Pietromonaco and Barrett posit “A central tenet of attachment theory is that people develop mental representations, or internal working models that consist of expectations about the self, significant others and the relationship between the two. ” (Pietromonaco and Barrett, 2000, 4:2, p156). I illustrate how this internal working model is developed via the relationship between infant and primary caregiver, demonstrating that maternal deprivation can create a ‘faulty’ internal working model, which may lead to psychopathology in later life.
I also demonstrate how these internal working models influence our reactions to loss and bereavement in adulthood and their potential impact on the counselling relationship. In addition, I explore the multi-layered losses experienced by HIV+ gay men and finally draw some conclusions. Freud’s view on the infant’s attachment to its mother was quite simple “the reason why the infant in arms wants to perceive the presence of its mother is only because it already knows by experience that she satisfies all its needs without delay.” (Freud, 1924, p188 cited in Eysenck, 2005, p103).
In contrast, behaviourists believed that feeding played a central role in the development of attachment. (Pendry, 1998; Eysenck, 2005). These theories were termed ‘secondary-drive theories’. In 1980, Bowlby recalled “this [secondary drive] theory did not seem to me to fit the facts…. but, if the secondary dependency was inadequate, what was the alternative? ” (Bowlby,1980, p650 cited in Cassidy and Shaver, 1999, p3).
Bowlby’s theory was influenced by his paper “Forty Four Juvenile Thieves”, where he concluded a correlation exists between maternal deprivation in infancy, leading to affectionless psychopathology and subsequent criminal behaviour in adolescents. (Bowlby, 1944, 25, p19-52). This led to him researching the impact of loss on children displaced through war and institutionalisation, resulting in ‘Maternal Care and Mental Health’ (1952), where he confirms a link between ‘environmental trauma’ and resultant disturbances in child development.
As a result of this research, Bowlby concluded “it is psychological deprivation rather than the economic, nutritional or medical deprivation that is the cause of troubled children. ” (Bowlby in Coates, 2004, 52, p577). He was further influenced by Lorenz who found that goslings would follow and ‘attach’ themselves to the first moving object they saw. This following of the first moving object was called ‘imprinting’. (Lorenz, 1937 cited in Kaplan, 1998, p124).
Clearly babies cannot follow at will – to compensate for this, ”Bowlby noted that ‘imprinting’ manifested itself as a spectacularly more complex phenomenon in primates, including man, which he later labelled ‘attachment’. ” (Hoover, 2004, 11:1, p58-60). He also embraced the work of Harlow and Zimmerman who worked with infant rhesus monkeys demonstrating that not only did the need for attachment give them security, but that this need took priority over their need for food. (Harlow and Zimmerman 1959 cited in Green and Scholes, 2003, p9).
Dissatisfied with traditional theories, Bowlby embraced new understandings through discussion with colleagues from such fields as developmental psychology, ethology, control systems theory and cognitive science, leading him to formulate his theory that the mechanisms underlying the infant’s tie to the mother originally emerged as a result of evolutionary and biological pressures. (Cassidy and Shaver, 1999; Green and Scholes, 2003).
Defining his attachment theory as “a way of conceptualising the propensity of human beings to make strong affectional bonds to particular others.” (Bowlby, 1979 cited in Green and Scholes, 2003, p7), he posited “that it is our affectional bonds to attachment figures that engage us in our most intense emotions. ” and that “this occurs during their formation (we call that ‘falling in love), in their maintenance (which we describe as ‘loving) and in their loss (which we know as ‘grieving’), (Green and Scholes, 2003, p8), thereby replacing the secondary-drive theory with a model emphasising the role relationships play in attachment and loss. (Waters, Crowell, Elliott et all, 2002, 4, p230-242).
Disregarding what he called Freud's ‘cupboard love’ theory of attachment, he believed instead that a child is born ‘biologically pre-disposed’ to become attached to its mother, claiming this bond has two essential features: the biological function of securing protection for survival and the physiological and psychological need for security. (Green and Scholes, 2003; Schaffer, 2004). Sonkin (2005) describes four features to this bond: secure base, separation protest, safe haven and proximity maintenance.
