Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?

SCHRES-08569; No of Pages 8 Schizophrenia Research xxx (xxxx) xxx

Contents lists available at ScienceDirect

Schizophrenia Research journal homepage: www.elsevier.com/locate/schres

Assessment of adherence to oral antipsychotic medications: What has changed over the past decade? Dawn I. Velligan ⁎, Natalie J. Maples, Josie J. Pokorny, Candace Wright University of Texas Health Science Center at San Antonio, San Antonio, TX, USA

a r t i c l e

i n f o

Article history: Received 11 March 2019 Received in revised form 10 October 2019 Accepted 12 November 2019 Available online xxxx Keywords: Adherence Antipsychotic medication Adherence assessment Compliance Schizophrenia

a b s t r a c t Introduction: In a previous review, spanning 3 decades, we found that self-report and other non-objective measures were the primary means of assessing adherence to oral antipsychotic medications for individuals with schizophrenia. Moreover, consensus regarding the definition of adherence was completely lacking. Here, we examined the next decade of studies to determine what may have changed. Method: We searched the peer reviewed literature published between January 1, 2007 and December 31, 2017 using Google scholar, Science Direct, CINAHL, PsychINFO, PsychARTICLES and Medline. Search terms were medication adherence or medication compliance or medication acceptance or medication follow-through or medication concordance or medication persistence AND schizophrenia. We included articles that assessed adherence behavior. Results: The search yielded 663 articles, 363 of these were eliminated. Included studies represent over 560,000 individuals. Definitions of adherence remain variable with cutoffs from 67% to 95%. Subjective measures of adherence remain the most commonly used. However, the use of objective measures has significantly increased, as has the use of electronic claims data. However, the absolute number of studies using objective measures remains low and very few approaches identify the amount of medication actually taken. Conclusions: Some movement toward more standardization and the use of more objective measures of adherence has been made over the past decade. However, objective measures continue to be underutilized and definitions remain variable. Assessing adherence in less than optimal ways calls into question the results of studies purporting to identify reasons for problem adherence and to elucidate the relationships among adherence and other variables. © 2019 Published by Elsevier B.V.

1. Introduction Following a prescribed antipsychotic medication regimen is essential to maximizing outcomes for individuals with schizophrenia (Morant et al., 2016). Poor adherence to oral antipsychotic medications is arguably one of the most important modifiable risk factors contributing to relapse of psychotic symptoms and rehospitalization (Velligan et al., 2009a). As previous surgeon general C Everett Coop stated, “Medicines don't work in patients who don't take them (Lindenfeld and Jessup, 2017).” Particularly in an illness such as schizophrenia in which the consequences of poor adherence can cause so much emotional pain for those with the illness and their families and where the costs of caring for the illness are more than $155.7 billion annually in the US, improving adherence and outcomes is an important public health priority (Cloutier et al., 2016). ⁎ Corresponding author at: Department of Psychiatry and Behavioral Sciences, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Dr., San Antonio, TX 78229-3900, USA. E-mail address: [email protected] (D.I. Velligan).

Researchers have worked for decades to better understand the causes of problem adherence and to design and study interventions to improve the taking of oral medication (Granholm et al., 2012; Velligan et al., 2008, 2009b; Velligan and Kamil, 2014; Velligan et al., 2013). However, without a common framework including standardized definitions of adherence, non-adherence and partial adherence, as well as best practices to measure adherence behavior, a better understanding and better treatments are likely to elude us (Velligan et al., 2006, 2009a). Changing the term “adherence” to “concordance” and then to “medication interest and follow through” may have helped to promote the use of a shared decision making framework in prescribing practices and to highlight the relationship between doctor and patient as a key factor in developing a collaborative medication treatment plan (Goff et al., 2011). However, these efforts have done little to improve adherence study methodology and do not address the fact that half of all medications prescribed are not taken (Berger et al., 2012; Goff et al., 2011; Kuwabara et al., 2015). In 2006, we published a review of three decades of research examining adherence to oral medications in schizophrenia outlining the

https://doi.org/10.1016/j.schres.2019.11.022 0920-9964/© 2019 Published by Elsevier B.V.

Please cite this article as: D.I. Velligan, N.J. Maples, J.J. Pokorny, et al., Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.11.022

2

D.I. Velligan et al. / Schizophrenia Research xxx (xxxx) xxx

problems with the lack of an agreed-upon set of definitions and standards for the measurement of adherence to oral antipsychotic medications (Velligan et al., 2006). In that review of 161 studies, we found that self-report and other non-objective measures, often assessing only attitudes toward medication rather than behavior, were the primary means of evaluating adherence to oral medications, and that consensus regarding the definitions of adherence was completely lacking. Recommendations from that effort included: 1) increasing the use of objective or direct measures to assess adherence behavior, 2) using an agreed upon cutoff for defining problem adherence, which may need to be illness- or medication-specific, and 3) employing multiple measures of adherence within the same study. In 2009 consensus guidelines, experts in serious mental illness recommended when using a cutoff, greater than or equal to 80% of medication taken should be used as delineating individuals who can be considered adherent. Literature examining administrative claims for multiple illnesses provides support for selecting 0.80 as a reasonable cutoff for stratification based upon predicting subsequent hospitalization in persistent illnesses including schizophrenia (Hardy et al., 2018; Karve et al., 2009; Weiden et al., 2004). It is unclear whether studies are using these recommendations. It is unclear whether there has been an increase in the use of more objective adherence measures of adherence, since the published 2006 review. Subjective measures of adherence are based upon the report of an individual (patient, caregiver, provider) or consist of reviewing chart documentation on patient adherence or prescriptions written and integrating data to form an impression of adherence. Objective measures require the measurement of a behavior performed by the individual related to adherence (openings of a bottle or ingesting a pill with electronic data capture or filling a prescription) or involve the collection and measurement of a biological sample (e.g., plasma, urine). Objective measures include electronic claims data, pill counting, plasma levels and electronic monitors. Unless medication ingestion is being directly observed, all measures of adherence whether they are objective or subjective can either over or underestimate the adherence of specific individuals. For example, claims data can underestimate adherence if individuals have had episodes of poor adherence in the past and have stockpiled medication. Plasma levels are impacted by the behavior in the days immediately preceding the blood draw, and pills can be dumped. However, objective measures are less biased by recall, clinical state, and assumption, and are therefore thought of as standards against which to judge other measures (Velligan et al., 2006, 2009a). The current study is a systematic review of the past decade of research, beginning roughly where the former review ended. The goal of the present review is to determine whether measurement of medication adherence for patients with schizophrenia has become more objective and standardized in the past decade of published literature, and to make recommendations to advance the field.

