Current State of Medication Adherence: Challenges and Solutions

[Dr. Nilsen] Hello. Good afternoon or good
morning depending on where you are on this beautiful day. This is Wendy Nelson from the
Office of Behavioral and Social Sciences Research here at the National Institute of Health.
Welcome to the Adherence research network distinguished speaker series. I am absolutely
thrilled today to introduce our speaker who is Hayden Bosworth who is here from Duke University.
Dr. Bosworth is a health services researcher who focuses on patient and organization level
factors to improve treatment adherence. He is the associate director of the center for
health services research and primary care and a career award scientist at the Durham
VA. I think you’re going to find that this is just going to be an amazing presentation
and I know you don’t want to hear me, so with that I’m going to switch to Dr. Bosworth so
welcome everyone. [Dr. Bosworth] Thank you Dr. Nelson for the
opportunity, I’m very thrilled to be here. Today we’re going to talk about the
current state of medication adherence. Some
of the challenges and some of the solutions. Um, sorry, I’m trying to find a way to. Alright we’re having a little technical difficulty. So, just the outline I’ll talk a little bit
about what the problem is current policy in the landscape of issues. Talk briefly about
medication adherence, how do we measure it. I’ll talk a little bit about patient/provider
communication. I’ll add a little bit on mobile health in regards to medication adherence
and then summarize with recommendations. So kind of a whirlwind tour, but I’ll try to
hit some of the high points through this afternoon. So, this is not any surprise, but adherence
is a huge issue and I would argue is probably one of the largest public health problems
that we are experiencing. [All guests have been muted] Determine it’s effectiveness. In general the
prevalence data suggests that about fifty percent of people take the medication as prescribed
appropriately. This translates into approximately 300 billion dollars in health care costs.
And I would also include in that 300 billion there’s patient safety that contributes to
that huge number there. I also think, again, in light of the policy issues that I’ll mention
a little bit, there has been significant changes in the use of generics and current legislation
changes are going to result in a number of people entering the health care system and
how that will impact medication adherence. So it doesn’t surprise me anymore that across
any disease, any problem there is always issues about adherence. This is an article that was
first transplanted hand and unfortunately the individual wasn’t able to adhere to his
drug regimen and the hand despite it successfully working, was not able to follow the treatment
and they had to remove the hand. We see this in transplants and other various things and
we’ll talk a little bit about the challenges, but it’s not just the simple chronic disease
medications, but it’s across multiple areas and we see problems there with adherence. So, some of the problems facing payers if
we also put this into the context of policies CMS has instituted a 5-star quality system
to rate how well pill refill, how we’re doing with adherence to a health care plan. And
the legislation is changing such that by 2014, a 3-star medication advantage plan stands
to lose an average of $16 per member which translates to almost $200 a year, so where
there was currently incentives, benefits of need pill refill, there will be penalties
as we move to 2014, so there’s a lot of challenges in the healthcare environment of how do we
improve the adherence rate. And I think that’s changing the overall landscape and how we’re
focusing on that. So these motivations are resulting how do we improve these star ratings?
How do we increase market opportunities, membership, these quality bonus payments, and then basically
how does that result in increased plan profitability. So for those of us who have been doing adherence
for a while now, early on it was a little challenging because we didn’t get as much
attention, but with these policy changes that are aligning, a much more focus in interest
in these topics are becoming more of an opportunity for us to collaborate and add to. So, just quickly the star rating system diabetes,
the other thing to frame this is that it’s diabetes hypertension cholesterol, and I think
it’s really important to highlight why those three drugs or classifications were chosen.
Part of it is because those are lower cost medications with a pretty good track record
of effectiveness, so a lot of the adherence literature looks at the evaluations are still
reasonably inadequate and I think that’s an area that we do need to focus more on and
in the sense that adherence is a proxy and that even when you have somebody adhering
it doesn’t necessarily correlate with good outcomes. So, this is where we’re starting
from, I think there’s room for us to improve but if we were to look at, for diabetes here,
the PDC is the possession, the proportion of days covered and you can look at the 1,
2, 3, 4, and 5 star and so, in general if you’re at about 71 to 75% you’re a 3-star,
so the goal is to try to move it to a 4 star, which is a slight improvement in adherence
defined by pill refill. And then the 5-star. The point too is that the 5-star is still
only at 79% pill refill adherence rate. So, it’s interesting because there’s a disconnect
in the literature because when we look at adherence we arbitrarily define as 80% or
greater and we just talked briefly, Dr. Nelson and I, and others in the room, that may be
adequate for certain drugs, but definitely would not be adequate for HIV or other medication. So, again, this is I think a way of defining
it’s where the movement is, but it also does take from a health care plan significant changes
to go from a 71 to a 75, a 4% improvement in adherence. So, just to summarize I think
what the goal of medication adherence is achieving optimal medication adherence depends on the
patient being prescribed the right medication, filling it, and taking it correctly over time
and it requires the appropriate prescribing, effective patient-provider communication,
coordination and active engagement participation by the patient. There are a lot of pieces
there and the point too is simply it’s just not the patient popping a pill. There’s a
lot of parts that need to be considered and put into place to ensure that adequate adherence
and treatment outcomes are met. Another way to operationalize this, this is
a quick schematic that I put together, that if we define adherence in the context of medication,
it is a truly complex behavior. So, the assumption is that you have a problem, you make or you
need to make an appointment, you need to accept and believe that there is a need for medication,
accept that medication and then go fill it. Now, right off the bat there we often times
in literature look at primary non-adherence and 20 to 25% of people don’t even fill the
prescription so when we’re looking at the fill rate and then taking medication and then
pill refill, we’re usually citing 50%, but I would argue that it’s really 50% is assuming
you’ve already filled it and it has already excluded that 25% who may not even have taken
the prescription and filled it. And then this cycle continues and in returning
to the provider for the monitoring and getting another prescription or refilling it. Now,
that’s typically one drug, now we times it by 9 or 10 drugs we can start seeing the complexity
or the challenges that may exist. In terms of just medication adherence, again we were
briefly talking before we started and talking about some of the factors that we think are
relevant to non-adherence, but if you look across the literature, there have been about
over 100 factors at this point that have been identified that have been predictive of non-adherence.
