Abstract
Background
Insurance status and SES are associated with the stage of melanoma at diagnosis. However, the influence of Medicaid enrollment on melanoma stage has not been studied in detail. This study examined the effect of Medicaid enrollment status and duration on melanoma stage at diagnosis in a large, multi-ethnic California population.
Methods
California Cancer Registry records were linked with statewide Medicaid enrollment files to identify 4558 men and women diagnosed with invasive cutaneous and metastatic melanoma during 1998–1999. Multivariate logistic regression was used to evaluate the association between prediagnosis Medicaid enrollment status and late-stage diagnosis and tumor depth at diagnosis.
Results
Late-stage disease was diagnosed in 27% of Medicaid and 9% of non-Medicaid melanoma patients. Those enrolled in Medicaid at diagnosis and those enrolled intermittently during the year prior to diagnosis had significantly greater covariate-adjusted odds of late-stage cancer than those not enrolled in Medicaid (OR 13.64, 95% CI=4.43, 41.98, and OR 2.77, 95% CI=1.28, 5.99, respectively). Participants continuously enrolled during the previous year were not at increased odds for late-stage disease. An increased likelihood of late-stage melanoma was also associated with low SES (p<0.05) and non-Hispanic black race/ ethnicity (p<0.10) after covariate adjustment.
Conclusions
Men and women intermittently enrolled in Medicaid or not enrolled until the month of diagnosis had a significantly increased likelihood of late-stage melanoma. Greater education and outreach, particularly in low-SES areas, are needed to improve melanoma awareness and access to screening.
Introduction
Melanoma outcomes differ strongly by tumor thickness at diagnosis1; thus, understanding the disparities of SES and race/ethnicity that affect the stage at diagnosis is of considerable clinical and public health importance. While an increased incidence of melanoma is associated with markers of higher SES, a more-advanced stage at diagnosis is associated with lower SES.2 The risk of late-stage melanoma also varies by race/ethnicity, with African Americans, Asians, and Hispanics at higher risk than non-Hispanic whites,3–5 although melanoma is much less likely to occur in these groups. However, race/ethnicity does not appear to entirely mitigate socioeconomic disparities in late-stage melanoma.2,6–8
A major determinant of melanoma stage at diagnosis in the U.S. is health insurance. A 1994 study of Florida melanoma patients found a four-fold higher risk of late stage at diagnosis for those covered by Medicaid compared to those with commercial indemnity insurance.7 More-recent studies have reported associations between the status and duration of Medicaid enrollment and the late-stage diagnosis of breast, colorectal, pancreatic, and cervical cancers,9–12 implying that people not enrolled in Medicaid or those enrolled for shorter periods prior to cancer diagnosis are at greatest risk for more-advanced cancer. These associations have not yet been examined for melanoma.
To better understand disparities in late-stage melanoma, this study jointly examined the effects of SES, race/ethnicity, and the timing of Medicaid enrollment on two measures of melanoma progression (stage of disease and tumor depth), taking advantage of a linked cancer registry–Medicaid resource available for the large, multi-ethnic population of California.
Methods
Melanoma Case Ascertainment
Population-based melanoma case information was obtained from the California Cancer Registry (CCR), which has conducted statewide cancer surveillance since 1988.13 All cases of incident invasive cutaneous and metastatic melanoma diagnosed between January 1, 1998, and December 31, 1999, among California residents aged <65 were considered.
Data included patient demographic characteristics, tumor characteristics, and treatment for 4 months following diagnosis. Demographic measures were categorized according to (1) race/ ethnicity—non-Hispanic white (white); non-Hispanic black (black); Hispanic; and non-Hispanic Asian (Asian); (2) age at diagnosis (15–29, 30–39, 40–49, 50–59, and 60–64)11; (3) marital status (married or not married); and (4) date of diagnosis (typically from the date of first positive microscopic confirmation).
The CCR does not routinely collect patient education or income information, but prior studies have derived a composite index of neighborhood-level SES from census data geocoded to patient address at diagnosis. Census block group-level information on income, poverty, education, home value, and blue-collar status were used to compute a composite neighborhood-level SES score. This score is divided into quintiles (Quintile 1=residence in lowest-SES census block group; Quintile 5=residence in highest-SES block group). Neighborhood-level SES has been associated previously with cancer, cardiovascular disease, and other health outcomes, although it may represent somewhat different constructs than individual-level SES.11,14–16
Stage at diagnosis was defined according to Surveillance, Epidemiology, and End Results (SEER) summary stage 1977 guidelines as localized (skin only); regional (lymph node involvement); or distant (distant sites or nodes involved). Stage was grouped into two categories: early (localized cutaneous disease) and late (regional and visceral metastasis). One hundred sixty patients with unknown stage—eight of whom were enrolled in Medicaid—were excluded from this analysis. Tumor histologic subtypes were classified according to the ICD, Oncology (2nd ed.): nodular melanoma (NM, morphology code 8721); lentigo maligna melanoma (LMM, 8742); superficial spreading melanoma (SSM, 8743); acral lentiginous melanoma (ALM, 8744); spindle cell melanoma (8772); and melanoma not otherwise specified (NOS, 8720).17 Desmoplastic melanomas were excluded due to insufficient numbers.
