The practices and perceptions of black, primary care physicians who treat predominantly adult, black patients in the US were evaluated in a survey-based research study to identify key barriers and facilitators of HIV testing. There were 502 surveys included in the analysis, out of >34,000 invitations. Response rates were 0.8% [101/12665] for Panel 1, 5% [46/919] for the post-conference emails sent to those on the NMA master file, and 0.6% [124/20463] for Panel 2 (0.6% [44/7261] among NMA members and 0.6% [80/13202] among non-NMA members). There were 231 intercept surveys completed at the NMA Annual Convention. In all, 321 (64%) of the surveys were completed by NMA members. Characteristics of the respondents and their practices are described in Table . Physician-reported demographics of patient base: 56% black race; 67% women; 87% ≥18 years of age; 24% on Medicaid (13% uninsured); 31% with low SES (defined as ‘poor’ on the survey); 6% with HIV/AIDS.
Physician and practice characteristics of respondents who met inclusion criteria (physician reported)
More physicians perceived the HIV/AIDS epidemic to be a crisis in the black population than in the general US population (55% vs. 14% respectively), and physicians generally severely over-estimated local county and state HIV prevalence to be 13-14% on average (the highest actual local prevalence is approximately 3% in Washington DC [17
] and nationwide it is <1% [18
]) (Table ). If physicians believe that up to 14% of the local population has HIV, then HIV testing should be as common and frequent as testing for hypertension (30% in adults ≥18 years old, 2005–2008) [19
] or diabetes (8% in all patients, 2010) [20
], which occur during nearly every doctor’s visit. Moreover, 80% of physicians responded that HIV testing should be routine and 20% reported that it should be risked based, yet routine HIV testing (55%) was reported as less of a primary reason for testing than risk-based testing (70-89%) (Table ). Furthermore, physicians reported testing only about one-third (34%) of their patients for HIV in the past year, and much of the testing was risk-based instead of routine. The gap between the percentage of physicians who say that testing should be routine (80%) and those who actually reported testing patients for HIV in the past year (34%) indicates that a disconnect exists which prevents physicians from taking action on their beliefs. Physicians seem to understand that they should test more routinely, but translating that into practice is difficult. Physicians were more likely to test at-risk patients, presumably because it is easier for physicians to point to documented risk factors as justification for recommending a test. In the context of a patient with no discernible risk factors (other than sexual activity), however, physicians may fear appearing judgmental of the patient’s behavior and avoid recommending an HIV test.
HIV testing perceptions and practices of black, primary care physicians (physician reported)
Physicians in this sample reported that among patients who were tested for HIV, 67% was due to their or another physician’s recommendation. Physician recommendation for an HIV test was driven most significantly by key factors such as specialty, comfort raising the issue with patients perceived not to be at risk for HIV, age of physician, and percentage of black, low-SES, and Medicaid patients in the practice (Figure ). Significant characteristics (P<0.05) that distinguished the “more” and “less” routine testers are listed in Table . Ninety-five percent of the “more routine” testers believed that HIV testing should be routine for all patients compared to only 64% of the “less routine” testers, but reported actual testing rates fell far short of these values.
Figure 1 Percentage of patients reportedly tested for HIV according to key drivers of physician recommendations. aSignificantly higher (P<0.05) than other group(s) in category; Overall testing rate was 34%; “High” physician comfort level (more ...)