So, the challenge for us both in the research environment as well as in the implementation
environment is how do we create programs that improve adherence, but that can target or
tailor despite all of these potential factors? So how do we create something that’s scalable,
that’s cost effective, and then also, I’ll come back to this in a moment, but how do
you initiate and maybe create habits in long term maintenance? And that’s something really
important because for many drugs that’s beyond acute we’re trying to get people to either
take them for long periods of time in the context of cardiovascular drugs or so forth.
And then we also could talk about other drugs where we want to decrease the use of it, pain
medications and so forth. I won’t talk too much about those types of drugs, I’m talking
more in the context generally of chronic long-term medication. So, this is just a theoretical framework that
helps us justify where are we starting, what are we looking at, and understanding the moving
parts there. Again, it’s easy to assume that it’s the patient characteristics, there’s
the social cognitive theory where you have to look at the perceived risk and the benefit,
understanding those cognition, forgetfulness is really key aspect that most of us will
focus on. The other part, cognition, is inductive reasoning. So, if I wake up at 8 o’clock and
I have to take the medication 3 times a day, I have to be able to do the calculations in
my head that that is 8 and then 8 hours later is 4 o’clock in the afternoon. So, being able
to do that type of math is important. Literacy, I won’t spend too much time on it. But, yet
25-40% of our primary care setting that we see at least in our context are functionally
illiterate, which means that they really have a hard time understanding the back of a Tylenol
bottle let alone the handouts that we often give our patients. So, literacy is a big issue
we need to think about. I’ll come to the coping and stress in just
a moment, but daily hassles and aggravations really long term have huge implications. Side
effects are obviously key aspects. Some of our own work, what we find is that allowing
patients and expectations of what the side effects are and how to handle them can actually
be really helpful to alleviate those issues. Mental health, big aspect there. Particularly
depression, one of the most consistent predictors of non-adherents are depressed individuals.
And then comorbidities the more complex the, it’s not the number of drugs, it’s the complexity
of the drugs that really has an impact on non-adherence. So those are some of the patient characteristics,
then there’s obviously the provider characteristics which I’ll come to in a moment, but communication
style. There’s a lot of work that the certs are doing in shared decision making that’s
really important. Intensity of therapy, there’s a lot of work early on clinical inertia where
providers aren’t necessarily following guidelines or there’s a little bit of slippage. Understanding
the medication regimen that could also probably cause error. Care coordination, you know when
there’s multiple providers involved, how does that all handle? Then there’s the policy that
I mentioned briefly the CMS 5-Star rating earlier, but there’s also the medical environment,
the community. And then all of this has an impact on treatment adherence. And the important
part of all this is often times adherence isn’t really our primary outcome, it’s the
proxy and in the end we’re trying to really improve outcome, but this is just a framework
for us to try to figure out what do we think are the major issues if we were to look at
all those factors that have been shown to be related to medication adherence. You’ll also notice that I did not put in demographics
or social or financial aspects, those are key aspects, but often times those are, can’t
modify someone’s race or necessarily education level. For us we’re looking at modifiable
factors. So, I want to just briefly one of the things that were recently just published
that I think is kind of exciting and, again, it’s good to have this cross-pollination,
this is work that initially we took from the HIV literature, but we just completed a study
of post-MI patients who were getting a web-based intervention or a nurse-based intervention
with the web. We’re still, and I’ll come back to some of the data because I think that using
technology is helpful, but there are some challenges depending on population and I’ll
show that data, but for this part here I just want to, an aspect that we haven’t really
tapped into a lot of is chaos. And so with this we looked at post-MI patients, about
400 individuals, they were recruited from 3 main hospitals and if you look to the left
at that 1.07 that is life chaos, this is an adapted measure, but it’s just in questions
like I’m always late, there’s always something going on, and it’s something we can administer
relatively quickly, but the take home message is despite putting all of these other factors
that we would envision into a model, we still find that life chaos is still significantly
projective of outcomes. And I think the important part for us is the modifiable factor that
we could perhaps look at. So, I think it’s an interesting topic that’s
we’re further exploring and, again, coming from the HIV literature. Just briefly again,
we found that 43% despite having a heart attack recently reported that they were non-adherent
to their cardiovascular medication in the last month and that independently life chaos
was an important factor for non-adherence. So, this kind of leads to the issues regarding
medication measures of adherence and I just want to highlight these. We can use cheap,
easy-to-use methods, then there can be more expensive, relevant, and more reliable measures
and I think there’s a place for all of these. The question is where, what’s the question
we’re asking, who’s using it, and what do we need to know? So, from a clinical perspective
if I’m just simply as a provider trying to identify whether or not someone’s adhering,
simply asking them in a non-judgmental way, are you having any problems with your medication?
