Colorectal Cancer: Differences B e t w e e n C o m m u n i t y and Geographically Distant Patients Seen at an Urban Medical Center Alfred I. Neugut, M.D., Ph.D., Dennis Timony, B.S., Todd Murray, B.A. From the Department of Medicine and Comprehensive Cancer Center, College of Physicians and Surgeons, Columbia University, New York, New York Neugut AI, Timony D, Murray T. Colorectal cancer: differences between community and geographically distant patients seen at an urban medical center. Dis Colon Rectum 1991;34:64-68.
ost studies of survival and clinical characteristics in the o n c o l o g y literature describe experiences with patients seen in hospital or medical center settings. Usually the results observed with these patients are either not c o n t r o l l e d or compared with statistics from p o p u l a t i o n - b a s e d t u m o r registries. The major p r o b l e m with this approach is that it is clear that multiple selection factors go into the utilization of a specific hospital by a specific patient. Thus, clinical series derived from hospitals may be highly biased, and one must wonder h o w generalizable their findings may be. Despite the limitations, these are the easiest kinds of studies to do, and we must be prepared to try to derive as m u c h information as possible from them. Several studies in the past have attempted to analyze patient populations seen at hospitals, focusing on differences b e t w e e n the local c o m m u nity p o p u l a t i o n and patients referred to the hospital from more distant c o m m u n i t i e s for s e c o n d a r y or tertiary care. Referral patients may be sicker or have more advanced disease than their local c o m m u n i t y counterparts, and thus, if a large fraction of the hospital patient p o p u l a t i o n is referral, the results seen may be relatively poor. Another possibility is that referral patients c o m e from a m o r e affluent and u p p e r class b a c k g r o u n d and thus may have better clinical o u t c o m e s than patients from the local c o m m u n i t y . A study p e r f o r m e d at the Mayo Clinic I c o m p a r e d referral patients and local residents of O l m s t e d County with endometrial carcinoma. The authors reviewed the medical records for over 1200 w o m e n seen at the Mayo Clinic b e t w e e n 1953 and 1972 and f o u n d that referral patients were m o r e likely to present with advanced disease than local comm u n i t y patients. W h e n five-year survival rates were
M
Many studies in clinical oncology rely on hospital-derived patients. Hospitals vary in the proportions of patients from the local catchment area vs. those from more distant places, of whom a larger proportion are presumably referrals. To study the differences between these two types of patients, we analyzed 1,245 colorectal cancer patients seen at a large urban medical center over a seven-year period. Three hundred ninety-eight patients were from the local community (32 percent), 489 were from the extended community (39.3 percent), and 358 from more distant communities (28.8 percent). The patients from the local community tended to be older and from minority ethnic groups. In addition, the local community patients were more likely to have advanced disease at the time of presentation. The grade of the tumor and its site distribution within the large bowel were similar for the three groups. After adjusting for age, sex, race, and stage of disease, the survival was somewhat better for the distant community patients as compared with the local and extended communities (P < 0.02). Overall, in our patient population, the distant patients tended to have more favorable socioeconomic factors and less advanced disease, and these differences may account, in large part, for a better prognosis for these patients. Particularly in large cooperative trials, studies may need to take into account the respective proportions of local community and geographically distant patients in analyzing and generalizing treatment outcomes. [Key words: Colorectal cancer; Hospital patients; Patient referral; Survival]
Dr. Neugut is the recipient of a Preventive Oncology Academic Award (K07-CA01211)from the N.I.H. and is an Andrew Mellon Fellow in Epidemiology and Medicine. Partial support was provided by a grant from the Aaron Diamond Foundation. Address reprint requests to Dr. Neugut: Division of Oncology, VC 12-225, Columbia University College of Physicians and Surgeons, 630 West 168th Street, New York, New York 10032. 0029/-7828/91/4306-0319/$3.00 Diseases of the Colon & Rectum Copyright 9 1991 by The American Society of Colon & Rectal Surgeons 64
Vol. 34, No. 1
COLORECTAL CANCER
compared, adjusting for age and stage, the local residents did slightly better than referral patients. A similar studye was done for prostate cancer cases seen at the Roswell Park Memorial Institute in Buffalo, New York, comparing patients diagnosed at RPMI to patients diagnosed elsewhere. Again, referral patients were more likely to present at a more advanced stage than patients diagnosed at RPMI. The age distribution of the two groups was similar. We recently compared local community and geographically distant patients seen at Columbia-Presbyterian Medical Center with breast cancer. 3 This study of 2750 newly incident cases seen between 1980 and 1986 showed that the distant patients tended to have less advanced disease than the local community patients. The local patients were more likely to be of a lower socioeconomic class or of a minority ethnic group, but this did not account for the differences in disease stage. When five-year survival rates were compared, adjusting for race, age, and stage of disease, there were no significant differences. In this current report, we use the ColumbiaPresbyterian Medical Center experience to compare local and nonlocal patients presenting with colorectal cancer.
