Flu Shots and Cancer
Introduction
Previous studies have demonstrated influenza immunization to be among the best processes for managing cancer (Pearce, O’Laughlin, & Hayney, 2008). They have also demonstrated that person diagnosed with cancer could respond well to influenza immunization without severe adverse events (Berglund et al., 2014). This notwithstanding, studies have also shown the rate of influenza immunization to be low among adults (Privileggio et al., 2013). They have also shown mandatory immunization to be an ethical issue (Steckel, 2007). Based on these facts, influenza immunization is an important topic in public health worth to be evaluated deeper than it has been evaluated this far (Doganis et al., 2013).
The retrospective research design used by Locher and her colleagues enabled the researchers to determine the purpose of their study, which was to establish the likelihood of receiving influenza immunization among older adults before and after they were diagnosed or not diagnosed with cancer. In contrast to this research design, the researchers could have also used prospective research design to evaluate the purpose of their study. However, the following are the pros and cons for the two study designs.
Table 1: study designs comparison
| Study design | Pros | Cons |
| Retrospective design | Inexpensive and less time consuming than prospective design. | Inferior to prospective design, culpable to errors due to bias and confounding. |
| Prospective design | Superior to retrospective design because there are fewer sources of confounding and biases errors. | Time consuming, costly to conduct |
The main research question for the study was whether older adults diagnosed with cancer were likely to receive influenza immunization than their counterparts that were not diagnosed with cancer. The null hypothesis associated with this research question could have been that older adults not diagnosed with cancer were not likely to receive influenza immunization than their counterparts diagnosed with cancer. In terms of effects, the study was designed to measure the effects of cancer diagnosis on influenza immunization.
Methods
The population researched in this study included Medicare beneficiaries fifty years and over that, had either been diagnosed or not with cancer. Nothing has been commented regarding the sampling procedures that were used in the study. Accordingly, it is not possible to determine whether the sample was collected randomly or non-randomly (Ojha et al., 2015). This notwithstanding, a sample of 35,257 persons diagnosed with cancer, and another sample of 203,124 persons not diagnosed with cancer were used in the study. These samples were later on matched using propensity matching score to produce two samples of sizes 35,229 and 35,257. As the researchers demonstrated, the two samples increased the statistical power of the results together with the likelihood of rejecting the null hypothesis (Locher et al., 2012). As a result, it was possible to generalize the results of the study to the whole population considered in the study. With regard to data collection, the researchers have not indicated the tools they used to collect data, but they have indicated that they used health common procedure coding system codes to identify influenza immunization. The main variables that were included in the study were cancer specifics (stage of cancer, site, and date of diagnosis), demographics (race, sex and age) and influenza immunization. These variables are highlighted in the table below.
Table 2: variables of the study
| Variable type | Variable name | Source | Potential responses | Level of measurement |
| Dependent | Influenza immunization | SEER | Yes/no | Binomial/Dichotomous |
| Independent | Cancer diagnosis | SEER | Yes/no | Binomial/Dichotomous |
| Independent | Age | SEER | Actual number from 50 and above | Continuous |
| Independent | Sex | SEER | Male or female | Binomial/Dichotomous |
| Independent | Stage of cancer | SEER | Early, distant, intermediate and onstage | Binomial/Dichotomous |
| Independent | Race | SEER | White, black and other | Binomial/Dichotomous |
| Independent | Site | SEER | Breast, colon and prostate | Binomial/Dichotomous |
The Stata/MP programme was used to analyze the data. With regard to sample description, descriptive statistics together with sex, race and age were used to describe the sample. By definition, univariate statistics focus their attention on one descriptive variable whereas multivariate statistics focus their attention on more than two descriptive variables (Sullivan, 2012). Point estimates together with 95% confidence intervals for various proportions were calculated to describe the characteristics for the sample. Multivariate analyses were also conducted to adjust for demographic characteristics with an aim of determining statistical significance for various demographic factors.
Despite the fact that the use of confidence intervals and proportions was enough to determine the significance of various variables used in the study, correlation coefficients could have also been used to determine the relationship between cancer diagnosis and influenza immunization. This could have been done because the focus of the study was on the association between cancer diagnosis and influenza immunization, which could be done using correlation coefficients.
Confounding variables refer to the variables that affect the results of the study. These variables need to be identified so that their influence on the results in terms of bias can be minimized. The potential confounders in this study included gender, race, cancer’s stage and age. Other possible confounders not evaluated in the study include clinicians’ decisions on influenza immunization and accessibility to facilities that offer immunization.
Results
The authors noted that majority of Medicare beneficiaries diagnosed with cancer that were likely to receive influenza immunization were male, white and older adults.
