The current study examined non-adherence to HIV primary care visits among MSM in care at Fenway Health (FH). During the year following study assessment, almost one-third of participants missed at least one appointment without cancelling. This rate is consistent with prior work showing that appointment non-adherence is common in HIV-infected groups (Giordano et al., 2003
; Israelski et al., 2001
; Mellins et al., 2003
; Mugavero et al., 2009a
; Sohler et al., 2007
). This study is also the first to extend findings to HIV-infected MSM specifically.
Controlling for college degrees, non-adherent MSM had higher prevalence of income falling at or below $20,000 and lower prevalence of private/HMO insurance coverage relative to adherent MSM. These findings are not surprising, given that socioeconomic factors are key predictors of missed appointments (Israelski et al., 2001
; Mugavero et al., 2009a
; Neal et al., 2001
; Palacio et al., 1999
; Trenouth & Hough, 1991
; Van der Meer & Loock, 2008
). However, HIV appointment non-adherence persists even in programs that reduce financial barriers to care (Giordano et al., 2007
; Kissinger et al., 1995
; Sohler et al., 2007
), highlighting the need to identify additional risk factors in the context of socioeconomic disadvantage.
Among demographic factors, younger age and racial/ ethnic minority status (African American and Hispanic/ Latino versus non-Hispanic White) independently predicted elevated risk for appointment non-adherence. These findings support that the risks associated with age and racial/ethnic minority status cannot be solely explained by socioeconomic factors. The association of younger age with appointment non-adherence has been reported in prior work (Catz et al., 1999
; Israelski et al., 2001
); researchers have speculated that younger patients may be less adherent due to irregular schedules or lower perceived susceptibility to sickness (Catz et al., 1999
), although these hypotheses have not been tested. Determinants of racial/ethnic group differences in appointment non-adherence also merit more attention. Mugavero et al. (2009a)
reported that higher rates of appointment non-adherence among African Americans relative to non-Hispanic Whites partially explained racial disparities in rates of virologic failure. Other work has suggested that elevated rates of disease progression among African Americans and Hispanics/ Latinos reflect both greater delays in initiation of ART and lower treatment adherence (Giordano et al., 2010
). As African Americans and Hispanic/Latinos are disproportionately infected with HIV and underrepresented in clinical trials (CDCP, 2008
; Gifford et al., 2002
), strategies to increase access and engage patients in treatment early on may be critical to reducing disparities in HIV outcomes.
The current results further substantiate that depression may be a key factor in care disengagement (Cashman et al., 2004
; Weinger et al., 2005
; Weiser et al., 2006
), independent of socioeconomic status. MSM who screened positive for clinical depression were approximately two times as likely to be non-adherent than those with negative screens. Depression assessment and management is part of current standard care guidelines at FH. However, it is plausible that hopelessness and self-isolation may lead MSM with depression to disengage from care. Missed appointments, in turn, decrease opportunities for providers to detect and treat depression, creating a cycle with downstream effects on HIV health. Prior work has shown that cognitive-behavioral therapy can be adapted to improve both depression and medication adherence in HIV (Safren et al., 2009
). Current findings show that this dual target approach may be relevant to a range of self-care behaviors including care engagement.
In comparison to screens for depression, positive PTSD screens were more prevalent in the non-adherent group relative to the adherent group but did not independently predict risk for non-adherence in the multivariable model. This study was the first to test this relationship, and more work is needed to understand how PTSD may impact barriers to HIV care. For instance, it is possible that some symptoms (e.g., hypervigilance) may preserve health behaviors whereas others (e.g., withdrawal) may increase risk. The importance of this work is highlighted by the high prevalence of trauma history and PTSD among HIV-infected MSM; in the current sample, 41% of non-adherent MSM and 32% of adherent MSM screened positive for this disorder.
With regard to patient perceptions, low appointment expectancy predicted higher risk for non-adherence. This finding extends prior work examining perceptions of care (Beach et al., 2006
; Fischer et al., 2009
). Results suggest that education about the purpose of HIV care may hold benefit as a target for change. Findings also showed that more than half of the sample endorsed shame about being HIV-infected. However, the overwhelming majority felt the clinic staff understood the challenges of living with HIV, and neither factor independently predicted adherence status. Importantly, while MSM were highly likely to endorse that they would be able to keep all scheduled appointments during the next year, this belief was only marginally related to subsequent risk for non-adherence. Intervening to enhance problem-solving skills may increase care engagement for patients whose perceived health efficacy does not translate to actual behaviors.
Current findings should be interpreted with attention to study limitations. This study focused on MSM in care at FH, and it is unknown whether results would generalize to other groups or to MSM in other types of care settings. Also, since recruitment was conducted on-site at FH during a 3.5-year recruitment window, more frequent visit attendance might have increased the likelihood of being approached for recruitment, leading to possible underestimation of non-adherence in this study sample. Within the current sample, patient perceptions were such that most MSM endorsed good health, positive attitudes about staff, and strong sense of appointment expectancy and self-efficacy. This relative lack of variability, as well as limited psychometric data supporting these items may have limited our ability to detect actual risks associated with these perceptions.
For the current analysis, hypothesized risk factors were selected based on prior research but do not represent an exhaustive list. Other suggested influences such as coinfection with hepatitis C (Giordano et al., 2003
) or practical barriers such as lack of transportation to clinic visits were not addressed. Furthermore, we did not collect data on patient-reported reasons for non-adherence, and some missed appointments may have reflected specific circumstances such as inpatient hospitalizations of which the clinic was not advised. It is possible that risk factors for appointment non-adherence may vary among subsets of patients such as those with or without serious health problems or cognitive or functional impairment related to HIV and comorbid conditions. These hypotheses should be tested in future work identifying risk factors for appointment adherence and subsequent HIV outcomes.
In sum, findings indicate that missing HIV medical appointments is prevalent among HIV-infected MSM, even in settings where clinic staff members are perceived as supportive and sexual minority stigmatization is minimized. Socioeconomic barriers to care need to be addressed but represent only a portion of risk for care disengagement. MSM who are younger or of racial/ethnic minority status are at greater risk for missing appointments, independent of socioeconomic factors. Depression and low appointment expectancy and self-efficacy represent risks that may be amenable to intervention. The current findings help identify patients who may benefit from more intensive assessment and intervention to reduce depression symptoms and problem-solve healthcare barriers.