65
local community hospital for a large, urban, relatively indigent population, while also serving as a referral center for patients from t h e rest of New York City, Westchester County, Long Island, and New Jersey. The Columbia Tumor Registry registers approximately 1,800 newly incident cancer cases each year. It has been in operation since late 1979. Colon and rectal cancer together comprise approximately 10 percent of new cancer cases seen at the Medical Center. From January 1980 through D ecem ber 1986, 1,308 new colorectal cancer cases were registered. These cases were subdivided into three groups, based on the patient's zip code of residence. Group 1, local patients, comprised patients from the six zip code areas near CPMC in northern Manhattan. Group 2 was composed of patients from other parts of New York City and Long Island. The third group consisted of patients residing in Westchester, New Jersey, or points more distant. Frequency distributions, survival analyses, and other statistics were calculated for each patient group. Survival analyses used Cox proportional-hazards model analysis, 4 with appropriate adjustments as indicated for age, sex, race, and stage. For 63 of the 1308 cases, the zip code was not available and they were excluded from these analyses.
METHODS The Columbia University Comprehensive Cancer Center, located at the Columbia-Presbyterian Medical Center (CPMC), is one of 22 such centers in the United States. It is in the Washington Heights section of upper Manhattan in New York City. The local community contains approximately equal proportions of blacks, whites, and Hispanics. CPMC is the only hospital serving this area of Manhattan since the closing of two small community hospitals in the early 1980s. Thus, it is the
RESULTS There were 398 patients in the local community (32.0 percent). Four hundred and eighty-nine patients with colorectal cancer were in the extended community (39.3 percent) and 358 were in the distant communities (28.8 percent). Sixty-two percent of the local community patients were over age 70 at the time of diagnosis whereas 51,9 percent were over age 70 in the extended community and 39.5 percent were over age 70 in the more distant
Table 1 Age Distribution for CPMC Colorectal Cancer Patients by Community Status Age in Years (%) Local community Extended community
0-29
30-39
40-49
50-59
60-69
70-79
80-89
>90
3 (0.8) 1
6 (1.5) 7 (1.4) 5 (1.4)
16 (4.0) 27 (5.5) 18 (5.0)
43 (10.8) 76 (15.5) 76 (21.2)
83 (20.9) 124 (25.4) 115 (32.1)
147 (36.9) 162 (33.1) 103 (28.8)
88 (22.1) 85 (17.4) 36 (10.1)
12 (3.0) 7 (1.4) 2 (0.6)
(o.2) Distant community
3
(0.8)
Total 398 (100.0) 489 (100.0) 358 (100.0)
66
NEUGUT E T A L
communities (Table 1). In addition, patients from the distant group were more likely to be male (62 percent) than those from the local (45.2 percent) or extended (44.6 percent) communities (Table 2). The patients from the distant group were also more likely to be white (91 percent) than patients who presented themselves from the local community (50.5 percent) or extended community (74.0 percent) (Table 3). This is consistent with census figures which show that 58.5 percent of the local community is black or Hispanic, whereas NewYork City as a whole is 39 percent black or Hispanic. 5 Table 2
Overall, local community patients presented with more advanced disease than did patients from the other two groups (Table 4). Overall, 26.9 percent of the local community group had advanced disease at diagnosis compared with 18.0 percent and 19.3 percent in the extended and distant communities. The three groups had similar distributions for the grade of the tumor (Table 5), and the site distribution within the large bowel (Table 6). Survival curves were calculated for the three groups, adjusting for age, sex, race, and stage of disease (Fig. 1). Survival was better for distant patients as compared with the local and extended communities (P < 0.02).
