Health disparity: time delay in the treatment of breast cancer in Louisiana.
Ferdaus, Riaz ; Kim, Min Su ; Larson, James S. 等
INTRODUCTION
Since the early 1990s, breast-conservative surgery (BCS) and
radiotherapy have increased significantly. The NIH Consensus Statement
on breast cancer treatment in 1990 recommended BCS as a preferable
alternative to mastectomy, and as the primary treatment of localized
breast cancer for most of the women with early stage breast cancer
(Riley et al., 1999). This change in preferred treatment approach has
increased the importance of radiotherapy as a component of standard
treatment. An NIH fact sheet, entitled "Cancer Advances in Focus:
Breast Cancer," says that breast conservative surgery followed by
local radiotherapy has replaced mastectomy as the preferred surgical
approach for treating women with localized breast cancer. Radiotherapy
given after surgery significantly reduces the risk of local and regional
recurrences, and is a positive influence on overall survival.
(Poortmans, 2007).
In localized invasive breast cancer cases in general, radiotherapy
is given after most cancer-directed surgery. The importance of
radiotherapy after surgery to prevent loco-regional recurrence and
thereby improve survival is demonstrated in many studies (Joslyn &
West, 2000; Poortmans, 2007). Radiotherapy reduces the loco-regional
recurrences by 70% (Abner et al., 1991). For every four loco-regional
recurrences prevented at five years, one life at fifteen years will be
saved (Abner et al., 1991). Delayed initiation of radiotherapy has been
shown to increase the incidence of loco-regional recurrences (Huang et
al., 2003). According to the Huang et al., the 5-year loco-regional
recurrence rate is significantly higher in patients treated with
radiotherapy more than 8 weeks after surgery, compared with those who
started treatment within 8 weeks of surgery. Mikeljevic et al. (2004),
in a retrospective study on 7,800 patients, found that delaying the
initiation of radiotherapy for 20-26 weeks after the surgery is
associated with a decreased survival after breast conservative surgery.
The interval between the surgery and the radiotherapy depend on
various factors. These factors can be categorized into patient, surgeon
and facility characteristics. Race is a significant factor in many
studies. Black women are more likely to experience treatment delays,
particularly very long delays, within any given stage (Lund et al.,
2007). These delays are not explained by other demographic factors
(Caplan et al., 2000; Dennis et al., 1975). Age also affects the
interval between breast conservative surgery and radiotherapy. Women
younger than 50 years are more likely to experience treatment delay than
the women 50 years or older (Dennis et al., 1975, Gwyn et al., 2004).
Other patient factors responsible for treatment delay include marital
status, insurance status and other socioeconomic characteristics.
Substantial research has been done on the role of the facilities in
providing treatment to cancer patients. The hospital factors associated
with the BCS include university or teaching hospitals or both, hospitals
located in urban areas, and large hospitals (Kotwall et al., 1996).
However, little is known about the impact of facility characteristics on
the interval between breast conservative surgery and radiotherapy. The
number of beds in the facility, teaching status, the American College of
Surgeons' Commission on Cancer (ACoS CoC) approval status and
ownership classification--these characteristics are the most important
facility characteristics affecting receipt of guideline concordant
cancer treatment in a timely fashion (Kotwall et al., 1996).
The physician characteristics affecting cancer treatment also have
been analyzed. Physician's age, sex, race, medical school,
graduation year, license issue year and residency hospital are the
characteristics found to be associated with breast cancer treatment
(Cyran et al., 2001; Gilligan et al., 2007). Several surgeon
characteristics are significantly associated with the variations in the
breast cancer treatment received (Gilligan et al., 2007). But, in
examining the effects of the physician characteristics on the interval
between surgery and radiotherapy in localized invasive female breast
cancer patients, the existing literature offers little information.
This research analyzes the factors affecting the interval (number
of days) between the breast conservative surgery and radiotherapy.
Although the maximum safe time interval between breast conservative
surgery and radiotherapy has not yet been established (Mikeljevic et
al., 2004), some previous studies have tried to explain the phenomenon
based on recurrence rate and survival rate. For example, one study
suggested that breast cancer patients with a delay of more than 90 days
have lower survival rates than those without delay (Hershman et al.,
2006). Another study found that long delays in beginning postoperative
radiotherapy are linked to increased risk of local recurrence (Huang et
al., 2003).
