Poor medication adherence is common across medicine (
1). Long-acting medication formulations are a potentially powerful support for adherence, but these formulations are underutilized both for contraceptive (
2) and antipsychotic medications. Differences in rates of use of long-acting injectable (LAI) antipsychotics across countries have long been noted (
3), with the United States consistently having lower utilization than several other developed nations (
4–
7). Barriers to use of LAI antipsychotics include clinicians unnecessarily limiting their discussions of LAI formulations to patients with demonstrated nonadherence (
8), patients not receiving information about LAI antipsychotics (
9,
10), clinician concerns about the effects of LAI formulations (
5,
8,
11–
13), clinician perceptions that patients will refuse LAI administration (
14), and patient concerns about injections (
13). These insights have primarily come from surveys (
5,
6,
9,
11,
12,
15–
18), with less frequent use of focus groups (
8,
10) and direct observation of clinical medication sessions (
13,
19).
To further understand barriers to use of LAI antipsychotics, we used a different method for information gathering than prior work. We adopted a focused ethnography approach (
20), which involves detailed interviews with participants to study their shared experiences and perspectives within a particular culture and setting, such as a mental health clinic. The interviewing method was contextual inquiry, in which the interviewer introduces several discussion topics and probes the interviewee to delve deeper to identify underlying problems. In prior work on barriers to use of LAI antipsychotics, researchers obtained insights only from a limited number of interested parties (e.g., patients, psychiatrists). We expanded the range of interviewees to include patients, psychiatrists, nurses, therapists, clinic administrators, payers, national organizations, and the pharmaceutical industry. We anticipated that these methods might complement prior work by eliciting a more expansive or different range of information about barriers to use of LAI antipsychotics than previous methods.
Methods
Context
The project was a performance improvement initiative of the Care Transitions Network (CTN) for People With Serious Mental Illness, a Practice Transformation Network funded by the Centers for Medicare and Medicaid Services. The CTN provides training for member facilities. The project’s goals were to identify barriers to LAI use at CTN sites, to provide sites with information about available CTN resources relevant to identified barriers, and to determine topics for future CTN training. The CTN framework allowed the project to be conducted at multiple sites with a central team that performed all interviews and analyses. Facilities received a report presenting the interview insights from their facility and from across the network and information on CTN resources for overcoming barriers identified at the site. Reports presented solely group data to allow participants to raise issues anonymously. Data collection occurred from October 11, 2017, to February 16, 2018.
Participants
Clinical sites.
Five outpatient sites and one inpatient site in New York State participated. Sites included those with relatively low and relatively high utilization of LAI antipsychotics. Patient participants had schizophrenia-spectrum disorders and were taking antipsychotic medications. Current or past LAI use was not required. Psychiatrists, nurses, and therapists were currently treating patients with schizophrenia-spectrum disorders; there were no restrictions on the frequency that prescribers recommended LAI formulations. All clinic administrators were eligible. Participants at each site were selected by a single clinic administrator at each site. Participants were fully informed of the nature of the project, that their participation was voluntary, and that their data would be anonymized. Patient participants provided written consent and were given the option of being interviewed alone or accompanied by a family member, friend, or clinician.
Stakeholders.
To obtain insights from stakeholders, we recruited medical representatives from insurance firms, pharmaceutical companies, and a national professional organization. Stakeholder participants were required to have an awareness of LAI antipsychotics and other treatment options for psychotic disorders. Participants were assured that their responses and organizations would be confidential. The stakeholder participants were not informed of any of the activities involving the clinical sites or the sponsorship of the project. Off-site participants provided written consent for the project.
Interviews.
On-site interviews were conducted by the second, third, and fourth authors. Interviewers were not affiliated with any of the sites so that interviewees would feel comfortable expressing negative opinions. Interviews followed a semistructured format, with interviewers following an interview guide. Stakeholder interviews were conducted by telephone [for the focus of the interview questions by participant category, see online supplement]. Notes were taken during all interviews, and interviews with clinical staff were audio recorded.