Sex Distribution for CPMC Colorectal Cancer Patients by Community Status Male (%) Female (%) Total (%) Local community Extended community Distant community Total
180 (45.2) 218 (44.6) 222 (62.0) 620 (49.8)
218 (54.8) 271 (55.4) 136 (38.0) 625 (50.2)
398 (100.0) 489 (100.0) 358 (100.0) 1245 (100.0)
Dis Colon Rectum, January 1991
DISCUSSION Observations made in hospitaLbased studies are often limited in generalizability because of different patient mixes between hospitals. In large part, this may reflect a variation in local vs. distant referral patient makeups. Thus, it would be useful to learn in what way referral patients may differ from patients derived from local catchment areas.
Table 3
Race Distribution for CPMC Colorectal Cancer Patients by Community Status Local Extended Distant Community (%) Community (%) Community (%) White Black Hispanic Other Total
201 (50.5) 131 (32.9) 61 (15.3) 5 (1.3) 398 (100.0)
361 (74.0) 82 (16.8) 31 (6.4) 14 (2.8) 488 (100.0)
325 (91.0) 13 (3.6) 10 (2.8) 9 (2.5) 357 (100.0)
Total (%) 887 (71.4) 226 (18.2) 102 (8.2) 28 (2.2) 1243 (100.0)
Table 4
Stage Local Regional Distant Unknown Total
Stage Distribution for CPMC Colorectal Cancer Patients by Community Status Local Extended Distant Community (%) Community (%) Community (%) 144 (36.2) 124 (31.2) 107 (26.9) 23 (5.8) 398 (100.0)
203 (41.5) 178 (36.4) 88 (18.0) 20 (4.1) 489 (100.0)
153 (42.7) 116 (32.4) 69 (19.3) 20 (5.6) 358 (100.0)
Total (%) 500 (40.2) 418 (33.6) 264 (21.7) 63 (5.1) 1245 (100.0)
Vol. 34, No. 1
67
COLORECTAL CANCER
Table 5 Grade Distribution for CPMC Colorectal Cancer Patients by Community Status Grade
Local Community (%)
Extended Community (%)
Distant Community (%)
Total (%)
50 (12.6) 198 (49.9) 69 (17.4) 80 (20.2) 397 (100.0)
65 (13.3) 267 (54.6) 67 (13.7) 90 (18.4) 489 (100.0)
64 (17.9) 180 (50.3) 50 (14.0) 64 (17.9) 358 (100.0)
179 (14.4) 645 (51.8) 186 (15.0) 234 (18.8) 1244 (100.0)
Well differentiated Moderately differentiated Poorly differentiated Unknown Total
Table 6 Site Distribution for CPMC Colorectal Cancer Patients by Community Status Local Extended Distant Primary Site Community Community Community Cecum and ascending colon Hepatic flexure Transverse colon Descending and splenic flexure Sigmoid Rectosigmoid Rectum Anus Colon NOS Total
93 (23.3) 16 (4.0) 28 (7.0) 40 (10.1) 111 (27.9) 26 (6.5) 67 (16.8) 2 (0.6) 15 (3.8) 398 (100.0)
In our analysis of 1,245 colorectal cancer patients s e e n at the C o l u m b i a - P r e s b y t e r i a n Medical Center, we f o u n d that distant c o m m u n i t y patients are m o r e likely to be male, younger, white, a n d to have localized disease. More esoteric clinical characteristics, such as t u m o r grade and site of the lesion w e r e similar a m o n g the three g r o u p s of patients. W h e n survival was adjusted for these differences b e t w e e n the groups, there was no overall survival difference a m o n g t h e m . Malkasian and Annegers 1 f o u n d that referral patients to the Mayo Clinic with e n d o m e t r i a l c a n c e r t e n d e d to have m o r e a d v a n c e d disease than patients from the s u r r o u n d i n g O l m s t e d County p o p ulation. In our o w n earlier study of breast cancer, 3
106 (21.7) 14 (2.9) 29 (5.9) 51 (10.4) 156 (31.9) 40 (8.2) 71 (14.5) 8 (1.6) 14 (2.9) 489 (100.0)
81 (22.6) 5 (1.4) 27 (7.5) 24 (6.7) 129 (36.0) 27 (7.5) 49 (13.7) 4 (1.2) 12 (3.4) 358 (100.0)
Total 280 (22.5) 35 (2.8) 84 (6.7) 115 (9.2) 396 (0.0) 93 (7.5) 187 (15.0) 14 (1.2) 41 (3.3) 1245 (100.0)
as well as in this study of colorectal cancer, we f o u n d that the g e o g r a p h i c a l l y distant patients w e r e m o r e likely to have localized disease. This m a y reflect referral patterns that are peculiar to our m e d i c a l center or m a y be m o r e characteristic of m e d i c a l centers in urban settings. Also, as in our earlier study of breast cancer, the distant patients t e n d e d to have a better prognosis, t h o u g h this was p r i m a r i l y a reflection of differences in d e m o graphic and clinical presentation. Thus, there are clearly differences in prognosis and clinical o u t c o m e b e t w e e n local and g e o g r a p h ically distant patients. T h e s e differences, however, m a y s i m p l y reflect differences in d e m o g r a p h i c and clinical characteristics. T h e specific patterns are