2. Method We searched the English language peer-reviewed literature published between January 1, 2007 and December 31, 2017 using Google scholar, Science Direct, CINAHL, PsychINFO, PsychARTICLES and Medline search engines to find articles specifically measuring adherence behavior to oral antipsychotic medications. Search terms were medication adherence OR medication compliance OR medication acceptance OR medication follow-through OR medication concordance OR medication persistence AND schizophrenia. The key inclusion criterion was whether an article described a study in which medication adherence behavior (rather than attitudes only) was assessed with at least one methodology. Duplicates were removed by the search engine. Selected articles were then searched for overlapping samples. Only when sample overlap could be clearly established, the study with the higher number of participants was included in tallies. When samples

were exact duplicates with different foci but the same adherence measure, only one of the studies was included. The articles meeting all inclusion criteria were entered into a database with fields including the authors, year of publication and number of the article on the search output. In addition, each reviewer made a determination as to whether the article represented an adherence study or other type of study. An investigation was classified as an adherence study if the goal was to identify predictors of adherence to oral medication, to examine relationships between adherence and other variables, to investigate the assessment of medication adherence, or to examine the effects of a treatment specifically designed to improve medication adherence. Using these criteria, studies examining the effects of psychosocial or medication treatment on a variety of outcomes, or studies examining multiple predictors of outcome or relapse were not classified as adherence studies. This classification procedure was developed and reliably used in our previously published review (Velligan et al., 2006). Each reviewer recorded the sample size. When individuals with multiple diagnoses were included in the sample, where possible, only the number of individuals with schizophrenia/schizoaffective disorders was recorded in the database as the size of the sample studied. Moreover, when studies included some individuals taking only depot and others taking only oral, when possible, only those on oral medication were included in the sample size for our purposes. Therefore, the sample size recorded for some articles did not agree with that reflected in the original published manuscript. There were a number of studies reporting on individuals with serious mental illness on antipsychotic medication in which numbers with specific diagnoses were not reported. In these cases, the entire sample was reported. Reviewers also entered a brief narrative description of the study, the type or types of adherence measures utilized and the specific definition of adherence (e.g. cut off scores, percentage of medication taken, number of days gap). Each study was reviewed by two reviewers and disagreements for any type of classification were resolved based upon consensus. There were a total of 24 disagreements across 300 articles (5 regarding whether the study was an “adherence” study as defined; 11 regarding the sample size specific to oral antipsychotics and/or schizophrenia/schizoaffective disorder, and 8 in which one of the measures of adherence was missed (e.g., provider report (based upon selfreport and report of significant other labeled as only self-report))). Most of the last category were a result of methodology being unclear in the publication. We followed PRISMA guidelines where they were relevant for the purpose of this review. For example, we did not assess the risk bias nor methodological quality of included studies because outcomes were not a focus of this review. We were interested only in whether an assessment of adherence was collected and if so, what type of adherence assessment was used. The review was not registered but was an attempt to replicate our previous work. We defined categories of adherence measures to correspond to the previous review. Definitions and examples follow. Standard self-report - Validated scales such as the Brief Adherence Rating Scale (BARS) (Byerly et al., 2008); the Medication Adherence Rating Scale (MARS) (Thompson et al., 2000),1 the Mediation Adherence Questionnaire (Morisky et al., 2008), and the Clinician Rating Scale (CRS) (Kemp et al., 1996). (e.g., Chang et al., 2016; Montes et al., 2012). Ad Hoc self-report - Questionnaires or interviews developed by an investigator for the purposes of a particular study such as the Medication Adherence Behavior Scale (Yen et al., 2005). (e.g., Hamann et al., 2014; Yasuhiko et al., 2012).

1 There have recently been copyright issues with this self-report measure which have been detailed elsewhere “https://www.sciencemag.org/news/2017/09/pay-or-retractsurvey-creators-demands-money-rile-some-health-researchers” “https:// retractionwatch.com/2017/01/26/use-research-tool-without-permission-youll-hear/”.

Please cite this article as: D.I. Velligan, N.J. Maples, J.J. Pokorny, et al., Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.11.022