Tumor thickness was categorized as thin (≤2 mm) or thick (>2 mm) because survival differences among cutaneous melanoma stages (as determined by the American Joint Committee on Cancer) are most apparent at this tumor cut-off point.18 Data on tumor thickness were available for 4153 (91%) participants; analyses on tumor thickness were limited to this subset. Patients missing stage or tumor-depth data did not differ significantly from those with known outcomes with respect to study variables (data not shown).
Linkage with Medicaid Enrollment Files
All protocols were approved by the State of California IRB. Medicaid enrollment status was determined from linkage with monthly Medicaid (termed Medi-Cal in California) enrollment files provided by the Medical Care Statistics Section of the California Department of Health Services Methods for 1995–1999.11,12 Standard probabilistic linkage software (Integrity) was used to match Medicaid enrollment files to CCR patient files according to Social Security number; first, middle, last, and alias names; birth day, month, and year; and ZIP code of residence. Manual review was performed to confirm all possible matches.
Two measures were used to characterize Medicaid enrollment: (1) a yes/no measure of enrollment status at time of diagnosis, and (2) a categorical variable with four levels. Those levels were (1) first enrolled in Medicaid at the month of diagnosis, (2) enrolled during the month of diagnosis and 1–11 months prior to diagnosis, (3) enrolled during the month of diagnosis and for at least 12 months prior to diagnosis, and (4) not enrolled in Medicaid at diagnosis. These categories of Medicaid enrollment permitted assessment of the potential influence on outcome of being screened through Medicaid or other health insurance services during the year prior to melanoma diagnosis. Sensitivity analyses were performed to take into account the potential effects of misclassifying patients who may have retroactively enrolled in Medicaid after their diagnosis. Patients enrolled in Medicaid within 3 months before diagnosis were alternately categorized into Levels 1 and 2 of the 4-level Medicaid enrollment variable. Model results were not significantly different using these two categorizations, so these patients remained in their original category.
Month of melanoma diagnosis was known for 5781 matched melanoma patients aged 15–64. Of these, 1223 were excluded either because data were missing for one or more of the measures of melanoma stage (n=160); race/ethnicity (n=525); gender (n=10); or SES (n=573), or because of unusual/infrequent histologic subtype (n=165), for a sample size of 4558.
Statistical Analyses
Demographic and melanoma characteristics of the study participants were evaluated and compared using T-tests and chi-squared statistics. Multivariate logistic regression was utilized to examine the associations between Medicaid status and melanoma stage. Two sets of models were run, examining the impact of the dichotomous Medicaid variable and the effect of the 4-level Medicaid variable. Models were adjusted for neighborhood SES, race/ethnicity, age, gender, histology, year at diagnosis, and marital status. Effect modification was evaluated by assessing the significance of interaction terms between Medicaid variables and covariates, using a cut-off of p=0.2 for stratification. In the subset of 4153 participants with tumor-depth data, logistic regression was used to examine the effect of the dichotomous Medicaid variable on melanoma depth at diagnosis. The reference group in all models was patients not enrolled in Medicaid. Analyses were conducted in 2007, using SAS 9.1. All tests of significance were two-sided; significance was defined as p<0.05.
Results
Of the 4558 California residents with melanoma, 117 (2.6%) were enrolled in Medicaid during the month of melanoma diagnosis (Table 1). Women and unmarried individuals were significantly more likely to be covered by Medicaid at melanoma diagnosis than were men and married participants. Age and Medicaid enrollment status were not strongly associated, but decreasing neighborhood SES was associated with a graded, significant increase in Medicaid enrollment. Black and His-panic melanoma patients were significantly more likely to be enrolled in Medicaid than whites and Asians.
Table 1.
