Colorectal, Breast, and Prostate Cancers in Medicaid: METHODS

26 Dec

Study Population and Data Sources

This study utilized an historical cross-sectional study design. The data source for this study was Maryland Medicaid administrative claims data (including demographic, eligibility, managed care organization (MCO) enrollment data, medical, and institutional fee-for-service claims) and MCO encounter data. In accordance with patient confidentiality concerns, this study was approved by the State of Maryland (Protocol # 01-16). It has also been reviewed and deemed to be exempt by the Institutional Review Board of the University of Maryland (Exemption No. CDM-040101).

To be included in our analysis, individuals needed to be Maryland Medicaid recipients 18 and older, with encounters, medical or institutional claims based on International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9 CM) diagnosis codes; prescription drug National Drug Codes (NDCs) for chemotherapy drugs, tamoxifen, analgesics, hematopoietics and Xeloda; or current procedure technology (CPT) codes for lung, colorectal, breast. Furthermore, they must have been continuously eligible for Medicaid between January 1,2000 and December 31,2000 to be includ ed in the study cohort. See for the ICD-9CM Diagnosis Codes that were used to identify the cancers of interest. Demographic and enrollment information was extracted from each source file.


Frequencies and crosstabulations were performed on all data to validate the completeness and integrity of the data. Algorithms were developed to evaluate claims for adjustment and duplications. Validation of these algorithms was conducted by reviewing raw claims for randomly selected recipients. The resultant data were unique with no duplication.
canadian drugstore online

The prevalence for each cancer was calculated. The prevalence rates reflect the period from January 1, 2000 to December 31, 2000. The number of eligible Maryland Medicaid recipients 18 and older on January 1, 2000 was used as the denominator. Prevalence was calculated by race, region, age, and gender.

Prevalence rates across regions were estimated to explore reasons behind the different patterns of racial disparities across regions reported by previous researchers. Prevalence rates across age and gender groups were calculated so that they could be compared with estimates of previous studies at the national level so that our study results could be validated.

There were three racial groups: black, white, and other. The racial group “other” was comprised of Hispanics, Asians, Native Americans, Pacific Islanders/Alaskans, and those of unknown ethnicity/race. Since each of these “other” racial groups individually accounted for less than 4% of the total Maryland Medicaid population, we decided that it was not appropriate to calculate prevalence rates for each individual group (combined, they account for less than 10% of the study population). Thus, only the differences between whites and blacks were analyzed in this study.

We defined geographic region as urban (Baltimore city); rural (Allegany, Garrett, Washington, Kent, Queen Anne’s, Caroline, Talbot, Dorchester, Somerset, Wicomico, and Worchester counties); and suburban (the rest of Maryland) based on the proportion of agricultural populations in the total population in the regions. Each person was categorized to a geographic region (urban, suburban, rural) according to his/her county of residence on January 1, 2000. The patients were categorized into two age groups, those under 65 and 65 and older. The differences in subgroups were tested using Chi-squared tests.

Cancer prevalence data were age-adjusted using the direct standardization method. This was done by multiplying the age-specific rates in the target population by the age distribution of the standard population. Maryland Medicaid enrollees were used as the standard population.