The concept of a secure base is fundamental to attachment theory and is used to describe a dependable attachment to a primary caregiver. This secure base is established by providing consistent levels of safety, responsiveness and emotional comfort from within which the infant can explore his or her external and internal worlds and to which they can return, thus providing a sense of security. Separation protest is exhibited as a sign of the distress experienced upon separation from an attachment figure, who may also be used as a safe haven to turn to for comfort in times of distress.
When safety is threatened, infants attract the attention of their primary caregiver through crying or screaming. Maintaining attention and interest, e. g. vocalising and smiling, and seeking or maintaining proximity, e. g. following or clinging, all serve to promote the safety provided by the secure base (providing of course that parents respond appropriately). (Holmes, 1993; Cassidy and Shaver 1999; Becket, 2002; Green and Scholes, 2003).
Proximity seeking is a two way process, for example child seeking parent or parent seeking child. (Weiss in Murray-Parkes, Stevenson-Hinde and Marris, 1991; Becket, 2004; Sonkin, 2007). Bowlby also recognised ‘unwilling’ separation caused by parents who were phsycially present but not able to respond, or who deprived infants of love or ill-treated them, left them with a sense of immense deprivation and that this unwilling separation and resultant loss leads to deep emotional distress. (Green and Scholes, 2003).
At a recent conference, the Centre for Attachment based Psychoanalytical Psychotherapy (CAPP) asserts “Early interactions with significant others in which there are fundamental failures of empathy, attunement, recognition and regulation of emotional states, have been shown to cause the global breakdown of any coherent attachment strategy, thus engendering fears of disintegration and threatening psychic survival. In the face of such experiences, powerful dissociative defences may be employed, encapsulating overwhelming feelings of fear, rage and shame. ” (CAPP, 2007).
Together with Robertson and Rosenbluth, Bowlby demonstrated that even brief separation from the mother has profound emotional effects on the infant. Their research highlighted a three stage behavioural response to this separation: protest - related to separation anxiety; despair - related to grief and mourning; and detachment - related to defences. (Robertson, Rosenbluth, Bowlby, 1952 in Murray-Parkes, Stevenson-Hinde and Marris, 1991). Ainsworth, Blehar, Waters and Wall (1978) later established the inter-relatedness between attachment behaviour, maternal sensitivity and exploration in the child.
Under clinical settings, they sought to observe the effects of temporary separation from the mother, which was assessed via the ‘strange situation’ procedure. This study involved children between the ages of 12 to 18 months who experienced separation from their mother, introduction to an unfamiliar adult and finally reunion with their mother. Ainsworth et al reasoned that if attachment was strong, mother would be used as a secure base from which the infant could explore, thereby promoting self-reliance and autonomy. Upon separation, infants usually demonstrated separation anxiety.
Upon re-union, the mother’s maternal sensitivity and the child’s responses were observed, thus providing a link between Bowlby’s theory and its application to individual experience. The trust/mistrust in the infant’s ability to explore their world from the secure base is re-inforced by Erikson’s (1965) examination of early development and the child’s experiencing of the world as a place that is nurturing, reliable and trustworthy (or not). Influenced by Ainsworth’s previous work in Uganda, the ‘strange situation’ led to the classification of secure or insecure attachment styles in infants.
Insecure styles were further grouped into insecure/avoidant and insecure/resistant (ambivalent). (Pendry, 1998; Holmes, 2001; Eysenck, 2005). Main and Solomon later added a fourth attachment style - insecure/disorganised. (Main and Solomon, 1986 in Cassidy and Shaver, 1999, p290). Throughout all of these interactions, an ‘internal working model’ is developed, the cultivation of which relies on the dyadic patterns of relating between primary caregiver and infant (Bretherton, 1992, 28, p759-775), comprising the complex monitoring of internal states of primary caregiver and infant. (Waters, Crowell, Elliott et al, 2002, 4, p230-242).
According to Schore “These formative experiences are embedded in the developing attachment relationship – nature and nurture first come together in mother-infant psychobiological interactions. ” (Schore, 2001, 17, p26). Over time, this leads to the infant’s ability to self-monitor their emotions (affect regulation), but until such time, Bowlby posited the mother acts as the child’s ego and super-ego ”She orients him in space and time, provides his environment, permits the satisfaction of some impulses, restricts others.