68
NEUGUT 100 ----
90
local community extended
community
....... distant community
80 R
70 . ~
\ \ \
60 "0
N
ETAL
Dis Colon Rectum, January 1991
u n i q u e to each hospital and m e d i c a l center. Studies, particularly w h e n u n d e r t a k e n by c o o p e r a t i v e groups with multiple hospitals, w h i c h utilize varying m i x e s of g e o g r a p h i c a l l y distant and local patients, m u s t b e careffflly c o n t r o l l e d for variables such as age, sex, race, and stage. It m a y also b e w o r t h w h i l e to use local c o m m u n i t y vs. geographically distant referral as a control variable in studies of clinical o u t c o m e . Further studies are n e e d e d to clarify this in o t h e r t u m o r types.
N
50
REFERENCES
E ~,
40
3o 20 10 I I I I 10 20 30 40
I I I I 50 60 70 80
I I 90 100
Months Figure 1. Estimated survival function for three groups of colorectal cancer patients, adjusting for age, sex, stage, and race (see text for details).
1. Malkasian GD, AnnegersJF. Endometrial carcinoma: comparison of Olmsted County and Mayo Clinic referral patients. Mayo Clin Proc 1980;55:614-8. 2. Slack NH, Lane WW, Priore RI, Murphy GP. Prostatic cancer treated at a categorical center, 1980-1983. Urology 1986;27:205-13. 3. Timony D, Neugut AI. Differences between community and geographically distant breast cancer patients treated at an urban medical center. In: Anderson PN, Engstrom P, Mortenson LE, eds. Advances in cancer control, vol. 6. New York: Alan R Liss, 1989,6:373-8. 4. Cox DR. Regression models and life-tables. J Stat Soc (B) 1972;34:187-220. 5. New York State Statistical Yearbook, 1986-1987. 13:19-21.
ost studies of survival and clinical characteristics in the o n c o l o g y literature describe experiences with patients seen in hospital or medical center settings. Usually the results observed with these patients are either not c o n t r o l l e d or compared with statistics from p o p u l a t i o n - b a s e d t u m o r registries. The major p r o b l e m with this approach is that it is clear that multiple selection factors go into the utilization of a specific hospital by a specific patient. Thus, clinical series derived from hospitals may be highly biased, and one must wonder h o w generalizable their findings may be. Despite the limitations, these are the easiest kinds of studies to do, and we must be prepared to try to derive as m u c h information as possible from them. Several studies in the past have attempted to analyze patient populations seen at hospitals, focusing on differences b e t w e e n the local c o m m u nity p o p u l a t i o n and patients referred to the hospital from more distant c o m m u n i t i e s for s e c o n d a r y or tertiary care. Referral patients may be sicker or have more advanced disease than their local c o m m u n i t y counterparts, and thus, if a large fraction of the hospital patient p o p u l a t i o n is referral, the results seen may be relatively poor. Another possibility is that referral patients c o m e from a m o r e affluent and u p p e r class b a c k g r o u n d and thus may have better clinical o u t c o m e s than patients from the local c o m m u n i t y . A study p e r f o r m e d at the Mayo Clinic I c o m p a r e d referral patients and local residents of O l m s t e d County with endometrial carcinoma. The authors reviewed the medical records for over 1200 w o m e n seen at the Mayo Clinic b e t w e e n 1953 and 1972 and f o u n d that referral patients were m o r e likely to present with advanced disease than local comm u n i t y patients. W h e n five-year survival rates were
M