Huang et al. (2003) also suggest a negative impact of radiotherapy
delay on local control probability. Their study of 46 breast cancer
patients treated with breast conservative surgery analyzed the effect of
delayed initiation or radiotherapy with respect to chemotherapy status.
Poorer survival was associated with delays in initiating radiotherapy
longer than seven weeks in the patients with no chemotherapy, while no
significant effects on survival were found in patients who also received
chemotherapy when their radiation treatment was initiated more than 24
weeks postoperatively (Ampil et al., 1999). Mikeljevic et al. (2004)
suggest that adjuvant radiotherapy in breast conservative surgery
patients should not be unnecessarily delayed and certainly should not
commence later than 20 weeks following surgery.
METHODS
To analyze the factors affecting the interval (number of days)
between the breast conservative surgery and radiotherapy, this study
used the data on patient characteristics derived from the database of
the Louisiana Tumor Registry (LTR). Data on (census tract level)
socioeconomic status came from the U.S. Bureau of Census 2000 (Census
data). And, data on provider and facility characteristics were obtained
from the Louisiana State Board of Medical Examiners Physician Database,
as well as published American Hospital Association data.
This study population includes women who are both Louisiana
residents, and who are newly diagnosed with localized invasive breast
cancer. From the LTR database, a sample of 567 cases was selected.
Missing data and other considerations relating to incomparable types of
radiation treatment resulted in a total sample of 497 cases.
To determine which factors influence the treatment of localized
invasive breast cancer in female patients, the interval between breast
conservative surgery and radiotherapy is estimated as shown in the
following equation:
Logit (P) = Log [P/ (1-P)] = [[beta].sub.0] +
[[beta].sub.1][x.sub.1](Age at Diagnosis) + [[beta].sub.2][x.sub.2]
(Race) + [[beta].sub.3][x.sub.3] (Insurance Status) +
[[beta].sub.4][x.sub.4] (Marital Status)+ [[beta].sub.5][x.sub.5]
(Education) + [[beta].sub.6][x.sub.6] (Urban-Rural status) +
[[beta].sub.7][x.sub.7] (Practice Age in Louisiana) +
[[beta].sub.8][x.sub.8] (Graduation School) + [[beta].sub.9][x.sub.9]
(Practice Region) + [[beta].sub.10][x.sub.10] (Bed Size) +
[[beta].sub.11][x.sub.11] (Ownership Status) +
[[beta].sub.12][x.sub.12] (Teaching Status) + [[beta].sub.13][x.sub.13]
(ACoS CoC Approval Status) + [e.sub.i]
The dependent variable in this study is the number of days between
the breast conservative surgery and radiotherapy. Although the maximum
safe time interval between breast conservative breast conservative
surgery and radiotherapy has not been established (Mikeljevic et al.,
2004), some studies have examined the relationship of treatment delay
with recurrence and survival rates (Ampil et al., 1999; Hershman et al.,
2006; Huang et al., 2003).
Hershman et al. (2006) suggest that breast cancer patients with a
delay of more than 90 days have lower survival rates than those without
delay. This study adopts the Hershman et al. (2006) findings and
constructs the interval between breast conservative surgery and
radiotherapy into a dichotomous variable, with a 90-day cut-off. Out of
the 497 patients, 341 (68.6%) received radiation within 90 days of
surgery and the remaining 156 (31.4%) had delays over 90 days. Table 1
provides a description of all variables.
The independent variables in the study include patient, surgeon and
facility factors. The patient variables include age at diagnosis, race,
marital status, insurance status, census tract level, socioeconomic
factors (poverty level, education and occupation), and census tract
level urban-rural status. The surgeon variables include the medical
school from which the surgeon graduated, the number of years the surgeon
practiced in Louisiana, and the geographical region in Louisiana where
the surgeon practiced. The facility variables include the size of the
facility, the ownership status, the teaching approval status, and the
Commission on Cancer approval status.