Characteristics of 4558 California melanoma patients by Medicaid enrollment status
| Not enrolled in Medicaid (n=4441) n (%) | Enrolled in Medicaid (n=117) n (%) | Total (N=4558) n (%) | |
|---|---|---|---|
| SEER summary stage* | |||
| Localized | 4020 (90.5) | 85 (72.7) | 4105 (90.1) |
| Regional/distant | 421 (9.5) | 32 (27.4) | 453 (9.9) |
| Tumor depth (mm), mean ± SD* | 1.12 ± 1.33 | 1.98 ± 2.22 | 1.14 ± 1.37 |
| Tumor deptha,* | |||
| ≤2 mm depth | 3545 (87.3) | 65 (67.7) | 3610 (86.8) |
| >2 mm depth | 512 (12.7) | 31 (32.3) | 543 (13.2) |
| Gender* | |||
| Male | 2397 (54.0) | 41 (35.0) | 2438 (53.5) |
| Female | 2044 (46.0) | 76 (65.0) | 2120 (46.5) |
| Age at diagnosis, mean ± SD** | 46.2 ± 11.0 | 43.2 ± 11.7 | 46.1 ± 11.0 |
| Age at diagnosis (years) | |||
| 15–34 | 714 (16.1) | 28 (23.9) | 742 (16.3) |
| 35–44 | 1139 (25.6) | 30 (25.6) | 1169 (25.6) |
| 45–54 | 1433 (32.3) | 37 (31.6) | 1470 (32.3) |
| 55–64 | 1155 (26.0) | 22 (18.8) | 1177 (25.8) |
| Neighborhood SES quintile* | |||
| 1 | 215 (4.8) | 38 (32.5) | 253 (5.6) |
| 2 | 503 (11.3) | 28 (23.9) | 531 (11.7) |
| 3 | 897 (20.2) | 27 (23.1) | 924 (20.3) |
| 4 | 1235 (27.8) | 14 (12.0) | 1249 (27.4) |
| 5 (highest) | 1591 (35.8) | 10 (8.6) | 1601 (35.1) |
| Race/ethnicity* | |||
| Non-Hispanic black | 21 (0.5) | 3 (2.6) | 24 (0.5) |
| Non-Hispanic white | 4137 (93.2) | 89 (76.1) | 4226 (92.7) |
| Non-Hispanic Asian | 33 (0.7) | 1 (0.9) | 34 (0.8) |
| Hispanic | 250 (5.6) | 24 (20.5) | 274 (6.0) |
| Marital status* | |||
| Married | 2633 (59.3) | 45 (38.5) | 2678 (58.8) |
| Not married | 1808 (40.7) | 72 (61.5) | 1880 (41.2) |
| Histologic subytpe (%) | |||
| Superficial spreading melanoma | 1735 (39.1) | 30 (25.6) | 1765 (38.7) |
| Nodular melanoma | 342 (7.7) | 15 (12.8) | 357 (7.8) |
| Lentigo maligna melanoma | 96 (2.2) | 3 (2.6) | 99 (2.2) |
| Acral lentiginous melanoma | 38 (0.9) | 3 (2.6) | 41 (0.9) |
| Spindle cell melanoma | 43 (1.0) | 2 (1.7) | 45 (1.0) |
| Melanoma, NOS | 2187 (49.3) | 64 (54.7) | 2251 (49.4) |
Tumor-depth data were available on a subset of 4153 patients.
p<0.001;
p<0.01
mm, millimeter; NOS, not otherwise specified; SEER, Surveillance, Epidemiology, and End Results
Among the 117 men and women enrolled in Medic-aid at the time of their melanoma diagnosis, 16 (14%) were enrolled for the first time during the month of their diagnosis; 37 (32%) had been enrolled for a period of from 1 to 11 months (continuously or non-continuously) during the year prior to diagnosis; and 64 (55%) had been enrolled for the entire past year (Table 2). Individuals living in higher-SES neighborhoods were equally likely to have been enrolled in Medicaid for a longer period of time compared with those in lower-SES neighborhoods. However, white race/ethnicity was associated with a longer period of Medicaid enrollment during the 1-year period prior to diagnosis when compared to Hispanic race/ethnicity; sufficient data were available to compare only these two groups.
Table 2.