If they say yes, it’s a good indicator that there’s a problem, but, you know, for more
complex drugs or where there’s concerns that people may not be taking them, we can rely
on technology and other various things. But again costs are issues that I think we have
to balance, both from the time and from the implementation of the devices themselves. I would argue that we’re still grappling with
the measurement of adherence, I think the hundred pound gorilla is the pill refill adherence
that CMS is using, but I think that in general we really are looking at multiple aspects
of adherence more, measuring it. A paper a junior faculty member published that we were
involved with was looking at self-report and pill refill adherence and the question was:
how related are they, or are they actually two distinct behaviors? So, just looking at
that, if you look at 30-day, 60-day, or 90-day refill relative to a four-item question that
asks: have you been forgetful, are you careless, do you stop when you feel good, do you stop
when you feel bad? The Kappa statistics are pretty poor. Ranging from 0.1 to 0.13, so
you could say “well what am I supposed to do at this point?” So the take home message
is actually both were predictive of blood pressure, so both the pill refill whether
under supply or over supply so, over supply is, kind of, would be pill hoarding where
you have 120% or greater adherence and then under supply would be less than 80% and then
the self-reported were both. So all three were indicative of poor blood pressure in
this study, so I think that the take home message that we argued were they are both
important, pill refill and self-report, but they’re measuring different aspects of medication
adherence. Oh, excuse me. So, the other part is you know, this is a
study that Ryan Shaw, who is Junior faculty, who’s post doc bid was with me where he looked
at baseline and then 24 month outcome and again in a self-reported adherence quickly
assessed was a pretty good indicator upwards of 24 months. After 24 months, it was no longer
predictive of blood pressure control and so I think we have to continuously assess adherence,
but at the end of the day, how frequently do we need to do that? And the data suggests
perhaps, you know, on average 6 to 12 months is a good indicator of poor outcome at least
in the context of blood pressure control. And lastly I just want to, this is an example
of a trial that we published earlier. And this was a good example of also the transition
from effectiveness to implementation. So in this trial among 558 Medicaid patients we
gave our general self-management program, which did have a significant focus on medication
adherence, but also trust issues like alcohol and smoking, but in summary the adherence
rates the medical using medication possession ratios went to 55% in the 9 to 12 months prior
to the initiation of the program to 77%. You know, this was a quasi-experimental design
and we didn’t have randomization, but when we looked at the comparison of 4,000 individuals
who would’ve been eligible for the program, they stayed consistent at 55%. So, just the
point is that in general too was we do our trials to think about who the stakeholders
are, engaging them and so that we can have these transitions where we can then take our
projects and hopefully successfully implement them into the real world. So, I want to transition a little bit to where
is the role of the clinician and patient and the role of communication. So, effective net
therapies really involve patient adherence and effective disease management. I think
patient adherence is one part of the larger self-management issue and I think, I often
say it takes two to tango at least. So, the patient is usually the focal point, but a
lot relies on how they interact with the health care system, particularly the provider or
providers and how they work with them. So, I think the patient-centered medication management
this kind of model that we’ve been moving towards involves the shared decision making
to reach provider-patient concordance. It doesn’t work often at times when the provider
just writes a prescription without engaging the patient and seeing where they are, their
beliefs on medication, their thoughts regarding that. It also assumes there’s effective prescribing
by the providers, so are they using evidence-based guidelines? Are they dealing with clinical
inertia? And really focusing on watching and doing what’s appropriate based upon the current
guidelines. And then tracking the feedback by the patient’s self-monitoring and then
basically the medication taking behavior, so there’s this cyclical fashion that we really
need to think about. One of the challenges again as we move from
an effectiveness randomized smoke control trial model to an implementation is eventually
looking at the reimbursement and how do we do all this. It’s easier in a research environment,
but often times it’s a disconnect to what we do in the research and how we get that
into the real world and, again, as I keep pointing it out, that I think we need to keep
that mind as we do this type of work. Provider communication is I think a really
powerful tool that can help improve outcomes in both enhancing patient knowledge, addressing
their beliefs, and also improving satisfaction of treatment. An important thing to keep in
mind too is that often times multiple studies have shown that patients only recall about
50% of their information that’s provided. So, we need to think a little bit about how
we are conveying this information and I’ll turn shortly to the role of mobile health
but I think that’s a particular area where mobile health can be very effective in helping
not only recall, but also help reassure and comprehend information and that this poor
patient comprehension recall contributes to unwitting or non-adherence. So, some of the collaborative communication
provide clear instructions. Now this may seem pretty obvious, but often times we’ve heard
people say “well, take the medication twice a day”. Well, many of us would assume that
means I took it at 8 and I take it 12 hour later, but there are a good proportion of
individuals who may interpret that twice a day is I take two pills in the morning and
I’m good for the next day. So defining very clearly how the medication should be taken
is one important part. I mentioned earlier the side effects and being able to really
give the patient some clue of what the expectations are and what to do if you experience these
side effects. I think that really understanding where the patients are, do they want a paternalistic,
being told what to do, or do they really want to involve and share decision making. There’s
a lot of complementary medicine that occurs in that and we don’t look at those things
which have tremendous possibilities of interactions with the particular medications, but I think
there’s also the data suggests there’s kind of a mispercept9ion unfortunately that providers
don’t want to open these conversations because it may lead to a 20 minute conversation which,
with a 15 minute visit is challenging. So, a lot of the programs that we’ve worked on
are trying to use and do what we can outside the actual clinic visit, but yet thinking
about how reimbursement can be incorporated into that. So, again just you know needing to measure
correctly, consistently, and adequate timeframes. So, the other part of this is that it’s not
just adequate to have this initially with the start of a new dosing, or new medication.
