariano WA, Ragland DR: The effect of comorbidity on 3-year survival of women with primary breast cancer. Ann Intern Med 1994; 120:104-110
(15.) West DW, Satariano WA, Ragland DR, et al: Comorbidity and breast cancer survival: a comparison between black and white women. Ann Epidemiol 1996;6:413-419
(16.) Crawford J, Cohen H: Aging and neoplasia. Ann Rev Gerontol Geriatr 1984; 4:3-32
(17.) Steel K Caring for the elderly person: focusing on cancer. Perspectives on Prevention and Treatment of Cancer in the Elderly. Yancik R, Carbone PP, Patterson WB, et al (eds). New York, Raven Press, 1983, pp 51-61
(18.) Jaen CR, Stange KC, Nutting PA: Competing demands of primary care: a model for the delivery of clinical preventive services. J Fam Pract 1994; 38:166-171
(19.) Cooper GS, Fortinsky RH, Hapke R, et al: Primary care physician recommendations for colorectal cancer screening. patient and practitioner factors. Arch Intern Med 1997; 157:1946-1950
(20.) Shoen R, Marcus M, Braham R: Factors associated with the use of screening mammography in a primary care setting. J Community Health 1994; 19:239-252
(21.) Burack R, Liang J: The acceptance and completion of mammography by older black women. Am J Public Health 1989; 79:721-726
(22.) Kiefe CI, Funkhouser E, Fouad MN, et al: Chronic disease as a barrier to breast and cervical cancer screening. J Gen Intern Med 1998; 13:357-365
(23.) Stange K, Flocke S. Goodwin M: Opportunistic preventive services delivery. J Fam Pract 1998; 46:419-424
(24.) Grady K, Lemkau JP, McVay JM, et al: The importance of physician encouragement in breast cancer screening of older women. Prev Med 1992; 21:766-780
(25.) Bostick R, Sprafka JM, Virnig BA, et al: Predictors of cancer prevention attitudes and participation in cancer screening examinations. Prev Med 1994; 23:816-826
(26.) Chao A, Paganini-Hill A, Ross RK, et al: Use of preventive care by the elderly. Prev Med 1987; 16:710-722
(27.) Satariano WA: Comorbidity and functional status in older women with breast cancer: implications for screening, treatment and prognosis. J Gerontol 1992; 47:24-31
(28.) Catalano RA, Santariano WA: Unemployment and the likelihood of detecting early-stage breast cancer. Am J Public Health 1998; 88:586-590
(29.) Mettlin CJ, Murphy GP, McGinnis LS, et al: The National Cancer Database report on prostate cancer. American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer l995; 76:1104-1112
(30.) Mettlin C, Murphy GP, Babaian RJ, et al: The results of a five-year early prostate cancer detection intervention. Investigators of the American Cancer Society National Prostate Cancer Detection Project. Cancer l996; 77:150-159
(31.) Morrison A: Review of evidence on the early detection and treatment of breast cancer. Cancer 1989; 64(suppl):265l-2656
(32.) Shapiro 5, Venet W, Strax P, et al: Periodic Screening for Breast Cancer. The Health Insurance Plan Project and Its Sequelae, 1963-1986. Baltimore, Johns Hopkins University Press, 1988
(33.) Newcomb P, Norlfleet RG, Storer BE, et al: Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst 1992; 84:1572-1575
(34.) Selby J, Friedman GD, Quesenberry CP, et al: A case-controlled study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 1992; 326:653-657
