A systematic review of service-user reasons for adherence and nonadherence to neuroleptic medication in psychosis

A systematic review of service-user reasons for adherence and nonadherence to neuroleptic medication in psychosis

Clinical Psychology Review 51 (2017) 75–95 Contents lists available at ScienceDirect Clinical Psychology Review journal homepage: www.elsevier.com/l...

585KB Sizes 6 Downloads 67 Views

Clinical Psychology Review 51 (2017) 75–95

Contents lists available at ScienceDirect

Clinical Psychology Review journal homepage: www.elsevier.com/locate/clinpsychrev

Review

A systematic review of service-user reasons for adherence and nonadherence to neuroleptic medication in psychosis Miriam Wade a, Sara Tai b,*, Yvonne Awenat b, Gillian Haddock b a

Division of Psychology and Mental Health, School of Health Sciences, Faculty of Biology, Medicine and Health, 2nd Floor Zochonis Building, Brunswick Street, University of Manchester, M13 9PL United Kingdom University of Manchester, Manchester, United Kingdom

b

H I G H L I G H T S • Service-users with psychosis often discontinue neuroleptic medication. • Service-user reasons for adherence and non-adherence are reviewed. • Several themes of reasons are identified across studies of differing methodologies.

a r t i c l e

i n f o

Article history: Received 20 June 2016 Received in revised form 9 September 2016 Accepted 26 October 2016 Available online 28 October 2016 Keywords: Neuroleptic medication Antipsychotic medication Adherence Non-adherence Psychosis Schizophrenia

a b s t r a c t People diagnosed with psychosis, such as those with schizophrenia-related disorders, are routinely prescribed neuroleptic medication as a primary treatment. Despite reported benefits of neuroleptic treatment for symptom remission and relapse prevention, discontinuation rates are high. Research examining factors associated with neuroleptic non-adherence report inconsistent findings. Reasons for adherence to neuroleptic medication are under-researched. The current review aimed to synthesise evidence exploring service-user self-reported reasons for adherence and non-adherence to neuroleptic medication. A systematic literature search of databases and reference list searching identified 21 studies investigating service-user accounts of reasons for adherence and/or non-adherence to neuroleptic medication. Qualitative, quantitative and mixed-method studies were included in the review. Several themes of reasons were identified. Reasons for both adherence and non-adherence were largely similar; medication efficacy, compatibility with personal medication or religious beliefs, side-effects and the influence of relationships with other people. Experiences of stigma and economic difficulties were generally identified as reasons for non-adherence only while experiences of fear and coercion were identified as reasons for adherence only. The review identified crucial factors which may aid service providers in bettering treatment for people with psychosis and will provide evidence which could contribute to future prescribing guidelines. © 2016 Elsevier Ltd. All rights reserved.

Contents 1. 2.

3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 1.1. Aims . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Search strategy . . . . . . . . . . . . . . . . . . . . . 2.2. Eligibility criteria. . . . . . . . . . . . . . . . . . . . . 2.3. Quality appraisal . . . . . . . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Study characteristics . . . . . . . . . . . . . . . . . . . 3.2. Measurement of adherence and non-adherence perspectives .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

. . . . . . . . .

76 77 77 77 78 78 78 78 79

⁎ Corresponding author at: Dr. Sara Tai Senior Lecturer in Clinical Psychology & Hon. Consultant Clinical Psychologist School of Psychological Sciences, The University of Manchester, Coupland I Building, Oxford Road, Manchester M13 9PL UK. E-mail addresses: [email protected] (M. Wade), [email protected] (S. Tai).

http://dx.doi.org/10.1016/j.cpr.2016.10.009 0272-7358/© 2016 Elsevier Ltd. All rights reserved.

76

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

3.3. 3.4. 3.5. 3.6.

Participant characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . Overall quality assessment . . . . . . . . . . . . . . . . . . . . . . . . . . Data extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.6.1. Reasons for adherence and non-adherence to neuroleptic medication . . . 3.7. Medication efficacy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.1. Quantitative findings . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.2. Qualitative findings. . . . . . . . . . . . . . . . . . . . . . . . . . 3.7.3. Benefits associated with symptoms - quantitative findings . . . . . . . . 3.8. Personal beliefs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8.1. Medication beliefs . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8.2. Religious beliefs . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.9. Relationships with other people . . . . . . . . . . . . . . . . . . . . . . . . 3.9.1. Quantitative findings . . . . . . . . . . . . . . . . . . . . . . . . . 3.9.2. Qualitative findings. . . . . . . . . . . . . . . . . . . . . . . . . . 3.10. Fear and coercion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.10.1. Quantitative findings . . . . . . . . . . . . . . . . . . . . . . . . 3.10.2. Qualitative findings . . . . . . . . . . . . . . . . . . . . . . . . . 3.11. Side-effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.11.1. Quantitative findings . . . . . . . . . . . . . . . . . . . . . . . . 3.11.2. Qualitative findings . . . . . . . . . . . . . . . . . . . . . . . . . 3.12. Stigma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.12.1. Quantitative findings . . . . . . . . . . . . . . . . . . . . . . . . 3.12.2. Qualitative findings . . . . . . . . . . . . . . . . . . . . . . . . . 3.13. Economic issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.13.1. Quantitative and qualitative findings . . . . . . . . . . . . . . . . . 3.14. Other reasons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.14.1. Quantitative and qualitative findings . . . . . . . . . . . . . . . . . 3.15. Subgroup differences . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.15.1. Participants from low-income countries . . . . . . . . . . . . . . . . 3.15.2. Cultural differences in the influence of family and professionals . . . . . 3.15.3. Further subgroup differences . . . . . . . . . . . . . . . . . . . . . 4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. What reasons do service-users report for non-adherence to neuroleptic medication? 4.2. What reasons do service-users report for adherence to neuroleptic medication? . . 4.3. Are there any sub-group differences in reasons for adherence and non-adherence? . 4.4. Views on illness causation, recovery and treatment priorities . . . . . . . . . . . 4.5. The importance of relationships . . . . . . . . . . . . . . . . . . . . . . . . 4.6. Limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.6.1. Less reporting of non-adherence reasons . . . . . . . . . . . . . . . . 4.6.2. Non-representative participants . . . . . . . . . . . . . . . . . . . . 4.6.3. Measurement differences . . . . . . . . . . . . . . . . . . . . . . . 4.6.4. Conclusions and clinical implications . . . . . . . . . . . . . . . . . . References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. Introduction Schizophrenia affects N 21 million people globally (World Health Organisation, 2016) and is often described as a long-term chronic illness (World Health Organisation, 2011). The prevailing view on treatment is that continuous long-term use of neuroleptic medication is required for symptom control and relapse prevention (American Psychiatric Association, APA, 2006). Within the United Kingdom, the National Institute for Health and Care Excellence (NICE, 2014) recommends neuroleptic medication as a primary treatment for individuals diagnosed with psychoses or schizophrenia spectrum disorders. Continuous use of neuroleptic medication has been associated with positive clinical and functional outcomes (Dunayevich et al., 2007) and relapse prevention (Leucht et al., 2012). Non-adherence in the form of unilateral medication discontinuation (as opposed medication switching in consultation with a prescriber) is associated with poor outcomes in relation to symptoms, increased risk of relapse and hospitalisation, greater care costs and suicide (Higashi et al., 2013; Sullivan, Wells, Morgenstern, and Leake, 1995; Weiden and Olfson, 1995). However, acceptability of neuroleptic treatment for serviceusers is often problematic. Individual studies report discontinuation

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

79 82 82 82 82 85 85 85 88 88 88 89 89 89 89 89 89 89 90 90 90 90 90 90 90 90 91 91 91 91 91 91 91 92 92 92 92 92 93 93 93 93 94 94

rates of between 75% and 90% within 12 to 24 months of hospital discharge (Mullins, Obeidat, Cuffel, Naradzay, and Loebel, 2008; Weiden and Zygmunt, 1997), while clinical trial data demonstrate 74% of patients discontinue medication within 18 months (Lieberman et al., 2005). Partial medication non-adherence occurs as frequently as complete medication discontinuation (Svestka and Bitter, 2007) and can take the form of both increasing and decreasing recommended dosages (Rogers et al., 1998). Subsequently, adherence and non-adherence are no longer regarded as dichotomous states but as two ends of a spectrum (Julius, Novitsky, and Dubin, 2009). Neuroleptic adherence has been described as one of the most important issues in the long-term treatment of schizophrenia (Vauth, Löschmann, Rüsch, and Corrigan, 2004) while non-adherence has been described as one of the most challenging aspects of treating schizophrenia (World Health Organisation, 2003). Understanding what influences non-adherence is considered important for improving the welfare of people experiencing schizophrenia and subsequently reducing healthcare costs associated with acute psychotic episodes (Byerly, Nakonezny, and Lescouflair, 2007; Higashi et al., 2013). Research has sought to identify factors associated with non-adherence, often in an attempt to define targets for adherence interventions (Sendt, Tracy, and Bhattacharyya, 2015).

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

Despite a wealth of studies seeking to identify factors associated with neuroleptic non-adherence few consistent predictors have been determined. A recent review (Sendt et al., 2015) seeking to identify ‘robustly implicated factors emerging from methodologically rigorous studies’ (p. 14) reported contradictory results. Only insight into illness and positive attitudes towards medication were consistently associated with adherence. Inconclusive results were reported for factors such as side-effects, symptom severity and socio-demographic characteristics. Similarly, contradictory findings exist regarding efficacy of interventions designed to improve neuroleptic adherence. Whilst some studies report positive effects (Roberts and Velligan, 2011; Zygmunt, Olfson, Boyer, and Mechanic, 2002), others have found little improvement in adherence (Barkhof, Meijer, de Sonneville, Linszen, and de Haan, 2012; McIntosh, Conlon, Lawrie, and Stanfield, 2006). Inconsistencies may result from overemphasising clinician or researcher understanding of factors influencing adherence and underrepresenting service-user voices in the literature (Geekie, 2004). Research predominantly targets factors statistically associated with non-adherence (e.g. Lambert et al., 2004; Perkins et al., 2006; Rungruangsiripan, Sitthimongkol, Maneesriwongul, Talley, and Vorapongsathorn, 2011; Weiden, Mackell, and McDonnell, 2004) while clients' subjective reasons for medication non-adherence are infrequently examined (Löffler, Kilian, Toumi, and Angermeyer, 2003). Such bias is potentially problematic, especially in light of discrepancies between clinician and service-user perspectives on factors influencing neuroleptic use (e.g. Barbui et al., 2009; Pyne et al., 2006). Research exploring factors considered important by researchers and clinicians may overlook potentially important factors for service-users. Furthermore, efforts to improve adherence are not likely to be effective if they focus on factors that are not relevant to the client, and may serve to negatively impact trust between service-users and clinicians (Pyne et al., 2006), potentially impacting engagement with services. Continued prescription of neuroleptics, alongside efforts to improve adherence in the face of high rates of non-adherence could further highlight misunderstanding between service-users and clinicians. Understanding reasons for discontinuation is pivotal in treatment processes (Weiden and Ross, 2002) and may be essential in enabling clinicians to support clients in making collaborative treatment decisions, including whether or not to take neuroleptic medication, as recommended in clinical guidance (NICE, 2014). Morrison, Hutton, Shiers, and Turkington (2012) note that the risks associated with neuroleptic use, such as physiological side-effects, may influence non-adherence and that decisions to discontinue medication may, at times, represent a rational informed choice rather than an irrational decision due to symptom experience or poor illness insight. Furthermore, collaborative prescribing and facilitating patients' treatment choices may help to engage people who might otherwise reject services. Attention in the literature is largely dedicated to identifying reasons for neuroleptic non-adherence, whereas reasons for adherence are under-researched (Ascher-Svanum et al., 2010; Chen et al., 2011; Moritz, Hünsche, and Lincoln, 2014). Understanding reasons why people choose to continue medication could facilitate clinicians in better supporting service-users in patient-led care and collaborative decision-making (Chang, Tao, and Lu, 2013; Morrison et al., 2012). Therefore this is an area warranting further exploration. 1.1. Aims The current review aimed to address gaps in the literature related to understanding people's subjective reasons for neuroleptic continuation and discontinuation. Specifically, paucity of research in this area is addressed by exploring the question ‘What do service-users self-report as reasons for adherence and non-adherence to neuroleptic medication for psychoses?’ A systematic search strategy enabled the identification and synthesising of research addressing this question. For the purposes of the review, only studies targeting service-user reports of reasons for

77

neuroleptic adherence and non-adherence were included in order to ensure that a comprehensive service-user view was obtained. The review also sought to explore whether there were differences in reasons between sub-groups of service-users, such as those who were experiencing first episodes of illness compared to those with a more chronic illness course, or those taking first generation neuroleptic medication compared with those taking second-generation medication.

2. Methods 2.1. Search strategy A protocol and systematic search strategy were developed following guidance from the PRISMA statement (Moher, Liberati, Tetzlaff, and Altman, 2009). Statement guidelines include an evidence-based 27item checklist for systematic review reporting. The current review was registered with PROSPERO (Centre for Reviews and Dissemination (CRD), 2016), registration number: CRD42016033440. To capture relevant studies for inclusion, five search categories were developed. Medical Subject Headings (MeSH) were used to identify additional key subject headings. Within each category, the Boolean operator ‘OR’ was used, while the Boolean operator ‘AND’ was used to combine categories. A Boolean phrase combining all within and between category search terms was run within each platform. Wildcard asterisks were used to capture related terms. Categories were as follows: 1. Terms relating to service-users (patient*; client*; service-user*; consumer*; mental health patient*). 2. Terms associated with perspectives, including measure names identified during a scoping search used to assess perspectives (perspective*; views; reason*; attitude*; beliefs; Drug Attitude Inventory; Rating of Medication Influences Scale; Attitudes Towards Neuroleptic Treatment; Beliefs about Medication Questionnaire; Health Belief Dialogue; self-report; decision-making). 3. Terms to capture adherence (adherence; compl*; medication-adherence; medication-compliance; non-compl*; patient adherence; patient compliance; concordance; non-adher*; treatment refusal). 4. Terms related to neuroleptic medication (antipsychotic medication; antipsychotic*; antipsychotic treatment; neuroleptic*; neuroleptic treatment; neuroleptic drugs; neuroleptic medication; neuroleptic agent; atypical antipsychotic). 5. Terms associated with schizophrenia (schizophrenia-spectrum disorders; psycho*; schizophren*; schizoaffective disorder; first episode psychosis; schizoaffective psychosis). The Boolean phrase incorporating the above terms was used to systematically search Web of Science, Embase, Medline, Psychinfo and CINAHL Plus up to October 27th 2015. A search start date was not specified in order to capture as many service-user perspectives as possible. A limit of English language was applied to all searches, while a limit of human populations was applied to all searches other than that in CINAHL Plus where this limit was not available. In Web of Science, Psychinfo and CINAHL Plus, additional limits of ‘article’, ‘peer reviewed journal’ or ‘peer reviewed’ and ‘research article’ were applied respectively. Similar limits were unavailable for Embase and Medline. Searches in Embase, Medline and Psychinfo were performed using Ovid while CINAHL Plus was searched via EBSCO Host. For searches using Ovid, searches were performed using the ‘keyword’ function. This option was not available in Web of Science or Ebsco Host, therefore the ‘topic’ field and the ‘Boolean Phrase’ option were selected respectively. Reference list searching was also conducted on three recent reviews in the area in order to identify further articles for inclusion (Higashi et al., 2013; Richardson, McCabe, and Priebe, 2013; Sendt et al., 2015).

