In 2010, about 50,000 new human immunodeficiency virus (HIV) infections occurred in the United States, and more than 1.1 million people were living with HIV infection (1,2). About 16% of HIV infections go undetected (2). Since 2006, the Centers for Disease Control and Prevention (CDC) has recommended HIV testing as a routine part of medical care for all persons aged 13 to 64 years (3). Compared with the general population, people with HIV infection have increased cancer risks; Kaposi sarcoma, non-Hodgkin lymphoma, and cervical cancer are AIDS-related cancers (4,5). Because of co-infections with hepatitis B and C and Epstein-Barr virus, HIV-infected people also have elevated risks for non-AIDS–related cancers, such as liver cancer and Hodgkin disease (6).

Driven by improvements in early detection and treatment, cancer survival rates have increased substantially in the past 30 years (7). By 2007, nearly 12 million US cancer survivors were living, and the number is projected to reach about 18 million by 2022 (8,9). Given the high prevalence of both cancer and HIV infection and the natural linkage between the two, occurrence of HIV infection in cancer survivors is not rare. Patel et al reported that for HIV-infected persons, having AIDS-related cancers, a lower cluster of differentiation (CD4) cell count, and high viral load at cancer diagnosis were positively associated with poor 5-year survival (10). Undiagnosed HIV infection among cancer patients may contribute to poor cancer treatment outcomes and delay timely HIV care and treatment, resulting in increased risks for death (11). Identification of cancer survivors who have never been tested for HIV infection can reduce that risk.

To our knowledge, little is known about HIV testing among cancer survivors (11). This study’s objective was to describe the proportion of cancer survivors tested for HIV by demographic and health-related variables, to provide state-specific estimates of HIV testing, and to characterize factors associated with HIV testing.

Of 407,402 persons interviewed during 2009 who answered the questions about cancer history, 59,173 reported a history of cancer. Of these, 24,485 were aged 18 to 64 years and were included in our analysis. Most cancer survivors were women (62.3%), non-Hispanic white (83.5%), aged 45 to 64 years (74.6%), and had health insurance (89.5%). About 13% of cancer survivors had AIDS-related cancers: 11.5% had cervical cancer and 1.9% non-Hodgkin lymphoma. Approximately 3% of cancer survivors reported at least 1 HIV risk behavior in the past year.

Of cancer survivors aged 18 to 64 years, 40.8% reported having ever had an HIV test (Table 1). The proportion who reported HIV testing varied significantly by race/ethnicity, age, education, cancer type, time since cancer diagnosis, and HIV risk behaviors. Within these respective categories, those most likely to have been tested included non-Hispanic blacks (59.5%), patients aged 25 to 34 years (72.2%), patients with less than a high school education (45.0%), cervical cancer survivors (51.2%), survivors 5 to 9 years post cancer diagnosis (42.9%), and those who reported engaging in at least 1 HIV risk behavior in the past year (58.0%). Measures of emotional support, life satisfaction, employment status, marital status, disability status, health insurance, and medical cost concern also were significantly associated with HIV testing. History of HIV testing varied significantly by state, ranging from 68.3% in the District of Columbia to 24.1% in Nebraska (Table 2). Among all states, the District of Columbia (68.3), Maryland (50.6%) and California (48.7%) had the highest proportions of cancer survivors who reported having been tested for HIV.

The multivariable logistic regression analysis (Table 3) revealed that women with an AIDS-related cancer, persons who had ever had a disability, and patients with medical cost concern were more likely to report having had an HIV test than men with cancer and women with non-AIDS–related cancer, those who did not have a disability or those who did not have medical cost concerns. Other significant findings included higher prevalences of testing among non-Hispanic blacks and Hispanics, patients with at least some college education, and patients who were not currently married or living with a partner. After adjusting for all variables, health insurance, emotional support, life satisfaction, employment status, and HIV risk behaviors were no longer significantly associated with HIV testing. There was a significant nonlinear association between age and HIV testing. Increasing age was significantly associated with a decline in HIV testing above age 35 (Figure 1). HIV testing prevalence tended to be slightly higher with increasing time since diagnosis up to around 10 years with slightly lower testing with increasing years since diagnosis beyond 20 years (Figure 2). However, this relationship was not significant (P = .07).