The research shows that age affects the interval between breast
conservative surgery and the radiotherapy. Women younger than 50 years
are more likely than the older women to experience treatment delays
(Dennis et al., 1975; Gwyn et al., 2004). Adaptation and quality of life
after breast cancer diagnosis are more difficult for younger women
compared to the older women (Wenzel et al., 1999). Because of
child-rearing activities and employment outside the home, younger women
experience more difficulties and disruptions from the disease and its
treatment (Ganz et al., 1998). On the other hand, older women have
public insurance coverage and tend not to have the same social
responsibilities as young women. Although the some variation is lost,
many similar studies have categorized the age variable into age groups
where patients under age 50 were grouped together (Dennis, Gardner,
& Lim, 1975; Ganz et al., 2002; Wenzel et al., 1999). In this study,
we group age into two categories: patients 50 years and younger in one
group (younger patients) coded as and patients 50 years and older in
another group (older patients) coded as '0'.
Significant racial differences exist in the treatment for women
with early-stage breast cancer (Joslyn, 2002). Black women with breast
cancer are more likely than their White counterparts to experience
treatment delays, particularly very long delays, within any given stage
(Lund et al., 2007). Racial distribution in Louisiana is limited mainly
to White and Black with a population distribution of 63.9% Whites and
32.5% Blacks (US Census Bureau). Other races contribute to the
population in minor fractions. The Hispanic population is very small in
Louisiana, as only 2.4% are of Hispanic origin. As this study population
consists of Whites and Blacks only, a dichotomous variable was created
where Whites were coded as '0' and Blacks were coded as
The marital status of the patient at the time of initial diagnosis
and/or treatment also is used to estimate the interval between breast
conservative surgery and radiotherapy. For the patients who are
suffering a wide array of illnesses, marriage is often associated with
longer life and better quality of life (Gore, Kwan, Saigal, &
Litwin, 2005; Sorlie, Backlund, & Keller, 1995). Findings from
recent studies on cancer patients are controversial in determining
relationship between marital status and cancer survival. A study on
bladder cancer found that marriage was associated with improved survival
in patients with bladder carcinoma, independent of other factors known
to influence mortality in this population (Gore et al., 2005). Another
study on non-small cell lung cancer patients found no survival
difference based on marital status, even after adjusting for variables
of prognostic significance (Jatoi et al., 2007).
Insurance status is defined as the primary payer/insurance career
at the time of initial diagnosis and/or treatment. The Insurance
variable in this research is grouped into no insurance, public insurance
and private insurance categories. In Louisiana, patients with no
insurance usually receive treatment from the Charity Hospitals in
different cities. Patients who do not have any insurance and are not
eligible for Medicare or Medicaid, can apply for an insurance program
called "Free Care." If approved, this covers all medical
expenses. This group of no insurance is combined with public insurance
to form a new group of public insurance and no insurance.
Education is a geographic-based proxy indicator. According to
Census 2000 Brief on educational attainment (Bauman & Graf, 2003)
80.4% of the US population 25 years and older at least have a high
school diploma, whereas, in Louisiana approximately 75% (74.8%) of the
population 25 years and older at least have a high school diploma. Based
on the percent of population with a high school diploma or higher level
of education, all the census tracts in Louisiana were divided into high
literacy and low literacy areas. The census tracts where 75% or more of
the census tract population who are 25 years and older having at least
high school equivalency were categorized as 'high literacy
area' and coded as '0'. The census tracts with less than
75% of the 25 years or older population having high school diploma or
more were categorized as 'low literacy area' and coded as
'1'. Using the census tract of the patient at diagnosis, these
data were assigned to each patient.
The urban-rural status of a patient is derived from the Census
Bureau database, using the census tract of the patient at diagnosis.
This variable is a geographic-based proxy indicator. While 20% of the US
population lives in rural areas, only 9% of the physicians practice
there (van Dis, 2002). Because rural patients have fewer healthcare
providers per capita and increased transportation barriers, it is common
for them to visit physicians less often than urban patients (Hughes
& Rosenbaum, 1989). The percent of population living in urban areas
was collected for each of the census tracts in Louisiana from the US
Census Bureau 2000 Census database. The tracts where 100% of the
population was dwelling in urban area were considered 'urban'
and coded as '1', and the remaining tracts were categorized
'non-urban,' which includes the mixed tracts as well and were
coded '0'.
The years in practice in Louisiana variable denotes the number of
years the surgeon practiced in Louisiana. The main purpose of this
variable is to examine if the young surgeons adhere to standard
treatment. Those surgeons who practiced six years or more were termed
experienced doctors and were coded '0', and those who
practiced five years or less were termed as young doctors and were coded
as '1'.