78

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

2.2. Eligibility criteria The current review sought to identify any studies that presented service-user reports on reasons for adherence or non-adherence to neuroleptic medication; thus papers were not excluded on the basis of publication period or methodology. The inclusion criteria for studies were: (a) available in English language; (b) used quantitative and qualitative methodology; (c) presented original data in a peer-reviewed empirical journal; (d) reported on people previously or currently taking neuroleptic medication (e) defined by study authors as having a schizophrenia spectrum disorder; (f) having a primary focus of exploring service-user self-reported reasons for neuroleptic adherence and/or non-adherence, through any measure, with at least one third of study participants being service-users; (g) aged 16 years or older. Exclusion criteria for the study were: (a) studies reporting on reasons for psychotropic medication adherence without detailing neuroleptic findings separately; (b) studies reporting only total scale or subscale scores without detailing specific reasons for adherence/ non-adherence; (c) studies not directly detailing service-user reported reasons for neuroleptic adherence/non-adherence including those reporting only on factors related to adherence, e.g. socio-demographic factors; those exploring reasons in multiple samples, e.g. service-user and clinician, without reporting service-user reasons independently; those detailing non-direct patient reports, e.g. reasons extracted from case notes; those detailing associations between adherence and other factors based on statistical relationships in the absence of service-user self- report; (d) studies including a non-schizophrenia spectrum population where results were not reported separately according to diagnosis; (e) studies involving participants aged under 16 years; (f) conference abstracts, reply to article papers, correction papers, review articles. The first author (MW) screened articles for eligibility. Uncertainties regarding article inclusion were discussed and resolved amongst all authors. In addition, 20% of included articles (n = 5) were discussed by all authors to ensure that they met inclusion criteria. For six of the included papers, information regarding participants' age was not detailed, or was provided only in the form of mean age (Kikkert et al., 2006; Matza et al., 2012; Moritz et al., 2009, 2013; Teferra, Hanlon, Beyero, Jacobsson, and Shibre, 2013; Tranulis, Goff, Henderson, and Freudenreich, 2011; Tunis, Faries, Stensland, Hay, and Kinon, 2007). In these cases, the study authors were contacted via e-mail to ask for clarification regarding participant ages. Five authors provided an e-mail response to confirm that study participants were aged 16 years or older. For one study (Tunis et al., 2007), attempts to contact authors were unsuccessful, but as the average age of participants was reported as 43 years old, this study was retained.

specific criteria determined by the methodological design of the study. Questions are answered ‘yes’, ‘no’ or ‘can't tell’. For each ‘yes’ response a star is awarded with a total of four stars possible for each study. Stars can be converted into percentages for reporting, such that studies gaining one star would receive a score of 25%, two stars would score 50%, three stars would score 75% and four stars would score 100%. Criteria for qualitative studies relate to relevance of data source, relevance of data analysis, appropriate consideration of how findings relate to the study context and appropriate consideration of how findings relate to researcher influence. For quantitative studies, appraisal was based on the sampling strategy, representativeness of sample/comparability of samples being compared, appropriate measurement and acceptable response rates. The first author conducted the quality appraisal for each study. Uncertainties were discussed and resolved amongst all authors following clarification from the developers of the MMAT via e-mail correspondence. A sample of 25% of the papers (n = 6) were rated by an independent colleague. Inter-rater reliability calculations were completed revealing a Kappa of 0.71, demonstrating substantial agreement (Landis and Koch, 1977). Studies were not excluded on the basis of quality as the review aimed to include as many service-user perspectives as possible. Instead, quality ratings were used as a means of evaluating the strength of presented evidence. As the included studies employed diverse methodologies, neither a meta-synthesis nor a meta-analysis were possible (Popay et al., 2006). Instead data analysis and reporting were based on narrative synthesis techniques and drew upon the narrative synthesis guidance of Popay et al. (2006). This guidance recommends the use of text and words to summarise and explain findings and suggests organising data synthesis to ‘tell a trustworthy story’. Guidance from the PRISMA statement (Moher et al., 2009), including the PRISMA checklist, were also used to ensure appropriate and transparent reporting.

3. Results In line with PRISMA guidelines, the selection of studies is detailed in a flowchart in Fig. 1. Database searching produced 4171 records, 49 additional records were identified through Google Scholar and reference list searching. All records were exported to Endnote online and duplicates removed, leaving a total of 2381 records. Screening led to the exclusion of 2309 records that did not meet inclusion criteria. Seventy-two remaining articles were retrieved for full-text screening and 51 were excluded for not meeting inclusion criteria (exclusion reasons are detailed in Fig. 1). The final review comprised of 21 studies.

2.3. Quality appraisal

3.1. Study characteristics

The Mixed Methods Appraisal Tool (MMAT, Pluye et al., 2011) was used to assess methodological quality as it allows for parallel assessment of qualitative, quantitative and mixed methods research. It was felt that this might provide fairer comparison of methodological quality than using separate tools for qualitative and quantitative studies, which could potentially skew results due to the differing nature or number of criteria being assessed. The MMAT is a relatively new appraisal tool, used to assess the methodological quality of studies in over 100 published review papers, including research on schizophrenia (e.g. Engels et al., 2014) and treatment adherence (e.g. Puts et al., 2013). The MMAT demonstrated high reliability (Pace et al., 2012) and validity (Pluye, Gagnon, Griffiths, and Johnson-Lafleur, 2009) in pilot testing. The MMAT contains five sets of methodological quality criteria applicable to studies of differing methodology, including; qualitative studies, mixed methods studies, quantitative randomized controlled studies, quantitative non-randomized studies and quantitative descriptive studies. For each study there are two screening questions followed by four

Descriptive information on included studies is presented in Table 1. Key non-adherence and adherence reasons are presented in Tables 2 and 3 respectively. Data is presented in author order, separated according to methodology. Of the included studies, 13 were quantitative (Ascher-Svanum et al., 2010; Chen et al., 2011; Hudson et al., 2004, Löffler et al., 2003; Matza et al., 2012; Moritz et al., 2009, 2013, 2014; Patel, de Zoysa, Bernadt, and David, 2008; Rosa, Marcolin, and Elkis, 2005; Sapra et al., 2014; Tunis et al., 2007; Yamada et al., 2006) seven were qualitative (Chang et al., 2013; Kikkert et al., 2006; Meshach, King, and Fulton, 2014; Pyne et al., 2006; Rogers et al., 1998; Teferra et al., 2013; Tranulis et al., 2011) and one employed a mixed methodology (Borras et al., 2007). All studies were published between 1998 and 2014 and were conducted in several countries; four in Germany, six in the United States, two in the United Kingdom, one in Japan, one in Taiwan, one in Brazil, one in Switzerland, one in Ethiopia and one in Nigeria. Three studies were conducted across multiple countries.

Identification

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

79

Additional records identified through other sources (reviews, reference list searching, google scholar searches) (n=49)

Records identified through database searching (n=4171 )

Records after duplicates removed

Records screened (n=2381)

Full-text articles assessed for eligibility (n=72)

Included

Eligibility

Screening

(n=2381)

Studies included in qualitative synthesis (n=21)

Records excluded (n=2309)

Full text articles excluded (n=51) Study does not detail specific adherence reasons (n=3) Study includes groups other than patients with schizophrenia spectrum disorder and does not report findings separately (n=8) Study does not detail diagnoses of participants or non-clinical populations are used (n=3) Study results based on statistical relationships between variables rather than participant self-report of variables as reasons (n=14) Primary study focus is not to explore patient self-reported reasons for adherence (n=18) Study reasons for adherence are not provided by patients or study does not detail who provided reasons (n=2) Study explores adherence to more than one type of psychotropic medication and does not report findings on neuroleptics separately (n=2) Study participants aged 16< (n=1)

Fig. 1. PRISMA flowchart detailing study selection.

3.2. Measurement of adherence and non-adherence perspectives Amongst the quantitative research, the Rating of Medication Influences Scale (ROMI, Weiden et al., 1994) was used in five studies and a Japanese version (Rating of Medication Influences Scale, Japanese version (ROMI-J, Yamada, Watanabe, Yagi, and Asai, 1999) was used in one study. Three studies employed the Reasons for Antipsychotic Discontinuation Interview (RAD-I, Matza et al., 2007), three studies used an online survey and one study used a ‘checklist’ completed during patient interviews (see Tables 2 and 3 for details of studies employing each measurement). The majority of studies measured reasons cross-sectionally. However, Ascher-Svanum et al. (2010) and Chen et al. (2011) presented discontinuation data from clinical trials measured at study entry and end/discontinuation, these are detailed separately in Table 3. Tunis et al. (2007) measured reasons five times over twelve months but presented only one dataset. Löffler et al. (2003) measured reasons at baseline and every six months over two years. For the purposes of the current review, only baseline data was extracted as this dataset had the largest sample size (286 compared to 143 with complete two year data). Quality appraisal indicated that measurements with a clear origin and good reliability/validity were used in nine of the 13 quantitative studies. In the remaining studies (Hudson et al., 2004; Moritz et al., 2009, 2013, 2014), the authors did not detail measure development or

information on reliability and validity, therefore the quality appraisal criterion could not be rated. Amongst qualitative studies, various methodologies were employed to gather data. Five studies used semi-structured interviews (Chang et al., 2013; Meshach et al., 2014; Pyne et al., 2006; Rogers et al., 1998; Tranulis et al., 2011), one used interviews plus focus group discussions (Teferra et al., 2013) and one used group discussions plus tasks to sort reasons according to importance and relation to other reasons (Kikkert et al., 2006). Data were analysed using content analysis (Chang et al., 2013; Pyne et al., 2006), thematic analysis (Meshach et al., 2014; Teferra et al., 2013), narrative analysis with grounded theory (Tranulis et al., 2011) and concept mapping (Kikkert et al., 2006). Rogers et al. (1998) did not detail the type of data analysis, but presented results in themes. The mixed method study of Borras et al. (2007) analysed data gathered in a semi-structured interview using content analysis. Visual analogue scales were used to collect quantitative data. Quality appraisal indicated that in all eight studies, data analysis was appropriate to the research question. 3.3. Participant characteristics For studies with more than one participant group (for example, service-user and clinician/carer samples) data pertaining only to service-users was extracted, unless results specified that a third party

80

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

Table 1 Study descriptive information.

Authors

Country Setting Frequency of reason measurement

Quantitative Ascher-Svanum Country: Multiple countries: et al. (2010) United States (64 study sites) Argentina (6 sites) Russia (4 sites) Puerto Rico (2 sites) Setting: Data taken from clinical trial. In/outpatient status not detailed. Measurement: Reasons measured at baseline and study end/ discontinuation Chen et al. Country: International. Individual countries not (2011) detailed. Setting: Data taken from parent study Multicentre randomized open label international trial. Outpatients who were not acutely ill. Measurement: Cross-sectional

Participants (n) Age (mean, SD, range) Gender (Male n, %) Ethnicity (n, %)

Participants: 199, data available for 191, study completion = 139 Age: 38.6 (12.0), 18–65 yrs Gender: Male (n = 115, 60.2%) Ethnicity: Caucasian 84 (44.0%) African descent 14 (7.3%), East Asian 38 (19.9%) Hispanic 50 (26.2%) Native American 1 (0.5%) West Asian 4 (2.1%) Participants: 153 complete data from 173 total (84 high barriers group, 69 low barriers group) Age: 45.7 yrs (high barriers group), 44.6 yrs (low barriers group) Gender: Male: 144 (94%) Ethnicity: White: 34 (22%). No further ethnicity information detailed. Participants: 307 (286 in baseline analysis) Age: 44.2 (11.2)18–64 yrs Gender: Male = 154 (50.2%) Ethnicity: Not detailed

50% Diagnosis: Schizophrenia or schizoaffective disorder Meet DSM-IV-TR diagnostic criteria Duration: Not detailed Medication description: Olanzapine in main study. Medications taken prior to study included Risperidone, Typical neuroleptics, Quetiapine & Aripiprazole.

Löffler et al. (2003)

Country: Germany Setting: Conventional oral, atypical, depot or combined neuroleptic treatment groups measured. Inpatient, outpatient and day care patients. Measurement: Reasons measured at baseline, then every 6 months for 2 years.

Matza et al. (2012)

Country: USA Setting: Patients recruited from 17 clinical sites: Veteran Administration Hospitals (3 sites), University Hospitals/Medical Centres (7 sites), private healthcare organisations (3 sites), private medical research organisations (2 sites), private psychiatric inpatient/outpatient services (2 sites) Measurement: Cross-sectional

Participants: 121 Age: Mean 41.6 (11.3), minimum age 16 yrs Gender: Male: 81 (66.9%) Ethnicity: White: 84 (69.4%) Black: 14 (11.6%) Asian: 5 (4.1%) Hispanic: 14 (11.6%) Other: 4 (3.3%)

Moritz et al. (2009)

Country: Germany Setting: Anonymous internet survey posted on German discussion forums for psychosis and other Psychiatric disorders Measurement: Cross-sectional Country: Germany Setting: Anonymous internet survey posted on German discussion forums for people with psychosis Measurement: Cross-sectional Country: Germany Setting: Anonymous internet survey posted on German discussion forums for people with psychosis Measurement: Cross-sectional

Participants: 69 Age: 35.13 (9.56), minimum age 16 years Gender: Male = 33 Ethnicity: Not detailed Participants: 113 used for analysis Age: 37.15 (9.55), minimum age 16 years Gender: Male = 39 Ethnicity: Not detailed Participants: 91, adherent group: 69 non-adherent group: 22 Age: 37.31 (8.94), 20–65 yrs Gender: Adherent group: Male = 35 Non-adherent group: male = 5 Ethnicity: Not detailed

Country: UK Setting: 2 sites, inner city London (Norwood & Brixton) and suburban area on London outskirts (Bromley). All voluntary outpatients. Measurement: Cross-sectional

Participants: 73/102 randomly selected from larger study of 222 Age: 18–34 yrs (n = 16, 21.9%) 35–44 yrs (n = 25, 34.3%) over 45 yrs (n = 32, 43.8%) Gender: Male (n = 43, 58.9%) Ethnicity: White (n = 30, 48.4%)

Patel et al. (2008)

50%

Diagnosis: Schizophrenia/Schizophreniform 450 (75.5%) Schizoaffective bipolar 94 (15.8%) Schizoaffective depression 52 (8.7%) Meet DSM-IV diagnostic criteria Duration: Median of 15.1 since illness onset Medication description: Risperidone or Olanzapine

Country: USA Setting: Data analysed from inpatients and outpatients taking part in Schizophrenia Guidelines Project, a multisite study funded by Dept. For Veteran Affairs involving interventions designed to improve adherence. Measurement: Cross-sectional