Characteristics by enrollment history of 117 melanoma patients enrolled in Medicaid at diagnosis
| Enrolled at month of diagnosis (n=16) n (%) | Enrolled 1–11 months prior to diagnosis (n=37) n (%) | Enrolled for all 12 months prior to diagnosis (n=64) n (%) | |
|---|---|---|---|
| SEER summary stage* | |||
| Localized | 5 (31.3) | 27 (73.0) | 53 (82.8) |
| Regional/distant | 11 (68.7) | 10 (27.0) | 11 (17.2) |
| Tumor depth (mm), mean ± SD* | 3.4 ± 3.2 | 1.7 ± 1.9 | 2.0 ± 2.2 |
| Tumor depth* | |||
| ≤2 mm depth | 3 (42.9) | 24 (70.6) | 38 (69.1) |
| >2 mm depth | 4 (57.1) | 10 (29.4) | 17 (30.9) |
| Gender* | |||
| Male | 8 (50.0) | 13 (35.1) | 20 (31.3) |
| Female | 8 (50.0) | 24 (64.9) | 44 (68.7) |
| Age at diagnosis, mean ± SD** | 46.7 ± 11.2 | 41.4 ± 11.5 | 43.3 ± 11.8 |
| Age at diagnosis (years) | |||
| 15–34 | 3 (18.8) | 11 (29.3) | 14 (21.9) |
| 35–44 | 4 (25.0) | 10 (27.0) | 16 (25.0) |
| 45–54 | 5 (31.3) | 10 (27.0) | 22 (34.4) |
| 55–64 | 4 (25.0) | 6 (16.2) | 12 (18.7) |
| Neighborhood SES quintile* | |||
| 1 | 4 (25.0) | 12 (32.3) | 22 (34.4) |
| 2 | 4 (25.0) | 9 (24.3) | 15 (23.4) |
| 3 | 5 (31.3) | 6 (16.2) | 16 (25.0) |
| 4 | 2 (12.5) | 6 (16.2) | 6 (9.4) |
| 5 (highest) | 1 (6.3) | 4 (10.8) | 5 (7.8) |
| Race/ethnicity* | |||
| Non-Hispanic black | 1 (6.3) | 0 (0) | 2 (3.1) |
| Non-Hispanic white | 10 (62.5) | 28 (75.7) | 51 (79.7) |
| Non-Hispanic Asian | 0 (0) | 0 (0) | 1 (1.6) |
| Hispanic | 5 (31.3) | 9 (24.3) | 10 (15.6) |
| Marital status* | |||
| Married | 8 (50.0) | 16 (43.2) | 21 (32.8) |
| Not married | 8 (50.0) | 21 (56.8) | 43 (67.2) |
| Histologic subtype | |||
| Superficial spreading melanoma | 1 (6.3) | 12 (32.4) | 17 (26.6) |
| Nodular melanoma | 1 (6.3) | 5 (13.5) | 9 (14.1) |
| Lentigo maligna melanoma | 0 (0) | 0 (0) | 3 (4.7) |
| Acral lentiginous melanoma | 1 (6.3) | 0 (0) | 2 (3.1) |
| Spindle cell melanoma | 0 (0) | 1 (2.7) | 1 (1.6) |
| Melanoma, NOS | 13 (81.3) | 19 (51.4) | 32 (50.0) |
p<0.001;
p<0.01
mm, millimeter; NOS, not otherwise specified; SEER, Surveillance, Epidemiology, and End Results
Patients with late-stage melanoma were significantly more likely to be enrolled in Medicaid at the time of diagnosis than those with localized disease (p<0.0001). There was a significant association between less time enrolled in Medicaid during the year prior to diagnosis and later stage; 17% of individuals enrolled during the entire year had late-stage disease versus 69% of those enrolled during their month of diagnosis (p<0.001). Tumor depth was also strongly associated with a history of Medicaid enrollment; those enrolled in Medicaid at diagnosis were significantly more likely to be diagnosed with melanoma greater than 2 mm (p<0.001). The mean tumor depth was 3.4 mm among those newly enrolled in Medicaid, compared to 1.7 mm and 2.0 mm among those enrolled for 1–11 months and 12+ months, respectively.
Individuals enrolled in Medicaid in the month of their melanoma diagnosis were 21 times as likely to have late-stage disease compared to those not covered by Medicaid at all (OR 21.00, 95% CI=7.26, 60.73; Table 3). Those enrolled in Medicaid for 1–11 months and 12+ months before diagnosis had ORs for late-stage diagnosis of 3.54 (95% CI=1.70, 7.36) and 1.98 (95% CI=1.03, 3.82), respectively, when compared to those not covered by Medicaid. Associations were attenuated by multivariate adjustment, but remained significant for individuals enrolled in Medicaid only at diagnosis (OR 13.64, 95% CI=4.43, 41.98) and for 1–11 months before diagnosis (OR 2.77, 95% CI=1.28, 5.99).
Table 3.