She is his ego and his super-ego. ” (Bowlby, 1951, p53 cited in Bretherton, 1992, 28, p765). Bowlby concluded a healthy internal working model is “a working model of an attachment figure who is conceived as accessible, trustworthy and ready to help when called upon”, whilst a ‘faulty’ model is “a working model of an attachment figure to whom are attributed such characteristics as uncertain accessibility, unwilllingness to respond helpfully, or perhaps the likelihood of responding hostilely. ” (Bowlby, 1979, p141).
Ainsworth suggests that positive attachment is more than explicit behaviour “it is built into the nervous system, in the course and as a result of the infant’s experience of his transactions with the mother. ” (Ainsworth, 1967, p429), thus supporting Bowlby’s theory. Later descriptions of attachment styles describe secure attachment as “the development of the basic machinery to self-regulate affects later in life”, (Fonagy, Gergely and Jurist, 2002 cited in Sarkar and Adshead, 2006, 12, p297), whilst insecure attachment “prevents the development of a proper affect regulatory capacity.” (Sarkar and Adshead, 2006, 12, p297).
This is supported by Schore (2003) who alludes to developmental affective neuroscience to set out a framework for affect regulation and dysregulation. Based on research into the development of the infant brain, he reviews neuro-scientific evidence to confirm the infant’s relationship with the primary caregiver has a direct effect on the development of brain structures and pathways involved in both affect regulation and dysregulation.
The research and evidence suggests the internal working model begins as soon as the child is born and is the model upon which future relationships are formed. The quality of the primary caregiver’s response to infant distress provides the foundation upon which behavioural and cognitive strategies are developed, which in the longer term influence thoughts, feelings and behaviours in adult relationships. (Cardwell, Wadeley and Murphy, 2000; Pietromonaco and Barrett, 2000, 4:2, p155; Madigan, Moran and Pederson, 2006, 42:2, p293).
A healthy, secure attachment to the primary caregiver would therefore appear essential for a child’s social, emotional and intellectual development, whilst interruption to this attachment would appear to promote the premise of psychopathology in later life. Whilst some evidence exists to demonstrate internal working models can be modified by different environmental experiences, (Riggs, Vosvick and Stallings, 2007, 12:6, p922-936), the extent to which they can change remains in question.
Bowlby himself postulated “clinical evidence suggests that the necessary revisions of the model are not always easy to achieve. Usually they are completed but only slowly, often they are done imperfectly, and sometimes done not at all. ” (Bowlby, 1969, p83). Whilst change may be possible, the unconscious aspects of internal working models are deemed to be specifically resistant to such change. (Prior and Glaser, 2006). We can safely assume therefore, that in the majority of cases, internal working models tend to persist for life.
I concur with Rutter’s criticism of Bowlby's concept of ‘monotropy’, i.e. Bowlby’s belief that babies develop one primary attachment, usually the mother, (Rutter, 1981 cited in Lucas, 2007, 13, p156 and in Eysenck 2005), accepting instead that infants form multiple attachments. This is supported by a study by Shaffer and Emmerson (1964) who concluded infants form a ‘hierarchy’ of attachments, often with the mother as the primary attachment figure, although nearly a third of children observed highlighted the father as the primary attachment figure. (Schaffer and Emmerson, 1964 in Cassidy and Shaver, 1999, p44-67).
Collins, Dunlop and Chrysler criticise Bowlby’s ‘lens’ in that it was “limited by his own cultural, historical and class position. Bowlby’s culturally biased assumptions and empiricist methods of inquiry concentrated on individualised detachment and loss as part of the normal course of mourning loss, which perpetuated the Western tradition of preserving the autonomous individual self as the normal goal of development. ” (Collins, Dunlop and Chrysler, 2002, p98), leading them to conclude Bowlby’s assumptions ignored other cultural practices (as did Ainsworths), with which I agree.
They also suggest Bowlby’s concept of maternal deprivation was perhaps exploited to get women to return to the home post World War II - “Characterised as a choice, this ‘homeward bound’ movement was supported by the various governments, whose maternalist and pronatalist ideology of the 1930s continued into the post-war period to provide a rationale for sending women home to reproduce … maternalism and the maternal deprivation hypothesis provided one conceptual framework for pronatal ideology as it intersected with the demands of governments and industrialists. ” (Collins, Dunlap and Chrysler, 2002, p102).