Moritz et al. (2014)

Overall quality assessment score (%)

Participants: 596 Age: 41.8 (11.0) 18–65 yrs Gender: Male = 371 (62.2%) Ethnicity: Caucasian 263 (44.1%) African descent 267 (44.8%) Hispanic 58 (9.7%) Other 8 (1.3%)

Hudson et al. (2004)

Moritz et al. (2013)

Diagnosis (n, %), method of diagnosis Illness duration (Mean years) Medication description

Diagnosis: Schizophrenia 75% Meet ICD-9 code for schizophrenia Duration: High barriers: 22.3, low barriers: 20.8 Medication description: Antipsychotic medication

Diagnosis: Schizophrenia Meet ICD-10; F20 criteria Duration: 16 (10) between first hospital admission and baseline assessment Medication description: Conventional oral/depot neuroleptic, atypical neuroleptic, combined neuroleptic medication. Clozapine, Risperidone, Olanzapine, Sertindole Diagnosis: Schizophrenia (58.7%) Schizoaffective disorder (41.3%) Clinical diagnosis from chart Duration: 17.7 (SD: 10.5) Medication description: Quetiapine: 26, Ziprasidone: 17, Risperidone: 16, Aripiprazole: 15, Olanzapine: 14, Haloperidol: 8, Paliperidone: 6, Clozapine: 12, Fluphenazine: 4, Prochlorperazine: 1, Thioridazine: 1, Thiothixene: 1 Diagnosis: Schizophrenia spectrum disorder Diagnosis self-disclosed by participant Duration: Not detailed Medication description: Neuroleptic medication Diagnosis: Schizophrenia Diagnosis self-disclosed by participant Duration: Not detailed Medication description: Antipsychotic medication Diagnosis: Schizophrenia (81.2%), another psychotic disorder (16.4%), schizoaffective disorder (2.2%), delusional disorder (1.1%) diagnosis self-disclosed by participant Duration: Not detailed Medication description: First/second generation. Adherent = 4 first generation, 65 second generation. Non-adherent = 1 first generation, 21 second generation Diagnosis: Schizophrenia (n = 65), schizoaffective disorder (n = 8) Meet ICD-10 criteria Duration: b15: 50.7%, 15–29: 39.7%, N30: 9.6% Medication description: Typical depot A.P: 24, atypical oral A.P: 42, typical oral A.P: 3, Clozapine: 4

100%

50%

50%

50%

50%

75%

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

81

Table 1 (continued)

Authors

Country Setting Frequency of reason measurement

Participants (n) Age (mean, SD, range) Gender (Male n, %) Ethnicity (n, %) Black/Black British (n = 26, 41.9%) Asian/Asian British (n = 4, 6.5%) Chinese/other (n = 1, 1.6%) Mixed (n = 1, 1.6%) Participants: 50 Age: 36.0 (6.82), 18–45 yrs Gender: Male 20 (40%) Ethnicity: Caucasian: 40 (80%) Black 5 (10%) Asian 5 (10%) Participants: 49 (FE: 33, ME: 16) Age: FE = 25.8 (6.7), 16–40, ME = 33.7 (9.3), 18–65 Gender: FE: Male = 73% ME: Male = 68% Ethnicity: FE = Caribbean (63.6%) African American (33.3%) Other (3.1%) ME = Caribbean or African American (88%) Other (12%)

Rosa et al. (2005)

Country: Brazil Setting: Outpatients at 2 public health clinics in Sao Paulo Measurement: Cross-sectional

Sapra et al. (2014)

Country: United States Setting: Data obtained from 2 studies, first episode (FE) sample (33), recently relapsing multi-episode (ME) patients (16). Both samples = inpatient & outpatient Measurement: Cross-sectional

Tunis et al. (2007)

Country: USA Setting: Outpatients (95%) recruited from an open-label trial comparing antipsychotic treatment options across 21 sites. Measurement: Reasons measured 5 times over 12 months.

Participants: 440 Age: Average age: 43 yrs Gender: Males = 63% Ethnicity: Caucasian: 54% African American: 34% Other: 12%

Yamada et al. (2006)

Country: Japan Setting: Outpatients recruited from 6 Hospitals in Japan who had good compliance for at least previous 3 months Measurement: Cross-sectional

Participants: 90 Age: 40.9 (12.8), 16–69 yrs Gender: Male = 53 Ethnicity: Japanese (100%)

Country: Switzerland Setting: Outpatients from public psychiatric outpatient facility Measurement: Cross-sectional

Participants: 103 Age: 18–65 yrs Gender: Male 70% Ethnicity: White European 79% Arabian 7% African 7% Asian 7%

Chang et al. (2013)

Country: Taiwan Setting: Psychiatric day care centre at Armed Forces Hospital Measurement: 3 interviews over 3 months

Participants: 10 Age: Mean 42.4, range 31–52 Gender: Male 70% Ethnicity: Chinese 100%

Kikkert et al. (2006)

Country: England, Germany, Italy, Netherlands Setting: Multisite, RCT to assess effectiveness of adherence therapy (QUATRO) study Measurement: Several sessions

Meshach et al. (2014)

Country: Nigeria Setting: Psychiatric hospital outpatients Measurement: Cross-sectional

Participants: 27 patients, 29 carers, 28 professionals Age: Not detailed Gender: 91 pps (41% male) attending brainstorming sessions, 89 (44% male) attending prioritizing & clustering sessions Ethnicity: Not detailed Participants: 11 Age: 22–49 Gender: Male 64% Ethnicity: (Presumed) Nigerian 100%

Pyne et al. (2006)

Country: USA Setting: Veterans and non-veterans from standard outpatient settings or intensive case management setting Measurement: 2 weeks (initial interview followed by short telephone interview within 2 weeks)

Qualitative/ mixed method Borras et al. (2007)

Participants: 26 Age: 44.3 (9.9), 20–70 yrs Gender: Male 23 (89%) Ethnicity: Caucasian: 15 (58%), African American: 11 (42%)

Overall quality assessment score (%)

Diagnosis (n, %), method of diagnosis Illness duration (Mean years) Medication description

Diagnosis: Schizophrenia Meet DSM-IV criteria Duration: 6.94 (SD: 6.32), range 0–22 Medication description: Typical neuroleptic agents or Risperidone

100%

Diagnosis: FE: Schizophreniform disorder, schizophrenia or schizoaffective disorder. ME: schizophrenia or schizoaffective disorder Meet DSM-IV criteria (confirmed by SCID) Duration: FE: 0.8(SD: 1.4) ME: 8.9(SD: 6.0) Medication description: FE: Risperidone, oral or depot. ME: Haloperidol, Fluphenazine, Risperidone, Ziprasidone Olanzapine, Aripiprazole Diagnosis: Schizophrenia (65%) Schizoaffective disorder (34%) Schizophreniform disorder (b1%) Meet DSM-IV criteria Duration: Not detailed Medication description: Olanzapine: 222 Risperidone: 218 Diagnosis: Schizophrenia Meet DSM-IV criteria Duration: 14.3 (SD: 9.9), range 1–38 Medication description: Neuroleptic medication including atypical, depot and oral

50%

100%

50%

75% Diagnosis: Schizophrenia 80% Schizoaffective disorder 18% Psychotic disorder not otherwise specified 2% Meet ICD-10 diagnostic criteria Duration: 14 (SD: 11) Medication description: Oral antipsychotic medication 83% Depot medication 17% Atypical antipsychotic 66% Typical antipsychotic 5% Association 28% Diagnosis: Schizophrenia (90%) Schizoaffective 75% disorder (10%) Method of diagnosis not detailed Duration: 2.95, range: 0.5–6 yrs (duration of admittance to day care) Medication description: Antipsychotic medication 50% Diagnosis: Schizophrenia Meet ICD-10 criteria for schizophrenia Duration: Not detailed Medication description: Antipsychotic medication

Diagnosis: Schizophrenia (100%) Meet ICD-10 criteria Duration: 1.5–25 yrs Medication description: Typical and Atypical antipsychotic medication Diagnosis: Schizophrenia/schizoaffective disorder Method of diagnosis not disclosed Duration: Not detailed Medication description: Oral antipsychotic medication (excluding Clozapine)

75%'

50%

(continued on next page)

82

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

Table 1 (continued)

Authors Rogers et al. (1998)

Teferra et al. (2013)

Tranulis et al. (2011)

Country Setting Frequency of reason measurement Country: UK Setting: Patients recruited from different points in mental health system including voluntary organisations, day centres, outpatient and in-patient facilities Measurement: Cross-sectional Country: Ethiopia Setting: Participants involved in ongoing population-based cohort study based in rural Ethiopia Measurement: Cross-sectional Country: USA Setting: Participants referred from community mental health centre outpatient clinic Measurement: Cross-sectional

Participants (n) Age (mean, SD, range) Gender (Male n, %) Ethnicity (n, %) Participants: 34 Age: 18–56 yrs Gender: Male 22 Ethnicity: Not detailed

Participants: 24 patients, 19 caregivers, 7 research field workers, 1 health worker Age: Not detailed Gender: Not detailed Ethnicity: (Presumed) Ethiopian 100% Participants: 20 Age: Mean 39 (SD: 14), 18 years minimum Gender: Male 85% Ethnicity: Not detailed

Diagnosis (n, %), method of diagnosis Illness duration (Mean years) Medication description Diagnosis: Schizophrenia/schizoaffective disorder Meet DSM-IV criteria Duration: Not detailed Medication description: Oral neuroleptics with or without maintenance depot neuroleptics Diagnosis: Schizophrenia (100%) Method of diagnosis not detailed Duration: Not detailed Medication description: Antipsychotic medication Diagnosis: Schizophrenia (70%), Schizoaffective disorder (30%) Chart diagnosis Duration: 15 (SD: 12), range: 1–34 Medication description: Antipsychotic monotherapy (90%), Clozapine (40%)

Overall quality assessment score (%) 50%

75%

50%

(e.g. professional/carer) was relaying information reported directly by a service-user. This was only relevant in Teferra et al. (2013) where professionals made a small number of diect quotations from service-users. Within service-user samples, numbers ranged from 49 to 596 participants in quantitative studies and 10 to 34 participants in qualitative studies. The mixed method study had a sample size of 103. For quality assessment of sampling procedures, the MMAT has differing criteria depending on study design (please see Methods section). For studies rated on the relevance of the sampling strategy (n = 11), nine were rated as having an appropriate sampling strategy. For those rated on whether participants were recruited in a manner that would minimise selection bias (n = 3), all three met criteria. The qualitative and mixed method studies (n = 8) were rated on whether selected participants could address the research question. All eight studies met this criteria. Most studies detailed where participants were recruited from or whether they were in/outpatients. Recruitment sources included inpatient, outpatient and daycare facilities, intensive case management settings and voluntary organisations. Participants were recruited from both private and public facilities. In addition, three studies recruited participants via online discussion forums (Moritz et al., 2009, 2013, 2014). All participants were described as experiencing a first episode of psychosis or as having a diagnosis of a schizophrenia-related disorder (see Table 1 for method of diagnosis). Prescribed neuroleptics included typical and atypical preparations. Medication was taken orally, by depot injection, or in combination. Some studies did not detail medication specific information, describing them only as ‘antipsychotic’ or ‘neuroleptic’ medication. Twelve studies detailed average illness duration of participants. These ranged from 0.8 years in a first episode sample to N30 years in chronic samples.

Data were extracted in a manner which aimed to make results from different studies as comparable as possible. For the quantitative studies, five reasons most frequently cited for adherence and non-adherence were extracted in the form of percentages. All studies used measures that contained more than five items, however, the number of reported items varied, with some studies detailing responses to all items and others detailing only two items. Most studies detailed at least five most frequently endorsed reasons and therefore this number was chosen to reflect a range of reasons whilst still being a manageable number that would allow for comparison between groups. For studies using the ROMI, items rated as having a mild to strong degree of influence were combined and extracted where possible. Sapra et al. (2014), only reported items with a strong degree of influence on adherence and thus only this data was extracted. For studies using the RAD-I (Ascher-Svanum et al., 2010; Chen et al., 2011; Matza et al., 2012) it was not possible to be as consistent. In the RAD-I, there is the option to rate both the ‘single most important’ reason for adherence and nonadherence, and/or the top primary reasons. In Ascher-Svanum et al. (2010), data was presented on both and thus this information was combined and extracted. However, Chen et al. (2011) reported only a small number of ‘single most important’ reasons; thus only the primary reasons were extracted. In the study of Matza et al. (2012), information was not presented on the single most important or primary reasons, thus the five most frequently cited reasons were extracted. Within the qualitative studies, all themes and/or percentages of endorsement of themes were extracted. Across all studies, any ‘additional points of interest’ felt to be relevant were also extracted.

3.4. Overall quality assessment

3.6. Data synthesis

The total quality appraisal scores indicated that 12 of the studies scored 50% for methodological quality (Ascher-Svanum et al., 2010; Chen et al., 2011; Kikkert et al., 2006; Matza et al., 2012; Moritz et al., 2009, 2013, 2014; Pyne et al., 2006; Rogers et al., 1998; Sapra et al., 2014; Yamada et al., 2006), six studies scored 75% (Borras et al., 2007; Chang et al., 2013; Hudson et al., 2004; Meshach et al., 2014; Patel et al., 2008; Teferra et al., 2013) and three studies met all criteria for methodological quality scoring 100% (Löffler et al., 2003; Rosa et al., 2005; Tunis et al., 2007). This indicates that all studies had fair to excellent methodological quality.

3.6.1. Reasons for adherence and non-adherence to neuroleptic medication Fourteen studies explored reasons for both adherence and nonadherence. Two studies (Chang et al., 2013; Tunis et al., 2007) explored adherence only and five studies explored non-adherence only (Hudson et al., 2004; Meshach et al., 2014; Moritz et al., 2009, 2013; Teferra et al., 2013). Several themes were identified relating to both adherence and non-adherence while some themes applied to either adherence or non-adherence only. Themes are discussed in turn below with relevance to adherence and/or non-adherence indicated.