This needs to reoccur frequently to not only ensure that the person is doing what, but
down the road continuing that they’re following through and assessing patient’s adherence
before making changes. So there’s a great body of work that the group at University
of Michigan and Eve Curr and Michelle Eiser have done where they’ve looked at clinical
inertia and medication adherence and often times what we see is a poor outcome may not
necessarily be patient adherence, it could be attributed to poor clinical inertia. So,
that’s basically providers not being assertive or aggressive enough in treatment or also
getting caught up in other comorbidity issues. So, some of our own work right now is looking
at pain and hypertension and so often times a patient will come in, they want to talk
about pain, but the blood pressure is a little poor and you can imagine that you know which
really gets focused on. So that, again, goes back to the patient-provider negotiating.
What’s discussed and where things are. I mentioned already the expected side effects. Alternative
provider support. So, not just looking to the physicians but alternative case managers,
social workers, and family are key to a lot of this. So, I just want to briefly also put this in
a larger context. I think medication adherence can define as perhaps a complex behavior but
I think it ends up being in a larger context of disease management and so when we’re implementing
these programs what we find is that groups don’t want to see silos. They want an all-encompassing
platform where issues regarding weight, diet, and medication can be all found simultaneously.
Now, that’s not necessarily from a research perspective, but as we think about moving
these into the real world and implementing them, we need to think of it in the context
of larger context of all the other issues going on. And then again, envisioning that, how do we
implement these in non-traditional settings outside the clinic and how can we do this
as easily administered and tailoring to the patient’s needs and the most cost-effective
ways of doing it. In the context of disease management, looking at multiple behaviors.
So, quickly we’ve done this across a number of different diseases and different issues,
but in general there’s a lot of the same underlying contexts whether it’s in diabetes or it’s
in sickle cell and part of that is patients understanding the risk and the benefit. What
is the disease? Why am I taking the medications? What are the risks of doing this? And if I
don’t do treatment, as well as the risk associated with the medication. Understanding the side
effects, knowing when, if I experience something, what do I do? Again, underlying of that you
can see a lot of common frameworks that can help generalize across multiple behaviors
and diseases. So, I’ve already alluded to this, I wanted
to also highlight it. I think there’s three stages to medication adherence. There’s the
initiation, there’s the sustaining, and the behavior maintenance and I think it’s really
important to keep all three in your mind. And I think there’s a lot more work done in
initiation and perhaps sustaining, but very little on the maintenance and you know and
so, in a trial we recently published in archives of internal medicine a telemedicine project
for hypertension. And one of the lessons we learned was that there was a huge proportion
of individuals who, they measured their blood pressure, they were doing everything that
they needed, but we didn’t really need to do anything except send them eventually reminders
every 6 months because they were following the recommendations, their blood pressure
was in control, there wasn’t anything for us to do. And that was approximately about
40% of the population and so, you know what our goal then eventually translated into was
maintaining that behavior. So we kept sending them positive reinforcements reminding them
that they’re still in the project, but it was a lesson for us to think about that it
wasn’t just simply getting their blood pressure under control. Once they were successful,
what do we do with those people? I also want to highlight I think some of the
techniques that are involved with what we’re talking about. I don’t think that there’s
going to be one particular mode to prove outcomes in the context of adherence. I think we need
to think of almost a toolbox since our own work looked at level of motivation, the role
of self-efficacy, problem solving, positive and negative framing, queuing and chaining
particularly, I give the example I have to take a particular medication in the evening
and it’s just once a day and it drove me nuts, so here I am doing work in adherence and I’d
be tired at the end of the day and I’d forget to take my medication, so it wasn’t until
I realized that I needed to brush my teeth before I got to bed because my wife would
kick me out of bed with stinky breath, so I put the medication right by the toothbrush.