The Graduation School variable codes the medical school where the
surgeon graduated. The patients treated by surgeons graduating from
Louisiana State University Medical Schools were coded as '1',
the patients treated by surgeons graduating from Tulane Medical School
were coded as '2', and the patients treated by surgeons
graduating from out-of-state or foreign medical schools were coded as
'3'.
The Practice Region in Louisiana variable was derived from the
business phone number of the surgeon performing the surgery. The
telephone area code was used to identify regions of the state where they
practiced, namely Southeast region (area codes 504 and 985), Central
region (area codes 225 and 337), and North region (area code 318). The
Bed Size variable is defined as the number of beds in the facility for
the most definitive treatment. This variable was coded into a
dichotomous variable where facilities with less than 300 beds were
termed as small and medium-sized facilities and coded as '0'
and the facilities with 300 or more beds were termed as large facilities
and coded as '1'.
The Ownership variable is defined as the ownership status of the
facility for the most definitive treatment. Patients treated in
privately owned facilities were coded as '0', and the patients
treated in government owned facilities were coded as '1'. The
Teaching Status variable is defined as the teaching status of the
facility for the most definitive treatment. Facilities which did not
have approval for MD or DO programs during the year 2004 were coded as
'0', and the facilities approved for MD or DO programs during
the year 2004 were coded as '1'.
The American College of Surgeons Commission on Cancer (ACoS CoC)
Approval Status variable is defined as the ACoS CoC approval status of
the facility where the patient received the most definitive treatment.
These facilities must meet certain criteria and perform certain duties
set forth by the ACoS CoC to win and maintain this approval status. In
general, this status should imply better cancer care, supervised by the
ACoS CoC. Facilities which had ACoS CoC approval during the year 2004
were coded as '0,' and the facilities did not have ACoS CoC
approval during the year 2004 were coded as '1'.
REGRESSION RESULTS
Table 2 provides frequencies for all variables. As shown in Table
2, out of the 497 patients, 341 (68.6%) received radiation within 90
days of surgery and the remaining 156 (31.4%) had delays over 90 days.
The mean age at diagnosis for the study population was 61.04 years,
range from 28 to 94, standard deviation of 12.6 years. Among the study
population (N = 497) 102 (20.5%) patients were younger than 50 years at
the time of diagnosis and 369 (74.2%) were White. A total of 71 or 14.6%
of the population were never married, 277 (57.0%) were married
(including common law), 7 (1.4%) were separated, 37 (7.6%) were
divorced, and 94 (19.3%) were widowed.
Table 3 provides the logistic regression results for the surgery
and radiotherapy interval model shown in equation (1). Out of 497 cases
in the study population, 456 cases were included in the final analyses.
For the remaining 41 cases, one or more fields were missing or unknown.
The overall percentage of patient, surgeon and facility variables
correctly predicted 75.2% of the variation in the dependent variable-
the interval between surgery and radiotherapy. Five (age, race, bed
size, residency approval status and Louisiana regions) out of the
thirteen independent variables were found to be statistically
significant.
Examination for multicollinearity among the independent variables
was necessary to perform the analyses. Tolerances and Variance Inflation
Factors for each variable are suggestive of no statistically significant
multicollinearity. In the age at diagnosis variable, the patients with
age 20 to 49 years at diagnosis were coded as '1' and the
patients with age 50 years or over were coded as '0'. It was
hypothesized that younger patients were more likely to have delays in
treatment. For this variable, the logistic coefficient was 1.391, the
Wald statistic was 27.524 and the age at diagnosis was significant at p
<.05. The odds ratio tells us that as age changes from 50 years or
over (0) to less than 50 years (1) the changes in the odds of getting
treatment after 90 days compared to getting treatment within 90 days is
4.02. The odds of a young patient having delays in treatment compared to
no delay are 4.02 times more than for an old patient. In short, younger
patients are more likely to have delay in treatment.