3.5. Data extraction

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

83

Table 2 Service-user reasons for non-adherence to neuroleptic medication. Author

Measure (analysis)

Quantitative Ascher-Svanum et al. RAD-I (2010) (frequencies of endorsed items)

Chen et al. (2011)

RAD-I (frequencies of endorsed items)

Hudson et al. (2004) Checklist of adherence barriers (frequencies of endorsed items)

Löffler et al. (2003)

ROMI (frequencies of endorsed items)

Matza et al. (2012)

RAD-I (frequencies of endorsed items)

Moritz et al. (2009)

Internet survey (frequencies of endorsed items)

Moritz et al. (2013)

Internet survey (frequencies of endorsed items)

Key non-adherence reasons (top five plus additional points of interest) Study drug discontinued by 49% of participants. Reasons for discontinuation prior to trial (sum of primary reasons & single most important reason): Positive symptoms not sufficiently improved/made worse (49%) Adverse effects (24%) Cost of medication (14%) No longer feel neuroleptic is needed (11%) Medication not covered by insurance (6%). For trial drug: Positive symptoms not sufficiently improved/made worse (41%) Adverse effects (42%) Told by another person to discontinue (14%) Other reasons (15%) Negative symptoms not sufficiently improved or made worse (9%) Additional points of interest: N/A Mean number of reasons = 4.1 (SD: 3.2) Reasons for discontinuation prior to trial (primary reasons): Positive symptoms not sufficiently improved or made worse (15.3%) Wishing to try a new antipsychotic (11.8%) Adverse events (10.6%) Functional status not sufficiently improved or made worse (10.6%) Negative symptoms not sufficiently improved or made worse (approx. 9%). For trial drug: Mean number of reasons = 2.3 (SD: 2.9). Positive symptoms not sufficiently improved or made worse (25%) Adverse events (16.7%) Negative symptoms not sufficiently improved or made worse (12.5%) Functional status not sufficiently improved or made worse (approximately 9%) Other reasons (approximately 8%) NB: approximate values calculated from figure, exact values not detailed. Additional points of interest: N/A Mean number of barriers: 1.8 Stigma (36%) Adverse drug reactions (34%) Other barriers (homelessness, substance abuse, 29%), Memory problems (28%) Lack of social support (27%). Additional points of interest: 11% reported being afraid of medication. NB: All percentages are approximate, calculated from figure provided. Mild to strong degree of influence (all medication groups combined): Distressed by side effects (49%) Medication felt to be currently unnecessary (40%) No perceived daily benefit of medication (31%) Denial of illness (27%) Embarrassment or stigma over medication or illness (12%). Additional points of interest: Negative relationship ship with prescriber (6.9%) Family/friends opposed to medications (6.5%) Mean number of reasons = 2.8 (SD: 1.8) Reasons endorsed with a ‘yes’ response: Adverse events (65.4%) Did not improve positive symptoms (22.2%) Another person told individual to stop taking medication (14.8%) Did not improve mood (11.1%) Wished to try an antipsychotic new to market (3.7%). Additional points of interest: N/A 71% reported periods of non-adherence: Too many side effects (79%) Stigma (48%) Did not feel neuroleptic medication was needed (47%) Forgot to take medication (32%) Distrust of physician/therapist (29%). Additional points of interest: 44% reported positive aspects of psychotic symptoms as reasons for non-adherence: (feelings of importance 18%, becoming another person when ill and needing this state at times 15%, missing voices 11%). 64% reported periods of non-adherence. Mean number of reasons = 3.83 (SD: 1.97). Too many side-effects (80%) Did not feel neuroleptic medication was needed (58%) Stigma (31%) Distrust of physician/therapist (31%) Medication intake seen as negating the validity of personal experience (28%)/Rejection of medication in general (28%). Additional points of interest: rated as important by at least 1/5th of participants: Forgetfulness (21%) Advice of relatives/friends against taking A.P (20%). 28% reported positive aspects of psychotic symptoms as reasons for non-adherence: (feelings of importance & power 18%, becoming another person when ill and needing this state at times 18%, missing voices 7%) (continued on next page)

84

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

Table 2 (continued) Author

Measure (analysis)

Key non-adherence reasons (top five plus additional points of interest)

Moritz et al. (2014)

Internet survey (frequencies of endorsed items)

Patel et al. (2008)

ROMI (frequencies of endorsed items)

Rosa et al. (2005)

ROMI (frequencies of endorsed items)

Sapra et al. (2014)

ROMI (frequencies of endorsed items)

Mean number of reasons = 6.23 (SD: 4.05). Intolerance, e.g. Side effects (71.4%) Afraid of potential side effects (71.4%), Recovery experienced after short time (52.4%) Not wanting to rely on medication (52.4%) Forgot the intake (28.6%). Additional points of interest: Feeling able to manage mental health problems without medication/did not need neuroleptic medication in own view (28.6%/19%) Friends/relatives advised against taking (28.6%) Medication does not fit symptoms (23.8%) Stigma (23.8%). 52.4% reported positive aspects of psychotic symptoms as reasons for non-adherence: without psychosis life is meaningless (19%), becoming another person when ill and needing this state at times (14.3%), feelings of importance and power which I do not want to miss (9.5%), attention deserved (4.8%), not having to work when ill (4.8%). Very few specific reasons detailed. Denial of illness (participants prescribed oral neuroleptic medication 33%, depot medication 14%) Medication unnecessary (oral, 24%, depot, 42%). Detailed as ‘common reasons’. Additional points of interest: N/A Non-compliance rate over 12 months = 48%. Open question- main reasons for medication non-adherence: ‘None’ (40%) Side-effects (36%) Desire to be normal (10%) Embarrassment/stigma (4%). Closed question (mild & strong influence combined): Side-effects (61%) Denial of disease (40%) Medication felt to be currently unnecessary (32%) Embarrassment/stigma associated with treatment (28%) No perceived day-to-day benefit (16%). Additional points of interest: 22% reported own or other (health professional or family/friends) opposition to the use of meds. Reasons rated as having a strong influence on non-adherence: First episode sample (FE): Distress from side-effects (32%) No benefits from the medication (27%) Denial of illness (27%) Stigma (27%) Interferes with life goals (21%) Multi-episode sample (ME): Distress from side-effects (44%) Medication felt to be unnecessary (31%) Stigma (25%) Access barriers (25%) Change in appearance (25%). Additional points of interest: Change in appearance due to medication more often endorsed by ME than FE (25% vs 0%, ×2 = 9.2, p b 0.04) and medication unnecessary (31.3% vs 8.8% ×2 = 4.1, p b 0.04) Not measured

Tunis et al. (2007)

ROMI (frequencies of endorsed items) Yamada et al. (2006) ROMI-J (frequencies of endorsed items)

Qualitative/mixed method Borras et al. (2007) Semi-structured interviews plus visual analogue scales (content analysis)

Chang et al. (2013) Kikkert et al. (2006)

Semi-structured interviews (content analysis) Group discussion, clustering and prioritizing tasks (concept mapping)

Anticipated reasons, mild and strong influence combined: Embarrassment or stigma over medication/illness and distressed by side effects, each (34.4%) Access to treatment problems (33.3%) No perceived daily benefit and interferes with life goals, each (14.4%). Additional points of interest: N/A

Religious rather than medical representations of illness were more prominent in participants who were non-adherent. 31% of non-adherent and 27% of partially adherent participants discussed an incompatibility/contradiction between religion and adherence to medication/had a negative spiritual vision of medication compared to 8% of adherent participants (e.g. medication hinders spiritual growth). Negative spiritual vision of illness and medication more likely in non-adherent than adherent participants (25% vs 3%). Additional points of interest: 44% of non-adherent & 27% partially adherent participants did not feel at ease to discuss religion with psychiatrist (compared to 15% of adherent participants). Not measured 5 clinically relevant themes identified from 10 clusters of statements as influencing adherence (results not separated according to adherence/non-adherence): Medication efficacy, e.g. relief of symptoms, rated as most important factor influencing adherence, 54% of patients rated items in this cluster as important. External factors which establish favourable conditions for adherence, e.g. increasing patient understanding and trust with clinician (professional & non-professional support, rated as important by 42%) & (information & involvement rated as important by 39%). Insight, e.g. accepting medication is needed, rated as important by 44% of participants. Side-effects, e.g. being embarrassed about visible movement disorders, including perceived side-effects (rated as important by 40%), side-effect self-management (rated as important by 48%) and social aspects of extrapyramidal side-effects (rated as important by 22%). Medication attitudes including positive medication attitudes & expectations, e.g. reducing coercion/hospitalisation as a result of illness (rated as important by 42%), negative expectations (rated as

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

85

Table 2 (continued) Author

Measure (analysis)

Meshach et al. (2014) Semi-structured interviews (thematic analysis)

Pyne et al. (2006)

Semi-structured interviews (content analysis)

Rogers et al. (1998)

Semi-structured interviews (presumed thematic analysis)

Teferra et al. (2013)

Semi-structured interviews plus focus group discussion (thematic analysis)

Tranulis et al. (2011) Semi-structured interviews (narrative analysis with grounded theory)

Key non-adherence reasons (top five plus additional points of interest) important by 36%) and negative medication attitudes, e.g. feeling better without medication (rated as important by 30%). Additional points of interest: N/A Themes identified: Experience of medication: Discontinuing medication when feeling well, frustration at prospect of unending treatment. Side-effects: Sedation and weakness leading to an inability to function at work/school. Increased appetite difficult when there was a lack of money to spend on buying more food. Poverty: lack of money to buy food, not wanting to take medication on an empty stomach, poorer relatives unable to afford medication & transportation costs to collect medication, those with low social support unable to afford medication, poorer relatives becoming hostile and critical or abandoning towards patients leading to non-adherence, frustration at poverty leading to low mood & lost interest in taking medication, poverty caused by being adherent as side effects impact ability to work. Barriers to adherence organised into 8 categories: Environment (e.g. Lack of support, cost of medication too high) Side-effects (e.g. Impaired work/interpersonal functioning) Relationship between provider and family (e.g. anger at family/provider) Insight and knowledge (e.g. medication is harmful/not effective) Symptoms and outcomes (e.g. being easily distracted/medications don't help) Substance abuse (e.g. use of alcohol/street drugs instead) Stigma (e.g. embarrassment) Dosing (e.g. missing doses/sleeping through dose times) Additional points of interest: N/A Themes of: Increasing/decreasing dose to cope with arising distress Side effects Inhibited everyday social interactions/opportunities caused by side effects Interference with social interaction/enjoyment (e.g. pub/leisure activities) Non-adherence to prove normalcy/avoid stigma. Additional points of interest: N/A Themes of: Poverty: good food required to counterbalance strong effects of medication on body, medication increases appetite, increased appetite burdens family/leads to hunger, thus lack of food leads to non-adherence (patient choice or family making patient stop medication) Lack of family support: e.g. family inability/unwillingness to take patient to get medication due to competing work demands. Illness perspective: Stop taking medication when feeling better (concept of continuing medication when better is alien) Stigma: Stopping medication to prevent others finding out about mental illness, stigma interfering with work and marriage prospects Health care provider issues: Fear of negative response from health worker after missed appointment/long waiting times, inflexibility of appointment system Side effects: Distressing effects which interfere with functioning Substance use: Not taking medication due to incompatibility with chewing khat and alcohol use. 95% reported non-adherence for extensive periods in past. Themes of: No perceived need for medication (reported by 60%) Pressure from others not to take the medication Side-effects (reported by 35%) Stopping medication to reduce stigma Not wanting to damage body. Additional points of interest: Despite side-effects reported as non-adherence reason, these were tolerated for years when medication brought about benefits.

3.7. Medication efficacy 3.7.1. Quantitative findings Medication efficacy was the most frequently cited reason for nonadherence and adherence. Poor medication efficacy was reported as a non-adherence reason in 10 studies (see Table 2) while medication benefits were also cited as adherence reasons in 10 studies (see Table 3). Participants reported non-adherence or adherence in relation to the efficacy of medication on symptom control (Ascher-Svanum et al., 2010; Chen et al., 2011; Matza et al., 2012; Moritz et al., 2014), health and well-being (Löffler et al., 2003), mood (Ascher-Svanum et al., 2010; Matza et al., 2012), functioning (Ascher-Svanum et al., 2010; Chen et al., 2011; Matza et al., 2012; Moritz et al., 2014), achievement with life goals (Sapra et al., 2014; Tunis et al., 2007; Yamada et al., 2006), and general improvements/daily benefits (Chen et al., 2011; Löffler et al., 2003; Rosa et al., 2005; Sapra et al., 2014; Tunis et al., 2007; Yamada et al., 2006). Benefits to cognition (Ascher-Svanum et al., 2010; Matza et al., 2012), relapse prevention (Löffler et al., 2003; Patel et al., 2008;

Rosa et al., 2005; Sapra et al., 2014; Yamada et al., 2006) and wishing to lead a normal life (Moritz et al., 2014) were reported as adherence reasons only. Up to 49% of participants reporting on nonadherence gave poor medication efficacy as a non-adherence reason (Ascher-Svanum et al., 2010) while up to 90% of those reporting on adherence (Yamada et al., 2006) gave good medication efficacy as an adherence reason. 3.7.2. Qualitative findings Only one study reported on medication efficacy as a reason for nonadherence (Pyne et al., 2006). Consistent with quantitative studies, barriers to medication adherence included feeling that medications were harmful or did not help. Good medication efficacy was reported as an adherence reason in five studies (Table 3). Again, findings largely supported those of quantitative studies. Participants reported adherence to prevent relapse and hospitalisation (Chang et al., 2013; Kikkert et al., 2006; Rogers et al., 1998), to improve functioning (Kikkert et al., 2006), to allow goal achievement such as leaving hospital or living

86

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

Table 3 Service-user reasons for adherence to neuroleptic medication. Author

Measure (analysis)

Quantitative Ascher-Svanum et al. RAD-I (frequencies of endorsed items) (2010)

Chen et al. (2011)

RAD-I (frequencies of endorsed items)

Hudson et al. (2004) Checklist of adherence barriers (frequencies of endorsed items) Löffler et al. (2003) ROMI (frequencies of endorsed items)

Matza et al. (2012)

RAD-I (frequencies of endorsed items)

Moritz et al. (2009)

Internet survey (frequencies of endorsed items) Internet survey (frequencies of endorsed items) Internet survey (frequencies of endorsed items)

Moritz et al. (2013) Moritz et al. (2014)

Patel et al. (2008)

ROMI (frequencies of endorsed items)

Rosa et al. (2005)

ROMI (frequencies of endorsed items)

Sapra et al. (2014)

ROMI (frequencies of endorsed items)

Key adherence reasons (top five plus additional points of interest) Study drug continued by 51% of participants. Reasons for continuation (Sum of primary reasons & single most important reason): Benefits positive symptoms (86%) Perceived improvement (40%) Benefits functioning (19%) Benefits cognition (12%) Benefits mood (8.3%). Additional points of interest: N/A Mean number of reasons = 8.5 (SD = 4). Primary continuation reasons: Perceived improvement (29%) Benefits functioning (24.6%) Benefits positive symptoms (24.6%) Benefits mood and already tried another neuroleptic medication approximately 11.5% each. NB: approximate values calculated from figure, exact values not detailed. Additional points of interest: N/A Not measured Mild to strong degree of influence (all medication groups combined): Relapse prevention (88%) Perceived daily benefit (improved health & well-being, 79%) Fear of rehospitalisation (69%) Positive relationship with prescriber (41%) Positive family belief (23%). Additional points of interest: Pressure/force was the 6th most imp reason (7%) Mean number of reasons = 3.4 (SD: 1.8) reasons Reasons endorsed with a ‘yes’ response: Benefits positive symptoms (65%) Benefits functional status (55%) Benefits mood (30%) Benefits cognition (22.5%) Mild, tolerable side effects, another person told individual to continue medication & person already tried less effective/tolerable antipsychotics (20% each). Additional points of interest: N/A Adherence reasons not assessed, however, 78% of psychosis sample reported only taking A.P because their physician wants them to. 17% reported just pretending to take A.P Not measured Mean number of reasons = 3.49 (SD: 1.74) Wish to lead a normal life (74.6%) Afraid of worsening mental health (70.1%) Medication helps to better control symptoms (59.7%) Fear of psychotic symptoms (49.3%) Symptoms lead to impairment (43.3%) Additional points of interest: Do not want to be called crazy (31.3%), advice of family/friends to take medication properly (20.9%) Very few specific reasons detailed. Fear of rehospitalisation (participants prescribed oral neuroleptic medication 47%, depot medication 67%), Relapse prevention (oral 78%, depot 71%). Detailed as ‘common reasons’ Additional points of interest: N/A Open question - main reasons medication adherence: Perceived day to day benefit (88%) Relapse prevention (6%) Respect for authority (4%) Fear of hospitalisation (2%). Closed question (mild & strong influence combined): Positive relationship with physician (82%) Perceived day to day benefit (80%) Positive family belief (72%) Relapse prevention (64%) Supervision (54%) Respect for authority (52%). Additional points: Pressure from family or health system was rated as important by 52% collectively. Reasons rated as having a strong influence on adherence: First episode (FE): Doctor-patient relationship (77%) Relapse prevention (74%) Day to day benefit (44%) Helps with life goals (32%) Family support (32%). Multi-episode (ME): Relapse prevention (63%) Day to day benefit (56%) Family support and fewer side effects of current neuroleptic (19% each) Doctor-patient relationship (13%)