So there can be very simple things that we can do as well as get into really complex
things. So, but anyway just having, setting realistic goals, action steps, self-reward,
rehearsal, cognitive reframing and shared decision making. And then one other part I’ll
get to at the end is I think behavioral economics are a key growing area that we should think
about as well. So, just in general, again, focusing on some
recommendations regarding maintenance and sustainability is that long-term sustained
adherence is often times only obtained by a small group of patients. And we just don’t
have as much work as I think we need to look at, to focus on this. It’s an area that will
hopefully continue growing further. And we can look to our social psychology colleagues
because what we’re really trying to do is create habits, so if I could see you all I
could ask you how many of you brushed your teeth today, how many of you plan to exercise?
We all get into specific habits and so the goal is that at least in the cardiovascular
disease medication where there’s good advocacy of the drugs, we want to try to create this,
just a habit that if you forget to take your medication, you realize it and you feel uncomfortable.
It’s like patting your wallet or you realizing your keys are missing. So, that’s what we’re trying to translate
to. So, again here just making habits, routinizing ways, and it’s forming these habits over time.
So, one thing too is in terms of the programs, you use terms like tailoring and targeting
and I want to just put this in a larger context of the patients that are in care. You know
you move from the left to the right in terms of complexity and effectiveness I think, but
you can get into personalized words just simply the name of the person single characteristics,
targeted maybe just so that certain groups, older adults, younger adults may get relevant
information. And then we get into individualization and tailoring where we can get not only targeting,
but identifying perhaps levels of motivation, beliefs, and when we get more into those levels
we generally see that it’s more effective and we can get to the point now with technology
to make those scalable to try to individualize and tailoring these programs. So just quickly a little bit about tailoring.
It’s collected so that, it is dependent upon collecting new data so that you can create
and tailor the message. It can be on cultural, personal beliefs and past behaviors are just
some examples. And then you know in terms of under use, we can look at why, if we’re
targeting underuse we can look at perhaps contributing factors poor communication with
doctors, mistrust, side effects concerns, just a couple examples and then depending
upon if the problem is poor communication, then some of the stuff we’ve done for example
would be perhaps role playing with the individual. How could they improve those relationships
so if they feel rushed, we can work with them and provide some examples of entering the
room with a list and saying “here are the two things that I’d like to be able to address”
and role playing mistrusting the doctors some other issues that we’ve worked on as well.
Mentioned the side effects. I just briefly again of the overuse, you know when we talk
about adherence we have to be careful to not just focus on utilizing and increasing utilization.
There are areas particularly with pain medication that we want to decrease utilization, so it
changes the whole definition of adherence upside down, but it’s understanding why there’s
a greater need, or perception for the need a history of substance abuse, poor communication
with the doctor, and addressing those. So, often times depending upon what we’re
focusing on, we need to be careful if we’re looking at underuse or overuse in the context
of adherence, but can tailor the messages to address those issues. I also think, I want
to just highlight stepped level of care. I think that you can almost tailor the level
of intervention to the dosage of what the patient needs. I mentioned the example trial
where we had a group that were self-monitoring and they didn’t need anything beyond positive
reinforcement because they were doing everything that they needed. So, you know, hopefully,
and we’re seeing this is that perhaps creating programs and interventions and studying them,
where say 50% get a low-intensive program and then they trigger something that’s more
intensive and that may involve some type of nurse-case management or higher level mobile
technology, but the goal is that the high-touch, in-person, or pharmacy level which is more
costly may only be a certain group of individuals. So our analytics and our analyses need to
get better and there’s good data out there that may be able to allow us to utilize that,
but our goal too is not to wait until the train wreck happens, but can we use, collect
data that identifies who needs to be triaged, if you will, into the different levels. So
I think these adapted tandigms and trying to create the right, tailoring the right dosage
of interventions are going to be really key as we move forward. I just wanted to start, transition briefly
to mobile health, there’s probably anybody affiliated with Dr. Nelson in this group this
isn’t anything perhaps new, but I think this is fast and furious coming down the pike and
we need to be aware of this and how do we incorporate this into care? But it’s not just
simply the cost, it’s the worldwide use and from an international perspective some of
the changes that we’re seeing are quite remarkable in countries like Africa, areas like Africa
and other places where it’s almost leapfrogging the web to using the mobile technology and
so how do we create programs to utilize those infrastructures and I’ll talk a little bit
about that. But I think at this point right now, that in general, you know the data is
getting better and some of it has been, early on, based on small trials, some of it’s from
industry so I think that, you know, the more we push it towards using very good research
methodology will be really key to better understanding what’s going on. And the goal is to move this
towards scientific evidence, populations of patients and putting this into the real healthcare