D.I. Velligan et al. / Schizophrenia Research xxx (xxxx) xxx

Electronic Self-report - A novel approach to self-report in which the participant responds to a text message or to a query from a “smart” pill container asking whether or not a specific medication was taken on a specific day or time (Granholm et al., 2012; Velligan et al., 2013). Significant other report - report of individuals close to the person including family members and other typically unpaid caregivers (Togay et al., 2015; Üçok et al., 2011). Clinician or treatment team report the opinion of a clinical provider or researcher based upon all available information. Often this included the report of the person and family members. When the article included specific sources on which the opinion of the clinician report was based, these were added as additional measures of adherence for that study (Jordan et al., 2014; Kopelowicz et al., 1998). Chart review abstraction only from the medical chart which may include prescriptions but not evidence of filling those prescriptions (Chien et al., 2015; Hilton et al., 2015). Electronic claims data. Data from payer organizations of claims made for medication prescriptions. Typically, these examine variables such as percentage of days covered by medication or mean gap in which medications were not available (Domino et al., 2014; Hansen et al., 2012). Traditional pill count - individual brings pills to the clinic to be counted (Alptekin et al., 2010). In-home pill counts - pills counted on home visits that are unannounced and/or occur at random intervals (Farooq et al., 2011; Velligan et al., 2013). Electronic monitoring - pill bottles or smart pill containers with the capability of storing and/or downloading information regarding openings and often obtaining electronic self-report of side effects or other factors (Gutiérrez-Casares et al., 2014; Misdrahi et al., 2018; Velligan et al., 2013). Plasma levels - quantify drug or active metabolites in plasma (Beebe et al., 2016; Jonsdottir et al., 2013; Velligan et al., 2007a). Urine analysis examines metabolites in urine. Tracer substances or electronic tracers - include substances to make urine fluoresce or microchips ingested with pills that monitor ingestion (Kane et al., 2013; Man et al., 2013). Direct observation using an electronic device phone or other device “watches” ingestion and may work with a computer algorithm that flags suspicious behavior (turning away, spitting out) etc. (e.g., Bain et al., 2017). All source verification - uses all data available to arrive at a percent adherence. Some articles using this approach do not say how data were combined whereas others are clear about primary measures or weighting of information (Ouellet-Plamondon et al., 2015; Weiden et al., 2012). Unknown Classification - There were some studies that did not list their methodology for measuring adherence (e.g., Baby et al., 2009; Balikci et al., 2013).

2.1. Data analysis Much of the data presented is descriptive with counts and percentages presented for the different methods used to assess adherence. In addition, we examined differences between our prior review (2006) and the current literature in terms of the proportions of use for all types of adherence assessment methods using Fisher's Exact test (Velligan et al., 2006). Each test was a 2 × 2 (i.e., 2006 review vs. present review by number of times a specific method was used or not used). In addition, we used a Fisher's Exact Test to compare the proportion of STUDIES in each review that included only subjective measures of adherence versus those that included at least one objective measure. We did not calculate statistics if there were expected frequencies lower than five. Due to multiple comparisons we selected a p value of b.01 to identify significant differences between the frequencies of specific methodologies across the two reviews.

3

3. Results The electronic database search yielded 663 articles. Of these, a total of 363 were eliminated from examination for failing to meet inclusion/exclusion criteria. Our flow diagram for selection of articles appears in Fig. 1. Due to the volume of included articles we elected not to list all references in the article, but these are available in supplemental material. We identified selected articles for illustrative purposes above. The included articles represent approximately 560,490 individuals with schizophrenia and schizoaffective disorders on oral antipsychotic medication. We say approximately because there were studies in which a diagnostic specific breakdown of the sample was not presented. Because of the number of articles, contacting authors for each identified problem would have made this review an unmanageable task. Instead, we report the final sample size as a close approximation to the number of individuals with schizophrenia/schizoaffective disorders on oral medications actually studied. The largest samples were from studies of claims data in which samples ranged from 43 to 102,884 individuals. One hundred fifty-two (152) articles were classified as adherence studies and 148 classified as more general studies. As in the 2006 review (Velligan et al., 2006), results were nearly identical when looking at articles classified as adherence papers versus those that were more general. Therefore, in the interest of space we present the analysis for the literature as a whole. 3.1. Numbers of studies assessing adherence The first notable finding is the number of studies in which adherence behavior has been measured has increased by over five- and one-half times in the past decade. In the 2006 review, there was an average of 53.6 articles per decade compared to a total of 300 in the one decade covered by the current study. 3.2. Types of measures used Fig. 2 presents the types of adherence measures and the number of times each was utilized. A total of 457 methods were used across 300 studies, with 85 studies employing more than one method. Only two studies used a specific strategy for all source verification. Only one used a novel electronic means to watch ingestion. As can be seen from the figure, the most frequently used methodologies continue to be subjective methods of assessment including the report of the person and the report of the treatment team or researcher. We compared the current findings (n = 457 methods used) with those of the earlier 2006 review (n = 258 methods used) examining the proportion of times each specific methodology was used compared to all methodologies used (n = 715). These results appear in Fig. 3. The use of electronic claims data has increased significantly, representing only 3.10% (8/258) of the methods used in publications in the three decades prior to 2006 and 12.69% (58/457) of the methods used in more recent publications (Fisher's Exact Test two-sided p b .0001). There was a strong trend suggesting provider report was used less in the past decade (13.35%; 61/457) than in the previous three decades (19.77% (51/258); Fisher's Exact Test two-sided p = .03). Similarly, there was a strong trend suggesting that the report of significant others was used less often in the recent literature (8.32%; 38/457) compared to the previous three decades (13.57% 35/258; Fisher's Exact Test twosided p = 03). The use of chart data did not differ significantly compared with the previous review (Fisher's Exact Test two-sided p = .40). There was no significant difference over time in the use of self-report (Fisher's Exact Test two-sided p = .94). However, there were five studies that utilized a novel electronic form of self-report in which individuals reported whether or not they took their medication in response to either a text message or smart pill container. This reflects self-report in real time or on the day the medication should be ingested versus reflecting

Please cite this article as: D.I. Velligan, N.J. Maples, J.J. Pokorny, et al., Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.11.022

4

D.I. Velligan et al. / Schizophrenia Research xxx (xxxx) xxx

Fig. 1. Flow diagram of studies included in the systematic review.