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

87

Table 3 (continued) Author

Tunis et al. (2007)

Measure (analysis)

ROMI (frequencies of endorsed items)

Yamada et al. (2006) ROMI-J (frequencies of endorsed items)

Qualitative/mixed method Borras et al. (2007) Semi-structured interviews plus visual analogue scales (content analysis)

Chang et al. (2013)

Semi-structured interviews (content analysis)

Kikkert et al. (2006)

Group discussion, clustering and prioritizing tasks (concept mapping)

Meshach et al. (2014) Pyne et al. (2006)

Semi-structured interviews (thematic analysis) Semi-structured interviews (content analysis)

Rogers et al. (1998)

Semi-structured interviews (presumed thematic analysis)

Key adherence reasons (top five plus additional points of interest) Additional points of interest: Relationship with therapist rated as important influence on adherence by 17.7% of FE group but 0% of ME group (p b 0.04). Doctor-patient relationship rated as important by 76.5% of FE group versus 12.5% of ME group (p b 0.01) Help with life-goals reported as strong influence by 32% of FE group versus 0% ME group (p b 0.02). Pressure/coercion reported by more of ME group than FE group (6.3% vs 2.9%) Mild & strong influence combined: Perceived daily benefit (87%) Fear of relapse (77%) Side effect relief (63%) Fulfilment of life-goals (60%) Deference to authority (53%). Additional points: Items 6–9 are all related to relationships or the influence of others (positive relationship with clinical staff 48%, outside positive opinion about medication 47% outside opinion that current medications are better 35%, pressure of staff 14%) Mild & strong influence combined: Relapse prevention (90%) Positive family belief (68.9%) Positive relationship with prescribing physician (60%) Perceived daily benefit (56.7%) Fear of rehospitalisation/fulfilment of life goals (each 54.4%). Additional points of interest: 6th most important reason for compliance rated at baseline was deference to authority (45.6%).

Medical rather than religious representations of illness more prominent in adherent participants. Religion gave meaning to medication for some (i.e. Gift from God). For others illness but not medication seen spiritually (i.e. Illness sent by God but medication helps to overcome illness). 37% of adherent participants and 35% of those on depot medication reported a positive spiritual vision of medication compared to 27% of partially adherent and 13% of non-adherent participants. Positive spiritual visions of illness and medication more likely in adherent than non-adherent participants (29% vs 13%). Additional points of interest: N/A Themes of: Benefits of medication (symptom improvement, relapse prevention, living a normal life, satisfaction with life) Family support (e.g. taking medication due to care and support of family) Chinese concept of filial piety (notion that Chinese children are supposed to show extreme respect for parents, achieving to bring honour to the family & not being a burden to family, e.g. taking medication to repay parents) Hope for the future (e.g. taking medication due to hope that it will cure illness and allow a job to be secured, or allow people to be able to support children in their futures, e.g. with careers and marriage which allows participants to adhere to the Chinese traditional belief of ‘fulfilling their parental duty’) 5 clinically relevant themes identified from 10 clusters of statements as influencing adherence (results not separated according to adherence/non-adherence): Medication efficacy, e.g. relief of symptoms, rated as most important factor influencing adherence, 54% of patients rated items in this cluster as important. External factors which establish favourable conditions for adherence, e.g. increasing patient understanding and trust with clinician (professional & non-professional support, rated as important by 42%) & (information & involvement rated as important by 39%). Insight, e.g. accepting medication is needed, rated as important by 44% of participants. Side-effects, e.g. being embarrassed about visible movement disorders, including perceived side-effects (rated as important by 40%), side-effect self-management (rated as important by 48%) and social aspects of extrapyramidal side-effects (rated as important by 22%). Medication attitudes including positive medication attitudes & expectations, e.g. reducing coercion/hospitalisation as a result of illness (rated as important by 42%), negative expectations (rated as important by 36%) and negative medication attitudes, e.g. feeling better without medication (rated as important by 30%). Additional points of interest: N/A Not measured Motivators for taking medication - 3 categories identified: Environment (e.g. seeing others do well on medication) Symptoms and outcome (e.g. avoid worsening of symptom severity/interpersonal relationships/living situation, avoid losing freedom) Provider/family relationships (e.g. improved relationships) Additional points of interest: N/A Themes of: Preventing relapse/hospitalisation Calming effect of medication Control/reduce symptoms (voices/auditory hallucinations/untoward thoughts) Adherence allows distressing symptoms to be coped with Adherence allows coping with difficult situations, e.g. those causing anxiety/agitation Countering the adverse effects of recreational drugs Power of family/Mental Health professionals Fear of sanctions Adherence due to practical help of others Pressure/coercion (continued on next page)

88

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

Table 3 (continued) Author

Teferra et al. (2013) Tranulis et al. (2011)

Measure (analysis)

Semi-structured interviews plus focus group discussion (thematic analysis) Semi-structured interviews (narrative analysis with grounded theory)

Key adherence reasons (top five plus additional points of interest) Fear of losing future care Deference to authority Lack of perceived autonomy to stop taking medication Benefit for self and others Additional points of interest: None Not measured Themes of: Deference to authority/trusting clinician/family advice Adherence due to external pressure (family/clinicians) Adherence to achieve goals (e.g. to get out of hospital, to be allowed to live with family) Adherence due to re-experiencing of symptoms following non-adherence Adherence due to symptom reduction/control once neuroleptic medication is reinstated following period of non-adherence.

with family (Tranulis et al., 2011), to live a normal life (Chang et al., 2013) and to control, reduce or cope with symptoms (Chang et al., 2013; Kikkert et al., 2006; Pyne et al., 2006; Rogers et al., 1998; Tranulis et al., 2011), “I do not get sent to the hospital (acute ward) anymore since taking medication” [(preventing hospitalisation, Chang et al., 2013, p. 274).] “This medication keeps me in the best mental health. I can live a normal life every day by taking the medication” [(living a normal life, Chang et al., 2013, p. 274).] “Some of them are really funny voices but er sometimes I get the shits up, you know the telly and er, you know, you get frightened but if you take your drugs you cope and dismiss it” [(coping with symptoms, Rogers et al. 1998, p. 1317).] Additionally, participants reported adherence to avoid a worsening of interpersonal relationships or living situations (Pyne et al., 2006), to allow coping with situations causing anxiety and agitation (Rogers et al., 1998), to feel satisfaction with life (Chang et al., 2013), and because adherence provided hope for a future cure and the prospect of work and means to support children (Chang et al., 2013). 3.7.3. Benefits associated with symptoms - quantitative findings Although many studies reported on the importance of medication efficacy on symptom control, an additional finding detailed in three studies concerned non-adherence due to individuals not wanting symptom reduction (Moritz et al., 2009, 2013, 2014). This occurred as people experienced positive aspects of psychotic symptoms including feelings of importance and/or power (Moritz et al., 2009, 2013, 2014), becoming another person and needing this state at times (Moritz et al., 2009, 2013, 2014), feeling that attention is deserved (Moritz et al., 2014) and not having to work when ill (Moritz et al., 2014). People reported non-adherence due to missing voices (Moritz et al., 2009, 2013) and feeling that life is meaningless without psychosis (Moritz et al., 2014). Between 28% and 52.4% of participants reported non-adherence due to at least one positive aspect of symptom experience. Similar findings were not reported in any of the qualitative studies. 3.8. Personal beliefs 3.8.1. Medication beliefs 3.8.1.1. Quantitative findings. Medication beliefs in relation to illness were reported as a reason for non-adherence in nine studies (see Table 2). Service-users reported discontinuing medication due to beliefs that it was currently unnecessary (Ascher-Svanum et al., 2010; Löffler et

al., 2003; Moritz et al., 2009, 2013; Patel et al., 2008; Rosa et al., 2005; Sapra et al., 2014), because they feared dependence on the medication (Moritz et al., 2014), because they experienced recovery or felt able to manage mental health difficulties without medication (Moritz et al., 2014), because taking medication was seen as negating the validity of personal experience (Moritz et al., 2013), because they were afraid of neuroleptic medication (Hudson et al., 2004), or because they rejected medication generally (Moritz et al., 2013). Non-adherence due to medication beliefs was reported by 11% (Ascher-Svanum et al., 2010; Hudson et al., 2004) to 58% (Moritz et al., 2013) of participants. Medication beliefs were not reported as an adherence reason in any of the quantitative studies. 3.8.1.2. Qualitative findings. Most findings concerned non-adherence with five studies reporting reasons in this domain (Table 2). In Kikkert et al. (2006), two clusters of reasons were of relevance, described as ‘negative medication attitudes’ and ‘negative expectations’. Example quotes revealed decisions to discontinue medication due to feeling better without it, believing that medication is unnatural or would cause harm, or a traumatic experience when first given medication. In Meshach et al. (2014) and Teferra et al. (2013), participants reported discontinuation when feeling better; “The reason I stopped taking the medication was that I felt I was now well and wanted to see how I could cope without the drugs” [(Meshach et al., 2014, p. 252).] “Many of them tell me that they stopped taking the medication because they felt better. When they get better and start to work and lead a normal social life, they don't see the need to continue taking the medication” [(Teferra et al., 2013, p. 4, field worker quote following discussion with service-users).] Related findings included not taking neuroleptic medication due to feeling it was unneeded (Tranulis et al., 2011), not believing in medication continuation once health is restored (Teferra et al., 2013) and feeling frustrated at unending treatment (Meshach et al., 2014). Rogers et al. (1998) reported participants being non-adherent through increasing and decreasing prescribed medication doses as a way of coping; reflecting beliefs around usefulness of medication as determined by current situation and level of distress. Kikkert et al. (2006) was the only study to address medication beliefs in relation to adherence, and reported two clusters, described as ‘insight’ and ‘positive medication attitudes and expectations’. Example quotes revealed that adherence occurred if people perceived medication to be effective, when they believed it was needed, and following previous good experiences with medication.

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

3.8.2. Religious beliefs 3.8.2.1. Mixed methods findings. Borras et al. (2007) reported results supporting previous findings that medication beliefs and also religious beliefs influence adherence. Amongst non-adherent participants, religious rather than medical representations of illness were more prominent and 58% of non/partially-adherent participants reported conflicts between religious beliefs and neuroleptic use: “My illness is an ordeal sent by God, medication is not part of God's plan and I will not take it” [(Borras et al., 2007, p. 1242).] “Medication hindered me to grow in my spiritual life” [(Borras et al., 2007, p. 1242).] Furthermore, 71% of non/partially-adherent participants reported feeling uneasy discussing religion with psychiatrists. Conversely, medical rather than religious representations of illness were more prominent amongst adherent participants. Religion gave meaning to medication, for example, people saw medication as being a gift from God. For others, medication was seen as separate from religion: “My illness is God's punishment for my sins. God and medication are two different things. Medication helps me to overcome my illness”. [(Borras et al., 2007, p. 1242).]

3.9. Relationships with other people 3.9.1. Quantitative findings The influence of relationships with other people was of importance in adherence. Eight studies reported relationships with others leading to non-adherence (see Table 2). Relationships with others were also reported as a reason for adherence in eight studies (see Table 3). Participants reported relationships with prescribers and clinical staff as important; for non-adherence when these relationships were negative or distrustful (Löffler et al., 2003; Moritz et al., 2009, 2013) and adherence when these relationships were positive (Löffler et al., 2003; Rosa et al., 2005; Sapra et al., 2014; Tunis et al., 2007; Yamada et al., 2006). Three studies reported deference or respect for authority as an adherence reason (Rosa et al., 2005; Tunis et al., 2007; Yamada et al., 2006). Moritz et al. (2009) did not use standardised adherence measures but asked participants a question about compliance and 78% reported taking their medication only because their physician wanted them to. Relationships with family and friends influenced adherence. Serviceusers reported non-adherence due to a lack of social support (Hudson et al., 2004) and due to having family, friends or health professionals opposed to neuroleptic medication or who advised discontinuation (Ascher-Svanum et al., 2010; Löffler et al., 2003; Matza et al., 2012; Moritz et al., 2013, 2014; Rosa et al., 2005). Conversely, adherence was reported due to family support (Sapra et al., 2014) and a positive family belief in medication/family or friend advice to take medication (Löffler et al., 2003; Matza et al., 2012; Moritz et al., 2014; Rosa et al., 2005; Tunis et al., 2007; Yamada et al., 2006). Between 6.5% (Löffler et al., 2003) and 31% (Moritz et al., 2013) rated the influence of relationships as a non-adherence reason while between 4% and 82% (Rosa et al., 2005) reported it as an adherence reason. 3.9.2. Qualitative findings Service-users in three studies reported non-adherence due to relationships with other people (Table 2), while participants in five studies reported that relationships with other people influenced adherence (Table 3).