ORs for late-stage melanoma diagnoses among 4558 men and women in California, 1998–1999
| Crude OR (95% CI) | Adjusteda OR (95% CI) | |
|---|---|---|
| Medicaid enrollment and coverage | ||
| Enrolled at diagnosis, not before | 21.00 (7.26, 60.73) | 13.64 (4.43, 41.98) |
| Enrolled for 1–11 mo. before | 3.54 (1.70, 7.36) | 2.77 (1.28, 5.99) |
| Enrolled for 12 or more mo. before | 1.98 (1.03, 3.82) | 1.30 (0.64, 2.64) |
| Not covered by Medicaid | 1.0 | 1.0 |
| Race/ethnicity | ||
| Asian/Pacific Islander | 1.64 (0.63, 4.26) | 1.60 (0.59, 4.31) |
| Hispanic | 1.58 (1.11, 2.25) | 1.11 (0.74, 1.65) |
| Non-Hispanic black | 3.92 (1.62, 9.51) | 2.45 (0.92, 6.55) |
| Non-Hispanic white | 1.0 | 1.0 |
| SES quintile | ||
| 1 | 3.04 (2.12, 4.36) | 2.40 (1.61, 3.58) |
| 2 | 2.06 (1.52, 2.80) | 1.69 (1.23, 2.34) |
| 3 | 1.58 (1.20, 2.08) | 1.49 (1.12, 1.98) |
| 4 | 1.10 (0.83, 1.45) | 1.10 (0.83, 1.46) |
| 5 (highest) | 1.0 | 1.0 |
| Age at diagnosis (years) | ||
| 15–34 | 1.18 (0.86, 1.61) | 1.24 (0.89, 1.72) |
| 35–44 | 1.0 | 1.0 |
| 45–54 | 1.23 (0.95, 1.60) | 1.23 (0.94, 1.61) |
| 55–64 | 1.20 (0.91, 1.58) | 1.15 (0.86, 1.53) |
| Marital status | ||
| Unmarried | 1.10 (0.90, 1.33) | 1.07 (0.86, 1.31) |
| Married | 1.0 | 1.0 |
| Gender | ||
| Male | 1.62 (1.33, 1.98) | 1.66 (1.34, 2.05) |
| Female | 1.0 | 1.0 |
| Histologic subtype | ||
| Superficial spreading melanoma | 0.39 (0.30, 0.50) | 0.40 (0.31, 0.52) |
| Nodular melanoma | 2.59 (1.98, 3.40) | 2.39 (1.81, 3.16) |
| Lentigo maligna melanomab | — | — |
| Acral lentiginous melanoma | 2.82 (1.40, 5.70) | 2.34 (1.10, 4.99) |
| Spindle cell melanoma | 1.66 (0.77, 3.61) | 1.53 (0.70, 3.38) |
| Melanoma, NOS | 1.0 | 1.0 |
Adjusted for neighborhood SES, race/ethnicity, age, gender, histologic subtype, and marital status
Small cell sizes prevented the calculation of odds for this category.
mo., month; NOS, not otherwise specified
Neighborhood SES quintile was strongly associated with odds of late-stage melanoma. Even after risk-factor adjustment, individuals in the lowest three SES quintiles had ORs for late-stage diagnosis of 2.40 (95% CI=1.61, 3.58); 1.69 (95% CI=1.23, 2.34); and 1.49 (95% CI=1.12, 1.98), versus those in the highest SES quintile, respectively. In additional logistic models that included all covariates except for the SES measures, the adjusted ORs for the Medicaid variables increased by 20% to 35%, although little change was observed in the CIs. A further set of models included all covariates but excluded the Medicaid measures; in these models, the OR for the SES measures increased by 0% to 9%, while the ORs for Hispanic and black race/ethnicity increased by 6.6% and 8.2%, respectively.
Race/ethnicity was strongly associated with late-stage diagnosis. Hispanics and blacks had significantly higher unadjusted odds of having late-stage disease at diagnosis than whites. Risk-factor adjustment attenuated the OR to nonsignificance for Hispanics (OR 1.11, 95% CI=0.74, 1.65) and blacks (OR 2.45, 95% CI=0.92, 6.55).
Age and marital status were not associated with late-stage diagnosis, but men had significantly higher odds of having metastatic disease than women (adjusted OR 1.66, 95% CI=1.34, 2.05). Nodular melanomas were significantly more likely to be late-stage at diagnosis than malignant melanoma NOS, while SSM and LMM had lower odds of late-stage diagnosis.
The impact of Medicaid enrollment (ever, never) on stage at diagnosis was also examined. Individuals on Medicaid at diagnosis were >3 times more likely to have late-stage melanoma at diagnosis than those not on Medicaid (OR 3.60, 95% CI=2.37, 5.47). Multivariate adjustment attenuated the OR to 2.55 (95% CI=1.61, 4.04). The ORs for the other covariates differed only slightly from those presented in Table 3 (data not shown).
Among the subset of 4153 participants with tumor-depth data, similar analyses were carried out, using the dichotomous Medicaid variable and the thick (>2 mm) versus thin (≤2 mm) tumor-depth outcome variable. Results were comparable to those for tumor stage at diagnosis. Individuals with Medicaid coverage were significantly more likely to have thick tumors at diagnosis than those not on Medicaid coverage (OR 3.31, 95% CI=2.14, 5.13); this association remained (OR 2.37, 95% CI=1.42, 3.97) after multivariate adjustment.