Interview analysis.
Notes and audio recordings of interviews were transcribed into MindManager mind-mapping software. The interviews were organized by interviewee and question. The first step was to identify the underlying reason for a problem (i.e., the root cause) voiced by the interviewee. Root-cause analysis was performed by the second, third, and fourth authors. For each insight or comment, a series of underlying causes was documented, some voiced directly by interviewees and others uncovered through discussions within the assessment team. The team debated all of the potential root causes of an insight if there was a lack of available information. Root causes were later validated through the prioritization interviews. After isolating the root cause, a problem statement was written for each root cause in a specific syntax. (For an example of a root-cause analysis and problem statement, see online supplement.) All of the problem statements were given unique codes, and duplicated problem statements were removed.
The remaining problem statements were then rated on anchored scales that separately assessed the following domains: the problem’s impact on patient outcomes, the number of stakeholders affected, impact on the clinic’s business model, the frequency of the problem, and the degree to which current solutions existed for the problem. These ratings were combined to produce a composite prioritization score. The needs statements with prioritization scores in the top 30% were then presented in a validation exercise to three psychiatrists and three therapists who had not participated in the earlier interviews. Interviewees also provided their assessment of their current ability to address these problems. Identified needs with high priority were then grouped into themes.
Results
The 79 site interviews included 23 patients, 16 psychiatrists, three nurses, 23 therapists, and 14 clinic administrators. Stakeholder interviewees were four medical representatives of insurance companies, one representative of a pharmaceutical company, and one representative of a national professional organization. The most salient themes (themes accounting for the top 20% of need statements) contributing to the low usage of LAI antipsychotics are presented below.
Eligibility
The prevailing perception was that LAI antipsychotics are most suitable for patients who have a history of nonadherence with oral antipsychotic medications or who have a history of repeated hospitalizations. Most psychiatrists stated that they rarely bring up LAI options with patients, and they do so only if patients have a history of usage of LAI antipsychotics or significant nonadherence and resulting hospitalizations.
Most patients interviewed were unaware of LAI antipsychotics as a treatment option. A few patients knew someone using LAI antipsychotics (who were prescribed them primarily because of a history of nonadherence). In these situations, patients often reported that they observed improvement with LAI treatment. One patient said that observing improvement in others was the reason he was more open to try a LAI formulation when the doctor raised this possibility. When patients who were unaware of LAI options were presented with the features of this method of administration, many of them had a positive assessment of the features, citing specifically the reduced effort and reliance on family members or other support systems to keep track of oral medication intake.
Issues Related to Inpatient Use
Both psychiatrists and therapists stated that the best time to introduce a patient to a LAI antipsychotic is when he or she is admitted to a hospital because of a psychotic decompensation. Some patients had received short-acting injectable medications while in the hospital. The negative experience of being given a medication against their will generalized for them from short-acting injections to long-acting injections. These patients refused to use LAI antipsychotics and also informed their friends and acquaintances of their experience, thereby dissuading them from using this option as well.
Support When Making Medical Decisions
Psychiatrists reported spending 15 minutes on average with each patient. Most psychiatrists (and therapists and administrators) said that was not ideal but unavoidable with the high patient-to-psychiatrist ratio prevalent in mental health clinics. Therapists in the study reported spending 45–60 minutes with each patient. However, therapists did not engage much with patients about their medications. Many therapists were not aware of the benefits or risks of using LAIs, and a few did not even know it was a treatment for psychosis.
Patients reported not feeling emotionally supported when making a decision about starting a new medication. Overall, patients were reluctant to switch medications, and this transition became more difficult if they lacked support. Patients also perceived that they had limited choice in medication decisions. Some psychiatrists acknowledged that they do not have much time during their sessions with the patient to introduce or discuss new treatments, such as LAI antipsychotics. This shortcoming limited the patient’s access to information about LAI formulations.