Key characteristics of “more routine”a versus “less routine”b testers for HIV in the past yearc
Specialty was the most impactful factor affecting HIV testing rates, with OB/GYNs leading the way with 52% of patients tested within the last year (Figure ). It may be easier to raise the issue of HIV testing during a routine gynecological exam because testing for HIV can be placed in the context of general screening for sexually transmitted infections since sexual activity is assumed, and women may want to know their HIV status if they are planning to get pregnant or would like to prevent vertical transmission to the child, which align with previous observations of testing practices among OB/GYNs [21
]. Moreover, guidelines from the CDC, American College of Obstetrics and Gynecologists (ACOG), and many states recommend opt-out HIV testing for pregnant women, and some states even require it [1
]. In contrast to OB/GYNs, physicians in other specialties appear to find it harder to raise the issue of HIV testing in relatively comfortable contexts that would avoid any stigma or perceived judgments associated with HIV; in-office materials such as brochures or posters that educate patients on the importance of getting tested may enable more comfortable physician-patient interactions in these scenarios. While it is true that OB/GYNs see only female patients, the percentage of female patients in physician practices was not a driver of HIV testing due to physician recommendation. Also of note is that although female physicians reported testing a greater percentage of patients for HIV than their male counterparts (38% vs. 29%, P
<0.05), this may have been due in part to the over-representation of women as OB/GYNs (65%, n=84), and not just a gender-specific predilection. Finally, it should be acknowledged that OB and GYN data were pooled in this analysis, and the percentage of reported patients tested by OBs alone is likely higher than the rate reported for OB/GYNs for the reasons mentioned above related to mandatory testing during pregnancy.
The physician’s comfort level discussing HIV and age was also a predictor of testing frequency (Figure ) (see Additional file 2
). Physicians who rated themselves as highly comfortable with raising the issue of HIV with patients who they perceive to be not
at risk of contracting the virus tested more frequently than those who did not report being comfortable. Having conversations about HIV routinely rather than only with patients who exhibit risk factors such as multiple sex partners, injection drug use, previous incarceration, or suspected sex work should certainly improve testing rates. Physicians <40 years old tested patients more frequently than those ≥40 years old (39% vs. 32%, P
<0.05). Because they received medical training during the HIV era (post-1981), younger physicians may have greater education, social comfort, and tolerance with the risk factors associated with HIV [25
]. Previous studies have concluded that female and younger physicians obtained sexual histories to a greater extent in primary care settings, and physician discomfort (particularly with opposite gender, very young or old, and homosexual patients) reduced the likelihood of physicians obtaining sexual histories [26
]. This suggests that physicians, especially older, male physicians, may benefit from more practical clinical strategies and techniques to obtain a sexual history from all patients uniformly.
Physicians with more black (≥75%), low-SES (>45% ‘poor’), and Medicaid (≥30% enrolled) patients in their practice were more likely to test for HIV (Figure ). These groups are disproportionately affected by HIV and may be perceived to be at higher risk based on traditional risk-based testing recommendations provided by previous public-health and medical organizations. Physicians with more black patients believed that HIV was more of a crisis within the black community than the general US population, and physicians with more black and low-SES patients were more likely to over-estimate HIV prevalence at their county and state levels. Patients on Medicaid receive HIV testing coverage and physicians can use that as a reason to recommend a test.
Physicians with more HIV/AIDS (≥7%) patients tested more than those with fewer HIV/AIDS (≤1%) patients (39% vs. 32%, P<0.05), as would be expected.