Our study, based on the largest telephone survey of adults in the United States, found that only 41% of US cancer survivors under age 65 in 2009 had ever had an HIV test. The likelihood of having had HIV testing varied markedly by state. Factors associated with HIV testing included being non-Hispanic black or Hispanic, younger age, having higher education, not being married or living with a partner, being disabled, and having medical cost concerns. Having an AIDS-related cancer was associated with HIV testing only among women.

Although information on the proportion tested for HIV is lacking for cancer survivors, it has been widely reported for the US general population. In 2013, CDC reported, on the basis of the National Health Interview Survey (NHIS), that the percentage of adults who had ever been tested for HIV significantly increased from 36.6% in 2000 to 45.0% in 2010 (13). Similar to what we have observed among cancer survivors, NHIS analyses showed that among the general population, non-Hispanic blacks, people aged 25 to 34 years, and people with a reported risk for HIV infection had the highest proportion of HIV testing (13). A British study examined HIV testing among cancer patients in a clinic where HIV testing was recommended for all lymphoma patients (14). This study reported that only 59% of 113 lymphoma patients underwent an HIV test after their cancer diagnosis, and a high prevalence of HIV co-infection (8.0%) was observed (14). In 2013, the US Preventive Services Task Force also recommended that clinicians screen adolescents and adults aged 15 to 65 years who were at average risk for HIV infection (15). The low percentage of HIV testing reported by our study emphasizes the need for more efforts to promote HIV testing among cancer survivors. Our study may serve as a baseline measure to monitor changes in HIV testing among cancer survivors.

Compared with the overall population, adolescents and young adults are disproportionately affected by HIV (3). This finding may partially account for the decline in HIV testing with age, as suggested by our study. However, among persons living with a diagnosed HIV infection, the percentage aged 50 years or older increased from 28.6% in 2007 to 32.7% in 2009 (1). One possible explanation is that HIV testing increased over time among this older population (13). The positive associations between HIV testing and disabilities as identified in our study were consistent with a 2002 NHIS study among the general population (16). That study revealed that adults with physical and mental disability were more likely to report having had an HIV test than nondisabled adults (16). One explanation for the greater likelihood of HIV testing among physically and mentally disabled adults was their increased risk of rape and sexual abuse and a higher rate of participation in HIV risk behaviors (16).

Our study shows that about 60% of US cancer survivors under age 65 have never had an HIV test. Although non-Hodgkin lymphoma and cervical cancers are AIDS-related cancers, almost half of patients with these cancers had never been tested for HIV. The nontargeted “opt-out” HIV testing (the test will be performed unless the patient declines) in all health care settings, recommended by CDC in 2006, has not been widely implemented (11). This may, in part, contribute to the low reports of HIV testing among cancer survivors. Other contributing factors have not been reported in the literature. However, studies have shown that for the general population, fear, stigma, and discrimination associated with an HIV diagnosis; physicians’ attitudes; and the need for additional health care personnel and effort affect adoption of HIV testing (13,17,18). In addition, the cost or reimbursement of the cost for HIV testing are shared concerns of both patients and health care providers (19). In our study, we found that only 16.7% of cancer survivors reported having medical cost concerns, and cancer survivors who had medical cost concerns were more likely to have had an HIV test than cancer survivors who had no such concerns. In 2009, the Centers for Medicare and Medicaid Services allowed reimbursement for routine HIV testing (20). However, the changes in Medicare and Medicaid coverage may not have had a substantial effect on HIV testing in our study population because of the age limits of 18 to 64 years and the time the survey was conducted. One possible explanation is that medical cost concerns are a surrogate for having low income. Those who have low income may be perceived to be at higher risk for HIV by health care providers and, therefore, may be more likely to be tested. Additional research is warranted to further investigate this issue.