We must also remember that Bowlby’s observations “were based on children who had been separated from their primary caregivers during the Second World War” (Lemma, 2003 cited in Lucas, 2007, 13, p156), and that these procedures “were based on behaviours that occurred during stressful situations rather than under normal circumstances. ” (Lucas, 2007, 13, p156) [this latter criticism also applies to Ainsworth’s work].
Nonetheless, in highlighting the damaging effects of institutionalised care on young children, Bowlby’s strengths lie in drawing attention to the role attachment, attachment behaviour and attachment behavioural systems play in a child’s development and the subsequent potential consequences of disruption to the bond between infant and primary caregiver. I concur with Cassidy and Shaver’s (1999) criticisms of the strange situation in that there are too many unconsidered variables for a firm theory to be established at the time of Ainsworth’s writings, accepting their view that she did not consider the mood nor temperament of the child.
Nonetheless, Ainsworth et al have provided a tool with which to measure attachment styles in infants, which is still in use today. Later research by George, Kaplan and Main assesses adult internal models through the use of the Adult Attachment Interview. This classification of adult attachment styles promotes the idea of models extending into adulthood as a template for future relationships. (George, Kaplan and Main 1985 cited in Pendry, 1998).
Hazan and Shaver continued this line of research identifying patterns of attachment behavior in adult romantic relationships, concluding the same four attachment styles identified in infancy remain true for adult relationships. (Hazan and Shaver 1987 in Cassidy and Shaver, 1999, p355-377). Although theoretically rooted in the same innate system, adult romantic attachment styles differ from parent-child bonds to include reciprocity of attachment and caregiving, as well as sexual mating. (Hazan and Zeifman, 1999 in Cassidy and Shaver, 1999, p336-354).
The literature on bereavement has become inseparable from Bowlby’s theory of attachment and, following from this, the way in which people react to the loss of this attachment. On reflecting on losses in adult life, Weinstein (2008) observes Bowlby’s persistence of formative attachments and how the pattern of protest, despair and detachment that follows a baby’s separation from its primary caregiver is re-activated and presented in full force in adult loss.
Weinstein writes “The ability of the adult to cope with attachment in intimate relationships to negotiate independence, dependency and inter-dependency; and to manage loss is all about how successfully they coped with separation as an infant. As a baby they had to retain their sense of their mother even in her absence and now as adults, as part of the mourning process, they strengthen their own identity with the support of the internalised object. ” (Weinstein, 2008, p34).
According to Murray-Parkes (1996), the intensity and duration of this grief is relative to what is lost and the grief process is an emotional response to this loss. Murray-Parkes joined Bowlby at the Tavistock Centre in 1962. Together they presented a paper linking the protests of separation highlighted by Robertson, Rosenbluth and Bowlby (1952) in young children separated from their mothers, to grief in adults. (Bowlby and Murray-Parkes, 1970 in Murray-Parkes, Stevenson-Hinde and Marris, 1991, p20).
Around the same time, Murray-Parkes visited Kubler-Ross who was conducting her own research into death and dying. This work was later published in ‘On Death and Dying’ (1969) which examines the process of coming to terms with terminal illness or grief in five stages: denial; anger; bargaining; depression and acceptance. Murray-Parkes later produced a four-phase grief model consisting: shock or numbness; yearning and pining; disorganisation and despair; and re-organisation.
In contrast to the passive staged/phased approaches by Kubler-Ross and Murray-Parkes, and perhaps more in line with Freud’s concept of having ‘to do grief work’, Worden developed a four-staged, task-based grief model: “to accept the reality of the loss; to work through the pain of grief; to adjust to an environment in which the deceased is missing; and to emotionally relocate the deceased and move on with life. ” (Worden, 2003). All three models are deemed to be therapeutically useful in that they recognise grief as a process and provide a framework of descriptors for ‘normalising’ grief reactions.
That said, they are clearly prescriptive and caution should be exercised in taking any of these prescriptive stages, phases or tasks literally. It is equally important to recognise the uniqueness of individual responses to loss and to avoid prescribing where a client ‘ought’ to be in their grieving process. Since these models were never designed as a linear process, it is likewise important not to steer clients through these stages. This is supported by Schuchter and Zisook (1993), who assert “Grief is not a linear process with concrete boundaries but, rather, a composite of overlapping, fluid phases that vary from person to person.” (Schuchter and Zisook, 1993 in Stroebe, Stroebe and Hansson, 1993, p23).