Discussion
In a population-based series of 4558 melanoma patients, those enrolled in Medicaid were more likely to have advanced disease (measured both by stage of disease and tumor depth) than those with no evidence of Medicaid enrollment. The duration of Medicaid enrollment, however, strongly influenced this association. People newly enrolled in Medicaid were 13 times more likely to have later-stage disease, people intermittently enrolled in Medicaid were more than twice as likely to have later-stage disease, and people continuously enrolled in Medicaid were just as likely to have advanced disease than those not on Medicaid. These associations remained after adjustment for SES, race/ ethnicity, age, marital status, and histologic subtype. Low neighborhood SES and male gender were also independent risk factors for later stage at melanoma diagnosis.
The finding that men and women continuously enrolled in Medicaid during the year prior to diagnosis did not have greater odds for late-stage disease suggests that the screening services provided through Medicaid were adequate. Those who were not continuously enrolled, however, were at increased risk of later-stage disease, and the increasing OR gradient observed by decreasing the length of time enrolled indicates that continuous access to screening is essential for decreasing the likelihood of advanced stage at diagnosis.
The only prior study examining the impact of Medicaid enrollment on melanoma stage at diagnosis found an adjusted OR of late-stage diagnosis for Medicaid enrollees of 4.69 (95% CI=1.90, 11.56) compared to individuals with commercial indemnity insurance.7 In the present analysis, those enrolled in Medicaid (yes/no enrollment variable) also had significantly greater odds of late-stage disease, as well as greater odds of having a tumor depth >2 mm. The current study’s results vis-à-vis duration of Medicaid enrollment underscore the importance of including prior enrollment history in studies utilizing Medicaid data. Studies analyzing other types of cancer have also reported the existence of an inverse gradient between the length of previous Medicaid enrollment and the likelihood of advanced-stage disease.9–12
Several studies have described associations between low SES and an increased likelihood of late-stage melanoma at diagnosis, and report that the effect of SES on late diagnosis operates independently of race/ethnicity.2,6,8,19–22 The results of the current study also showed a clear inverse gradient by neighborhood SES, with individuals living in lower-SES neighborhoods being significantly more likely to have late-stage melanoma at diagnosis than those in high-SES neighborhoods, even after covariate adjustment. The finding that these neighborhood SES–late stage disease associations remained significant after adjustment for Medicaid enrollment further supports the notion that cancer screening in low-SES neighborhoods in California is inadequate, similar to prior analyses.11
Previous studies have reported that Hispanics, Asians, and blacks have higher rates of late-stage melanoma at diagnosis than whites.3,5,23 In this contemporary Californian population, Hispanics and blacks had significantly higher unadjusted odds of late-stage melanoma at diagnosis than whites. After covariate adjustment, the likelihood for late-stage disease among blacks remained marginally significant, while the likelihood among Hispanics was substantially attenuated. These results suggest that much of the difference in melanoma outcome between Hispanics and whites in California is due to the effect of SES and insurance coverage. These data are of interest, given a 2006 study reporting that recent increases in melanoma incidence in the Californian Hispanic population have been primarily in thicker (>1.5 mm) tumors.4 Together, these results indicate that increased education on sun avoidance and skin self-examination may be of particular importance in this population.
Older age was not associated with more late-stage melanoma in this analysis. As in previous studies, males were more likely to have late-stage disease.24 Similar to prior analyses, nodular melanomas were more likely to be later-stage than melanoma of unclassified histology,25 while superficial spreading and lentigo maligna melanomas were less likely to be late-stage. However, nearly 50% of melanoma subtypes were not specified, which limits the interpretation of this finding.
To the authors’ knowledge, this study represents the largest published person-year assessment of insurance status and melanoma outcome. Advantages include the use of a large linked resource from the multi-ethnic population of California and a previously validated measure of SES. Unlike prior studies, this analysis adjusted for the potential biological importance of histologic subtype and used two different outcome measures of melanoma progression. Study limitations include limited statistical power due to the small number (n=117) of melanoma patients enrolled in Medicaid at melanoma diagnosis, which discouraged race/ethnicity-specific Medicaid analyses among non-Hispanic Asians and blacks. Furthermore, individuals who were truly uninsured could not be distinguished from insured individuals who did not link with the Medicaid files, which may have attenuated the ORs for late-stage disease among those enrolled in Medicaid. It has been shown previously that cancer registry data regarding Medicaid status have only modest agreement with linkage,26 so cancer registry data sources regarding patient insurance status were not utilized to further explore this issue.
Additionally, Medicaid allows individuals to apply retrospectively for benefits up to 3 months prior to enrollment, and it was not possible to determine whether the 30 individuals enrolled in Medicaid from 1 to 3 months before diagnosis were retrospectively enrolled (and therefore misclassified). However, analyses of this group indicated that their odds of late-stage disease were lower than that of individuals enrolled at diagnosis, but higher than that of individuals enrolled 4–11 months prior to diagnosis. These findings suggest that the overall gradient of risk observed in this study is accurate.