Table 3: demographic characteristics for the sample population with and without cancer
| With cancer | Without cancer | ||
| Variable | Type of statistic | Statistic (%, mean, standard deviation) | Statistic (%, mean, standard deviation) |
| Age Mean Standard deviation | Range | 81.8 6.0 | 77.6 12.1 |
| Sex Male Female | Frequency | 46.7 53.3 | 40.2 59.8 |
| Race White Black Other | Frequency | 87.8 7.0 5.1 | 82.6 8.7 8.7 |
| Prior immunization Yes No | Frequency | 46.8 53.2 | 42.6 57.4 |
| Site Breast Colon Prostate | Frequency | 33.7 24.8 41.5 | |
| Stage Early Intermediate Distant Unstaged | Frequency | 40.3 72.1 2.6 2.3 |
The DOD analysis revealed that immunization rate among persons diagnosed with cancer increased over time from 46.8 percent to 50.8 percent. However, this change was slightly lower than the one witnessed among persons not diagnosed with cancer. The change witnessed among persons not diagnosed with cancer increased from 42.6 percent to 79.7 percent.
Table 4: DOD for influenza immunization rate
| Variable (Immunization rate) | Type of statistic | With cancer % (95% CI) | Without cancer % (95% CI) | Adj. | p-value |
| Immunization rate | Logistic regression | 0.034-0.045 | 0.369-0.373 | 0.331 | <0.001 |
Logistic regression demonstrated that persons that had not been diagnosed with cancer were 7.25 times likely to receive influenza immunization than their counterparts that had been diagnosed with cancer.
Table 5: association between immunization rate and cancer among the beneficiaries of Medicare
| Type of statistic: multivariate logistic regression | |||
| Cancer and immunization versus no cancer and immunization | |||
| Classification | aOR | 95% CI | p-value |
| Received immunization | 0.138 | 0.134-0.143 | <0.001 |
Conclusion/discussion
The main purpose of the study by Locher et al (2012) was to determine the prevalence for influenza immunization among older adults with and without cancer using a retrospective design to gather the data. Based on the data obtained from SEER tumor registry, there are strong indications that majority of old people diagnosed with cancer do not receive influenza immunization. In addition, the study established that men were more likely to be immunized than women with blacks and white people demonstrating a higher likelihood for immunization than other racial groups (Locher et al., 2012). The study’s findings are similar to the findings of other studies conducted previously. For example, with regard to racial difference in influenza vaccination, the study’s findings are similar to the findings of a study conducted by Egede and Zheng in 2003. The said study established that there were significant racial differences in influenza vaccination among high-risk individuals (Egede, & Zheng, 2003). In my professional practice, I would use these findings to advocate for immunization among the people that are less likely to receive immunization (Jones et al., 2008). This does not mean that I would not focus my attention on the male blacks and whites because they are more likely to receive immunization, but it means that I would be more concerned with the people that are less likely to receive immunization than those that are likely to receive it. This simply means that I wound use the study’s findings to guide my professional practice (Ortbals et al., 1977).
Limitation
Given that the data for the study came from a U.S based database, then its findings cannot be generalized to other parts of the world. However, they may be generalized in USA because they are representative of the U.S population. In spite of this fact, given that the data for the study evaluated cancer victims that benefitted from Medicare in 2001 only, then it may not be possible to generalize the study’s findings to other time points because past data were not taken into account in the study. At the same time, the study has not indicated whether research participants were identified randomly or non-randomly, and for this reason, the applicability of the findings may be limited because this issue is not clear.
Future work
Based on my literature review, future studies should focus their attention on the possible influence that clinicians may have on influenza immunization among cancer patients. Although this does not mean that clinicians may not be aware of its importance among cancer patients, it is possible that they focus much of their attention on cancer treatment and forget about influenza immunization (Locher et al., 2012).
References
Berglund, A. et al. (2014). The response to vaccination against influenza (H1N1) 2009, seasonal influenza and Streptococcus pneumonia in adult outpatients with ongoing treatment for cancer with and without rituximab. Acta oncologica, 53; 1212-1220.
Doganis, D. et al. (2013). Compliance with immunization against H1N1 influenza virus among children with cancer. Pediatric hematology and oncology, 30; 149-153.
Egede, L., & Zheng, D. (2003). Racial/ethnic differences in influenza vaccination coverage in high-risk adults. American journal of public health, 93(12); 2074-2078.
Jones, K. et al. (2008). Improving adult immunization rates in primary care clinics. Nursing economics, 26(6); 404-407.
Locher, J. et al. (2012). Influenza immunization in older adults with and without cancer. Journal of the American Geriatrics Society, 60(11), 2099-2103.
Ojha, R. et al. (2015). The impact of vaccine concerns on ratio/ethnic disparities in influenza vaccine uptake among health care workers. American journal of public health, 105(9); e35-e41.
Ortbals, D et al. (1977). Influenza immunization of adult patients with malignant diseases. Annals of internal medicine, 87; 552-557.
Pearce, C., O’Laughlin, J., & Hayney, M. (2008). Optimal cancer management includes influenza vaccination. Journal of the American pharmacist association, 48(6); 813-814.
Privileggio, L. et al. (2013). Rates of immunization against pandemic and seasonal influenza in persons at high risk of severe influenza illness: a cross-sectional study among patients of the French Sentinelles general practitioners. BMC public health, 13(246); 1-8.
Steckel, C. (2007).Mandatory influenza immunization for health care workers – an ethical discussion. AAOHN Journal, 55(1); 34-39.
Sullivan, L (2012). Essentials of biostatistics in public health. Sudbury: Jones & Bartlett Learning.