Loss to follow-up in understanding cravings: Tailoring effective public health messaging for HCV/HIV-at-risk populations.

Maxwell Gustin UA 25 and Jim Stellar

Can the psychological principles behind drug addiction research apply to at-risk health groups (e.g. HCV and HIV populations) to create a noticeable effect in the real-world? This blog is on a slightly different topic from what is usual for this blog series. It has in common that it applies the blog’s cognitive-emotional integration thinking and does it from the behavior of drug addiction to address a public health treatment issue – the loss to follow-up (LTFU) in patients who are undergoing treatment for certain medical conditions like HVC/HIV. But first let’s start with the ideas from drug addiction research.

Psych/neuro basis of addiction (even from laboratory rats): One of the good story lines in the recent history of neuroscience has been the advances in understanding drug addiction. It began with studies done on animals and the underlying brain systems that they have in common with humans. However,  first we must say that back in the day in the 1970s when JS got started graduate school, the idea was that addiction could be treated by detoxification. That idea was supported by the observation that many rewarding drugs like opiates also produced a painful withdrawal period during drug absence and that pain helped to sustain drug-taking behavior. However, the treatment also failed even if the addict made it through a withdrawal/detoxification period. What was needed, and is the more modern view, is that learning-based cravings for the drug played a large role in the relapse even after detoxification. Rats and humans learn and remember, essentially forever. That may be why today we think of drug addiction as a “chronic-relapsing disease.” The more modern idea is that stimuli like the drug paraphernalia or the environment in which the drug was taken or even the cologne that the drug pusher always wore, gets associated by learning with the drug effect and can become a trigger for later craving and relapse. These cravings can be very strong, sometimes called “volcanic” cravings. The addicts are then almost like Pavlov’s dogs that salivate to the bell only after the bell is conditioned by pairing it with food. Of course that conditioning grows stronger with repeated cue-drug pairings, and the triggering stimuli can cause a craving even well after the addict gives up the drug. It was this strengthening of drug sensitization or craving reactions to drug triggers in laboratory animals that  JS studied in rats in his professorial research career, before he went into academic administration, and before he returned to professorial teaching/writing and met MG.

We took you through the above discussion because, again, we thought it might be useful here to apply lessons from drug addiction behavioral neuroscience research to our discussion in this blog of LTFU in HVC/HIV treatment. LTFU is what the name suggests and we will come back to discussing it later. But first, what is the public health problem in the first place?

Treatment Relationship to LTFU and Substance Abuse Disorder (SUD)

Whether it is drug treatment programs or infectious disease clinics, clinical retention and dropout has become a growing concern in public health. An array of drug treatment programs can be found in New York State, and nationwide. Similarly, infectious disease clinics have become more prevalent and are readily available. If we have some tools to play a critical role in fighting drug addiction as well as the eradication of widespread pathologies, why are we not seeing more promising results?

More than 2.8 million New Yorkers and 48.4 million nationwide, suffer from SUD. This population, especially people who smoke/sniff drugs (PWSD) and/or people who inject drugs (PWID), are at risk of infectious diseases. Hepatitis C (HCV), Hepatitis B (HBV), Human Immunodeficiency virus (HIV), and Syphilis are some infections this population is at high risk of developing due to drug use lifestyle. New anti-viral therapies have become available in recent years, and have been shown to significantly decrease symptoms and in many cases, eliminate the virus entirely.

Yet, rates of HCV and other infectious diseases, that can otherwise be treated, have not dropped significantly. In fact, in some populations the rate of HCV has gone up (e.g. non-Hispanic American Indian/Alaska Native persons). Experts understand the danger this issue poses, as well as the potential benefits of these treatments. However, they argue it will be difficult for well developed countries, and near impossible for underdeveloped countries, to reach the World Health Organization’s (WHO) hepatitis elimination target of 2030. To address the at-risk populations and the prevalence of these diseases, many states including New York have allocated funds to provide these medications and treatments. Although these programs are often free and are highly effective, many patients do not finish treatment. A lack of retention and high rates of LTFU, ultimately affect us all and prevent those in need from curing the infection.