I agree with Servaty-Seib’s observations “the stage/phase approaches emerged solely from a death-loss focus … Worden’s work was an important development in the understanding of the process of coping adaptively with bereavement as each task is clearly defined in an action-oriented manner. ” (Servaty-Seib, 2004, 26:2, p125). Stroebe and Schut’s dual process model brings together death-loss focus and task-based models. (Stroebe and Schut, 2001 cited in Servaty-Seib, 2004, 26:2, p125).
In my work at Positive East, I work with HIV+ gay men experiencing multi-layered loss. My philosophy is to build and maintain a therapeutic relationship within a safe, confidential, contained space where clients can explore their issues. The archetype ‘working towards a model of gay affirmative therapy’ (Davies and Neal, 1996, p24-40) provides me with a framework within which to explore gay culture and to apply an assenting approach to the work, which I believe promotes empathy and helps me to work in the best interests of the client.
Conducting my own assessments, I complete a full client history, genogram and timeline, which provides a comprehensive insight into clients attachments and losses. It is important to acknowledge the social context within which multi-layered loss takes place (e. g. heterosexism, homophobia, HIV-related stigma) as well as recognising that individual attachment styles may influence individual reactions to these losses and may also impact on the counselling relationship.
Losses experienced by HIV+ gay men include loss of identity, health, appearance, mobility, self-respect, career, financial security, relationships and intimacy. (Riggs, Vosvick and Stallings, 2007, 12:6, p922-936; Koopman, Gore-Felton, Marouf et al, 2000, 12:5, p663-672; Fernandez and Ruiz, 2006, p356). Corr, Nabe and Corr (1997) describe these losses as the cognitive, affective and behavioural responses to the impact of the loss. In identifying attachment styles in HIV+ adults, Riggs, Vosvick and Stallings (2007) found that 90% of gay and bisexual HIV+ adults recruited into their study demonstrated insecure attachment.
They suggest the diagnosis of HIV produces a strong trauma reaction, impacting on adult attachment style. In the same study, they found that HIV+ heterosexual adults were more likely to be secure, whereas gay and bisexual adults were more likely to be fearful, preoccupied, avoidant; or dismissing, respectively. This led them to conclude that gay and bisexual people must therefore contend with societal forces that their heterosexual counterparts do not.
They hypothesise “A diagnosis of HIV may be reminiscent of the coming out process, particularly with respect to concerns regarding stigma and disclosure, and thus may provoke similar fears about rejection by loved ones and society as a whole that contribute to greater attachment insecurity. ” (Riggs, Vosvick and Stallings, 2007, 12:6, p931). This is supported by Koopman, Gore-Felton, Marouf et al (2000) who cite attachment style as a contributing factor associated with the high levels of stress experienced by HIV+ individuals.
They comment “From this perspective, perceived stress is likely to be greater among [HIV+] persons having a highly anxious attachment style because their hypervigilance in interpersonal relationships leads to misinterpreting others’ behaviours as rejecting or critical of themselves. ” (Koopman, Gore-Felton, Marouf et al, 2000, 12:5, p670). This would suggest that HIV+ gay men with insecure attachment style may experience difficulties in developing and maintaining relationships, which, in turn, may impact on the therapeutic relationship.
Additionally, according to Kelly, Murphy, Bahr et al “Dependable and supportive attachments play a crucial role in adjusting to HIV infection. Lack of such attachments and social support has been shown to be a significant predictor of emotional stress among HIV+ adults. ” (Kelly, Murphy, Bahr et al, 1993, 12:3, p215-219). This has significant implications for the psychological well being of HIV+ gay men whom, considering their perceived attachment difficulties, may experience difficulties in forming such supportive relationships.
In examining the suitability of the common grief models when working with this client group, I accept Copp’s criticism of the Kubler-Ross model for its focus on psychosocial dynamics “to the exclusion of physical, and to a lesser extent, spiritual dimensions. ” (Copp, 1998, 28:2, p383). I also agree with Knapp’s criticisms of the staged/phased grief models espoused by Kubler-Ross and Murray-Parkes. Knapp observes “while both of these models may be applicable to those experiencing a singular loss, neither model takes into consideration the multiplicity of losses thrust upon the seropositive gay male population.