The race of the patient significantly predicted whether they had a
time delay in treatment of breast cancer (coefficient = 0.654, Wald
statistic = 5.335, P < .05). The estimated odds ratio was 1.924,
which means that for the black women with localized breast cancer, the
odds of experiencing delay in treatment is almost twice as great as that
of their white counterparts. For race variable, this research fails to
reject the null hypothesis that there is no difference of treatment
delay between Black patients and White patients and draws inference that
Black patients are more likely to have delays 90 days or more between
breast conservative surgery and radiotherapy.
The education level of the patient, however, did not significantly
predict whether they had a time delay in treatment, and neither did the
urban-rural variable. Note that although these predictors are not
significant, the coefficients were the same sign. Contrary to
expectation, both marital status and insurance status variables were
shown to decrease delay in treatment. Although these variables do
exhibit the unexpected sign, the p-values of these effects suggest the
variables were not significant determinants.
Turning to the analysis of surgeon characteristics effect, Table 3
shows that the surgeon's practice region in Louisiana significantly
predicted whether patients had a time delay in treatment of breast
cancer. For the Southeast variable, the estimated odds ratio of 2.024
suggests that the odds to have delay in treatment for the patients
treated in southeast region are almost twice as great as that of the
patients treated in north regions. For the Central variable, the
estimated odds ratio of 2.426 means that the odds to have delay in
treatment for the patients treated in central region are almost twice as
great as that of the patients treated in north regions.
The number of years the surgeon practiced in Louisiana did not
significantly predict whether patients had time delays in treatment of
breast cancer. Among the facility characteristics variables, it was
hypothesized that delays in treatment would be different between these
two bed- size groups of hospitals. For this variable, the logistic
coefficient was negative and it was -1.548, the Wald statistic was
15.352, and it was significant at p < 0.05. The estimated odds ratio
of 0.206 suggests that the odds of having delay in treatment of patients
treated in hospitals with less than 300 beds are 1/0.206, almost five
times as great as that of the patients treated in hospitals with 300 or
more beds. In other words, the patients treated in smaller hospitals
have greater tendencies to experience delays in treatment.
For the teaching status variable, the estimated odds ratio was
2.727, which means that patients treated in teaching hospitals were
almost 2.7 times as likely to have delays than patients treated in
non-teaching hospitals. For teaching status variable, this research
fails to reject the null hypothesis that there is no difference between
the teaching and non-teaching hospitals and draws inference that
patients treated at teaching hospitals are more likely to have delays 90
days or more between breast conservative surgery and radiotherapy.
Finally, the ownership status of facility, however, did not
significantly predict whether patients had time delay in treatment of
breast cancer. The approval of the ACosCoc was not significant in this
crucial aspect of cancer treatment.
DISCUSSION AND CONCLUSION
This research attempts to examine health disparity using the
quantitative findings from the logistic regression analyses of the
interval between breast conservative surgery and radiotherapy in
localized female breast cancer patients diagnosed in 2004 in Louisiana.
Through the logistic regression analyses, we identify the specific
variables contributing inequality. As noted earlier, findings have shown
that the inequalities in receiving radiation on time after surgery were
identified in five areas--patient's age at diagnosis, race, size of
the facility where the patient received treatment, teaching status of
the facilities and geographical location where the patient received
treatment. According to results, young women who were diagnosed with
localized breast cancer and underwent breast conservative surgery and
radiotherapy were clearly susceptible to experiencing delays between
breast conservative surgery and radiotherapy. Younger women appear to
experience more difficulties and disruptions from the disease and its
treatments because of child-rearing activities and employment outside
the home (P. A. Ganz et al., 1998).
The study also supports other studies finding statistically
significant racial differences in the treatment of women with
early-stage breast carcinoma (Joslyn, 2002). Black women were more
likely to experience treatment delays, particularly very long delays,
within any given stage (Lund et al., 2007) and these delays were not
explained by other demographic factors (Caplan et al., 2000; Dennis et
al., 1975). Although numerous efforts have been made to ensure racial
equity and social justice, since the Civil Rights movement, the racial
disparity identified in this research conforms again to the findings of
other studies establishing disparity grounded in racial difference in
different areas of health and welfare.
Patients treated in teaching hospitals are more likely than those
treated in non-teaching hospitals to experience delays, which supports
the findings from literature. Although teaching hospitals are
historically well known for state of the art treatment of their
patients, many of the teaching hospitals are financially stressed
compared to the non-teaching hospitals (Ayanian & Weissman, 2002).