system and considering the reimbursement system, how do we figure out ways to sustain these
programs beyond the research environment so we don’t get treated or thought of as seagulls
where we pop in, utilize a research infrastructure and once the research is done, leave that
environment. I think that we need to be mindful of how do we sustain these, so reimbursement
is going to be key. So far from my look at the Mhealth trials,
they generally show the best outcomes are when delivered with some human support. Whether
it’s an opportunity for them to click on options to seek additional, if they have any questions,
or just knowing that there’s somebody behind the technology checking to make sure what’s
going on. I think that in general the trials are showing reasonably high attrition rates
early on and we’re not yet able to see the long-term implications, but hopefully that
will continue and we’ll be able to look at 6, 12, and 18 and 24 months, but right now
we’re seeing a decrease in engagement over time and how do we engage and keep them sticking
to the patient’s date and the individuals use these programs is a key area that we need
to focus more on. So, I just want to frame, you know, mobile
health we’re all set to have realistic expectations of where it is. I think it’s right in the
middle between increased reach and efficacy so again this goes back to the step level
you know that we may need phone counseling and group programs for certain people that
have pretty challenging treatment regimens and could be supported with the use of mobile
health, but not everyone needs phone counseling and individual visits. So the mobile health
I think is a great platform that can help mediate those relationships. But when we’re
looking at mobile health too, again from a scalability perspective it’s kind of like
a public health perspective, we expect to see huge changes in blood pressure, for example
or A1Cs. It’s probably not, but we have to maybe look at it from a public health perspective
particularly if we’re trying to frame the cost effectiveness of the return on investment. So, again, just to summarize here I think
it can capture some behavioral and physiological data in real time which is I think really
key. It can be utilized to reduce memory issues and capture longitudinal data I think the
geospatial technology and date and time stamping are kind of really exciting things that we
can utilize. So I think from a research perspective it’s creating an environment where we can
accept these changes and adapt on the fly if you will, but doing it in a rigorous way
to ensure fidelity. And this is just another schematic defining how I think mobile health
is. So if you look from the left to the right, patients prompted to capture data and then
data stream is collected, transferred to the mobile phone, can look at geocoding, time
stamps, self-reports, so pain measurements for individuals with asthma associated with
pollen counts and other things like that. So trying to really take advantage of perhaps
some environmental issues. And then processing these and looking at trends and patterns over
time and then analyzing and visualizing the feedback. Another example that we could use
is obtaining blood pressure values, trying to remind them to do it periodically and then
we’ve created algorithms that if you get you know 3 values in a 2 week time and those are
above a certain value, we can use those to actually make changes to medication, but it
works in a research environment, we’d also have to work through some of the HIPA and
the FDA regulations where mobile is and how we define them. This is also some work from our group. Ryan
Shaw this is part of his dissertation. I think just that it’s understanding the frequency
and timing of mobile health and this may be not new to anybody, but understanding how
do we convey that information? What’s the appropriate dose? When do we send it, how
frequently? So for a trial he had done weight loss that was just posted in the American
journal of medicine. What he found was that approximately one message per day at 8am was
the most ideal time or most successful for weight loss and outcomes. So just presenting
that as well. But I think we have to be cautious of the
technology again back to a prior study that I was mentioning before. It’s called SPRY
and it’s a post-MI patients with about 400 individuals. We had about 30% were African
American, 28% were reporting inadequate income and 61 was the age. These were all individuals
who agreed to participate in the study. They all had heart attacks recently and they all
had to have accessibility to technology web or computers. Despite that 66% reported an
interest in using a website to email their doctor, 66% would share the health information
with family members, only about a quarter were interested in using a website to send
blood sugar data to their doctor. 18% were interested in reading or tracking their health
conditions. I think these data translates to those who are using a computer more, there’s
better outcomes for their A1c, LDL, or total cholesterol, as well as blood pressure. But
I think a point too is that some of our older or sicker people are less likely to be using
the mobile technology and so I think what we’re looking at is a platform so that it’s
not just simply a phone app, it’s intervoice recognition and other ways and Gary Bennett
a colleague of mine has done a nice job of looking at weight loss and the control, the
self-efficacy of feeling like you can control or decide which mode of administration was
very successful in resulting in positive weight loss. So at the end of the day, for those
of us in research, we’re really hopefully focusing on the content where the technology
is just part of the tools that we’re using. The technology is going to continue changing.
The content and getting it right is I think the essential aspect in that. So, in terms of just some recommendations,
some critical issues to think about, I mentioned this before, just depends about what your
goals are, but if you’re quickly trying to assess whether or not someone is having a
problem, this is a question we’ve used before. Have you missed any pills in the last week?