LTFU and treatment dropout rates are affected by several critical factors. Most notably socioeconomic factors. Studies found that age, race, and financial status correlate to treatment dropout rates. Persons 12-19 in age, black, and reliant on public assistance, were found to correlate with increased likelihood of treatment dropout. Additionally, unstable situations such as homelessness, legal trouble, heightened drug use, and mental health crises are all contributing socioeconomic factors that increase one’s likelihood of dropping out of infectious disease or substance use treatments. Other factors that affect likelihood of treatment dropout are patient knowledge, provider empathy, and treatment accessibility.

What makes public health messaging effective? / Psychology of messaging experts agree, relevant research into public health messaging is extremely limited. Mostly due to difficulty in determining causality in multi-faceted campaigns as well as a lack of funding. Effective public health messaging requires a thorough understanding of the intended audience. The same experts argue the public (and sometimes the audience) can be divided into two distinct groups; “those inclined to agree with messages about social determinants of health and those inclined to disagree.” This distinction has the potential to improve and adapt public health messaging by making it more relevant and accurate.

What is LTFU and how can we prevent it?

In high-income countries, even the United States, patients are disengaging from life-saving treatments. LTFU in HIV treatments has become a growing trend across the country and the world. Hepatitis C (HCV) follows a similar pattern of LTFU that better captures barriers to care and inequities between countries. In high-income countries, LTFU in HCV treatments largely vary but have been found to be approximately 13%. Yet in low-income countries, Brazil in this example, LTFU in HCV treatments is approximately 59% after diagnosis and prior to treatment. It is likely high-income countries are more well-equipt for combating LTFU than low-income countries, but why are patients still becoming disengaged in the first place, and what can we do to prevent LTFU?

Worldwide, loss to follow-up (LTFU) is one of the leading causes affecting treatment outcomes (reference). Whether it is infectious disease clinics, SUD programs, or cancer treatments, it is crucial for the health and safety of the individual, and the community, that patients continue recommended treatments at their prescribed intervals. Many of these treatments require thorough and lengthy follow-up periods to ensure the treatment is effective and the patient’s condition is improving. LTFU fundamentally underscores much of the work being done in these programs. In New York State for example, many clinics (infectious diseases, SUD, and cancer) keep records of LTFU but do not have the resources to properly track outcome status for those patients. Studies show that the majority of LTFU occurs within the first 6 months of treatment. One of the most troubling practices frequently done in these studies and treatments is the delayed determination of LTFU. If a patient is undergoing a 12-month treatment program for example, many program procedures dictate waiting 12 months for the treatment to end to categorize the patient as LTFU. By doing this treatment programs and research studies overlook barriers to care and underscore the importance of patient engagement. So what can be done to both lower LTFU and raise data accuracy?

What many successful programs have utilized is a group of specialized health care workers to “trace” and follow up with patients regarding any missed treatment. This option allows for the program to better understand the barriers to providing care to their patients. Additionally, by keeping track of any missed appointments or medication refills, the data becomes more accurate, and better suited for addressing LTFU. Furthermore, large HIV prevention trials have instilled this tracing technique, among others, and have found that tracing patients is one of the most effective ways at preventing LTFU. In addition to patient tracing, these prevention trials have found two successful strategies for combating LTFU. Meeting patients who are LTFU at their homes was found to be a successful alternative that both located the majority of those LTFU, and significantly prevented further LTFU from that sample. Another strategy used in these trials was providing nutritional support and family supplements. This was accompanied with an association of lower LTFU in those with food insecurity. Techniques used to reduce LTFU often improve patient outcome and data accuracy.

How can an understanding of drug cues and triggers be used to better prevent LTFU?

Let’s take a little dive into neuroscience, on what might trigger LTFU by looking at what triggers craving in drug addicts and see how that plays out in some basic psychology research. We have three observations.