a longer duration (i.e., several weeks or months) typically captured by non-electronic self-report measures. The use of pill count did not differ across the reviews (Fisher's Exact Test two-sided p = .65). However, in the past decade the use of randomly timed and/or unannounced pill counts in the home was used in a total of six studies. This methodology did not appear in the previous review. There were no significant differences with respect to the use of electronic pill containers (Fisher's Exact Test two-sided p = .81), and no differences in the use of blood plasma levels (Fisher's Exact Test two-sided p = .13). Metabolites in urine were used in only one study in the 2006 review and in no studies in the current review. Tracer substances or devices were used in only one study in the earlier review. In the current study, this methodology was used once, as an electronic tracer imbedded in the pill itself. One new study used a developing technology in which a smart phone “watches” ingestion of medication and uses parameters to gauge whether suspicious behavior occurred that raise the possibility that medication was not taken (Kane et al., 2013). In the current review, there were four studies in which the method of assessing adherence could not be conclusively determined from the published article. The majority of studies using multiple methodologies employed two nonobjective measures. However, there were significantly fewer articles in the current review using multiple methodologies; only 28.67% (86/ 300), compared to the earlier review in which 40.37% (65/161) of the studies employed multiple adherence assessments. This difference was significant (Fisher's Exact Test two-sided p = .01). 3.3. Objective versus subjective measures and their combination We were specifically interested in comparing the current review with studies examined in the 2006 review with respect to the reliance on subjective measures of adherence. Both the previous review and the expert consensus guidelines recommended an increased use of objective assessments. Fig. 4 presents the proportion of studies using

subjective only methods across the two reviews. The remainder of studies used objective measures either alone or in combination with subjective measures. Results indicate a significant decrease in the use of subjective only measures in studies assessing adherence. In the previous review 80.7% (130/161) of articles used subjective only methods while in the current review, 67.3% (202/300) of articles used subjective only measures (Fisher's Exact Test two-sided p = .0023). 3.4. Variability in definitions of adherence Definitions of adherence found in the current literature appear in Table 1. Some definitions were qualitative with no anchors (e.g., “extent to which client agrees to take medication and will take it freely”). Adherence was often dichotomized but in very different ways across studies. Ordinal scales for categorizing adherence included everything from 3-point to 10-point scales. These were anchored differently and often dichotomized differently for different studies. For example, scores on the scale developed by Morisky et al. (2008; Medication Adherence Questionnaire) used dissimilar cut off scores in different studies. Different versions of this scale exist as well (4 items versus 10). Some scales were used in total when only a few items represent adherence behavior (e.g., PETIT). There were also multiple percentage cutoffs used for continuous data to dichotomize adherent versus adherent patients. These ranged from 67% to 90%. Moreover, there were inconsistent methods used to define partial adherence, breaking the continuum at different cut points (i.e., 50%–80% and 30%–89). Typically, there was no validation of this methodology or explanation for the selection of cutoffs. 4. Discussion The present study examined the last decade of research to determine the types of adherence assessments utilized in studies purporting

Please cite this article as: D.I. Velligan, N.J. Maples, J.J. Pokorny, et al., Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.11.022

D.I. Velligan et al. / Schizophrenia Research xxx (xxxx) xxx

5

Number of mes (percent) Adherence Methodolgies were used accross 300 publicaons from 2007-2017 out of 457 total methods 140

116 (25.4%) 120

100

80

58 (12.7%)

71 (15.5%)

61 (13.3%) 60

35 (7.7%) 38 (8.3%)

40

27 (5.9%) 19 (4.2%) 1 (0.2%) 5 (1.1%) 13 (2.8%) 6 (1.3%) 2 (0.4%) 1 (0.2%)

20

0

0

Fig. 2. Number of times adherence methodologies were used across 300 publications from 2007 to 2017.

the past decade to a report published in 2006 examining the previous three decades of adherence measurement. Results of the present review indicate that self-report, remains the most common method of assessing adherence to oral antipsychotics in

to measure adherence to oral antipsychotics and the consistency of definitions. The goal was to determine how the field has progressed in the past decade and whether there was greater use of objective adherence assessment and more standardization. We compared the literature of

Proporon of mes Various Adherence Assessments were used in Published Studies1 50 45

41.47 42.01

40 35 30 25 20 15

19.77

+

13.35

13.57

+

8.32

* 12.69

9.69

7.66

10

6.2 7.22 1.94 4.16

3.1

2.33 2.84

5 0.2 0 0 Self Report Treatment Significant Team Other

Chart

Electronic Pillcount Claims Data

Previous 3 Decades 2006 review

Plasma Levels

Urine

Electronic pill containers

2007-2017

Fig. 3. Proportion of various adherence measures in published studies.

Please cite this article as: D.I. Velligan, N.J. Maples, J.J. Pokorny, et al., Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.11.022

6

D.I. Velligan et al. / Schizophrenia Research xxx (xxxx) xxx

Proporon of Studies using Subjecve Only Methods for Assessing Adherence Across Decades1 100

80.7

90 80

*

Proporon of Studies

70

67.3

60 50 40 30 20 10 0

2006 Review of 3 previous decades

2007-2017

Fig. 4. Proportion of studies using subjective only methods for assessing adherence across decades.

the published literature. Use of both the report of the treatment team and of significant others has decreased in the recent decade although the difference was only at the trend level. There has been an increase in the use of electronic claims data as well as an increase in the use of more objective measures overall. Definitions of adherence remain numerous and not standardized. Continued use of self-report is concerning given evidence demonstrating that self-report measures underestimate adherence (e.g., Byerly et al., 2007; Stephenson et al., 2012; Velligan et al., 2007b). Its use reflects that self-report remains the easiest measure to obtain. In several of our own previous studies, we asked individuals to rate their adherence on the BARS after they watched research staff counting their pills on randomly scheduled, unannounced home visits. Participants, nevertheless, rated themselves as fully adherent almost 90% of the time, although pill count data revealed adherence rates closer to 60%. This may be because the individuals who participated in these adherence studies generally intended to take their medication as prescribed but were not consistently effective in doing so (Velligan et al., 2008). Interestingly, the accuracy self-reported medication adherence may depend upon the timeframe covered and the methodology used to collect the data. Novel strategies that collect self-report in response to a daily text message or to queries made after opening a smart pill device are being used; albeit infrequently. Self-report captured through electronic means across short time frames has been found to be more highly correlated with pill counts than self-report measures asking about a time span of two weeks (Velligan et al., 2007b, 2013). Most individuals can accurately recall whether or not they have taken their medication in a given day. The problem with using such strategies in studies