Many studies in clinical oncology rely on hospital-derived patients. Hospitals vary in the proportions of patients from the local catchment area vs. those from more distant places, of whom a larger proportion are presumably referrals. To study the differences between these two types of patients, we analyzed 1,245 colorectal cancer patients seen at a large urban medical center over a seven-year period. Three hundred ninety-eight patients were from the local community (32 percent), 489 were from the extended community (39.3 percent), and 358 from more distant communities (28.8 percent). The patients from the local community tended to be older and from minority ethnic groups. In addition, the local community patients were more likely to have advanced disease at the time of presentation. The grade of the tumor and its site distribution within the large bowel were similar for the three groups. After adjusting for age, sex, race, and stage of disease, the survival was somewhat better for the distant community patients as compared with the local and extended communities (P < 0.02). Overall, in our patient population, the distant patients tended to have more favorable socioeconomic factors and less advanced disease, and these differences may account, in large part, for a better prognosis for these patients. Particularly in large cooperative trials, studies may need to take into account the respective proportions of local community and geographically distant patients in analyzing and generalizing treatment outcomes. [Key words: Colorectal cancer; Hospital patients; Patient referral; Survival]
Dr. Neugut is the recipient of a Preventive Oncology Academic Award (K07-CA01211)from the N.I.H. and is an Andrew Mellon Fellow in Epidemiology and Medicine. Partial support was provided by a grant from the Aaron Diamond Foundation. Address reprint requests to Dr. Neugut: Division of Oncology, VC 12-225, Columbia University College of Physicians and Surgeons, 630 West 168th Street, New York, New York 10032. 0029/-7828/91/4306-0319/$3.00 Diseases of the Colon & Rectum Copyright 9 1991 by The American Society of Colon & Rectal Surgeons 64
Vol. 34, No. 1
COLORECTAL CANCER
compared, adjusting for age and stage, the local residents did slightly better than referral patients. A similar studye was done for prostate cancer cases seen at the Roswell Park Memorial Institute in Buffalo, New York, comparing patients diagnosed at RPMI to patients diagnosed elsewhere. Again, referral patients were more likely to present at a more advanced stage than patients diagnosed at RPMI. The age distribution of the two groups was similar. We recently compared local community and geographically distant patients seen at Columbia-Presbyterian Medical Center with breast cancer. 3 This study of 2750 newly incident cases seen between 1980 and 1986 showed that the distant patients tended to have less advanced disease than the local community patients. The local patients were more likely to be of a lower socioeconomic class or of a minority ethnic group, but this did not account for the differences in disease stage. When five-year survival rates were compared, adjusting for race, age, and stage of disease, there were no significant differences. In this current report, we use the ColumbiaPresbyterian Medical Center experience to compare local and nonlocal patients presenting with colorectal cancer.
65
local community hospital for a large, urban, relatively indigent population, while also serving as a referral center for patients from t h e rest of New York City, Westchester County, Long Island, and New Jersey. The Columbia Tumor Registry registers approximately 1,800 newly incident cancer cases each year. It has been in operation since late 1979. Colon and rectal cancer together comprise approximately 10 percent of new cancer cases seen at the Medical Center. From January 1980 through D ecem ber 1986, 1,308 new colorectal cancer cases were registered. These cases were subdivided into three groups, based on the patient's zip code of residence. Group 1, local patients, comprised patients from the six zip code areas near CPMC in northern Manhattan. Group 2 was composed of patients from other parts of New York City and Long Island. The third group consisted of patients residing in Westchester, New Jersey, or points more distant. Frequency distributions, survival analyses, and other statistics were calculated for each patient group. Survival analyses used Cox proportional-hazards model analysis, 4 with appropriate adjustments as indicated for age, sex, race, and stage. For 63 of the 1308 cases, the zip code was not available and they were excluded from these analyses.