Incubation of drug craving:

First, let’s consider how cues get established in the brain. We’d like to go back to a piece of research done in JS’s neuroscience laboratory some years ago in which he was teaching rats to be sensitized to the box in which they were given cocaine by measuring how much they got excited and walked around in that box. This was an animal model of craving development in humans. What they did in this experiment was also measure biochemical changes in the subregions of the nucleus accumbens, which is where cocaine is thought to act to produce reward and maybe craving. They noticed that in one of them there was an expression of a protein associated with an immediate early gene (delta fos-?) that gets expressed when neurons get activated. Immediate early genes are common and turn on when cells get activated and need to make more proteins, but this one is a bit different. It gets converted into an almost insoluble form and stays within this brain area for some time after cocaine exposure. They joked in the lab that it was a long-term record of cocaine administration. Second, this protein in this brain area, was thought to drive the neural connections then could underlie the learning between the context (cues) for cocaine taking (or our administration) and perhaps produce craving. The final interesting point is that if you stop giving the cocaine, the persistence of delta fos-? in the brain could actually still produce the neural changes, so the cravings could actually grow for the first weeks of drug absence. That incubation of craving behavior was actually shown by an NIH study. One wonders if a similar kind of a long-term process of cue growth from the process of seeking treatment could cause some semi-permanent learning-driven changes in the brain could contribute to the LTFU effect.

Application to LTFU: In LTFU we are thinking a sensitization process grows with time, and as it does, emotional reactions increase. For example, suppose the trip to the clinic to get the treatment is challenging, even stressful, for the patients. The impact of frustration or other negative attributes associated with making the trip could grow over repeated trips. The lesson here would be that what happens at later times in treatment and causes LTFU may be different from what is seen initially in treatment. In the above research we can see how chemical changes in the brain’s neurons tells us that later craving effects utilize more broad and deeper brain pathways. Furthermore, in later stages there is actually an increase in stress sensitivity and decrease in reward from non-drug origins. This often leads to disengagement. From a neurobiological perspective, this makes sense. Our brains continue to change even after drug cessation. It also explains why someone may feel optimistic or engaged in early stages but later disengages within a few weeks. Simply put, later stage withdrawal may be associated more with the “want/need” while early stages are associated with the “like” or reward/enjoyment of the drug. These are common terms in drug addiction research. In terms of LTFU, this relationship may help us understand why patients are not finishing treatment. Depending on the stage of withdrawal, a person will crave differently. The key here is to combat the likelihood of disengagement. To do this we must implement procedural changes using these known differences in cravings within programs with high LTFU.

Opponent Process theory:

Second, let’s look at an important but older behavioral theory of drug addiction by Solomon and Corbit, published way back in1974 and see how it might apply to LTFU. Here we go back to a paper in learning psychology which suggests that any emotional effect created in the brain (they called it the A state) gives rise to an opposite effect (the B state) that tends to reduce the impact of the first effect. This is said to be the brain’s way of keeping you on an even-keel. But that second effect or B state grows with time and moves forward in time to better counter the first emotional effect of the A state. In drug addiction, the first effect is pleasure, and the second effect reduces that pleasure. The trick here is that with repeated presentations of the pleasurable drug the after-effect grows stronger so that pretty soon the drug addicts are getting very little reward or pleasure out of the drug, but have a pretty substantial craving or even withdrawal reaction when the drug effect in the body dissipates. The authors generalized this process to other phenomena such as when you meet someone you like and really enjoy their company, but you miss them a little when they’re gone. That early stage of the relationship is compared to having a very long relationship with this person, perhaps a spouse, and just seeing them day-to-day seems normal. But now, if that person should go away (e.g. in the case of death of a spouse), you have a very long and strong unpleasant reaction which we would call grief. The point again is that the brain is changing its reaction to a stimulus over time and the question is what does that have to do with LTFU.

