In this study, we determined that the prevalence rates for lung, colorectal, breast, and among Maryland Medicaid enrollees were 75/10,000, 63/10,000, 92/10,000, and 45/10,000, respectively. Prevalence rates are measures of the burden of diseases in a community for the purpose of setting public policy and allocating resources.
Cancer generally requires treatment over a period of time and, thus, is suitable for care prevalence studies. Cancer incidence calculations provide only a part of the epidemiology and disease burden story. Incidence rates do not capture the complex features of the healthcare required for cancer patients.
Table 6. Estimates of Cancer Prevalence Rates in Maryland Medicaid versus Rates Based on the National Health Interview Survey (per 10,000)
Our Estimates Estimates from 1987 NHIS Estimates from 1992 NHIS1
|74.50 14.2C 1763.22 34.3 35.5|
91.51 133.2d 79.3d
44.56 32.4 37.1е
|a Byrne J, Kessler L, Devesa SS. The prevalence of cancer among adults in the United States: 1987.Cancer. 1992;69:2154-2159. b Hewitt M, Breen N, Devesa S. Cancer prevalence and survivorship issues: analyses of the 1992 NationalHealth Interview Survey. J Natl Cancer Inst. 1999;91:1480-1486. c Includes lung/larynx cancers. d Female breast cancer only.|
e Includes prostate, testes, and other male genital organ cancers.
Using claims data to estimate prevalence rates is one of many ways to estimate cancer prevalence rates. The traditional method of estimating cancer prevalence rates is based on the Connecticut Tumor Registry model, which has recorded patients diagnosed since 1935. Using the number of survivors of cancer who had been diagnosed during a certain period, and mortality rates, researchers can directly compute a prevalence rate. When applied to the total U.S. population for selected years, a national estimate of cancer prevalence could be obtained. The accuracy of estimation can be compromised; however, since patients can be lost to follow-up. In addition, the ascertainment of nonfatal cancers in the earlier years of the registry can be incomplete. Another inaccuracy occurs if people diagnosed before a registry began are still alive at the reference time when the prevalence rate is estimated. These limitations, if not offset by the inclusion of people who no longer reside in the registry area, would lead to underestimation of the prevalence rate. One study by Capocaccia and DeAngelis confirmed the incompleteness of prevalence based on cancer registry data alone.
buy antibiotics canada
Two other sources exist for estimating prevalence rates. The Surveillance, Epidemiology, and End Results (SEER) Program, an authoritative source of information from the National Cancer Institute (NCI) on cancer incidence and survival at the national level, is also a cancer registry program. It has been used to estimate cancer prevalence rates. The National Health Interview Survey (NHIS) is an alternative source for obtaining prevalence estimates. It has been used to estimate the prevalence of cancer among adults in the United States. The NHIS collects data from a nationwide probability sample of noninstitutionalized people. Its information on chronic conditions is obtained from patients’ self-reports, which is considered to be less accurate than the cancer registry. Both under- and over-reporting are possible.
We used an alternative to the aforementioned means for estimating cancer prevalence rates by measuring “care prevalence,” which is “an estimate of prevalent cases that are still under care.” The NCI is carrying out a project estimating care prevalence for colorectal cancer using SEER-Medicare data. The care prevalence is a valid prevalence measure though it has not been widely used. It capitalizes on criterion and construct validity—that is, the ability to predict events. Furthermore, care prevalence presents a good case for external validity, or generalizability. Estimating care prevalence is especially important when we are interested in estimating burden of cancer in a program such as Medicare or Medicaid. Other measures of prevalence attempt to include all patients (e.g., those in remission or “cured;” or those still under care) for a measure of total prevalence; however, insurers may only be concerned with the number of cancer patients within their plan and who are currently under some type of care. Other studies have used Medicare and Medicaid claims data to determine cancer prevalence rates. Using claims data to estimate care prevalence rates also avoids the difficulties in differentiating prevalent cases and incident cases that previous researchers had when they tried to use claims data to identify cancer incident cases.
kamagra soft tabs
One must be cautious when comparing our estimates of the cancer prevalence rates in Maryland Medicaid to estimates of the prevalence rates at the national level in previous studies. Cancer prevalence rate estimates can vary for multiple reasons. Prevalence rates depend on the disease parameters, population characteristics, medical services, and the means of data collection. Variation in prevalence might be due to differences in incidences of the diseases and/or the mortality rates of the population selected for the analysis. The difference in prevalence can also result from the extent of screening programs, the stage distribution at diagnosis, and access to treatment facilities. In addition, prevalence rates can vary according to differences in coding conventions and data quality—namely, the follow-up rates from survey, the proportion of patients identified only from death certificates, and the number of patients who migrated in and out of the population. Medicaid data differ from cancer registry and NHIS in several aspects. Medicaid enrollees are of uniformly lower socioeconomic status. In addition, some Medicaid-eligible enrollees are only enrolled in the program when a health problem is discovered. Others are enrolled only when they “spend down” their assets and become eligible for Medicaid. However, the cancer registry and NHIS included patients of various socioeconomic status. Moreover, the administrative database collects data following different conventions than those of the cancer registry and NHIS.
buy generic levitra
We need to consider a reference framework despite the caveats of comparing our estimates of cancer prevalence rates among Maryland Medicaid population with those from previous studies at the national level. We have compared our estimates of overall cancer-specific prevalence rates in the Maryland Medicaid populations with those in the national population. Table 6 presents the comparison between our estimates and those based on the 1987 and 1992 NHIS. Maryland Medicaid data generally indicated higher prevalence rates (1.2 to 5.2 times as high as those at the national level) for all four of these cancers, compared with the national level. The only exception was the estimate of breast cancer based on the 1987 NHIS, which was higher than our estimate. No previous prevalence studies on these four cancers in other Medicaid population are available for comparison to our estimates in Maryland. One anecdotal argument for high cancer prevalence rates in the Medicaid population is that some individuals become Medicaid-eligible because of their cancer diagnosis and the related physical and economic impact. However, because of the significantly higher cancer prevalence rates in this population, further research should be done to determine whether our findings can be replicated in another Medicaid population, and, if so, to identify the reasons for much higher rates for cancer in the Medicaid population. One potential contributing factor to the observed higher prevalence rates relates to whether cancer costs would result in individuals becoming impoverished to the extent that they qualify for Medicaid. A post hoc analysis of the Maryland Medicaid eligibility file does not support this hypothesis. In the overall Medicaid population, 9.48% qualified through “spend down”; in the cancer population, 8.17% qualified through “spend down.”