These men experience overlapping losses, resulting in them being at differing stages with respect to different losses. ” (Knapp, 2000, 6:2, p143). Knapp offers a similar criticism of the Worden model in that “task models fail to account for the continuity of loss in the lives of seropositive gay men. ” (Knapp, 2000, 6:2, p143), with which I also concur. In addition, all three models incorporate an end point, which suggests the completion of a cycle, thereby pre-supposing some sort of finality.
These models are therefore limited in their application to my own work, since, as new losses take the place of old, my clients find themselves in a continual cycle of loss without the comfort of such an end point. Processing the loss of the ‘pre-infected self’ and re-defining the ‘HIV+ self’ often means working with the stage of identity vs role confusion in Erikson’s (1965) psychosocial model. Additionally, where partners stay together, a revisiting and re-negotiation of the adult stage of intimacy may be required since intimacy is often disrupted and sometimes lost due to HIV infection.
This stage is also revisited by clients where a partner chooses to end the relationship with a HIV+ partner. Working through the loss of the partner (usually due to fear of infection); as well as other significant relationships (usually due to HIV related stigma); is also key to the work. To support this work, I use the ’multi-dimensional’ grief model by Schuchter and Zisook (1993), adopting four of their five dimensions: emotional and cognitive responses; emotional pain; changes in relationships and changes in identity. (Schuchter and Zisook, 1993 in Stroebe, Stroebe and Hansson, 1993, p26-43).
I have also used Worden’s grief model in supporting a HIV+ client whose HIV infected partner committed suicide. This work is clearly demanding and is informed by the client’s internal working model of self and other. Due to perceived stigma and fear of rejection, it is not unusual for the client’s attachment behavioural system to be activated throughout the therapeutic relationship. Recent research highlights the mirroring of Bowlby’s theory within such a relationship.
Parish and Eagle (2003) and Sonkin (2005) draw attention to the manifestation of clients seeking proximity maintenance to the therapist; experiencing distress when the therapist is not available; seeking a safe haven when in distress; and using the therapist as a secure base. To cater for this, I strive to provide a secure base in therapy, ensuring I remain boundaried, punctual and professional, informing clients of any breaks and provide opportunities for clients to explore their anxieties. Clients in particular distress may also contact the agency, who in turn may contact me.
My experience has taught me that clients with avoidant attachment styles take time to build trust in the therapeutic relationship. I have also found the avoidant attached usually need permission/re-assurance to grieve their losses, whilst the anxiously attached require permission/re-assurance to stop grieving their losses. I am cognisant that the therapeutic relationship promotes attachment yet at the same time acknowledge the paradox in severing this attachment at the end of therapy. Ending sensitively is therefore crucial. I recognise that clients may develop co-morbid conditions such as alcohol and recreational drug abuse.
In line with the BACP ethical principles of beneficence, non-maleficence and self-respect (BACP Ethical Framework, 2007), I use supervision to monitor any emerging signs of such abuse, where a decision may be reached to refer these clients to external agencies or other, suitably experienced, internal counsellors. In assessing my strengths and areas for development, I am now much more aware of how early attachment experiences and internal working models impact on how clients process their losses as well as their potential impact on the counselling relationship and process.
I have extensive experience of working with loss and bereavement, which is underpinned by my specialist training and practice at Positive East, as a bereavement counsellor with the Bereavement Service and as a counsellor providing support to those bereaved through homicide at Victim Support. I believe a healthy, secure attachment to a primary caregiver is necessary for a child’s social, emotional and intellectual development. In turn this promotes the development of a healthy internal working model, disruption to which may lead to psychopathology in later life.
Whilst the internal working model tends to persist through the life course, I believe it can be modified by divergent experiences, but acknowledge this change may be difficult. Whilst I have extensive experience of working with loss, I now appreciate how early formative attachments influence our reactions to such loss and how these reactions may impact on the therapeutic relationship. Popular grief models clearly fall short in addressing the multi-layered losses experienced by this client group, demanding instead the integration of what is currently available.
The high level of insecure attachment style demonstrated in HIV+ gay men may be due in part to the unique challenges they face within the context of HIV related stigma and negative social experiences. Finally, I believe my knowledge of theory and sensitive application of skills has proved to be an effective strategy in working competently, sensitively and safely with this client group. Nonetheless, I recognise the need for continuous professional development and aim to attend workshops on attachment; and mental health and HIV during the summer.