Application to LTFU: In Solomon and Corbit’s research we can see how addiction is a product of how the brain maintains an emotional equilibrium, which is anything but anomalous. As mentioned, the second effect (or B state) grows over time such as continued drug use. It becomes less of a craving and more of an avoidance to an emotional crash. This B state is often associated with emotional discomfort, irritability and the desire to return to the A state (drug reward). This highlights that while the first effect can come and go quickly, the second effect is stronger and longer lasting. LTFU mirrors much of this as many disengage not within the first few days, but within the first few weeks. Conditioned stimuli (cues) are especially effective in triggering this B state. Treatment sessions, clinics, follow-ups, and most other aspects of a treatment program all have the ability to activate the B state emotions (stress, active cravings, etc.). Those under the B state are highly motivated to remove its effects, this usually manifests itself as escape oriented behavior. Much of this can be seen in LTFU. Especially that of B state activation, even after drug abstinence, which can reoccur quickly with a powerful trigger/cue. This balance of effects seems similar to the battle between the emotional states of an addict to stay on the abstinence path or relapse by giving into the cravings. The idea is that this balance plays out in LTFU

Long-term changes in competition with other rewards:

Finally, the last point we want to make from drug addiction comes from a researcher in operant psychology, not neuroscience. Here the idea by Heyman is any behavioral choice (to take a drug or to say with treatment) depends not only on what you think you are getting out of that treatment but on what you are not doing with other alternatives in your environment. When you are first taking cocaine as a young person, you may not be taking care of your good job responsibilities. But with growing age other good responsibilities come into your life like taking care of your family and children as well as your job. The totality of these good responsibilities now compete with your drug choices and being bigger they may be more successful in doing just that. This paper essentially proposes that this effect is why some older drug addicts are more likely to simply give up the drugs. The idea is psychological, having to do with changing contingencies for one’s operant behavior that lead to rewards in life. Does a process like this underlie LTFU over the long term? Do their changing responsibilities outside of treatment contribute to LTFU?

Application to LTFU: Heyman’s idea that older drug addicts are more likely to give up drugs due to growing outside influences, is seen in LTFU. Younger individuals have been found to have higher rates of LTFU not only in SUD cases but also in post-op care and chronic pathology treatment . Younger people, especially those struggling with addiction, often have less responsibilities and in turn are less affected by non-drug alternatives in the environment. Older individuals however, have been conditioned for much longer where the effects of noncompliance, such as the absence of environmental alternatives, have accumulated. That is, the separation from family, work, and potential becomes more potent with age. LTFU is lower in older individuals for a variety of reasons, which Heymen’s psychological base considers and offers a potential rationale for this behavior pattern.

Here are some recommendations for reducing LTFU.

We are going to build off of the three points above, knowing full well that this is a simplification of the LTFU discussion.  But it is a place to start. The overwhelming conclusion from all of the above three stories is that the time course of treatment deserves consideration when looking at reasons for LTFU. Simply put, the vulnerability to LTFU and the reasons for it, change with time in treatment. To prevent LTFU one has to recognize and explore what are those changing reasons. This means that as treatment progresses, the reasons for LTFU change and attempts to keep the patients in treatment have to change.

Looking at those three reasons:

1. If getting to treatment involves obstacles (e.g. travel time, time away from work or family, expense and hassle), expect that to grow as the brain learns from experience. Like drug craving (but opposite as it opposes treatment) these effects may grow with exposure and be based in brain changes associated with learning, so they will be permanent effects that the clinic needs to somehow counter to prevent LTFU.
2. Getting treated is a good thing. Although the process might not be (e.g. getting an infusion), the idea that the patient is in treatment and under the care of the medical establishment should be a good state. It will not last. Any satisfaction or even relief that comes with treatment will be opposed by the brain’s B process that is trying to keep the patient on that even-keel mentioned above, and it grows with experience. That needs to be addressed to prevent LTFU.
3. Finally, life itself (and the rewards and challenges associated with it) typically change with years. If treatment is forever (as in HIV) then the clinic and the patient needs to be responsive to changes. Maybe treatment involves getting to the clinic and initially that was done by bus. Later as the patient grows, maybe they have a car. That sounds good. But maybe they also now have a family as well as a job and getting the kids to school or soccer games becomes another task for that car. It could make getting to the clinic harder. If that happens, LTFU could be the result.

Notice that all of these three factors above involve the passage of time. The treatment plan set up on the first visit to the clinic may not be the one that works the best later on. That is the simple lesson from our  cursory analysis of the drug addiction literature. So what is our advice to the clinician from this perspective?

Advice for the clinic to reduce LTFU

The procedural methods need to change, and below are some thoughts about how to do that based on the above principles.