In summary, late-stage melanoma was present in 9% of all newly diagnosed California melanoma patients but in 27% of those enrolled in Medicaid. Length of enrollment in Medicaid was a key variable. Individuals enrolled continuously in Medicaid for more than 1 year did not have a greater likelihood of late-stage disease than those not on Medicaid, whereas individuals enrolled intermittently had significantly higher odds of late-stage disease, and those enrolled at diagnosis had the greatest odds of late-stage disease. Access to skin screening for melanoma through Medicaid is thus apparently successful if individuals are continuously enrolled in the year prior to diagnosis. Hispanic individuals did not have significantly higher odds of having late-stage melanoma at diagnosis, a finding contrary to recent studies that requires confirmation. Finally, lower neighborhood SES was also a significant independent predictor of late-stage disease, emphasizing the need for improved education and screening in low-SES communities to promote the early detection of melanoma.
Acknowledgments
Dr. Pollitt was supported by a Stanford University Medical Scholars Fellowship. Dr. Clarke was supported by a Developmental Grant from the Stanford Cancer Center and the SEER Rapid Response Surveillance Study program of the National Cancer Institute Surveillance Epidemiology and End Results program through a supplement to contract N01-PC-35136. The collection of cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract N01-PC-35136, awarded to the Northern California Cancer Center, contract N01-PC-35139, awarded to the University of Southern California, and contract N02-PC-15105, awarded to the Public Health Institute; and the CDC’s National Program of Cancer Registries, under agreement U55/CCR921930-02, awarded to the Public Health Institute. The ideas and opinions expressed herein are those of the authors, and endorsement by the State of California, Department of Health Services, the National Cancer Institute, and the CDC or their contractors and subcontractors is neither intended nor should be inferred.
Footnotes
The full text of this article is available via AJPM Online at www.ajpm-online.net; 1 unit of Category-1 CME credit is also available, with details on the website.
No financial disclosures were reported by the authors of this paper.
References
- 1.Markovic SN, Erickson LA, Rao RD, Weenig RH, Pockaj BA, Bardia A, et al. Malignant melanoma in the 21st century, part 2: staging, prognosis, and treatment. Mayo Clin Proc. 2007;82:490–513. doi: 10.4065/82.4.490. [DOI] [PubMed] [Google Scholar]
- 2.Ortiz CA, Goodwin JS, Freeman JL. The effect of socioeconomic factors on incidence, stage at diagnosis and survival of cutaneous melanoma. Med Sci Monit. 2005;11:RA163–172. [PubMed] [Google Scholar]
- 3.Byrd KM, Wilson DC, Hoyler SS, Peck GL. Advanced presentation of melanoma in African Americans. J Am Acad Dermatol. 2004;50:21–4. doi: 10.1016/s0190-9622(03)02091-7. discussion 142–3. [DOI] [PubMed] [Google Scholar]
- 4.Cockburn MG, Zadnick J, Deapen D. Developing epidemic of melanoma in the Hispanic population of California. Cancer. 2006;106:1162–8. doi: 10.1002/cncr.21654. [DOI] [PubMed] [Google Scholar]
- 5.Hemmings DE, Johnson DS, Tominaga GT, Wong JH. Cutaneous melanoma in a multiethnic population: is this a different disease? Arch Surg. 2004;139:968–72. doi: 10.1001/archsurg.139.9.968. discussion 972–3. [DOI] [PubMed] [Google Scholar]
- 6.Chang AE, Karnell LH, Menck HR. The national cancer data base report on cutaneous and noncutaneous melanoma: a summary of 84,836 cases from the past decade. The American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer. 1998;83:1664–78. doi: 10.1002/(sici)1097-0142(19981015)83:8<1664::aid-cncr23>3.0.co;2-g. [DOI] [PubMed] [Google Scholar]
- 7.Roetzheim RG, Pal N, Tennant C, Voti L, Ayanian JZ, Schwabe A, et al. Effects of health insurance and race on early detection of cancer. J Natl Cancer Inst. 1999;91:1409–15. doi: 10.1093/jnci/91.16.1409. [DOI] [PubMed] [Google Scholar]
- 8.Van Durme DJ, Ferrante JM, Pal N, Wathington D, Roetzheim RG, Gonzalez EC. Demographic predictors of melanoma stage at diagnosis. Arch Fam Med. 2000;9:606–11. doi: 10.1001/archfami.9.7.606. [DOI] [PubMed] [Google Scholar]