(35.) Roush C, Berwick M, Koh HK, et al: Screening for melanoma. Cutaneous Melanoma. Balch C (ed). Philadelphia, JB Lippincott Co, 1992
(36.) Berwick M, Begg CB, Fine JA, et al: Screening for cutaneous melanoma by skin self-examination. J Natl Cancer Inst 1996; 88:17-23
(37.) Roetzheim R, Pal N, Tennant C, et al: The effects of health insurance and race-ethnicity on the early detection of cancer. J Natl Cancer Inst 1999; 91:1409-1415
(38.) Ayanian J, Kohler BA, Abe T, et al: The relation between health insurance coverage and clinical outcomes among women with breast cancer. N Engl J Med 1993; 329:326-331
(39.) Diez-Roux A: Bringing context back into epidemiology: variables and fallacies in multilevel analysis. Am J Public Health 1998; 88:216-222
(40.) Hofer T, Wolfe R, Tedeschi P, et al: Use of community versus individual socioeconomic data predicting variation in hospital use. Health Serv Res 1998; 33:243-259
(41.) Krieger N: Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Am J Public Health 1992; 87:703-710
(42.) Krieger N, Fee E: Social class: the missing link in US health data. Int J Health Serv 1994; 24:25-44
(43.) Shambaugh E, Weiss M: Summary Staging Guide: Cancer Surveillance Epidemiology and End Results Reporting. Bethesda, Md: US Department of Health and Human Services, Public Health Service, National Institutes of Health, Publication No. 86-2313, 1977
(44.) Deyo R, Cherkin D, Ciol M: Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992; 45:613-619
(45.) Charlson M, Pompei P, Ales KL, et al: A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis 1987; 40:373-383
(46.) Charlson M, Szatrowski TP, Peterson J, et al: Validation of a combined comorbidity index. J Clin Epidemiol 1994; 47:1245-1251
(47.) Hunter C, Redmond Ck, Chen V, et al: Breast cancer: factors associated with stage at diagnosis in black and white women, J Natl Cancer Inst 1993; 85:1129-1137
(48.) Mandelblatt J, Andrews H, Kerner J, et al: Determinants of late stage diagnosis of breast and cervical cancer: the impact of age, race, social class, and hospital type. Am J Public Health 1991; 81:646-649
(49.) Nayeri K, Pitaro C, Feldman JG: Marital status and stage at diagnosis in cancer. N Y State J Med 1992; 92:8-11
(50.) Hosmer D, Lemeshow S: Applied Logistic Regression. New York, John Wiley & Son, 1989
(51.) American Cancer Society: Survey of physicians' attitudes and practices in early cancer detection. Cancer 1990; 40:77-101
(52.) Czaja R, McFall SL, Warnecke RB, et al: Preferences of community physicians for cancer screening guidelines. Ann intern Med 1994; 120:602-608
(53.) US Preventive Services Task Force: Guide to Clinical Preventive Services. Washington, DC, US Department of Health and Human Services, 2nd Ed, 1996
(54.) Kalbfleisch J, Prentice R: The Statistical Analysis of Failure Time Data. New York, John Wiley & Sons, 1980
(55.) SAS/STAT User's Guide, Version 6. Cary, NC, SAS Institute, 4th Ed, 1990
(56.) Mandelblatt JS, Wheat ME, Monane M: Breast cancer screening for elderly women with and without comorbid conditions, a decision analysis model. Ann Intern Med 1992; 116:722-730
(57.) lezzoni LI, Foley SM, Daley J, et al: Comorbidities, complications, and coding bias, does the number of diagnosis codes matter in predicting in-hospital mortality? JAMA 1992; 267:2197-2203