With everything in place, patients treated at large hospitals are
experiencing shorter delays than small and medium hospitals. Such
finding may be explained by the fact that the level of care is not
optimal in small and medium hospitals for patients with breast cancer.
After surgery is performed in a small or medium sized hospital, the
patients are finding themselves in a comparatively difficult situation
in terms of receiving radiotherapy. The characteristics of the patients
who are getting treatment in large hospitals may be different from those
who are getting treatment from smaller hospitals. In addition, the
practice region was statistically significant in influencing interval
between breast conservative surgery and radiotherapy. Local
socioeconomic factors have an influence.
This research provides a cross sectional view of the inequalities
present in the area of breast cancer treatment. Women who are younger
than 50 years are extremely vulnerable to treatment delays. The safety
net we created for the poor in society does not seem to work for this
group, and our policies need to be revisited. To ensure fairness to all,
it is necessary to eliminate the barriers to proper care for young
women. Although we have numerous policies in place with a goal to
eliminate racial disparity, race is still playing a significant role in
healthcare delivery. At the providers end, the disparity between large
and small facilities and teaching and non-teaching facilities needs to
be explored further to identify why this is occurring. In institutional
settings, these problems can be solved efficiently with proper policy
recommendation and implementation.
In conclusion, a formal policy evaluation in each area where
significant inequality was identified would enable policymakers and
policy scientists to propose reforms to minimize or eliminate
inequalities in those areas. Geographic variation in treatment interval
also needs to be explored in future research. Although geographic
variation in treatment and treatment outcomes has been identified by
many studies, little policy research has been done on specific areas of
variation to identify why those anomalies are occurring. Further
research is necessary to answer why there is geographic variation in the
intervals between breast conservative surgery, radiotherapy and other
areas of breast cancer patterns of care.
REFERENCES
Abner, A. L., Recht, A., Vicini, F. A., Silver, B., Hayes, D.,
Come, S., et al. (1991). Cosmetic results after surgery, chemotherapy,
and radiation therapy for early breast cancer. International Journal of
Radiation Oncology, Biology, Physics. 21(2):331-8, 1991 Jul., 21(2),
331-338.
Ampil, F. L., Burton, G. V., Li, B. D., & Mills, G. M. (1999).
Radiotherapy with and without chemotherapy after breast conservation
surgery for early stage breast cancer: a review of timing. European
Journal of Gynaecological Oncology, 20(4), 254-257.
Ayanian, J. Z., & Weissman, J. S. (2002). Teaching Hospitals
and Quality of Care: A Review of the Literature. The Milbank Quarterly,
80(3), 569-593.
Caplan, L. S., May, D. S., & Richardson, L. C. (2000). Time to
diagnosis and treatment of breast cancer: results from the National
Breast and Cervical Cancer Early Detection Program, 1991-1995. American
Journal of Public Health, 90(1), 130-134.
Cyran, E. M., Crane, L. A., & Palmer, L. (2001). Physician Sex
and Other Factors Associated With Type of Breast Cancer Surgery in Older
Women. Archives of Surgery, 136(2), 185-191.
Dennis, C. R., Gardner, B., & Lim, B. (1975). Analysis of
survival and recurrence vs. patient and doctor delay in treatment of
breast cancer (Vol. 35, pp. 714-720).
Ganz, P. A., Rowland, J. H., Desmond, K., Meyerowitz, B. E., &
Wyatt, G. E. (1998). Life after breast cancer: understanding
women's health-related quality of life and sexual functioning.
Journal of Clinical Oncology, 16(2), 501-514.
Gilligan, M. A., Neuner, J., Sparapani, R., Laud, P. W., &
Nattinger, A. B. (2007). Surgeon Characteristics and Variations in
Treatment for Early-Stage Breast Cancer. Archives of Surgery, 142(1),
17-22.
Gore, J. L., Kwan, L., Saigal, C. S., & Litwin, M. S. (2005).
Marriage and mortality in bladder carcinoma (Vol. 104, pp.
1188-1194).
Gwyn, K., Bondy, M. L., Cohen, D. S., Lund, M. J., Liff, J. M.,
Flagg, E. W., et al. (2004). Racial differences in diagnosis, treatment,
and clinical delays in a population-based study of patients with newly
diagnosed breast carcinoma. Cancer, 100(8), 15951604.