Table 1 Multiple definitions of adherence across studies 2007–2017. “Extent to which client agrees to take medication and will take it freely.” Adherent versus non-adherent groups with no report of how this was defined Not taking medication as prescribed (no further description) “Yes” or “no” to the question as to whether medication was taken on a specific day Gap greater than or = to 3 months No claim for an antipsychotic in the past 30 days 3–10 doses missed in a month Failure to take it for 14 consecutive days 3-point; 4-point; 5-point; 6-point, 10-point categorical scales from never took to never missed or completely non-adherent to completely adherent. All different anchors. Many of these then dichotomized, usually with no rationale for the cut point selected. Plasma level was not in “expected” range Percentage cutoff for non-adherence included 90%, 80%, 75%, 70% and 67%.

examining level of adherence or treatments to improve level of adherence is that messages and devices also prompt the taking of medication. Whether this strategy is appropriate largely depends on the goals of a particular study. If medication is a baseline to which a psychosocial treatment is added, ensuring adherence with medication to the extent possible may be important. Certainly, in industry studies, the common use of pill counts at scheduled visits only identifies adherence “afterthe-fact.” It would be far more useful in such studies to prompt adherence with technologies or other methods to limit the number of missed doses. The increase in the use of electronic claims data represents an important advancement in the field. These studies typically include large numbers of beneficiaries from public or private payer systems. Research has demonstrated that gaps in which medications are not available increase the risk of hospitalization (Valenstein et al., 2002). It is important to keep in mind, however, that as with nearly all other methods of determining adherence to oral medications, refilling prescriptions does not indicate whether or not medication has been ingested. The decreased use of provider ratings of adherence in the more recent literature is supported by data. Multiple studies suggest that providers are poor at accurately identifying individuals who are nonadherent (Velligan et al., 2007b, 2009a). In the past decade, compared to the three previous decades, there is a decreased reliance on using only subjective measures for the assessment of oral antipsychotic adherence. Instead, studies are including more objective assessments such as pill counts, plasma levels, electronic claims data, electronic pill containers that record time and date of openings, and novel strategies capable of identifying ingestion of medication. While this is consistent with recommendations, the use of objective versus subjective measures remained relatively low over the past decade. As technologies advance, the use of these more direct and objective assessments is likely to increase. The use of a microchip in a pill detects actual ingestion and requires less use of research staff time than many objective approaches. Training of the participant is required, and use requires some comfort with technology. It is conceivable that such an approach may impact those with paranoia or health concerns. This approach may initially be cost-prohibitive for some studies. It is possible for patches to be removed by the participant. A novel smartphone platform is available that “watches” and records ingestion and uses computer software to ensure the correct pill is taken and swallowed (Bain et al., 2017). This approach requires less staff time than some objective approaches. Training of the participant is necessary, and devices could be lost or stolen. Initially this method may be cost-prohibitive for some studies.

Please cite this article as: D.I. Velligan, N.J. Maples, J.J. Pokorny, et al., Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.11.022

D.I. Velligan et al. / Schizophrenia Research xxx (xxxx) xxx

Slightly less techy options are available for direct observation of medication ingestion. For example, the use of applications such as face-to-face video available on many smart phones can allow a researcher to watch medication taking live and document whether or not ingestion occurred. The technology is readily available on phones possessed by large numbers of potential research participants and is inexpensive. However, using devices in this way may raise privacy concerns. However, since these platforms do not store Protected Health Information (PHI), they may meet the criteria for the HIPPA conduit exception rule (HIPPA Journal, 2018) https://www.hipaajournal.com/ hipaa-conduit-exception-rule/. This rule allows practitioners to use technologies that act as conduits. Conduits do not store PHI and can therefore be used without special business agreements that are needed for platforms that keep PHI. However, the rules regarding privacy are not clear on this point at this time. In general, these available technologies are relatively new and not well represented in the literature for the past decade. Definitions of adherence continue to vary with cutoffs to define adherent individuals ranging from 67% to 90%. This is despite expert consensus that 80% represents a reasonable adherence cutoff and should be consistently used in studies examining adherence to oral antipsychotics. More consistent with previous guidelines, all studies that dichotomized patients based upon electronic claims data used a cutoff of 80% to identify adherent patients. With standardized scales, different cutoffs were used to identify adherent versus non adherent participants. There continues to be numerous ordinal scales of differing ranges. The consequences of inadequate adherence measurement and a poor application of consensus cutoffs for identifying adherent individuals impacts the conclusions that can be drawn from research where adherence is measured. Much of the research describing treatment efficacy and effectiveness, reasons for problem adherence and the relationships among adherence and other variables (e.g., hospitalization, relapse, quality of life), is based on work that reflects these considerable problems with methodology. Conservative interpretation of these studies is important. The current review also had limitations. While agreement was good among raters, there is always room for error when judgement is involved. There are many non-English studies addressing adherence that were not included. Studies reviewed were necessarily limited by our search terms. We did not evaluate the methodological rigor of the studies examined. Recommendations from this review include further increasing the use of objective and direct methods to address adherence. We recommend that whenever self-report is utilized, it is used in conjunction with an objective measure of adherence where possible. This is supported by a recent study demonstrating very low concordance between adherence measures and substantially higher non-adherence with objective measures (Xu et al., 2018). We recommend that when reporting the use of an objective adherence measure, authors also report the details regarding how the measure was obtained. For example, for home-based pill counts, how initial counts are made, what happens when additional bottles turn up on subsequent counts, etc. This information can often be supplied in supplemental material. We strongly recommend increasing the use of digital technologies that can verify ingestion of medication where feasible. At present, some of these technologies are likely to be expensive. Other technologies that are less expensive, such as face-to-face video via smart phone, reside in a grey area in terms of compliance with HIPAA regulations. HIPPA guidelines for the use of face-to-face video platforms need to be clarified before practitioners and compliance officers will be comfortable supporting the use of these approaches to assessing adherence (US Department of Health and Human Services, 2013). The fact that some technologies both prompt and measure adherence can be a problem depending upon study design. There is no single, gold-standard for the assessment of adherence behavior. The measure chosen must be appropriate to the context of the study. However, objective measures such as digital pills,