As a reasonable, the sensitivity is not great but the specificity is fine. So at the end
of the day if they say yes they’re having a problem, you can pretty much rest assured
that there’s something you need to do. Common misperceptions should be anticipated and avoided
and this may seem like a no-brainer, but I use the example of the time my son was 6 years
old and asked me what the birds and the bees were. And for anybody who has kids it’s one
of those things where you know you prepare yourself, but you’re not necessarily ready
at 6-years old to answer that question. Fortunately my wife is smarter than I am and she said
before you answer that question, why don’t you ask him what he knows already? And so
when I asked him that, he said while I was reading something about Winnie the Pooh and
those birds and the bees and it was completely off-base from where I was. And so you know
it’s really important to see where the patients are, what their thoughts are before we just
start the whole rip and roll of giving information. I could’ve emotionally scarred myself as well
as perhaps my 6 year old if I didn’t stop and do that. So these are some just quick
things to consider as we’re moving forward. Missing appointments, you know someone doesn’t
show up, to have that conversation with you, there’s no way you can have the talk about
adherence so particularly if you’re looking in a psychiatric realm. Often times we have
titrate medication it’s not that we can just simply prescribe and get it right right out
of the bat. So I think there’s some real important key reminders and clues that there’s a problem
and we need to think about it. Technology, appointment reminders, you know these are
intervoice IVR is a quick way of getting this and we can do this. Technology mobile again
90% of all text messages are opened in 30 minutes or less, so I think that’s an important
thing for us to consider and again the scalability of using that is pretty, can be very powerful. I just want to highlight, I think education
and knowledge are important but those wont by themselves change behavior. And so again
I use the concrete examples. So we, people have to have a fundamental level of knowledge
to make informed decisions, but that’s not going to get us over the finish line. WE also
have to be mindful of the ways we present information. Instructions should be clear
and structured and picture charts, color-coded, so cognitive psychologists have done a lot
of work in how do we present information and I think we really need to take better use
of those and then there’s also an industry where they have done a nice job of doing printouts
that represent patient’s medications and additive information that can be provided to the patient.
If you envision you come in you see the doctor or the pharmacist, you get your prescription,
you get your drug and then you go and then you don’t see them for 30 to 90 days we need
to be able to have something that’s a quick reference to help identify that. I think the
other thing is, you know, pill reminders and blister packages those are some other technology
that’s coming out that I think can be potentially low cost ways of helping individuals to remind
to take the medication. Literacy is always an underlying issue that we have to be mindful
of. Then again some just behavioral aspects, identifying potential relapse into old behavior,
setting appropriate realistic goals, simplifying regimens. These, you know are very straightforward
nothing probably rocket science here, but they can accumulate and have large effects.
And reinforcing positive behaviors I think that sometime that again goes back to the
main issue that we need to not forget people who are doing okay. And you know ensure that
they continue to do the right thing. And then just quickly, operant conditioning
and behavioral economics are important aspects. A lot of work now is being done in this area.
Whether it’s incentivizing with financial reimbursement or decreasing copayments and
I think that it’s really important work, but again I think it falls within the larger context
of the multiple things that we need to use to move and improve adherence. So in summary,
medication adherence is a significant but complex problem and while I focus mainly on
the patient and a little on the provider, there are multiple levels from the healthcare
provider and policy that we need to keep. The times are changing, the policy landscapes
and incentives for ensuring medication adherence are changing. Data coming out looking at what
happens if you actually provide medication for free, what is that outcome? And those
types of things need to continue going forward I think those are important. Improving collaboration
around medication adherence. So the search movement towards shared decision making is
going to be key in that area. And again still not forgetting understanding mechanisms and
contributing factors to medication adherence. I feel like in some ways we’re, if you look
at healthcare disparities there’s really 3 levels, there’s the kind of descriptive studies
where I think we’ve done a good job of showing adherence is poor and across diseases and
medications we see that there’s some variability, but we have enough to do Cochran reviews and
meta-analyses. I think we’ve done a reasonable job of understanding some of the mechanisms,
but hopefully it conveyed that I think it’s not one mechanism, there’s likely to be multiple
mechanisms depending on the individual. Where we really need to focus on is trying and examining
the interventions parsing out what’s working and then how do we implement that into the
real world? And that I think is where the new future lies for us. And lastly some other things. One size doesn’t
fit all, so whether it’s step level, or creating the toolbox with different methods to improve
outcomes for individuals. I think we need to continue to look at alternative methods
of implementing intervention. So peers I didn’t really focus on that, but in certain communities
that can be very helpful and beneficial. It doesn’t necessarily always have to be a pharmacist.
I think technology is really going to be key, but I worry that if we get caught up in the
newest gadgets, it may not get us to where we need. So I think it’s not the panacea but
it’s definitely part of the toolbox. We really need to do a better job of costing and looking
at these programs and I think better identifying who the right programs are that we can better
utilize limited resources in the healthcare system. Methods for reimbursement, so how
do we incentivize providers to have, who only have 15 minutes to see a patient to gauge
medication reconciliation, where does that happen, who’s going to cover that and those
are some policy issues that we need to continue. I’ve harked on the initiation and maintaining
of adherence. I think those are really important. And considering the multilevels at ready.
And I think lastly we always need to consider stakeholders so, while we can’t do everything
I want it’s really important to identify who the stakeholders are and engaging them not
at the end but at the very beginning so that we avoid these long drawn out periods and
so I think movements like the CORI are really key to engage and learn how to do that with
the stakeholders. Particularly the patients which is a challenge as well. How do we engage
patients as the stakeholders? Okay, so and lastly just some of the challenges.