89

Non-adherence occurred when support from others was lacking (Pyne et al., 2006; Teferra et al., 2013) or when others pressurised the participant to discontinue (Tranulis et al., 2011), “My family used to take care of me when I was sick, but now when I ask them to accompany me here (to the hospital) for the follow-up, none of them are cooperative. I have to beg them and plead with them. Even my wife says she has a lot of work to do and can't come” [(Lack of family support, Teferra et al., 2013, p. 4).] “My boyfriend told him that he couldn't work with me anymore, cuz I got sick from the medication, and he poured it all down the toilet” [(Pressure to discontinue, Tranulis et al., 2011, p. 890).] Non-adherence also resulted from anger towards prescribers or family (Pyne et al., 2006), or due to attitudes and behaviours of health workers, which were experienced as off-putting (Teferra et al., 2013). Participants reported adherence due to deference to authority (Rogers et al., 1998; Tranulis et al., 2011), or trusting clinician or family advice (Tranulis et al., 2011), due to improved relationships with others when taking medication (Pyne et al., 2006; Rogers et al., 1998), due to seeing others do well on medication (Pyne et al., 2006) and due to professional and non-professional support (Kikkert et al., 2006; Rogers et al., 1998), including taking medication to repay the care and support of family (Chang et al., 2013), “My mother had been enduring physical toil and mental anguish for 10 years, such as working, cooking and looking after me…my mother's love and care makes me want to keep taking my medication regularly” [(repaying family care, Chang et al., 2013, p. 275)]

3.10. Fear and coercion 3.10.1. Quantitative findings Fear, pressure and coercion were identified as reasons for adherence in six studies (see Table 3). People used medication for fear of psychotic symptoms or worsening mental health (Moritz et al., 2014; Tunis et al., 2007), fear of hospitalisation/rehospitalisation (Löffler et al., 2003; Patel et al., 2008; Rosa et al., 2005; Yamada et al., 2006) or because of pressure or force from other people or the health system (Löffler et al., 2003; Rosa et al., 2005; Tunis et al., 2007). Moritz et al. (2009) found 17% of participants reported only pretending to take medication. 3.10.2. Qualitative findings Four qualitative studies reported similar findings (Table 3). Participants reported adherence due to external pressure and/or coercion from family and clinicians (Rogers et al., 1998; Tranulis et al., 2011) or due to a desire to reduce adverse consequences of illness such as hospitalisation and coercion (Kikkert et al., 2006). Similarly, seeking to avoid a loss of freedom was a motivator for medication adherence (Pyne et al., 2006). Rogers et al. (1998) reported several adherence reasons related to fear or coercion. These included taking medication due to a fear of sanctions or due to the power of family and clinicians, “I feel intimidated by him (psychiatrist) and he's over-bearing and he's got the power to section me, he's got the power to do just whatever he wants to me so I'm very careful what I say to him” [(Rogers et al., 1998, p. 320)] Participants also reported adherence due to a lack of perceived autonomy to stop taking medication and due to fearing a loss of future care if medication was refused, “I don't like the sort of stuff running through your blood, you know, and like feeling doped up and that. I've had it since I was thirteen and I'm thirty two now so I'm sick of it you know, and I know I could just

90

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

tell them to sod off you know but if I really needed help, if I really got ill, they wouldn't be there” [(Rogers et al., 1998, p. 320).]

3.11. Side-effects 3.11.1. Quantitative findings Experiences with side-effects were often reported as a non-adherence reason and at times, as an adherence reason. Side-effects were reported as a reason for medication discontinuation in 11 studies (see Table 2), while three studies reported adherence related to side-effect experiences (see Table 3). In relation to non-adherence, side-effects caused discontinuation because of adverse reactions (Ascher-Svanum et al., 2010; Chen et al., 2011; Hudson et al., 2004; Matza et al., 2012), because too many side-effects were experienced (Moritz et al., 2009, 2013) or were intolerable (Moritz et al., 2014). Participants also reported discontinuation due to fear of potential side-effects (Moritz et al., 2014), because side-effects caused distress (Löffler et al., 2003; Sapra et al., 2014; Yamada et al., 2006), interfered with life goals (Sapra et al., 2014; Yamada et al., 2006), or were inconvenient (Rosa et al., 2005). (Deleted last sentence here on changes in appearance). Not all of the studies reporting side-effects as a reason for non-adherence detailed the specific side-effects experienced (Hudson et al., 2004; Löffler et al., 2003; Matza et al., 2012; Moritz et al., 2013; Yamada et al., 2006). Those which did provide specific side-effects reported service-user experiences of weight gain, sedation, insomnia, akathisia, tremors and hyperprolactinemia (Ascher-Svanum et al., 2010), vomiting, hyperglycaemia, insomnia, nausea, oversleeping, akathisia and tremors (Chen et al., 2011), lack of energy (Moritz et al., 2009), reduced sexual drive, metabolic disorder, dysphoria, visual disorder and vegetative changes (Moritz et al., 2014), changes in appearance (Sapra et al., 2014), and anticholinergic or extrapyramidal side-effects (Rosa et al., 2005). In relation to adherence, people reported medication continuation when they experienced fewer side-effects compared with previous medications (Sapra et al., 2014; Tunis et al., 2007) and when side-effects were mild and tolerable (Matza et al., 2012). Side-effects were reported as a reason for non-adherence by between 10.6% (Chen et al., 2011) and 80% (Moritz et al., 2013) of participants while 19% (Sapra et al., 2014) to 63% (Tunis et al., 2007) reported side effect experiences as a reason for adherence. 3.11.2. Qualitative findings All six studies exploring non-adherence reported side-effects as important, echoing quantitative findings (Table 2). Participants discussed the social aspects of side-effects, particularly embarrassment about movement disorders (Kikkert et al., 2006) while others reported not wanting to damage their bodies (Teferra et al., 2013; Tranulis et al., 2011), “the drugs make you weak especially when they are not taken with proper food. Our minds get clearer but our bodies are getting weaker” [(Teferra et al., 2013 p. 4)] Participants also discussed inhibited functioning due to side-effects (Teferra et al., 2013) including work and interpersonal functioning (Meshach et al., 2014; Pyne et al., 2006) social opportunities and interactions (Rogers et al., 1998), and enjoyment of leisure activities (Rogers et al., 1998). “You're walking around like a zombie and you're like sort of you can't join in with things, I wouldn't be talking to you like what I'm talking now… when you're on Melleril you can't even be bothered holding a conversation you know, you're just sat there saying yes and no, so I won't take it I'm sorry but I'm not taking it”

[(Rogers et al., 1998 p. 1317).] Experiences with side-effects were not generally discussed as an adherence reason in the qualitative studies, the only exception being Kikkert et al. (2006) where participants reported side-effect self-management as important. 3.12. Stigma 3.12.1. Quantitative findings Stigma, embarrassment, and desire to be normal were reasons for discontinuation in eight studies (Table 2). Stigma was reported by 4% (Rosa et al., 2005) to 48% (Moritz et al., 2009) of participants and was the most frequently reported barrier to adherence in Hudson et al. (2004). People reported experiencing stigma or embarrassment in relation to both illness and medication. Stigma was rarely reported as an adherence reason, the only exception being Moritz et al. (2014) where 31.3% of participants reported taking medication to avoid being called ‘crazy’. 3.12.2. Qualitative findings Stigma was reported as a discontinuation reason in four studies (Table 2). Participants reported non-adherence to avoid stigma and prove normalcy (Rogers et al., 1998; Tranulis et al., 2011), especially when functioning well (Teferra et al., 2013). “I, like many other people, was trying to prove that I wasn't mad by getting off my injection, you know and that would prove it, I thought” [(Rogers et al., 1998, p. 1321).] Participants described fears that medication-induced involuntary limb movements would signal their status as a psychiatric patient to other community members (Rogers et al., 1998). Furthermore, they believed that others would interpret medication use as a sign of schizophrenia (Rogers et al., 1998) and thus discontinued to prevent others from discovering their illness (Teferra et al., 2013). Stigma could have a detrimental impact on people's self-esteem (Tranulis et al., 2011) job prospects (Rogers et al., 1998; Teferra et al., 2013) and marriage prospects (Teferra et al., 2013). “It's completely ruined my capability of holding down a job. You see I can't get a job because nobody's going to employ me with me taking my medication and my injections” [(Rogers et al., 1998, p. 1321).] Stigma was not reported as an adherence reason in the qualitative studies. 3.13. Economic issues 3.13.1. Quantitative and qualitative findings Economic issues were reported as non-adherence reasons in one quantitative study (20% Ascher-Svanum et al., 2010) and three qualitative studies (Meshach et al., 2014; Pyne et al., 2006; Teferra et al., 2013). In Ascher-Svanum et al. (2010) and Pyne et al. (2006), high medication costs were reported as a reason for non-adherence. Participants in Ascher-Svanum et al. (2010) also reported non-adherence due to medication not being covered by health insurance. Two of the studies (Meshach et al., 2014; Teferra et al., 2013) were conducted in low income countries and economic issues impacted adherence in numerous ways. Some people were unable to afford medication or transportation costs to collect medication, others relied on family for financial assistance, causing difficulties when family were unable to afford medication and related costs (Meshach et al., 2014). Increased financial burden on poorer family members could cause hostility, criticism and abandonment of the service-user, leading to non-adherence (Meshach et al., 2014). This could occur due to participant's own choice or because of

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

family pressure to discontinue (Teferra et al., 2013). In addition to medication and transportation costs, service-users reported increases in appetite whilst taking medication, leading to hunger and further financial burden (Meshach et al., 2014; Teferra et al., 2013), “Yes, there are people who have said they couldn't tolerate the hunger and stopped taking it (the medication). I have also met families who made them stop taking the drug because they can't afford to feed them when the patients take the drug, as it makes them eat a lot”. [(Psychiatric nurse, Teferra et al., 2013, p. 4)] Frustration caused by economic difficulties could lead to lowered mood causing participants to lose interest in taking medication (Meshach et al., 2014). Furthermore, medication adherence could lead to poverty as side-effects impacted ability to work (Meshach et al., 2014). “There are times I just get fed up with life when I think about all my sufferings and my inability to cater for my family. At such times I don't take my medicine” [(Meshach et al., 2014, p. 253)]

3.14. Other reasons 3.14.1. Quantitative and qualitative findings Finally, amongst 11 quantitative studies (see Table 2) and one qualitative study (Pyne et al., 2006) there were a number of ‘other’ non-adherence reasons reported. These included denial of illness (Löffler et al., 2003; Patel et al., 2008; Rosa et al., 2005; Sapra et al., 2014), access barriers (Sapra et al., 2014; Yamada et al., 2006), forgetting or memory problems (Hudson et al., 2004; Moritz et al., 2009, 2013, 2014), wishing to try a new neuroleptic (Chen et al., 2011; Matza et al., 2012), homelessness (Hudson et al., 2004), substance abuse/use of street drugs or alcohol instead (Hudson et al., 2004; Pyne et al., 2006), being easily distracted (Pyne et al., 2006) and missing/sleeping through doses (Pyne et al., 2006). Only two studies (Kikkert et al., 2006; Rogers et al., 1998) mentioned ‘other’ reasons for adherence. These included countering the adverse effects of recreational drugs (Rogers et al., 1998), good accessibility of medication, being accurately informed about side-effects, having an understanding of the illness and being given a choice regarding whether or not to take medication (Kikkert et al., 2006). 3.15. Subgroup differences 3.15.1. Participants from low-income countries A key difference between subgroups of participants in relation to non-adherence concerned the influence of economic issues. As detailed above, economic difficulties were one of the most influential reasons for non-adherence in two studies (Meshach et al., 2014; Teferra et al., 2013). Both studies were conducted in sub-Saharan African countries (Nigeria and Ethiopia respectively), a region where over 50% of people live below the poverty line (United Nations Development Programme (UNDP), 2011). In Teferra et al. (2013) service-users reported that neuroleptics which caused increased appetite were subjectively distressing in the absence of additional food. Increased food requirements also placed economic burden on families. Teferra et al. (2013) reported that chronic food insufficiency is common in rural areas where participants were drawn from and for households with food shortage; healthy family members are given priority over those with mental health difficulties. Teferra et al. (2013) also found severe malnutrition was the second most common cause of death in the cohort of schizophrenia patients from which their sample was drawn. Meshach et al. (2014), reported that schizophrenia patients in Nigeria are at an increased risk of poverty due to social exclusion, stigma and discrimination, which in turn impacts employment prospects. Amongst their participants, even those who were competent at work were dismissed or avoided by

91

business associates as they were perceived to be ‘mad’. Economic difficulties were particularly problematic for those who had poor relatives who were unable to afford the additional financial burden caused by the prescription of neuroleptic medication. Economic difficulties were discussed far less frequently in the other studies. This may have been due to the relative wealth of other countries where participants were drawn from or due to the provision of free medication to those on low incomes in some countries. 3.15.2. Cultural differences in the influence of family and professionals There were some cultural differences in relation to the degree of influence of family beliefs on adherence. For three studies in non-Western countries (Chang et al., 2013, Taiwan; Rosa et al., 2005, Brazil; Yamada et al., 2006, Japan), adherence was largely influenced by positive family beliefs about taking neuroleptic medication. This was reported as a top three reason for adherence in the quantitative studies; 72% of participants in Rosa et al. (2005) and 68.9% of those in Yamada et al. (2006) reported this as an adherence reason. Although the item ‘positive family belief’ was also rated as an adherence reason in one Western study (Löffler et al., 2003), it was only reported by 23% of participants. In Chang et al. (2013), participants reported adherence due to the Chinese concept of filial piety, the notion that Chinese children are supposed to show extreme respect to parents. This includes repaying parents for care and not being a burden to parents. Participants reported adherence in order to fulfil their obligation to filial piety. Such findings may reflect cultural differences in expectations on collective versus individualistic thinking and behaviour, which in turn may impact on reasons for adherence. However, this would need further exploration. 3.15.3. Further subgroup differences Further subgroup comparisons are difficult due to the relative homogeneity in factors of interest. For example, comparing in-patient with out-patient groups to explore whether reasons differed according to health setting could be useful. However, included studies recruited outpatients only or reported results based on combining groups. Similarly, it would have been useful to explore whether reasons differed between service-users prescribed first or second-generation medication. Again, studies either did not detail medication type or included participants prescribed both first and second-generation medication without separating results on this basis. Finally, it would have been useful to compare first episode and chronic illness service-users to explore whether reasons changed over the course of mental health difficulties and/or with prolonged contact with mental health services. However, not all studies detailed illness duration information. Those that did detail this information all reported fairly high mean illness durations indicating chronic illness experiences (range of 6.94–22.3 years with most studies reporting mean durations of 14 to 17 years). Some studies reported lower and higher illness durations but did not separate results on this basis. One study, Sapra et al. (2014) explicitly compared a first episode and multi-episode sample. Here, some differences were found; in relation to non-adherence, the multi-episode sample were significantly more likely to report non-adherence due to beliefs that medication was unnecessary or due to changes in appearance caused by medication. In relation to adherence, the first episode sample were significantly more likely to report adherence due to relationships with their prescribing doctor and non-prescribing therapist, and due to medication helping with life goals. The multi episode group were more likely to report adherence due to pressure and coercion. 4. Discussion This review sought to synthesise research exploring service-user reasons for adherence and non-adherence to neuroleptic medication in psychosis and schizophrenia-related disorders. With a specific focus on service-user accounts and with the aim to review reasons for