1. Interval/schedule adjustments to combat early LTFU/Craving changes: Now that we have an understanding of the underlying principles, adjustments can be made in treatment policy to better address LTFU. Our neuroscience background explains how cravings differ at different stages of addiction. To combat LTFU in those with SUD we must first distinctively separate patients into two subgroups; (1) early stage and (2) late stage addiction. We know that short-term drug cessation manifests cravings and ultimately affects the patients differently from long-term drug cessation. Additionally, those within the late stage subgroup are susceptible to different cue-induced cravings than those within the early stage subgroup. Applying these principles, we must first start with education. By educating patients as well as medical practitioners we take the first step in combating and addressing LTFU. Medical staff should focus their attention differently for patients in different stages. Especially that of those within their first 6 months of treatment. Addressing these patients and their cravings differently over time, and throughout treatment, could aid in understanding the barriers to care as well as overall patient engagement (or lack thereof). By doing so we gain a better perspective into additional obstacles and the growing effects late stage addicts experience. Conversely, if we identify these obstacles in early stage addicts we can better prevent their growth and manifestation in the late stage. Cravings are not the same in different stages, yet we currently treat them the same in different patients. We must address early stage and late stage addiction accordingly.

1. Cue/trigger awareness and prevention: Another critical point in addressing LTFU is the prevention and awareness of triggers and cues. In the above principles we discussed how something as unconscious as a smell or sound can create volcanic cravings (e.g. drug dealers cologne, flick of a lighter, etc.). Many people who struggle with addiction often feel powerless in their cravings. Those practitioners overseeing patients often overlook the environmental and experiential triggers unique to each person. While building awareness and educational campaigns could certainly help, we must also focus on cue prevention to appropriately address LTFU. Patients and practitioners alike, should be aware of common cues and triggers in order to ensure their absence before, during, and after treatment. In the form of education patients should be given a mental “toolbox” to better deal with and combat daily cues and triggers.

1. Younger safety net, older “fish net”: Similar to that of early and late stage addiction, another distinctive grouping should be made. To combat LTFU we must acknowledge that young addicts respond, act, and crave differently than older addicts. The above principles suggest much of this has to do with responsibilities as well as resources available. Life experience is another factor that can heavily influence this difference. Younger patients should take priority in LTFU prevention, as older patients are less likely to disengage. Additionally, older patients should be reminded of their responsibilities and resources available throughout treatment to ensure engagement and prevent LTFU. The key here is for treatment programs to change or highlight the contingencies for a patient’s behavior so it later leads to reward (e.g. cure, improved daily life, etc.).

1. Tracing (separate or combined? Maybe 3 + 2): We now understand that patient tracing is one of the best fundamental changes a treatment program can instill. While this effort requires an attentive staff and likely substantial funding, leaving patients to their own devices often leads to LTFU. Meeting those who have become disengaged at their homes or shelters is a major step in creating continuous treatment engagement. Tracing also helps in identifying specific barriers to care, which oftentimes includes socioeconomic hurdles. Addressing these barriers in the form of hygiene supplies, nutritional support, childcare resources, among many others, should be implemented. Whether it is substance use, infectious disease treatments or clinical trials; tracing should be integrated to cultivate engagement by preventing the found barriers to care.

Conclusion

LTFU is a growing issue in public health that must be addressed. To improve outcomes, changes must be made using an understanding of behavioral neuroscience principles. With these principles in mind we can see how minor changes have the potential to create a major impact. We have the tools to combat barriers to care, and in many cases cure infectious diseases; yet LTFU threatens these outcomes. We cannot provide high quality care for patients that are unable to engage in and finish treatment. If these gaps in policy and care continue to be neglected, the problems we see with LTFU and patient outcomes will continue. By grouping patients, tracing status, and educating those at-risk about influential factors (e.g. cues/triggers, socioeconomic barriers, etc.) we began tailoring treatment for those who need it most. Since HIV and HCV are disproportionately high in people who use drugs, adjustments in treatment policy should be made to better mitigate the subsequent LTFU that occurs due to SUD and other influential factors.