To separate wheat from the chaff, there’s a lot of work out there. Some of it’s on the
industry side so how do we really separate the wheat from the chaff and determine what’s
working the scalability, ROI, sustainability, and insuring that we’re doing a good job of
evaluating what we’re doing. So, I’ll stop for some questions. [Dr. Nilsen] Alright, thank you Dr. Bosworth.
I think this has been a really interesting and important webinar. We do have some questions
from the audience and if you have questions you can do it on Twitter #nihadherence. It
should be on the screen. Or email them to me at [email protected] But
we have some questions from the audience and
most of these are regimen questions so in reference to HIV that you talked about, one
of our listeners said: “those regimen simplifications which is one pill a day in HIV is helpful
for improving adherence as you noticed, but an alternative like a twice a day, often improves
the outcomes, the tolerability, the long term effects. While a single pill option receives
great marketing promotion and messaging, how do we really communicate to patients which
is the best option for them?” [Dr. Bosworth] That’s a great question and
so, I’m going to now, for the future try to incorporate that into any presentation. There
is always that challenge in perhaps oversimplifying things. Where we see the data is when you
go to 3 or 4 drugs is a huge precipitous drop in adherence. There isn’t that big of a difference
between once a day and twice a day and if you looked at medication for osteoarthritis
when it’s once a month, we also have some problems where people forget. So there’s this
optimal level and I would say that twice a day is probably in the ballpark because you’re
not necessarily having to interrupt your daily schedule to take those medications. So and
that goes back to the communication. So if you’re the provider, to have that communication,
say I can provide you an easier medication, but the effectiveness is not necessarily going
to be as good as twice a day. And I think these are some of the challenges Warfarin
for example the new drugs coming out. So, Warfarin, you know, has been a treatment for
50 years and we have these new drugs that are supposed to be easier once a day. And
we don’t know if that’s necessarily going to be the right outcome. WE need to continue
looking at that. So I think just to summarize I think that twice is generally not as problematic
as 3 or 4. And it also goes back to the patient/provider communication, having those discussions. [Dr. Nilsen] Great. We have another question
about regimen in terms of factors like color. Like the color of pills, the swallowability
of pills, the size. What do we know about factors, the tangible factors in medication
adherence? Do we know anything about that? [Dr. Bosworth] I personally haven’t done much
work in that, but my look at the literature is that that definitely has some challenges.
Well, we also, from diabetes and insulin I mean, patients want any opportunity to avoid
having to insert insulin is a big issue. So that’s injections that’s slightly different.
So, the formulation of the drugs I’m sure and the swallowing is always going to be a
problem and I don’t have a quick solution for this, but again having good communications
with the provider. Often times there are alternatives and maybe there are some solutions or ways
of addressing those concerns or problems. But, again, it goes back to identifying that
that clearly is a problem and figure out a way to resolve that. [Dr. Nilsen] Alright, we had a message too
that said “what about tasting kids?”and obviously context matters. [Dr. Bosworth] Yeah, well so it’s interesting
because that’s where I think the CVS and the Pharmacies you know our local pharmacists
charge $10 to put bubble gum flavor in amoxicillin so anybody who’s tried to get a kid to take
their medication you know the challenges of it. It tastes disgusting. And there’s a business
model where they actually get a little bit of extra money to do that. And we’ve also
looked at packaging as well. So you know taking your medication, putting it in blister packaging
so that you have multiple drugs in those blister packages. We don’t know what the cost effectiveness
of this is. You know the local retail pharmacists are going to have to figure out what their
business model is. And maybe those examples where you’re kind of almost a specialty pharmacy
trying to address the size of the pills, the flavor of the pills, or that older adult who
parents who’s on so many drugs, having somebody help them and not the wife or the parent,
the child doing the medication pill reminders because we also just because we see it consistently,
that doesn’t necessarily mean that they’re putting the medications in the right pill
box correctly either. So, maybe that’s where a role of pharmacists could be. [Dr. Nilsen] Okay, great. And we had one last
question. I know we are all very interested in your ideas about life chaos. But one of
our listeners was asking “how is it modifiable? Can you give a concrete example?” [Dr. Bosworth] Well, yes. So I think it’s
a more modifiable than social demographic factors like race. I think that where we could
do is providing some strategies on how to reduce the chaos. You know, looking to a mental
health therapist and things along those lines. Trying to simplify life a little bit. I don’t
mean to be superficial, but I think there’s a little bit more room to address those issues
than some of the difficult demographic factors. So, again this comes from the HIV literature
and there has been examples where, when people are struggling with mental health and all
of the other issues and getting to appointments, scheduling and those challenges. It’s helping
to facilitate that the role of the social workers could be really very beneficial. So
I think at a minimum at this point we’re looking at it as a key factor, for a risk factor but
trying to perhaps circle back with social workers and others that are. Please, offline
if you have suggestions or other things, it’s an area that we’re just starting to look at
that I think will be important [Dr. Nilsen] Alright. Well, thank you very
much I know we all really appreciated that. That was Dr. Hayden Bosworth from Duke University
and we thank you again for joining the NIH Office of Behavioral and Social Sciences Research
distinguished speaker series and adherence. So, thank you very much have a great day and
we’ll talk to you next month. Thank you.