Separate written consent for HIV testing among nonpregnant adults is another well recognized barrier. Twenty states had laws or regulations that required separate written consent when CDC released its 2006 HIV testing recommendation (17). By 2008, 11 states, including California and Maryland, enacted new legislation or regulations to remove this barrier (17). Our findings that in 2009, the District of Columbia, Maryland, and California had the highest reports of HIV testing among all the states may be partially due to the absence of requirements for separate consent in these states. Removal of requirements for a separate written consent for HIV testing in every state would likely promote HIV testing and reduce variations among states.

Kaposi sarcoma, non-Hodgkin lymphoma, and cervical cancer occur in excess in persons living with HIV (5). These cancers have been deemed AIDS-related conditions (5). Previous studies also indicate that HIV-infected persons are at higher risk than the general population for several non-AIDS–related cancers, such as anal cancer, Hodgkin disease, liver cancer, lung cancer, melanoma, and colorectal cancer (21,22). Besides the advent of highly active antiretroviral therapy, which has prolonged survival of HIV-infected patients, other factors that may account for increased cancer risk are immunosuppression (23,24), co-infection with an oncogenic virus (25–27), and unhealthy behaviors such as smoking (28). Awareness of HIV status is important at the time of cancer diagnosis or cancer recurrence to avoid treatment-related complications from co-infection with HIV, drug interactions, potential effects of chemotherapy on the CD4 cell count, and HIV viral load (14). Early identification of patients with HIV infection would allow early implementation of HIV infection prevention and control measures to reduce hospital-acquired infections among patients and health care providers (29,30). Coordination of anticancer and anti-HIV therapies could improve treatment outcomes among HIV-infected adults newly diagnosed with cancer or with cancer recurrence. Additionally, by knowing patients’ HIV status, health care providers can identify cancer survivors who are at elevated risk for a secondary cancer and educate them about the need for cancer screenings.

Our study has several limitations. First, BRFSS data were self-reported and subject to recall bias. Recall bias may be responsible for the slightly higher prevalence of certain cancers when compared with a study that used cancer registry data (8). HIV testing history data were self-reported and may be subject to social desirability or recall biases. Opt-out testing may also contribute to recall biases. For instance, patients may not have explicitly consented, so they may have been unaware when they were tested. Because of the limitations of recall bias and self-report, this study may seriously underestimate actual HIV testing among cancer survivors. Second, we are not able to determine whether the testing occurred before, at the time of, or after their cancer diagnosis. Third, because the 2009 BRFSS sampled only respondents with a landline telephone, persons without a landline telephone were excluded from this study. Fourth, BRFSS is limited to noninstitutionalized US citizens; cancer survivors who lived in nursing homes, long-term care facilities, or hospice were excluded. Last, the low response rate of 52.2% in the 2009 survey may limit the generalizability of the results to all cancer survivors living in the United States.

In conclusion, in 2009 about 60% of US cancer survivors under age 65 had never had an HIV test. The proportion of HIV testing varied substantially by demographic and health-related factors and by state. Increasing the proportion of new HIV infections diagnosed before progression to AIDS is a Healthy People 2020 (www.healthypeople.gov/2020/topicsobjectives2020/default) objective. One of the targets in the National HIV/AIDS Strategy is to increase the percentage of people who are living with HIV and know their serostatus from 79% to 90% by 2015. (13) Knowing the HIV serostatus at the time of cancer diagnosis or cancer recurrence is prerequisite to coordinating HIV and cancer treatments and improving treatment outcomes. Thus, our study points to the need for public health interventions to promote HIV testing among cancer survivors, especially among the demographic subgroups and states with a low proportion of HIV testing. With HIV testing, health care providers also can identify cancer survivors who are at elevated risk for a secondary cancer and educate them about the need for cancer screenings.