This study revealed that HIV-positive smokers reporting experience of physician-delivered smoking status assessment during any visit to a hospital or clinic in the past 12 months were over three times more likely to be in the contemplation or preparation stage of readiness to quit smoking. Outside of gender, lifetime history of injecting drug use showed the most robust association with current smoking status, followed by 30-day alcohol consumption, lack of formal education, and HIV symptom burden. Thus, optimal targeting of high-risk groups for smoking cessation efforts within the HIV-positive community would focus especially on males and recovering or active IDUs, incorporating the physician as a viable conduit for smoking cessation interventions.
As expected, current smoking was highly prevalent among PLWHA in our study, particularly among male participants. Though our data reveal a current smoking prevalence among HIV-positive women roughly comparable to the national rate for women in the region (16.0%), the prevalence of smoking among HIV-positive men was over twice the corresponding national rate (34.8%) . Identified predictors of current smoking status were generally consistent with results of previous studies conducted within HIV-positive populations in Western, developed settings [10, 33] and in Africa .
In terms of motivational readiness to quit, the proportion of participants still in the precontemplative stage of behaviour change was high relative to previous findings on HIV-positive smokers in developed countries [9, 33, 34]. Velicer et al.  reported typical proportions for readiness to quit as 40% in precontemplation, 40% in contemplation, and 20% in preparation among smokers in the general population. In our sample, in contrast, roughly two-thirds of smokers were precontemplative about quitting. This finding underscores the prime importance of exploring effective ways to build motivation for cessation in this important population of smokers.
Our study is the first to report on formal education as a potential barrier to quitting readiness among HIV-positive smokers. Given the strong positive linear correlation between education and the quit ratio (proportion of ever smokers who had quit smoking) demonstrated by Wetter et al. , it may be that those formally educated individuals still smoking at the time of our survey were less likely to be thinking seriously of quitting because they had already made a conscious decision to continue smoking, fully cognizant of the attendant risks. Further research would be warranted to elucidate the observed association between education level and readiness to quit smoking among PLWHA and in the developing country context.
Similarly, though data from developed countries suggest that smokers attempting to quit may be less successful and more susceptible to relapse with the presence of another smoker in the household [55, 56], our data indicate that cohabitation with at least one other smoker may contribute to greater quitting readiness in some populations. Indeed, social support networks comprise an important aspect of successful smoking cessation . Although this study did not specifically address the readiness to quit of smoking family members, we can reasonably assume that some discussion of quitting is likely to take place where more than one smoker inhabits a household. Future research might fruitfully explore cessation interventions designed to exploit existing support systems through soliciting the involvement of other household smokers.
We further found that history of at least one smoking cessation attempt was associated with greater quitting readiness. This finding is consistent with the TTM, by which the probability of successful change increases with the number of change attempts , and is also supported by results from previous studies of HIV-positive smokers . For the promotion of active progression through the stages of change toward cessation, this result recommends the importance of encouraging quit attempts even among those smokers demonstrating lower levels of motivational readiness.
Finally, our research opens up a new and previously unexplored space for the role and effectiveness of basic smoking status screening administered by physicians treating PLWHA. Because the HIV practice pattern typically consists of regular contact over an extended period of time, physicians treating PLWHA have enhanced opportunities to build rapport and trust and to better tailor the tone to the individual patient. Yet there is evidence that HIV care providers may actually be less likely to recognize current smoking in their patients [51, 58]. This points to a critical missed opportunity.
Notably, we found significantly higher motivational readiness for cessation among individuals reporting smoking status assessment by a physician, though not among individuals reporting such assessment by a paramedic or nurse. This implies that, in the context of this study setting, the words of a physician may be imbued with greater meaning or weight when it comes to issues of health behaviour change. Specifically, paramedics and nurses might be perceived more as peers, as they are frequently available for consultation in the NGO clinics. In contrast, interactions with physicians typically occur in more formal settings, where they may more likely be perceived as expert service providers and their words afforded greater value.
Drawing smokers' attention to their smoking behaviour is a form of feedback to the patient and might in itself be sufficient to effect behaviour change when administered by a trusted and valued health professional. While more intensive, multifaceted counselling interventions may be the ideal, they are not always feasible in busy primary care or clinical settings, particularly in low-resource environments, where health care providers' time is at a premium. From a public health perspective, even given a small effect size for facilitation of smoking cessation through brief physician intervention, the net effect can still be substantial, provided large numbers of physicians follow such a practice systematically .
These results should be considered within the context of several study limitations. First, although this study surveyed a large group of participants from multiple NGO outreach networks across the Kathmandu Valley, findings are specifically representative of PLWHA falling within the network of our partner NGOs. Importantly, our sample contained only two non-heterosexual participants, though men who have sex with men represent an important risk group in the HIV epidemic nationally. Second, this study relies on self-report of all measures, leaving room for several types of bias. Because information was collected through face-to-face interviews, a social desirability bias may have been introduced, though a confidential and sensitive method of survey administration was designed to limit this tendency. Although we did not verify smoking status biochemically for those who reported being nonsmokers, there is empirical evidence that self-report is a sufficient, reliable, and valid means to assess smoking status .
Third, measurement of smoking status assessment experiences in this study only covered the existence of such an interaction and did not address potential variability in the specific content or quality of the discussions on smoking taking place between patients and their physicians. Finally, though previous research has established that stage of behaviour change effectively predicts both smoking cessation attempts and actual cessation [29, 30], the tangible outcome variable of this study is readiness to quit rather than eventual quitting success. Further research will be needed to gain a better understanding of the different nuances to effective delivery of smoking status assessments in HIV care settings as well as the special needs HIV-positive smokers may have in actually enacting and maintaining smoking cessation.