7

electronic devices, tracer substances, and video verification are becoming possible options for a range of “gold-standard” approaches moving forward because they come closest to verifying ingestion. We further recommend use of the standardized cutoff of 80% for identifying those who are adherent to medication as laudable goal for the next decade. Multiple studies cited above support this cutoff. We recommend that percent adherence and the numbers of individuals meeting the cutoff for adherence be reported in studies whether or not these variables represent primary outcomes. The time frame that covers the adherence assessment should be well specified. We recommend clarification of the term “partial adherence”. Percentage of medication taken does not clarify the pattern of partial adherence. A person may take medication intermittently or may take a lower dose daily. These have different implications for intervention as well as for health belief which are purported to be related to adherence. Future studies may need to address these different meanings of partial adherence. Contributors Dr. Velligan wrote the manuscript. Co-authors Maples, Pokorny and Wright did the lit review and read the journal articles and selected the relevant articles for the review manuscript. Role of funding source This review has no funding to report. Declaration of competing interest The authors have no conflicts of interest to report. Acknowledgments There are no acknowledgements.

Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.schres.2019.11.022. References Alptekin, K., Aydin, H., Gucer, K.M., Dilbaz, N., Goka, E., Esendemir, C., Oguz, H., Arabul, G., Aslaner, F., Ayer, A., Celiker, A.R., Cirit, H., Pinar, Z., Unal, A., 2010. Patient adherence and efficacy of quetiapine treatment in schizophrenia: results of a multicentre, naturalistic 6-month follow-up study. Int. Clin. Psychopharmacol. 25 (6), 342–348. Baby, R.S., Gupta, S., Sagar, R., 2009. Attitudes and subjective reasons of medication compliance and noncompliance among outpatients with schizophrenia in India. Internet J. Epidemiol. 7. Bain, E.E., Shafner, L., Walling, D.P., Othman, A.A., Chuang-Stein, C., Hinkle, J., Hanina, A., 2017. Use of a novel artificial intelligence platform on mobile devices to assess dosing compliance in a phase 2 clinical trial in subjects with schizophrenia. JMIR mHealth and uHealth 5 (2), e18. Balikci, A., Erdem, M., Zincir, S., Bolu, A., Zincir, S.B., Ercan, S., Uzun, O., 2013. Adherence with outpatient appointments and medication: a two-year prospective study of patients with schizophrenia. Bull. Clin. Psychopharmacol. 23 (1), 57–64. Beebe, L.H., Smith, K., Phillips, C., 2016. Descriptions and correlates of medication adherence, attitudes, and self-efficacy in outpatients with schizophrenia spectrum disorders (SSDs). Arch. Psychiatr. Nurs. 30 (3), 400–405. Berger, A., Edelsberg, J., Sanders, K.N., Alvir, J.M.J., Mychaskiw, M.A., Oster, G., 2012. Medication adherence and utilization in patients with schizophrenia or bipolar disorder receiving aripiprazole, quetiapine, or ziprasidone at hospital discharge: a retrospective cohort study. BMC Psychiatry 12, 99. Byerly, M.J., Thompson, A., Carmody, T., Bugno, R., Erwin, T., Kashner, M., Rush, A.J., 2007. Validity of electronically monitored medication adherence and conventional adherence measures in schizophrenia. Psychiatr. Serv. 58 (6), 844–847. Byerly, M.J., Nakonezny, P.A., Rush, A.J., 2008. The Brief Adherence Rating Scale (BARS) validated against electronic monitoring in assessing the antipsychotic medication adherence of outpatients with schizophrenia and schizoaffective disorder. Schizophr. Res. 100, 60–69. Chang, W.C., Kwong, V.W.Y., Chan, G.H.K., Jim, O.T.T., Lau, E.S.K., Hui, C.L.M., Chan, S.K.W., Lee, E.H.M., Chen, E.Y.H., 2016. Prediction of functional remission in first-episode psychosis: 12-month follow-up of the randomized-controlled trial on extended early intervention in Hong Kong. Schizophr. Res. 173 (1), 79–83. Chien, W.T., Mui, J.H., Cheung, E.F., Gray, R., 2015. Effects of motivational interviewingbased adherence therapy for schizophrenia spectrum disorders: a randomized controlled trial. Trials 16 (1), 270. Cloutier, M., Aigbogun, M.S., Guerin, A., Nitulescu, R., Ramanakumar, A.V., Kamat, S.A., DeLucia, M., Duffy, R., Legacy, S.N., Henderson, C., Francois, C., Wu, E., 2016. The

Please cite this article as: D.I. Velligan, N.J. Maples, J.J. Pokorny, et al., Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.11.022