92

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

adherence as well as those for non-adherence, this review differed from other recent reviews in the area which have focused on factors associated with adherence and/or have focused on non-adherence only (e.g. Higashi et al., 2013; Richardson et al., 2013; Sendt et al., 2015). The following discussion is organised according to the research questions with possible explanations for reasons and clinical implications discussed. 4.1. What reasons do service-users report for non-adherence to neuroleptic medication? There were six main areas of reasons identified for neuroleptic nonadherence. These were; poor medication efficacy, personal beliefs that medication was unnecessary or that medication dependence was undesirable, the influence of relationships with other people including poor therapeutic relationships and beliefs of others that neuroleptics should not be taken, difficult side-effects, stigma at experiencing symptoms or taking neuroleptic medication and economic issues resulting in difficulties obtaining medication or coping with the economic impact of side-effects, such as the need for additional food. It may be important to consider that some of the reasons provided for non-adherence may be viewed similarly by both service-users and clinicians, while others may be understood differently. For example, both clinicians and service-users may consider adverse medication side-effects or poor medication efficacy to be reasons for nonadherence; indeed research has demonstrated this to be the case (Ascher-Svanum et al., 2010). However, the reasons described as ‘personal beliefs about medication’ which incorporated factors such as discontinuing medication due to beliefs that it was unnecessary, may be viewed as poor illness insight by clinicians. It is possible that previous research discussing the relationship between illness insight and adherence has indeed considered some of the same reasons as those in the current review. However, the emphasis on a clinician or researcher understanding of the reasons may have resulted in them being described as poor illness insight, a term that would be unlikely to be used by a service-user to describe their own understanding of their discontinuation reasons. 4.2. What reasons do service-users report for adherence to neuroleptic medication? There were three main areas of reasons identified for adherence to neuroleptic medication. These were; good medication efficacy, relationships with others including positive relationships with clinical staff, deference to authority and positive beliefs of friends and family regarding neuroleptic use. The third area related to experiences of fear and/or coercion including pressure or force from others to take medication and a fear of worsening symptoms, hospitalisation, loss of future care or freedom. In addition to the three main areas related to adherence, personal beliefs that neuroleptic intake was necessary was reported in one study and the experience of fewer or milder side-effects was reported as an adherence reason in three studies. 4.3. Are there any sub-group differences in reasons for adherence and non-adherence? Some sub-group differences were uncovered in the review, namely, differences between low and higher income countries in relation to the influence of economic difficulties on adherence, and cultural differences in relation to the degree of influence of family beliefs on adherence. One study (Sapra et al., 2014) reported differences between first-episode (FE) and multi-episode (ME) samples. However, these findings were based on samples taken from two different studies, the FE sample was over double the size of the ME sample, and the ME sample were recently relapsing. It is therefore difficult to ascertain whether such factors impacted on sample differences. Further exploration would be required to determine whether FE samples did indeed differ from chronic

samples. Additional sub-group comparisons were not possible due to the relative homogeneity of factors of interest or due to the lack of separation in reporting of reasons between subgroups. 4.4. Views on illness causation, recovery and treatment priorities The findings of the review may highlight the importance of beliefs regarding the causes of symptoms and the necessity of medication as well as the factors considered important for recovery by service-users. Such factors may influence neuroleptic adherence yet there may be a lack of understanding or consensus on the importance of these factors between service-users and mental healthcare providers. For example, research has demonstrated that service-users and members of the public tend to endorse psychosocial causes of schizophrenia (e.g. Pyne et al., 2006; Read, Magliano, and Beavan, 2013) while clinicians and mental health services may be more likely to favour biological causes (e.g. Coles, 2016; Kingdon, Sharma, and Hart, 2004). Differences in views on causation may impact treatment beliefs. In line with a biomedical model, the APA (2006) recommends long-term continuous treatment with neuroleptic medication to control symptoms and prevent relapse. However, if service-users favour psychosocial explanations for their difficulties, they may be less likely to view medication as beneficial in alleviating symptoms, particularly on a long-term basis, and may be more likely to value alternative methods of coping. Indeed, in Pyne et al. (2006), patients were less likely to feel that medication adherence was important to maintain health or control symptoms than providers were. The focus on a biomedical model of illness within mental health services may also contribute to differences in understanding of factors deemed important for recovery by service-users. Within such approaches, symptom reduction and reduced hospitalisation are regarded as key recovery outcomes (National Institute for Health and Clinical Excellence, 2010). However, in line with recovery literature, serviceusers within the current review reported that functioning, general health and well-being, achievement of life-goals, improved cognitive ability and being able to lead a normal life were important treatment targets. Thus the ability of medication to bring about benefits in these areas impacted people's willingness to comply with prescriptions. The importance of physical health, functioning and goal achievement may also have contributed to the finding that side-effects were reported as a non-adherence reason in 17 included studies. Often it was not the experience of side-effects per se that led to non-adherence but rather the severity or intolerability of side-effects or the impact of side-effects on other areas of life. In a related vein, stigma experiences were found to impact adherence with participants discontinuing medication due to a desire to feel ‘normal’ or due to fearing the impact of discrimination on self-esteem, job and relationship prospects; areas which were deemed important by service-users. Such findings highlight the complexity of the meaning of a positive outcome of treatment for service-users, which may go beyond symptom reduction to incorporate clinical, social and occupational domains (Cooke, 2014). Furthermore, in line with previous research (Klapheck, Nordmeyer, Cronjäger, Naber, and Bock, 2012), some service-users within the current review perceived benefits related to the experience of psychotic experiences and thus chose not to take neuroleptics as they did not desire symptom reduction. Again such findings may highlight misunderstanding between treatment priorities of patients with those of services that focus on symptom reduction as key in treatment. 4.5. The importance of relationships As detailed, relationships with other people influenced neuroleptic adherence. A key finding of the review was the influence of therapeutic alliance on adherence, in line with previous literature (Linden et al., 2001; Velligan et al., 2009). The mechanisms by which therapeutic alliance exerted an influence on adherence were not detailed in the studies

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

included in the review. However, there are a number of possible explanations. Moritz et al. (2013) discussed the importance for serviceusers of having difficult experiences, such as abuse and neglect, recognised by mental health professionals, while Klapheck et al. (2012) talked about the need for people with psychosis to be able to give subjective meaning to their experiences. Moritz et al. (2013) suggested that if clinicians overlook the importance of client's experiences in relation to their symptoms, then “non-adherence to medication might be a consequence” (pp. 920). If clinicians do tend to favour biological models of illness causation (see above) then it is indeed possible that the importance of client's experiences may be overlooked, impacting therapeutic alliance and adherence. Sapra et al. (2014) also discussed the fact that it is clinicians who ask service-users to take medication and it is therefore understandable that the perceived credibility of the prescriber and the strength of the relationship will be important influences on an individual's decision to continue or discontinue medication. Indeed, Löffler et al. (2003)found that 29% of those reporting a positive relationship with their prescriber as a reason for adherence also reported that perceived daily benefit of medication or relapse prevention were irrelevant influences on adherence, highlighting the importance of therapeutic relationships over symptom control for some people. In line with previous research (Santone et al., 2008), another key finding in relation to the influence of relationships concerned the views or support of family and friends on medication adherence and their instruction or advice to service-users on continuation or discontinuation. As family and friends are likely to be those who most closely witness the benefits or costs of neuroleptic use, it is perhaps understandable that they will voice their opinions on its use, which in turn may influence individual adherence. Furthermore, general perceptions on medication may influence how family and friends respond to an individual's neuroleptic use. Some research suggests that members of the general population may hold views against the use of medication generally, and neuroleptics particularly (Angermeyer and Matschinger, 1996). In such cases, family or friends may advise or direct service-users to discontinue neuroleptics. Given the above, it may be important for clinicians to consider the complexity of the people's relationships when discussing treatment choices with service-users. Unlike previous research (Sendt et al., 2015), the current review did not find evidence that social support and family involvement were more influential on adherence in younger study populations. Similar findings may have emerged had the current review included more first-episode samples. The influence of important others was also found to impact adherence via the experience of pressure or force from other people or services to take medication, or the fear of hospitalisation or the loss of future care. Within the U.K., The Mental Health Act (MHA, 1983, 2007) allows for compulsory hospitalisation and administration of medication without service-user consent. Given recent increases in the use of restricted freedoms imposed under the MHA (Health and Social Care Information Centre, 2016) it is not surprising that participants in the current review feared this potential consequence of medication discontinuation and adhered to medication accordingly. Swartz, Swanson, and Hannon (2003) found that 78% of clinicians reported believing that legal pressures made schizophrenia patients remain in treatment, and it is possible that this was the case for some participants in the current review. However, research has shown that fear of coerced treatment can lead people to avoid seeking help for mental health difficulties (Swartz et al., 2003) while feelings of empowerment and internal locus of control are associated with recovery (Harrow, Hansford, and Astrachan-Fletcher, 2009; Lloyd, King, and Moore, 2010; Morrison et al., 2013). Thus it may be important for clinicians and family members to develop understanding of the potential consequences of coercion and pressure for service-users and to develop more opportunities for patient choice in treatment (Morrison et al., 2012).

93

4.6. Limitations There were a number of limitations associated with the current review which could potentially affect interpretation or generalisibility of findings. These are outlined below. 4.6.1. Less reporting of non-adherence reasons In five studies, participants provided less non-adherence than adherence reasons (Chen et al., 2011; Matza et al., 2012; Rogers et al., 1998; Rosa et al., 2005; Yamada et al., 2006). For example, in Yamada et al. (2006), and Rosa et al. (2005), approximately 38%–100% of participants responded ‘none’ to each non-compliance item on the ROMI-J indicating that this item had no influence on non-adherence. Such findings could simply indicate fewer reasons existing for non-adherence, however, it is also possible that some participants feared discussions of non-adherence. This may especially be relevant as themes of coercion and pressure to take medication, and fear of sanctions for non-adherence, were reported in several studies. This pattern was reversed in Moritz et al. (2014), where more reasons for non-adherence than adherence were reported (mean of 6.23 and 3.49 respectively). This was an anonymous online study thus people may have felt more comfortable reporting non-adherence reasons. Teferra et al. (2013) conducted interviews and focus group discussions led by participant's treating physician, potentially reducing people's ability to openly discuss nonadherence. 4.6.2. Non-representative participants Two studies were considered to have recruited participants who were not representative of the intended population and thus did not score for this item during quality appraisal. In Yamada et al. (2006), only those with good compliance over the previous three months were recruited despite the study seeking to explore reasons for adherence and non-adherence. Therefore non-adherence reasons were only anticipated, rather than experienced. It is possible that such reasons might change during periods of actual medication discontinuation. In Sapra et al. (2014), all first-episode participants had to demonstrate a clinical response to oral neuroleptic medication and the multi-episode sample were only those willing to participate in psychosocial interventions. This may have resulted in samples who were more agreeable to services and thus less representative. In five studies (Ascher-Svanum et al., 2010; Chen et al., 2011; Moritz et al., 2009, 2013, 2014) it was unclear whether the samples were representative and thus they were rated as ‘can't tell’ during quality appraisal. 4.6.3. Measurement differences Not all studies employed the same measures to explore reasons for adherence and non-adherence, which may have affected comparison of results. Furthermore, not all measures used had known reliability and validity. Where studies did employ the same measure, differences existed in how reasons were reported. Definitions of adherence also varied amongst studies. For example, Yamada et al. (2006) defined non-adherence as one week of medication noncompliance over a twoyear period but Rosa et al. (2005) classified non-adherence as taking b75% of the prescribed dose over the last 30 days. Reasons could potentially have differed depending on whether adherence/non-adherence was short or long-term. Furthermore, for clinical trials, non-adherence was classified as complete discontinuation of the trial medication (Ascher-Svanum et al., 2010; Chen et al., 2011). Trial discontinuation may differ from real-world non-adherence (Rabinowitz, Levine, Barkai, and Davidov, 2009) and caution may be required in extracting data from clinical trials due to selection bias which may exclude potential participants, for example those not experiencing severe positive symptoms (Baloush-Kleinman et al., 2011). Finally, despite the best efforts of the research team, it is possible that not all relevant articles were retrieved and thus it is possible that differences or additional reasons for adherence and non-adherence may exist.

94

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95

Limitations considered, results of the current review demonstrate consistency in the findings of a variety of studies examining reasons for adherence and non-adherence. Data from clinical trials corroborate findings from real-world settings. This differs from other recent reviews, e.g. Sendt et al. (2015) which explored factors correlating with adherence and reported “mostly inconsistent results” (Sendt et al., 2015, p. 27). A potential reason for the greater consistency in findings in the current review may be due to the focus on exploring only service-user accounts rather than factors associated with adherence. It is possible that there are a number of commonalities amongst service-users with regard to personal influences on adherence. For example, medication efficacy, compatibility with personal beliefs and relationships with other people may be important influences for many people. 4.6.4. Conclusions and clinical implications The findings of the current review suggest that adherence and nonadherence to neuroleptic medication is a complex process influenced by a number of factors including medication efficacy, personal beliefs, relationships with others, side-effects, fear, coercion, stigma and economic issues. Such findings would suggest that decisions on neuroleptic use are not always related to symptoms of illness such as insight, as documented in previous literature (e.g. Higashi et al., 2013; Sendt et al., 2015), but instead reflect rational decision-making processes in which personal costs and benefits of neuroleptic use are considered (Morrison et al., 2012). The factors that are most salient for each person will vary according to their own specific set of needs and circumstances. In light of this, it may be useful for clinicians to initiate dialogue with service-users around how neuroleptic use may fit with their hopes and expectations for treatment, including a discussion of how neuroleptic use may be viewed and supported by family and friends, and how medication use fits with personal belief systems and resources, including economic resources. Currently, in line with clinical guidelines (NICE, 2014), conversations between clinicians and service-users on neuroleptic use may focus more on discussions of expected side-effects than compatibility with treatment expectations and life circumstances. This may potentially lead to later non-adherence when individuals find that treatment does not meet their needs and expectations. Indeed studies on shared decision making in psychiatry demonstrate that collaboration in decision-making is associated with increased adherence (Gonzalez, Williams, Noël, and Lee, 2005), greater client satisfaction (Malm, Ivarsson, Allebeck, and Falloon, 2003), increased participation of service-users and improved quality of decisions in relation to knowledge and values (Drake, Cimpean, and Torrey, 2009). It may also be important for conversations on medication use to incorporate discussions around whether neuroleptic use is the right treatment for the service-user, or whether alternative treatments, such as psychological therapy, may be more compatible with an individual's needs (Morrison et al., 2012). The development of open, collaborative discussions in routine practice may serve to promote treatment choice for service-users and engage more people in services. Furthermore, the provision of education, information and support may be useful at a societal level to help to address some of the difficulties highlighted by current participants, such as perceived stigma at taking medication. The development of better clinician-client dialogue combined with education and support may help to alleviate both personal distress and economic burden caused by the experience of psychotic symptoms. References American Psychiatric Association (2006). Evidence-based treatments for schizophrenia: Information for families and other supporters. Arlington, VA: American Psychiatric Association. Angermeyer, M. C., & Matschinger, H. (1996). Public attitude towards psychiatric treatment. Acta Psychiatrica Scandinavica, 94(5), 326–336. Ascher-Svanum, H., Nyhuis, A. W., Stauffer, V., Kinon, B. J., Faries, D. E., Phillips, G. A., ... Naber, D. (2010). Reasons for discontinuation and continuation of antipsychotics in the treatment of schizophrenia from patient and clinician perspectives. Current Medical Research and Opinion, 26(10), 2403–2410.