Hershman, D. L., Wang, X., McBride, R., Jacobson, J. S., Grann, V.
R., & Neugut, A. I. (2006). Delay in initiating adjuvant
radiotherapy following breast conservation surgery and its impact on
survival. International journal of radiation oncology, biology, physics,
65(5), 1353-1360.
Huang, J., Barbera, L., Brouwers, M., Browman, G., & Mackillop,
W. J. (2003). Does Delay in Starting Treatment Affect the Outcomes of
Radiotherapy? A Systematic Review. Journal of Clinical Oncology, 21(3),
555-563.
Hughes, D., & Rosenbaum, S. (1989). An Overview of Maternal and
Infant Health Services in Rural America. Journal of Rural Health, 5,
299-319.
Jatoi, A., Novotny, P., Cassivi, S., Clark, M. M., Midthun, D.,
Patten, C. A., et al. (2007). Does Marital Status Impact Survival and
Quality of Life in Patients with Non-Small Cell Lung Cancer?
Observations from the Mayo Clinic Lung Cancer Cohort. The Oncologist,
12(12), 1456-1463.
Joslyn, S. A., & West, M. M. (2000). Racial Differences in
Breast Carcinoma Survival. Cancer, 88(1), 114123.
Kotwall, C. A. M. D., Covington, D. L. P. H., Rutledge, R. M. D.,
Churchill, M. P. B. A., & Meyer, A. A. M. D. P. D. (1996). Patient,
Hospital, and Surgeon Factors Associated with Breast Conservation
Surgery: A Statewide Analysis in North Carolina. [Article]. Annals of
Surgery, 224(4), 419-429.
Lund, M., Brawley, O., Ward, K., Young, J., Gabram, S., & Eley,
J. (2007). Parity and disparity in first course treatment of invasive
breast cancer. Breast Cancer Research and Treatment.
Mikeljevic, J. S., Haward, R., Johnston, C., Crellin, A., Dodwell,
D., Jones, A., et al. (2004). Trends in postoperative radiotherapy delay
and the effect on survival in breast cancer patients treated with
conservation surgery. British Journal of Cancer, 90(7), 1343-1348.
Poortmans, P. (2007). Evidence based radiation oncology: Breast
cancer. Radiotherapy and Oncology, 84(1), 84-101.
Riley, G. F., Potosky, A. L., Klabunde, C. N., Warren, J. L., &
Ballard-Barbash, R. (1999). Stage at Diagnosis and Treatment Patterns
Among Older Women With Breast Cancer: An HMO and Fee-for-Service
Comparison. JAMA: The Journal of the American Medical Association,
281(8), 720-726.
Sorlie, P. D., Backlund, E., & Keller, J. B. (1995). US
mortality by economic, demographic, and social characteristics: the
National Longitudinal Mortality Study. American Journal of Public
Health, 85(7), 949-956.
van Dis, J. (2002). Where We Live: Health Care in Rural vs Urban
America. JAMA: The Journal of the American Medical Association, 287(1),
108.
Wenzel, L. B., Fairclough, D. L., Brady, M. J., Cella, D., Garrett,
K. M., Kluhsman, B. C., et al. (1999). Agerelated differences in the
quality of life of breast carcinoma patients after treatment. 86(9),
17681774.