METHODS The Columbia University Comprehensive Cancer Center, located at the Columbia-Presbyterian Medical Center (CPMC), is one of 22 such centers in the United States. It is in the Washington Heights section of upper Manhattan in New York City. The local community contains approximately equal proportions of blacks, whites, and Hispanics. CPMC is the only hospital serving this area of Manhattan since the closing of two small community hospitals in the early 1980s. Thus, it is the
RESULTS There were 398 patients in the local community (32.0 percent). Four hundred and eighty-nine patients with colorectal cancer were in the extended community (39.3 percent) and 358 were in the distant communities (28.8 percent). Sixty-two percent of the local community patients were over age 70 at the time of diagnosis whereas 51,9 percent were over age 70 in the extended community and 39.5 percent were over age 70 in the more distant
Table 1 Age Distribution for CPMC Colorectal Cancer Patients by Community Status Age in Years (%) Local community Extended community
0-29
30-39
40-49
50-59
60-69
70-79
80-89
>90
3 (0.8) 1
6 (1.5) 7 (1.4) 5 (1.4)
16 (4.0) 27 (5.5) 18 (5.0)
43 (10.8) 76 (15.5) 76 (21.2)
83 (20.9) 124 (25.4) 115 (32.1)
147 (36.9) 162 (33.1) 103 (28.8)
88 (22.1) 85 (17.4) 36 (10.1)
12 (3.0) 7 (1.4) 2 (0.6)
(o.2) Distant community
3
(0.8)
Total 398 (100.0) 489 (100.0) 358 (100.0)
66
NEUGUT E T A L
communities (Table 1). In addition, patients from the distant group were more likely to be male (62 percent) than those from the local (45.2 percent) or extended (44.6 percent) communities (Table 2). The patients from the distant group were also more likely to be white (91 percent) than patients who presented themselves from the local community (50.5 percent) or extended community (74.0 percent) (Table 3). This is consistent with census figures which show that 58.5 percent of the local community is black or Hispanic, whereas NewYork City as a whole is 39 percent black or Hispanic. 5 Table 2
Overall, local community patients presented with more advanced disease than did patients from the other two groups (Table 4). Overall, 26.9 percent of the local community group had advanced disease at diagnosis compared with 18.0 percent and 19.3 percent in the extended and distant communities. The three groups had similar distributions for the grade of the tumor (Table 5), and the site distribution within the large bowel (Table 6). Survival curves were calculated for the three groups, adjusting for age, sex, race, and stage of disease (Fig. 1). Survival was better for distant patients as compared with the local and extended communities (P < 0.02).
Sex Distribution for CPMC Colorectal Cancer Patients by Community Status Male (%) Female (%) Total (%) Local community Extended community Distant community Total
180 (45.2) 218 (44.6) 222 (62.0) 620 (49.8)
218 (54.8) 271 (55.4) 136 (38.0) 625 (50.2)
398 (100.0) 489 (100.0) 358 (100.0) 1245 (100.0)
Dis Colon Rectum, January 1991
DISCUSSION Observations made in hospitaLbased studies are often limited in generalizability because of different patient mixes between hospitals. In large part, this may reflect a variation in local vs. distant referral patient makeups. Thus, it would be useful to learn in what way referral patients may differ from patients derived from local catchment areas.
Table 3
Race Distribution for CPMC Colorectal Cancer Patients by Community Status Local Extended Distant Community (%) Community (%) Community (%) White Black Hispanic Other Total
201 (50.5) 131 (32.9) 61 (15.3) 5 (1.3) 398 (100.0)
361 (74.0) 82 (16.8) 31 (6.4) 14 (2.8) 488 (100.0)
325 (91.0) 13 (3.6) 10 (2.8) 9 (2.5) 357 (100.0)
Total (%) 887 (71.4) 226 (18.2) 102 (8.2) 28 (2.2) 1243 (100.0)
Table 4
Stage Local Regional Distant Unknown Total
Stage Distribution for CPMC Colorectal Cancer Patients by Community Status Local Extended Distant Community (%) Community (%) Community (%) 144 (36.2) 124 (31.2) 107 (26.9) 23 (5.8) 398 (100.0)
203 (41.5) 178 (36.4) 88 (18.0) 20 (4.1) 489 (100.0)
153 (42.7) 116 (32.4) 69 (19.3) 20 (5.6) 358 (100.0)
Total (%) 500 (40.2) 418 (33.6) 264 (21.7) 63 (5.1) 1245 (100.0)
Vol. 34, No. 1
67
COLORECTAL CANCER
Table 5 Grade Distribution for CPMC Colorectal Cancer Patients by Community Status Grade
Local Community (%)
Extended Community (%)