Most patients said they would prefer to be counseled about new treatments by their therapist, owing to the trust and time they had with their therapist, compared with the psychiatrist. When asked, therapists did not find the patients’ responses surprising and noted that it would certainly be beneficial to discuss matters as important as medication with the most trusted professional. Therapists also noted that they are more likely to detect patient nonadherence because they spend considerably more time with patients than do psychiatrists. Some therapists also noted that some of their patients feign adherence in front of their psychiatrist because they know psychiatrists are the ones charged with making medication changes.
Psychiatrists said they did not think therapists should be having medication conversations, primarily because that is the way it has always been. However, they acknowledged that therapists are in a better position to have the conversation if the therapist is properly trained and not prescribing (only discussing features and benefits). Therapists were generally very willing and felt that medication “consulting” could be part of their continuing education. In current practice, therapists limited their role to only encouraging patients to remain adherent with their medications and communicating incidences of nonadherence to the psychiatrist.
Inter-clinician Communication
The extent of communication between therapists and psychiatrists about patients varied considerably between clinics and often within clinics as well. The official communication route was through case notes written by therapists and psychiatrists. However, therapists did not know whether psychiatrists read their notes, and some assumed that psychiatrists did not read any notes. One of therapists’ biggest concerns was that psychiatrists would not be aware of a patient’s nonadherence if they did not read the therapist’s notes. Therapists noted that if the clinicians were encouraged to talk outside of work (e.g., in break rooms), then communication regarding patients was likely to improve.
Pharmacological Profile of LAI Antipsychotics
Many psychiatrists interviewed disliked the inability to vary the length of the effect of LAI antipsychotics. They were reluctant to start a trial of an LAI medication, especially if they were not aware of how the it would affect the patient. Psychiatrists noted that all patients do not react to antipsychotic medications in the same way; furthermore, they expressed the belief that patients may react differently to the LAI formulation than they would the oral form, even if both formulations contain the same core molecule. Psychiatrists reported that performing a trial of an LAI antipsychotic was more difficult than with oral medications. Psychiatrists said that it often takes a lot of trial and error to find an antipsychotic medication that patients respond to well. The benefit of oral medications is that the trials can be as short as a few days to evaluate side-effect tolerability. Patients could experience different or worse side effects with an LAI antipsychotic. Furthermore, psychiatrists did not want to wait 1 or 2 months before being able to make dose adjustments if needed to enhance efficacy. Overall, psychiatrists did not feel comfortable giving a patient an LAI antipsychotic unless the patient had used them previously.
Infrastructure
Only two of the five outpatient clinics had infrastructure that supported LAI utilization. Barriers identified included that without a sufficient volume of LAI prescriptions, pharmacies did not want to participate in LAI delivery systems, and clinics did not want to provide nurses to administer LAI antipsychotics. Only one of the psychiatrists interviewed administered LAI antipsychotics. The others said that it was not their job to administer medication, that psychiatry is a hands-off practice, and that they intentionally do not administer medications. Nurses had concerns about minimizing the effort required of them to keep track of LAI medications.
There was considerable variability (sometimes within the same clinic) in the perceived ease or difficulty of setting up an LAI program. Some administrators assumed that setting up an LAI program could be done quickly, whereas others described a long and arduous process. Clinics interested in setting up an LAI program had few examples to follow because of the paucity of clinics with LAI programs.
Payer Insights
When assessing a new drug, payers were primarily interested in efficacy, followed by price and molecule differentiation. Most payers believed that LAI antipsychotics reduce long-term hospitalization costs through increased adherence, but the substantial cost difference between LAI and oral antipsychotic medication is a barrier to widespread coverage. Payers were not willing to “buy futures” by spending significantly more on LAI antipsychotics to potentially save on later costs. They also commented that the process of helping patients switch to more premium drugs is cumbersome, requiring significant administrative burden on both payer and provider. Some payers also blamed providers for a lack of training or interest in advocating for LAI antipsychotics. Overall, payers believed that LAI formulations have an important role in the future of antipsychotic treatment, although they anticipated the growth to remain slow (largely because of the high treatment cost).