8

D.I. Velligan et al. / Schizophrenia Research xxx (xxxx) xxx

economic burden of schizophrenia in the United States in 2013. J. Clin. Psychiatry 77 (6), 764–771. Domino, M.E., Beadles, C.A., Lichstein, J.C., Farley, J.F., Morrissey, J.P., Ellis, A.R., Dubard, C.A., 2014. Heterogeneity in the quality of care for patients with multiple chronic conditions by psychiatric comorbidity. Med. Care 52 (Suppl. 3), S101–S109. Farooq, S., Nazar, Z., Irfan, M., Akhter, J., Gul, E., Irfan, U., Naeem, F., 2011. Schizophrenia medication adherence in a resource-poor setting: randomised controlled trial of supervised treatment in out-patients for schizophrenia (STOPS). Br. J. Psychiatry 199 (6), 467–472. Goff, D.C., Hill, M., Freudenreich, O., 2011. Treatment adherence in schizophrenia and schizoaffective disorder. J. Clin. Psychiatry 72 (4), e13. Granholm, E., Ben-Zeev, D., Link, P.C., Bradshaw, K.R., Holden, J.L., 2012. Mobile assessment and treatment for schizophrenia (MATS): a pilot trial of an interactive textmessaging intervention for medication adherence, socialization, and auditory hallucinations. Schizophr. Bull. 38 (3), 414–425. Gutiérrez-Casares, J.R., Cañas, F., Alonso-Escolano, D., Hidalgo-Borrajo, R., RodríguezMorales, A., 2014. Adherence to treatment and therapeutic strategies in schizophrenic patients: the ADHERE study. CNS Spectr 15 (5), 327–338. Hamann, J., Lipp, M.L., Christ-Zapp, S., Spellmann, I., Kissling, W., 2014. Psychiatrist and patient responses to suspected medication nonadherence in schizophrenia spectrum disorders. Psychiatr. Serv. 65 (7), 881–887. Hansen, R.A., Maciejewski, M., Yu-Isenberg, K., Farley, J.F., 2012. Adherence to antipsychotics and cardiometabolic medication: association with health care utilization and costs. Psychiatr. Serv. 63 (9), 920–928. Hardy, M., Jackson, C., Byrne, J., 2018. Antipsychotic adherence and emergency department utilization among patients with schizophrenia. Schizophr. Res. 201, 347–351. Hilton, N.Z., Ham, E., Lang, C., Harris, G.T., 2015. Weight gain and its correlates among forensic inpatients. Can. J. Psychiatr. 60 (5), 232–238. HIPPA Journal, 2018. The HIPAA conduit exception rule and transmission of PHI. 1/19/ 2018 ed. https://www.hipaajournal.com/hipaa-conduit-exception-rule/. Jonsdottir, H., Opjordsmoen, S., Birkenaes, A.B., Simonsen, C., Engh, J.A., Ringen, P.A., Vaskinn, A., Friis, S., Sundet, K., Andreassen, O.A., 2013. Predictors of medication adherence in patients with schizophrenia and bipolar disorder. Acta Psychiatr. Scand. 127 (1), 23–33. Jordan, G., Lutgens, D., Joober, R., Lepage, M., Iyer, S.N., Malla, A., 2014. The relative contribution of cognition and symptomatic remission to functional outcome following treatment of a first episode of psychosis. J. Clin. Psychiatry 75 (6), e566–e572. Kane, J.M., Perlis, R.H., DiCarlo, L.A., Au-Yeung, K., Duong, J., Petrides, G., 2013. First experience with a wireless system incorporating physiologic assessments and direct confirmation of digital tablet ingestions in ambulatory patients with schizophrenia or bipolar disorder. J. Clin. Psychiatry 74 (6), e533–e540. Karve, S., Cleves, M.A., Helm, M., Hudson, T.J., West, D.S., Martin, B.C., 2009. Prospective validation of eight different adherence measures for use with administrative claims data among patients with schizophrenia. Value Health 12 (6), 989–995. Kemp, R., David, A., Hayward, P., 1996. Compliance therapy: an intervention targeting insight and treatment adherence in psychotic patients. Beh. Cogn. Psychother. 24 (4), 331–350. Kopelowicz, A., Wallace, C.J., Zarate, R., 1998. Teaching psychiatric inpatients to re-enter the community: a brief method of improving the continuity of care. Psychiatr. Serv. 49 (10), 1313–1316. Kuwabara, H., Saito, Y., Mahlich, J., 2015. Adherence and rehospitalizations in patients with schizophrenia: evidence from Japanese claims data. Neuropsychiatr. Dis. Treat. 11, 935–940. Lindenfeld, J., Jessup, M., 2017. ‘Drugs don't work in patients who don't take them’ (C. Everett Koop, MD, US Surgeon General, 1985). Eur. J. Heart Fail. 19 (11), 1412–1413. Man, W.H., Ten Berg, M., Wilting, I., Huisman, A., Cahn, W., Douma, J.W., den Breeijen, H., Heerdink, E.R., Egberts, T., van Solinge, W., 2013. Fluorescence of neutrophil granulocytes as a biomarker for clozapine use. Eur. Neuropsychopharmacol. 23 (11), 1408–1413. Misdrahi, D., Tessier, A., Husky, M., Lange, A.C., Vrijens, B., Llorca, P.M., Bayle, F.J., 2018. Evaluation of adherence patterns in schizophrenia using electronic monitoring (MEMS®): a six-month post-discharge prospective study. Schizophr. Res. 193, 114–118. Montes, J.M., Medina, E., Gomez-Beneyto, M., Maurino, J., 2012. A short message service (SMS)-based strategy for enhancing adherence to antipsychotic medication in schizophrenia. Psychiatry Res. 200 (2–3), 89–95. Morant, N., Kaminskiy, E., Ramon, S., 2016. Shared decision making for psychiatric medication management: beyond the micro-social. Health Expect. 19 (5), 1002–1014. Morisky, D.E., Ang, A., Krousel-Wood, M., Ward, H.J., 2008. Predictive validity of a medication adherence measure in an outpatient setting. J. Clin. Hypertens. 10 (5), 348–354.