RIAZ FERDAUS
Our Lady Of The Lake College
MIN SU KIM
JAMES S. LARSON
Southern University
Table 1: Variables and Data Sources
Variable Description and data source
Dependent
variable
Interval between Dichotomous variable coded 1if the patients
surgery and who received radiotherapy after 90 days of
radiotherapy breast conservative surgery, and 0 otherwise;
Source: Louisiana Tumor Registry
Patient
characteristics
Age Dichotomous variable coded 1if the patient is
50 years old or younger at diagnosis, 0
otherwise; Source: Louisiana Tumor Registry
Race Dichotomous variable coded 1if the patient is
black, 0 otherwise; Source: Louisiana Tumor
Registry
Insurance Dichotomous variable coded 1if the patient has
public or no insurance, 0 otherwise; Source:
Louisiana Tumor Registry
Marital status Dichotomous variable coded 1if the patient is
not married, 0 otherwise; Source: Louisiana
Tumor Registry
Education Dichotomous variable coded 1if the patient
census track is in the area less than 75%
of the 25 years or older population having
high school diploma or more, 0 otherwise;
Source: U.S. Census Bureau
Urban-rural status Dichotomous variable coded 1if the patient
census track is in the area where 100% of the
population was dwelling in urban area, 0
otherwise; Source: U.S. Census Bureau
Surgeon
characteristics
Years of practice Dichotomous variable coded 1if the surgeon
practiced six years or more in Louisiana, 0
otherwise; Source: Louisiana State Board of
Medical Examiners The Medical School where the
surgeon graduated; Source: Louisiana
Graduation school State Board of Medical Examiners
Practice region The regions of the state where the surgeon
practiced; Source: Louisiana State Board
of Medical Examiners
Facility
characteristics
Bed size Dichotomous variable coded 1if the facilities
with 300 or more beds, 0 otherwise; Source:
American Hospital Association
Ownership Dichotomous variable coded 1if the patients
classification treated in government-owned facilities, 0
Teaching status otherwise; Source: American Hospital
Association Dichotomous variable coded 1 if
the facilities did not have approval for MD or
DO programs, 0 otherwise; Source: American
Hospital Association
AcoS CoC Dichotomous variable coded 1 if the facilities
approval status did not have American College of Surgeons
Commission on Cancer (ACoS CoC) Approval,
0 otherwise; Source: American Hospital
Association
Table 2
Frequency Distribution
Frequency
distribution
Variable Co
Dependent variable Within 90
days
Interval between surgery 341 (68.6%) After 90 days
and radiotherapy 156 (31.4%)
Patient characteristics
Age 50 yrs or Less than50
more yrs
395 (79.5%) 102 (20.5%)
Race White Black
369 (74.2%) 128 (20.5%)
Insurance Private Public or no
356 (17.6%) 133 (26.8%)
Maritial status Married Not married
284 (57.1%) 202 (40.6%)
Education High Low
education education
298 (60.0%) 199 (40.0%)
Urban-rural status Rural Urban
221 (44.5%) 276 (55.5%)
Surgeon characteristics
Years of practice 6 yrs or less than 6
more yrs
417 (83.9%) 55 (11.1%)
Graduation school Other LSU Tulane
190 (38.2%) 223 (44.9%) 59 (11.9%)
Practice region North Southeast Central
127 (25.6%) 178 (35.8%) 167 (33.6%)
Facility characteristics
Bed size Small or
Medium Large
269 (54.1%) 228 (45.9%)
Ownership classification Private Public
352 (70.8%) 145 (29.2%)
Teaching status No teaching Teaching
312 (62.8%) 185 (37.2%)
AcoS CoC approval No approval Approval
status 138 (27.8%) 359 (72.2%)
Table 3
Regression estimates for the surgery and radiotherapy interval
R Squared= .752
Coefficie Odd
Variable nt ratio S.E Wald P-Val
Patient
Characteristics
Age at Diagnosis 1.391 4.020 0.265 27.524 0.000 ***
Race 0.654 1.924 0.283 5.335 0.021 *
Marital Status -0.052 0.950 0.253 0.041 0.839
Insurance Status -0.374 0.688 0.302 1.533 0.216
Education 0.017 1.017 0.254 0.004 0.948
Urban-Rural
Status -0.014 0.987 0.258 0.003 0.958
Surgeon
characteristics
Graduation
School
LSU 0.148 1.160 0.274 0.292 0.589
Tulane 0.463 1.589 0.393 1.392 0.238
Practice Regions
Southeast 0.705 2.024 0.339 4.339 0.037 *
Central 0.886 2.426 0.348 6.501 0.011 *
Years of Practice 0.409 1.505 0.367 1.243 0.265
Facility
characteristics
Bed Size 0.000 **
-1.578 0.206 0.403 15.352 *
Teaching Status 1.003 2.727 0.380 6.980 0.008 **
Ownership Status 0.514 1.672 0.291 3.118 0.077
ACoS CoC
Approval 0.166 1.181 0.295 0.317 0.574
Constant 0.000 **
-1.921 0.146 0.410 22.011 *
Note: *** denotes significance at the 0.1% level, ** at the
1% * at the 5%.
* =