Discussion and Conclusions
Some of the barriers to use of LAI antipsychotics from our focused ethnographic approach replicate those from prior approaches. These barriers include clinicians unnecessarily limiting their consideration of whether to use LAI formulations to patients with demonstrated nonadherence (
8), patients not receiving information about LAI antipsychotics (
9,
10), and clinician concerns about LAI administration (
5,
8,
11–
13). The reluctance of payers to support interventions with immediate costs but only later cost savings affects medicine in general (
21). We also identified barriers, discussed below, which have not been a prominent topic of prior work.
Our clinicians believed that inpatient units had logistical advantages over outpatient settings for starting LAI treatment. However, there are important considerations for LAI initiation in inpatient settings. Patients may generalize negative experiences with immediate-release injectable antipsychotics, often given against patients’ will to manage agitation, to the different context of LAI treatment. This finding suggests that inpatient facilities should consider how to introduce LAI antipsychotics to patients in ways that respect their autonomy and provide education about the differences between immediate-release injectable versus LAI antipsychotics. Outpatient facilities should also consider the need to be able to continue education about LAI antipsychotics started on inpatient services or initiate these discussions in the outpatient setting.
Shared decision making (
22) has been defined as a care delivery process in which practitioners and clients seeking help for disorders collaborate to access relevant information and to enable client-centered selection of health care resources. Our data revealed several barriers to shared decision making about LAI antipsychotics. Regarding accessing relevant information, most of our patient participants, similar to those in other studies (
9,
10), had not been informed about the existence of LAI antipsychotics. Without this fundamental knowledge about options, patients cannot begin to make informed medication decisions. Psychiatrists had little time to inform patients of options, and the patients’ most trusted source of information, their therapists, lacked the knowledge to provide meaningful input.
Possible strategies to overcome these barriers include educating therapists about the fundamentals of LAI treatment, allocating prescriber time to permit discussion of medication options, or developing innovative ways to support patient education (such as web-based approaches). Prescribers may also benefit from additional training about LAI antipsychotics in order to give accurate information to patients. One example is that our prescribers, and also those in prior studies (
5,
8,
11–
13), had concerns about LAI side effects despite meta-analyses showing that the side effects of oral and LAI formulations are overall equivalent (
23).
After information is presented, patients need support for making their medication decisions. Our patients lacked support. Our results suggest that along with prescribers and nurses, therapists and other treatment team members can have important roles in supporting decisions about LAI antipsychotics. To provide this service, therapists and other staff need support to be able to address questions about LAIs within the expertise of nonprescribers. Our results also show the need for effective communication among team members (e.g., therapists being able to communicate their early identification of patients who need support for medication adherence).
Clinics must have the infrastructure to provide LAI administration if patients are to have the opportunity to use LAI antipsychotics. Our data highlighted the logistic challenges of an LAI program and the lack of administrative expertise for meeting these challenges. Resources for administrators setting up LAI services would be beneficial.
Our data had limitations, primarily related to the performance improvement context. The project sites were limited to CTN sites; thus, the facilities participating and the individuals interviewed may not be representative of the general community. Nevertheless, the barriers we identified should be present in at least a subgroup of clinics. The project also lacked a control group and explanatory assessments of factors contributing to barriers (e.g., health knowledge measures) that would be part of research. Our results can provide the basis for hypothesis generation for future research studies.
In summary, our data suggest that for effective shared decision making about LAI antipsychotics to occur, the entire treatment team should be involved in providing the accurate information and support needed by patients to make informed decisions about LAI treatment. The training needs and administrative support requirements of all the team members should be considered to provide patients with the information and support required.
Acknowledgments
The authors acknowledge the contributions of the individuals who agreed to participate in the ethnography interviews and the staff at Clinvue, who supported the analysis of the interviews.