- 9.Bradley CJ, Given CW, Roberts C. Late stage cancers in a Medicaid-insured population. Med Care. 2003;41:722–8. doi: 10.1097/01.MLR.0000065126.73750.D1. [DOI] [PubMed] [Google Scholar]
- 10.Koroukian SM. Assessing the effectiveness of Medicaid in breast and cervical cancer prevention. J Public Health Manag Pract. 2003;9:306–14. doi: 10.1097/00124784-200307000-00009. [DOI] [PubMed] [Google Scholar]
- 11.O’Malley CD, Shema SJ, Clarke LS, Clarke CA, Perkins CI. Medicaid status and stage at diagnosis of cervical cancer. Am J Public Health. 2006;96:2179–85. doi: 10.2105/AJPH.2005.072553. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Perkins CI, Wright WE, Allen M, Samuels SJ, Romano PS. Breast cancer stage at diagnosis in relation to duration of medicaid enrollment. Med Care. 2001;39:1224–33. doi: 10.1097/00005650-200111000-00009. [DOI] [PubMed] [Google Scholar]
- 13.McLaughlin C, Hotes J, Wu XC, editors. Cancer in North America, 1997–2001, Volume 1: Incidence. Springfield IL: North American Association of Central Cancer Registries; 2004. [Google Scholar]
- 14.Diez-Roux AV, Kiefe CI, Jacobs DR, Jr, et al. Area characteristics and individual-level socioeconomic position indicators in three population-based epidemiologic studies. Ann Epidemiol. 2001;11:395–405. doi: 10.1016/s1047-2797(01)00221-6. [DOI] [PubMed] [Google Scholar]
- 15.Krieger N, Williams DR, Moss NE. Measuring social class in U.S. public health research: concepts, methodologies, and guidelines. Annu Rev Public Health. 1997;18:341–78. doi: 10.1146/annurev.publhealth.18.1.341. [DOI] [PubMed] [Google Scholar]
- 16.Woolf SH, Johnson RE, Geiger HJ. The rising prevalence of severe poverty in America: a growing threat to public health. Am J Prev Med. 2006;31:332–41. doi: 10.1016/j.amepre.2006.06.022. [DOI] [PubMed] [Google Scholar]
- 17.Percy C, Van Holten V, Muir C. International classification of diseases for oncology. Geneva: WHO; 1995. [Google Scholar]
- 18.Balch CM, Buzaid AC, Soong SJ, et al. Final version of the American Joint Committee on Cancer staging system for cutaneous melanoma. J Clin Oncol. 2001;19:3635–48. doi: 10.1200/JCO.2001.19.16.3635. [DOI] [PubMed] [Google Scholar]
- 19.Carli P, De Giorgi V, Palli D, et al. Patterns of detection of superficial spreading and nodular-type melanoma: a multicenter Italian study. Dermatol Surg. 2004;30:1371–5. doi: 10.1111/j.1524-4725.2004.30434.x. discussion 1375–6. [DOI] [PubMed] [Google Scholar]
- 20.Roetzheim RG, Gonzalez EC, Ferrante JM, Pal N, Van Durme DJ, Krischer JP. Effects of health insurance and race on breast carcinoma treatments and outcomes. Cancer. 2000;89:2202–13. doi: 10.1002/1097-0142(20001201)89:11<2202::aid-cncr8>3.0.co;2-l. [DOI] [PubMed] [Google Scholar]
- 21.Singh GK, Miller BA, Hankey BF, Edwards BK. Area socioeconomic variations in US cancer incidence, mortality, stage, treatment, and survival, 1975–1999. Bethesda MD: National Cancer Institute; 2003. [Google Scholar]
- 22.Epstein DS, Lange JR, Gruber SB, Mofid M, Koch SE. Is physician detection associated with thinner melanomas? JAMA. 1999;281:640–3. doi: 10.1001/jama.281.7.640. [DOI] [PubMed] [Google Scholar]
- 23.Cress RD, Holly EA. Incidence of cutaneous melanoma among non-Hispanic whites, Hispanics, Asians, and blacks: an analysis of California cancer registry data, 1988–93. Cancer Causes Control. 1997;8:246–52. doi: 10.1023/a:1018432632528. [DOI] [PubMed] [Google Scholar]
- 24.Jemal A, Devesa SS, Hartge P, Tucker MA. Recent trends in cutaneous melanoma incidence among whites in the U. S J Natl Cancer Inst. 2001;93:678–83. doi: 10.1093/jnci/93.9.678. [DOI] [PubMed] [Google Scholar]
- 25.Demierre MF, Chung C, Miller DR, Geller AC. Early detection of thick melanomas in the U.S: beware of the nodular subtype. Arch Dermatol. 2005;141:745–50. doi: 10.1001/archderm.141.6.745. [DOI] [PubMed] [Google Scholar]
- 26.Chan JK, Gomez SL, O’Malley CD, Perkins CI, Clarke CA. Validity of cancer registry Medicaid status against enrollment files: implications for population-based studies of cancer outcomes. Med Care. 2006;44:952–5. doi: 10.1097/01.mlr.0000220830.46929.43. [DOI] [PubMed] [Google Scholar]