Among reported barriers to recommending an HIV test, three of the five most common barriers were associated with stigma, patient disclosure, and perceived value judgments; the remaining two reflected competing priorities on physicians’ time (Table ). Constraints such as shortage of resources or financial considerations (for the physician or patient) were ranked ninth and tenth, respectively. This indicates that physicians’ fear of an uncomfortable interaction with a patient may be a significant impediment to routine testing practices, and proper training (from medical school or continuing education) and sufficient time are important for physicians to routinely initiate the conversations that produce quality sexual histories [26
Most commonly reported barriers that limit physicians from recommending HIV testinga
Attitudinal barriers such as physicians’ perceptions of their patients’ reactions to a testing recommendation and reasons for possibly refusing an HIV test differed according to patient characteristics such as race and SES. Physicians reported that black patients were more likely than non-black patients to respond to an HIV testing recommendation with denial (52% vs. 45%), offense (24% vs. 18%), and anger (19% vs. 11%), and give reasons such as “I’m not gay” (55% vs. 47%), “I don’t feel sick” (42% vs. 35%), and “I don’t like needles/giving blood” (37% vs. 31%) for refusing an HIV test (all are P<0.05). This may reflect a belief among many physicians that patients may not accurately assess or convey their own risks for acquiring HIV during a clinical encounter. Given that the above responses could be the reasons why many patients refuse HIV testing in clinical settings, this represents an opportunity for developing physician training initiatives on constructive approaches to communicating the importance of testing to patients, particularly given that “physician recommendation” was the most commonly reported driver of HIV testing within this sample. Compared to physicians with fewer low-SES patients, those with more low-SES patients ranked ‘Patient would perceive test recommendation as accusatory or judgmental’ (51% vs. 65%) and ‘Patient wouldn’t want to be identified as HIV-positive’ (41% vs. 59%) as less of a barrier to HIV testing and ‘Physicians lack resources to treat patients with HIV/AIDS’ (27% vs. 17%) and ‘Physicians are ill-equipped to deal with emotional reaction’ (29% vs. 17%) as more of a barrier to testing (all are P<0.05)
Only 25% and 50% of physicians reported that they were tested for HIV themselves within the past 1 and 5 years, respectively, and 8% reported having never been tested for HIV themselves (mostly males [14% vs. 4% females, P<0.05] and physicians ≥55 years old [17% vs. 6% <55 years old, P<0.05]) (Table ). Physicians who were less likely to get tested for HIV themselves were less likely to recommend testing to their patients (Table ). There may be a correlation between personal and professional HIV testing practices, and physicians who do not view testing as important in their own lives may project that onto their patients.
Many of the resources that physicians reported would be important aids to help them test more frequently for HIV, such as in-office patient informational materials, increased media attention, additional education and training on HIV testing, government mandates requiring routine testing, and accurate pre-packed tests, are already available through the CDC, HIV/AIDS societies, and diagnostics manufacturers, but physician awareness and/or access may be limited (Figure ). The fact that physicians want more in-office informational materials and media attention reflects a desire to encourage patients themselves to raise the question of HIV testing or to increase patients’ comfort levels with the prospect of being tested for HIV. It has been reported that patient-focused rather than physician-focused materials would help primary care physicians more to improve HIV testing practices because physicians felt that resources facilitating communication with patients about HIV are more important than materials that enable physicians to be more proficient in treating HIV patients [29
]. Government mandates for routine testing would also lessen the burden on physicians to bring up the issue of testing by providing them with a non-judgmental reason for recommending an HIV test. In fact, many states still have statutes that are inconsistent with CDC guidelines regarding routine, opt-out testing for HIV in all healthcare settings [30
]. However, physicians should not place the responsibility of communicating about HIV testing on the patients alone. Those black, MSM, and low-SES patients who need testing the most are often already burdened with social oppressions and these patients may look to physicians to be a source of relief and understanding [31
Figure 2 Factors that would help physicians recommend more HIV testing (percentage who selected factor as important)a. aBased on the survey question: What would help you do more testing? (Factors were selected from a pre-defined list and multiple choices were (more ...)
This study has several limitations. Firstly, there is the potential for bias associated with volunteering to take the survey, recall, self-reporting, and overestimating or skewing certain responses to exhibit favorable, positive practices. Secondly, the small number of completed surveys (502) relative to the number of invitations distributed (>34,000), the high percentage of respondents who were NMA members, and incomplete geographical representation may mean that the data are not representative of all black, primary care physicians in the US. Thirdly, responses were physician-reported with no patient responses with which to compare, and no overall response rate was reportable due to an unknown response rate for the intercept surveys. Fourthly, some subgroup data may be confounded since OB and GYN data were combined, and OBs differ from all other specialties given their HIV testing requirements. Finally, there is little comparative data on the HIV testing attitudes and behaviors of physicians of other races and ethnicities to make meaningful comparisons using this study’s results. Despite these limitations, this is one of the first studies to examine the reported practices of black physicians with regard to HIV testing.