(58.) lezzoni LI: Assessing quality using administrative data. Ann Intern Med 1997; 127:666-674
(59.) Hughes JS, lezzoni LI, Daley J, et al: How severity measures rate hospitalized patients. J Gen Intern Med 1996; 11:303-311
(60.) lezzoni LI, Ash AS, Shwartz M, et al: Using severity measures to predict the likelihood of death for pneumonia inpatients. J Gen Intern Med 1996; 11:23-31
(61.) lezzoni LI, Ash AS, Shwartz M, et al: Predicting who dies depends on how severity is measured: implications for evaluating patient outcomes. Ann Intern Med 1995; 123:763-770
(62.) Rochon P, Katz JN, Morrow LA, et al: Comorbid illness is associated with survival and length of hospital stay in patients with chronic disability: a prospective comparison of three comorbidity indices. Med Care 1996; 34:1093-1101
(63.) Krousel-Wood MA, Abdoh A, Re R: Comparing comorbidillness indices assessing outcome variation: the case of prostatectomy. J Gen Intern Med 1996; 11:32-38
TABLE 1 Characteristics of Men and Women With Selected Cancers Diagnosed in Florida, 1994 (N = 32,074) Colorectal (n = 8,933) Characteristics [*] Median (SD) Age in years 71.5 (11.6) Household income $28,929 (10,593) No. (%) Sex Male 4,555 (51.0) Female 4,375 (49.0) Race White, non-Hispanic 7,626 (85.4) Black, non-Hispanic 534 (6.0) Hispanic 701 (7.8) Other 72 (0.6) Education High school graduate or less 4,162 (46.9) More than high school education 4,715 (53.1) Marital status Currently married 5,399 (60.4) Not married 3,534 (39.6) Insurance payer Medicare 5,736 (70.9) Medicaid 119 (1.4) Commercial insurance 709 (8.8) Commercial HMO 662 (8.2) Commerical PPO 484 (6.0) Uninsured 234 (2.9) Other 146 (1.8) Stage In situ 612 (6.9) Local 2,858 (32.0) Regional 3,977 (44.5) Distant 1,486 (16.6) Comorbidity Index 0 6,298 (70.5) 1 1,929 (21.6) 2+ 706 (7.9) Melanoma (n = 1,884) Characteristics [*] Median (SD) Age in years 62.1 (16.7) Household income $31,550 (12,193) No. (%) Sex Male 1,117 (59.3) Female 767 (40.7) Race White, non-Hispanic 1,763 (93.6) Black, non-Hispanic 15 (0.8) Hispanic 62 (3.3) Other 44 (2.4) Education High school graduate or less 650 (34.7) More than high school education 1,225 (65.3) Marital status Currently married 1,268 (67.3) Not married 616 (32.7) Insurance payer Medicare 713 (47.9) Medicaid 28 (1.9) Commercial insurance 274 (17.9) Commercial HMO 180 (11.8) Commerical PPO 186 (12.1) Uninsured 72 (4.7) Other 58 (3.8) Stage In situ 295 (15.7) Local 1,346 (71.4) Regional 129 (6.8) Distant 114 (6.1) Comorbidity Index 0 1,754 (93.1) 1 106 (5.6) 2+ 24 (1.3) Breast (n = 10,976) Characteristics [*] Median (SD) Age in years 64.0 (13.8) Household income $29,794 (10,913) No. (%) Sex Male - Female 10,976 (100) Race White, non-Hispanic 9,217 (84.0) Black, non-Hispanic 768 (7.0) Hispanic 830 (7.6) Other 161 (1.5) Education High school graduate or less 4,275 (43.5) More than high school education 5,557 (56.5) Marital status Currently married 6,188 (56.4) Not married 4,788 (43.6) Insurance payer Medicare 4,912 (49.7) Medicaid 249 (2.5) Commercial insurance 1,653 (16.7) Commercial HMO 1,081 (10.9) Commerical PPO 1,161 (11.7) Uninsured 472 (4.8) Other 360 (3.7) Stage In situ 1,391 (12.7) Local 6,372 (58.1) Regional 2,632 (24.0) Distant 581 (5.3) Comorbidity Index 0 9,601 (87.5) 1 1,120 (10.2) 2+ 255 (2.3) Prostate (n = 10,281) Characteristics [*] Median (SD) Age in years 69.8 (8.1) Household income $29,563 (11,201) No. (%) Sex Male 10,281 (100) Female - Race White, non-Hispanic 8,187 (79.6) Black, non-Hispanic 938 (9.1) Hispanic 1,066 (10.4) Other 90 (0.9) Education High school graduate or less 4,697 (43.1) More than high school education 6,217 (57.0) Marital status Currently married 8,068 (78.5) Not married 2,213 (21.5) Insurance payer Medicare 5,737 (65.7) Medicaid 96 (1.1) Commercial insurance 957 (11) Commercial HMO 864 (9.9) Commerical PPO 596 (6.8) Uninsured 245 (2.8) Other 233 (2.7) Stage In situ 60 (0.6) Local 8,254 (80.3) Regional 1,373 (13.4) Distant 594 (5.8) Comorbidity Index 0 8,876 (86.3) 1 1,099 (10.7) 2+ 306 (3.0) (*)Numbers for individual categories may not sum to total sample size because of missing data. TABLE 2 Bivariate Relationship of Comorbidity With Stage at Diagnosis No. (%) of Cancers Diagnosed at Late Stage Colorectal Melanoma (n = 8,933) (n = 1,884) Comorbid Index 0 3,748/6,298 (59.5%) [**] 214/1,754 (12.2%) [+] 1 1,264/1,929 (65.5%) 24/106 (22.6%) 2+ 451/706 (63.9%) 5/24 (20.8 %) Comorbidity None 3,748/6,298 (59.5%) [**] 214/1,754 (12.2%) [**] Any 1,715/2,635 (65.1%) 29/130 (22.3%) Breast Prostate (n = 10,976) (n = 10,281) Comorbid Index 0 2,762/9,601 (28.8%) [**] 1,626/8,876 (18.3%) [**] 1 356/1,120 (31.8%) 248/1,099 (22.6%) 2+ 95/255 (37.3%) 93/306 (30.4%) Comorbidity None 2,762/9,601 (28.8%) [+] 1,626/8,876 (18.3%) [**] Any 451/1,375 (32.8%) 341/1,405 (24.3%) (*)p [less than].05 for chi-square. (+)P [less than].01 for chi-square. (**)P [less than].001 for chi-square. TABLE 3 Effects of Comorbidity on the Odds of Late Stage Diagnosis [++] Colorectal Melanoma (n = 8,035) (n = 1,524) Comorbid conditions No 1.00 1.00 Yes 1.17 (1.06-1.29) [+] 1.62 (1.01-2.60) [*] Comorbidity Index 0 1.00 1.00 1 1.19 (1.06-1.32) [+] 1.73 (1.04-2.88) [*] 2+ 1.12 (0.95-1.33) 1.20 (0.40-3.62) Breast Prostate (n = 9,832) (n = 8,659) Comorbid conditions No 1.00 1.00 Yes 1.24 (1.09-1.41) [**] 1.30 (1.14-1.50) [**] Comorbidity Index 0 1.00 1.00 1 1.18 (1.02-1.35) [*] 1.19 (1.02-1.40) [*] 2+ 1.56 (1.20-2.03) [**] 1.75 (1.35-2.26) [**] (*)P [less than] .05 (+)P [less than] .01 (**)P [less than] .001 (++)Odds ratios for late stage diagnosis (regional/distant stage) are adjusted for age, sex (if appropriate), race-ethnicity, marital status, educational level, income level, urban residence, and insurance payer.
* Comorbidity was associated with later stage diagnosis for breast cancer, colorectal cancer, prostate cancer, and melanoma.
* Comorbidity was associated with higher mortality for breast cancer, colorectal cancer, prostate cancer, and melanoma.
* Higher mortality appeared to be the result of underlying illness rather than later stage diagnosis.
|Printer friendly Cite/link Email Feedback|
|Author:||ROETZHEIM, RICHARD G.|
|Publication:||Southern Medical Journal|
|Date:||Sep 1, 2001|
|Previous Article:||Prenatal Screening for Tay-Sachs Disease by Louisiana Obstetricians: A Survey Study.|
|Next Article:||Follow-up Survey of Children and Adolescents With Chest Pain.|