Ouellet-Plamondon, C., Rousseau, C., Nicole, L., Abdel-Baki, A., 2015. Engaging immigrants in early psychosis treatment: a clinical challenge. Psychiatr. Serv. 66 (7), 757–759. Stephenson, J.J., Tuncelli, O., Gu, T., Eisenberg, D., Panish, J., Crivera, C., Dirani, R., 2012. Adherence to oral second-generation antipsychotic medications in patients with schizophrenia and bipolar disorder: physicians' perceptions of adherence vs. pharmacy claims. Int. J. Clin. Pract. 66 (6), 565–573. Thompson, K., Kulkarni, J., Sergejew, A.A., 2000. Reliability and validity of a new Medication Adherence Rating Scale (MARS) for the psychoses. Schizophr. Res. 42 (3), 241–247. Togay, B., Noyan, H., Tasdelen, R., Ucok, A., 2015. Clinical variables associated with suicide attempts in schizophrenia before and after the first episode. Psychiatry Res. 229 (1), 252–256. Üçok, A., Serbest, S., Kandemir, P.E., 2011. Remission after first-episode schizophrenia: results of a long-term follow-up. Psychiatry Res. 189 (1), 33–37. US Department of Health & Human Services, 2013. Can a CSP be considered to be a “conduit” like the postal service, and, therefore, not a business associate that must comply with the HIPAA Rules? https://www.hhs.gov/hipaa/for-professionals/faq/2077/cana-csp-be-considered-to-be-a-conduit-like-the-postal-service-and-therefore-not-abusiness%20associate-that-must-comply-with-the-hipaa-rules/index.html Valenstein, M., Copeland, L.A., Blow, F.C., McCarthy, J.F., Zeber, J.E., Gillon, L., Bingham, C.R., Stavenger, T., 2002. Pharmacy data identify poorly adherent patients with schizophrenia at increased risk for admission. Med. Care 40 (8), 630–639. Velligan, D.I., Kamil, S.H., 2014. Enhancing patient adherence: introducing smart pill devices. Ther. Deliv. 5 (6), 611–613. Velligan, D.I., Lam, Y.W., Glahn, D.C., Barrett, J.A., Maples, N.J., Ereshefsky, L., Miller, A.L., 2006. Defining and assessing adherence to oral antipsychotics: a review of the literature. Schizophr. Bull. 32 (4), 724–742. Velligan, D.I., Diamond, P.M., Lopez, J., Castillo, D.A., Maples, N.J., Lam, F., Ereshefsky, L., Miller, A.L., 2007a. Cognitive Adaptation Training improves adherence to medication and functional outcome in schizophrenia. Schizophr. Bull. 33 (2), 608. Velligan, D.I., Wang, M., Diamond, P., Glahn, D.C., Castillo, D., Bendle, S., Lam, Y.W., Ereshefsky, L., Miller, A.L., 2007b. Relationships among subjective and objective measures of adherence to oral antipsychotic medications. Psychiatr. Serv. 58 (9), 1187–1192. Velligan, D.I., Diamond, P.M., Mintz, J., Maples, N., Li, X., Zeber, J., Ereshefsky, L., Lam, Y.W., Castillo, D., Miller, A.L., 2008. The use of individually tailored environmental supports to improve medication adherence and outcomes in schizophrenia. Schizophr. Bull. 34 (3), 483–493. Velligan, D.I., Weiden, P.J., Sajatovic, M., Scott, J., Carpenter, D., Ross, R., Docherty, J.P., 2009a. The expert consensus guideline series: adherence problems in patients with serious and persistent mental illness. J. Clin. Psychiatry 70 (S4), 1–51. Velligan, D.I., Diamond, P., Mueller, J., Li, X., Maples, N., Wang, M., Miller, A.L., 2009b. The short-term impact of generic versus individualized environmental supports on functional outcomes and target behaviors in schizophrenia. Psychiatry Res. 168 (2), 94–101. Velligan, D.I., Mintz, J., Maples, N., Li, X., Gajewski, S., Carr, H., Sierra, C., 2013. A randomized trial comparing in person and electronic interventions for improving adherence to oral medication in schizophrenia. Schizophr. Bull. 39 (5), 999–1007. Weiden, P.J., Kozma, C., Grogg, A., Locklear, J., 2004. Partial compliance and risk of rehospitalization among California Medicaid patients with schizophrenia. Psychiatr. Serv. 55 (8), 886–891. Weiden, P.J., Schooler, N.R., Weedon, J.C., Elmouchtari, A., Sunakawa-McMillan, A., 2012. Maintenance treatment with long-acting injectable risperidone in first-episode schizophrenia: a randomized effectiveness study. J. Clin. Psychiatry 73 (9), 1224–1233. Xu, D.R., Gong, W., Gloyd, S., Caine, E.D., Simoni, J., Hughes, J.P., Xiao, S., He, W., Dai, B., Lin, M., Nie, J., He, H., 2018. Measuring adherence to antipsychotic medications for schizophrenia: concordance and validity among a community sample in rural China. Schizophr. Res. 201, 307–314. Yasuhiko, H., Junji, U., Takaichi, M., Masaaki, K., Hiroaki, T., Masami, T., 2012. Effects of antipsychotic polypharmacy on side-effects and concurrent use of medications in schizophrenic outpatients. Psychiatry Clin. Neurosci. 66 (5), 405–410. Yen, C.F., Chen, C.S., Ko, C.H., Yeh, M.L., Yang, S.J., Yen, J.Y., Huang, C.F., Wu, C.C., 2005. Relationships between insight and medication adherence in outpatients with schizophrenia and bipolar disorder: prospective study. Psychiatry Clin. Neurosci. 59 (4), 403–409.

Please cite this article as: D.I. Velligan, N.J. Maples, J.J. Pokorny, et al., Assessment of adherence to oral antipsychotic medications: What has changed over the past decade?, Schizophrenia Research, https://doi.org/10.1016/j.schres.2019.11.022