Distant Community (%)
Total (%)
50 (12.6) 198 (49.9) 69 (17.4) 80 (20.2) 397 (100.0)
65 (13.3) 267 (54.6) 67 (13.7) 90 (18.4) 489 (100.0)
64 (17.9) 180 (50.3) 50 (14.0) 64 (17.9) 358 (100.0)
179 (14.4) 645 (51.8) 186 (15.0) 234 (18.8) 1244 (100.0)
Well differentiated Moderately differentiated Poorly differentiated Unknown Total
Table 6 Site Distribution for CPMC Colorectal Cancer Patients by Community Status Local Extended Distant Primary Site Community Community Community Cecum and ascending colon Hepatic flexure Transverse colon Descending and splenic flexure Sigmoid Rectosigmoid Rectum Anus Colon NOS Total
93 (23.3) 16 (4.0) 28 (7.0) 40 (10.1) 111 (27.9) 26 (6.5) 67 (16.8) 2 (0.6) 15 (3.8) 398 (100.0)
In our analysis of 1,245 colorectal cancer patients s e e n at the C o l u m b i a - P r e s b y t e r i a n Medical Center, we f o u n d that distant c o m m u n i t y patients are m o r e likely to be male, younger, white, a n d to have localized disease. More esoteric clinical characteristics, such as t u m o r grade and site of the lesion w e r e similar a m o n g the three g r o u p s of patients. W h e n survival was adjusted for these differences b e t w e e n the groups, there was no overall survival difference a m o n g t h e m . Malkasian and Annegers 1 f o u n d that referral patients to the Mayo Clinic with e n d o m e t r i a l c a n c e r t e n d e d to have m o r e a d v a n c e d disease than patients from the s u r r o u n d i n g O l m s t e d County p o p ulation. In our o w n earlier study of breast cancer, 3
106 (21.7) 14 (2.9) 29 (5.9) 51 (10.4) 156 (31.9) 40 (8.2) 71 (14.5) 8 (1.6) 14 (2.9) 489 (100.0)
81 (22.6) 5 (1.4) 27 (7.5) 24 (6.7) 129 (36.0) 27 (7.5) 49 (13.7) 4 (1.2) 12 (3.4) 358 (100.0)
Total 280 (22.5) 35 (2.8) 84 (6.7) 115 (9.2) 396 (0.0) 93 (7.5) 187 (15.0) 14 (1.2) 41 (3.3) 1245 (100.0)
as well as in this study of colorectal cancer, we f o u n d that the g e o g r a p h i c a l l y distant patients w e r e m o r e likely to have localized disease. This m a y reflect referral patterns that are peculiar to our m e d i c a l center or m a y be m o r e characteristic of m e d i c a l centers in urban settings. Also, as in our earlier study of breast cancer, the distant patients t e n d e d to have a better prognosis, t h o u g h this was p r i m a r i l y a reflection of differences in d e m o graphic and clinical presentation. Thus, there are clearly differences in prognosis and clinical o u t c o m e b e t w e e n local and g e o g r a p h ically distant patients. T h e s e differences, however, m a y s i m p l y reflect differences in d e m o g r a p h i c and clinical characteristics. T h e specific patterns are
68
NEUGUT 100 ----
90
local community extended
community
....... distant community
80 R
70 . ~
\ \ \
60 "0
N
ETAL
Dis Colon Rectum, January 1991
u n i q u e to each hospital and m e d i c a l center. Studies, particularly w h e n u n d e r t a k e n by c o o p e r a t i v e groups with multiple hospitals, w h i c h utilize varying m i x e s of g e o g r a p h i c a l l y distant and local patients, m u s t b e careffflly c o n t r o l l e d for variables such as age, sex, race, and stage. It m a y also b e w o r t h w h i l e to use local c o m m u n i t y vs. geographically distant referral as a control variable in studies of clinical o u t c o m e . Further studies are n e e d e d to clarify this in o t h e r t u m o r types.
N
50
REFERENCES
E ~,
40
3o 20 10 I I I I 10 20 30 40
I I I I 50 60 70 80
I I 90 100
Months Figure 1. Estimated survival function for three groups of colorectal cancer patients, adjusting for age, sex, stage, and race (see text for details).
1. Malkasian GD, AnnegersJF. Endometrial carcinoma: comparison of Olmsted County and Mayo Clinic referral patients. Mayo Clin Proc 1980;55:614-8. 2. Slack NH, Lane WW, Priore RI, Murphy GP. Prostatic cancer treated at a categorical center, 1980-1983. Urology 1986;27:205-13. 3. Timony D, Neugut AI. Differences between community and geographically distant breast cancer patients treated at an urban medical center. In: Anderson PN, Engstrom P, Mortenson LE, eds. Advances in cancer control, vol. 6. New York: Alan R Liss, 1989,6:373-8. 4. Cox DR. Regression models and life-tables. J Stat Soc (B) 1972;34:187-220. 5. New York State Statistical Yearbook, 1986-1987. 13:19-21.