Baloush-Kleinman, V., Levine, S. Z., Roe, D., Shnitt, D., Weizman, A., & Poyurovsky, M. (2011). Adherence to antipsychotic drug treatment in early-episode schizophrenia: A six-month naturalistic follow-up study. Schizophrenia Research, 130(1), 176–181. Barbui, C., Kikkert, M., Mazzi, M. A., Becker, T., Bindman, J., Schene, A., ... Tansella, M. (2009). Comparison of patient and clinician perspectives in the assessment of antipsychotic medication adherence. Psychopathology, 42(5), 311–317. Barkhof, E., Meijer, C. J., de Sonneville, L. M., Linszen, D. H., & de Haan, L. (2012). Interventions to improve adherence to antipsychotic medication in patients with schizophrenia – A review of the past decade. European Psychiatry, 27(1), 9–18. Borras, L., Mohr, S., Brandt, P. Y., Gillieron, C., Eytan, A., & Huguelet, P. (2007). Religious beliefs in schizophrenia: Their relevance for adherence to treatment. Schizophrenia Bulletin, 33(5), 1238–1246. Byerly, M. J., Nakonezny, P. A., & Lescouflair, E. (2007). Antipsychotic medication adherence in schizophrenia. Psychiatric Clinics of North America, 30(3), 437–452. Centre for Reviews and Dissemination (2016). PROSPERO. Retrieved January 21st, 2016, from http://www.crd.york.ac.uk/PROSPERO Chang, Y. T., Tao, S. -G., & Lu, C. L. (2013). Qualitative inquiry into motivators for maintaining medication adherence among Taiwanese with schizophrenia. International Journal of Mental Health Nursing, 22(3), 272–278. Chen, J., Ascher-Svanum, H., Nyhuis, A. W., Case, M. G., Phillips, G. A., Schuh, K. J., & Hoffmann, V. P. (2011). Reasons for continuing or discontinuing olanzapine in the treatment of schizophrenia from the perspectives of patients and clinicians. Patient Preference and Adherence, 5, 547–554. Cooke, A. (2014). Understanding psychosis and schizophrenia. British Psychological Society, Division of Clinical Psychologywww.bps.org Coles, S. (2016). Understanding biology in adult mental health. Clinical psychology bite-size (47). Nottingham Healthcare NHS Foundation Trust. Drake, R. E., Cimpean, D., & Torrey, W. C. (2009). Shared decision making in mental health: Prospects for personalized medicine. Dialogues in Clinical Neuroscience, 11(4), 455–463. Dunayevich, E., Ascher-Svanum, H., Zhao, F., Jacobson, J. G., Phillips, G. A., Dellva, M. A., & Green, A. I. (2007). Longer time to antipsychotic treatment discontinuation for any cause is associated with better functional outcomes for patients with schizophrenia, schizophreniform disorder, or schizoaffective disorder. The Journal of Clinical Psychiatry, 68(8), 1–478. Engels, G., Francke, A. L., van Meijel, B., Douma, J. G., de Kam, H., Wesselink, W., ... Scherder, E. J. (2014). Clinical pain in schizophrenia: A systematic review. The Journal of Pain, 15(5), 457–467. Geekie, J. (2004). Listening to the voices we hear: clients' understandings of psychotic experiences. In J. Read, L. R. Mosher, & R. P. Bentall (Eds.), Models of madness. Hove: Routledge. Gonzalez, J., Williams, J. W., Noël, P. H., & Lee, S. (2005). Adherence to mental health treatment in a primary care clinic. The Journal of the American Board of Family Practice, 18(2), 87–96. Harrow, M., Hansford, B. G., & Astrachan-Fletcher, E. B. (2009). Locus of control: Relation to schizophrenia, to recovery, and to depression and psychosis—A 15-year longitudinal study. Psychiatry Research, 168(3), 186–192. Health and Social Care Information Centre (2016). http://www.hscic.gov.uk/ (accessed on 22.04.2016) Higashi, K., Medic, G., Littlewood, K. J., Diez, T., Granström, O., & De Hert, M. (2013). Medication adherence in schizophrenia: Factors influencing adherence and consequences of nonadherence, a systematic literature review. Therapeutic Advances in Psychopharmacology, 3(4), 200–218. Hudson, T. J., Owen, R. R., Thrush, C. R., Han, X., Pyne, J. M., Thapa, P., & Sullivan, G. (2004). A pilot study of barriers to medication adherence in schizophrenia. The Journal of Clinical Psychiatry, 65(2), 1–478. Julius, R. J., Novitsky, M. A., Jr., & Dubin, W. R. (2009). Medication adherence: A review of the literature and implications for clinical practice. Journal of Psychiatric Practice, 15(1), 34–44. Kikkert, M. J., Schene, A. H., Koeter, M. W. J., Robson, D., Born, A., Helm, H., ... Gray, R. J. (2006). Medication adherence in schizophrenia: Exploring patients', carers' and professionals' views. Schizophrenia Bulletin, 32(4), 786–794. Kingdon, D., Sharma, T., & Hart, D. (2004). What attitudes do psychiatrists hold towards people with mental illness? The Psychiatrist, 28(11), 401–406. Klapheck, K., Nordmeyer, S., Cronjäger, H., Naber, D., & Bock, T. (2012). Subjective experience and meaning of psychoses: The German subjective sense in psychosis questionnaire (SUSE). Psychological Medicine, 42(1), 61. Lambert, M., Conus, P., Eide, P., Mass, R., Karow, A., Moritz, S., ... Naber, D. (2004). Impact of present and past antipsychotic side effects on attitude toward typical antipsychotic treatment and adherence. European Psychiatry, 19(7), 415–422. Landis, J. R., & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 159–174. Leucht, S., Tardy, M., Komossa, K., Heres, S., Kissling, W., Salanti, G., & Davis, J. M. (2012). Antipsychotic drugs versus placebo for relapse prevention in schizophrenia: A systematic review and meta-analysis. The Lancet, 379(9831), 2063–2071. Lieberman, J. A., Stroup, T. S., McEvoy, J. P., Swartz, M. S., Rosenheck, R. A., Perkins, D. O., ... Severe, J. (2005). Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. New England Journal of Medicine, 353(12), 1209–1223. Linden, M., Godemann, F., Gaebel, W., Köpke, W., Müller, P., Müller-Spahn, F., ... Tegeler, J. (2001). A prospective study of factors influencing adherence to a continuous neuroleptic treatment program in schizophrenia patients during 2 years. Schizophrenia Bulletin, 27(4), 585–596. Lloyd, C., King, R., & Moore, L. (2010). Subjective and objective indicators of recovery in severe mental illness: A cross-sectional study. International Journal of Social Psychiatry, 56(3), 220–229. Löffler, W., Kilian, R., Toumi, M., & Angermeyer, M. C. (2003). Schizophrenic patients' subjective reasons for compliance and noncompliance with neuroleptic treatment. Pharmacopsychiatry, 36(3), 105–112.

M. Wade et al. / Clinical Psychology Review 51 (2017) 75–95 Malm, U., Ivarsson, B., Allebeck, P., & Falloon, I. R. H. (2003). Integrated care in schizophrenia: A 2-year randomized controlled study of two community-based treatment programs. Acta Psychiatrica Scandinavica, 107(6), 415–423. Matza, L. S., Phillips, G. A., Revicki, D. A., Ascher-Svanum, H., Stauffer, V., Shorr, J. M., & Kinon, B. J. (2007). Developing a patient interview to assess reasons for antipsychotic discontinuation or continuation in the treatment of schizophrenia. Schizophrenia Bulletin, 33(2), 595–596. Matza, L. S., Phillips, G. A., Revicki, D. A., Ascher-Svanum, H., Malley, K. G., Palsgrove, A. C., ... Keefe, R. S. (2012). Validation of a patient interview for assessing reasons for antipsychotic discontinuation and continuation. Patient Preference and Adherence, 6, 521–532. McIntosh, A. M., Conlon, L., Lawrie, S. M., & Stanfield, A. C. (2006). Compliance therapy for schizophrenia. Cochrane Database of Systematic Reviews, 3. Meshach, O. E., King, K. M., & Fulton, J. A. (2014). Poor adherence to antipsychotics amongst schizophrenia patients in Nigeria. International Journal of Culture and Mental Health, 7(3), 246–258. Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. Annals of Internal Medicine, 151(4), 264–269. Morrison, A. P., Hutton, P., Shiers, D., & Turkington, D. (2012). Antipsychotics: Is it time to introduce patient choice? The British Journal of Psychiatry, 201(2), 83–84. Morrison, A. P., Shryane, N., Beck, R., Heffernan, S., Law, H., McCusker, M., & Bentall, R. P. (2013). Psychosocial and neuropsychiatric predictors of subjective recovery from psychosis. Psychiatry Research, 208(3), 203–209. Moritz, S., Favrod, J., Andreou, C., Morrison, A. P., Bohn, F., Veckenstedt, R., ... Karow, A. (2013). Beyond the usual suspects: Positive attitudes towards positive symptoms is associated with medication noncompliance in psychosis. Schizophrenia Bulletin, 39(4), 917–922. Moritz, S., Hünsche, A., & Lincoln, T. M. (2014). Nonadherence to antipsychotics: The role of positive attitudes towards positive symptoms. European Neuropsychopharmacology, 24(11), 1745–1752. Moritz, S., Peters, M. J., Karow, A., Deljkovic, A., Tonn, P., & Naber, D. (2009). Cure or curse? Ambivalent attitudes towards neuroleptic medication in schizophrenia and nonschizophrenia patients. Mental Illness, 1(1), 2. Mullins, C. D., Obeidat, N. A., Cuffel, B. J., Naradzay, J., & Loebel, A. D. (2008). Risk of discontinuation of atypical antipsychotic agents in the treatment of schizophrenia. Schizophrenia Research, 98, 8–15. National Institute for Health and Clinical Excellence (2010). Schizophrenia: Core interventions in the treatment and management of schizophrenia in adults in primary and secondary care (Updated edition ). Leicester and London: The British Psychological Society and the Royal College of Psychiatrists. National Institute for Health and Care Excellence (2014). Psychosis and schizophrenia in adults: Treatment and management (clinical guideline 178). London, UK: National Institute for Health and Care Excellencewww.nice.org.uk Pace, R., Pluye, P., Bartlett, G., Macaulay, A. C., Salsberg, J., Jagosh, J., & Seller, R. (2012). Testing the reliability and efficiency of the pilot mixed methods appraisal tool (MMAT) for systematic mixed studies reviews. International Journal of Nursing Studies, 49(1), 47–53. Patel, M. X., de Zoysa, N., Bernadt, M., & David, A. S. (2008). A cross-sectional study of patients' perspectives on adherence to antipsychotic medication: Depot versus oral. Journal of Clinical Psychiatry, 69(10), 1548–1556. Perkins, D. O., Johnson, J. L., Hamer, R. M., Zipursky, R. B., Keefe, R. S., Centorrhino, F., ... Tohen, M. (2006). Predictors of antipsychotic medication adherence in patients recovering from a first psychotic episode. Schizophrenia Research, 83(1), 53–63. Pluye, P., Gagnon, M. P., Griffiths, F., & Johnson-Lafleur, J. (2009). A scoring system for appraising mixed methods research, and concomitantly appraising qualitative, quantitative and mixed methods primary studies in mixed studies reviews. International Journal of Nursing Studies, 46(4), 529–546. Pluye, P., Robert, E., Cargo, M., Bartlett, G., O'Cathain, A., Griffiths, F., ... Rousseau, M. C. (2011). Proposal: A mixed methods appraisal tool for systematic mixed studies reviews. Retrieved on 14.01.2016 from http://mixedmethodsappraisaltoolpublic. pbworks.com Popay, J., Roberts, H., Sowden, A., Petticrew, M., Arai, L., Rodgers, M., ... Duffy, S. (2006). Guidance on the conduct of narrative synthesis in systematic reviews. ESRC Methods Programme, 15(1), 47–71. Puts, M. T. E., Tu, H. A., Tourangeau, A., Howell, D., Fitch, M., Springall, E., & Alibhai, S. M. H. (2013). Factors influencing adherence to cancer treatment in older adults with cancer: A systematic review. Annals of Oncology, 433. Pyne, J. M., McSweeney, J., Scott Kane, H., Harvey, S., Bragg, L., & Fischer, E. (2006). Agreement between patients with schizophrenia and providers on factors of antipsychotic medication adherence. Psychiatric Services, 57(8), 1170–1178. Rabinowitz, J., Levine, S. Z., Barkai, O., & Davidov, O. (2009). Dropout rates in randomized clinical trials of antipsychotics: A meta-analysis comparing first- and secondgeneration drugs and an examination of the role of trial design features. Schizophrenia Bulletin, 35(4), 775–788. Read, J., Magliano, L., & Beavan, V. (2013). Public beliefs about the causes of ‘schizophrenia’. In J. Read, & J. Dillon (Eds.), Models of madness (2nd ed.). Hove: Routledge. Richardson, M., McCabe, R., & Priebe, S. (2013). Are attitudes towards medication adherence associated with medication adherence behaviours among patients with

95

psychosis? A systematic review and meta analysis. Social Psychiatry and Psychiatric Epidemiology, 48(4), 649–657. Roberts, D. L., & Velligan, D. I. (2011). Medication adherence in schizophrenia. Drug Discovery Today: Therapeutic Strategies, 8(1), 11–15. Rogers, A., Day, J. C., Williams, B., Randall, F., Wood, P., Healy, D., & Bentall, R. P. (1998). The meaning and management of neuroleptic medication: A study of patients with a diagnosis of schizophrenia. Social Science & Medicine, 47(9), 1313–1323. Rosa, M. A., Marcolin, M. A., & Elkis, H. (2005). Evaluation of the factors interfering with drug treatment compliance among Brazilian patients with schizophrenia. Revista Brasileira de Psiquiatria, 27(3), 178–184. Rungruangsiripan, M., Sitthimongkol, Y., Maneesriwongul, W., Talley, S., & Vorapongsathorn, T. (2011). Mediating role of illness representation among social support, therapeutic alliance, experience of medication side effects, and medication adherence in persons with schizophrenia. Archives of Psychiatric Nursing, 25(4), 269–283. Santone, G., Rucci, P., Muratori, M. L., Monaci, A., Ciarafoni, C., & Borsetti, G. (2008). Attitudes toward medication in inpatients with schizophrenia: A cluster analytic approach. Psychiatry Research, 158(3), 324–334. Sapra, M., Weiden, P. J., Schooler, N. R., Sunakawa-McMillan, A., Uzenoff, S., & Burkholder, P. (2014). Reasons for adherence and nonadherence: A pilot study comparing firstand multi-episode schizophrenia patients. Clinical Schizophrenia & Related Psychoses, 7(4), 199–206. Sendt, K. V., Tracy, D. K., & Bhattacharyya, S. (2015). A systematic review of factors influencing adherence to antipsychotic medication in schizophrenia-spectrum disorders. Psychiatry Research, 225(1), 14–30. Sullivan, G., Wells, K. B., Morgenstern, H., & Leake, B. (1995). Identifying modifiable risk factors for rehospitalization: A case-control study of seriously mentally ill persons in Mississippi. American Journal of Psychiatry, 152(12), 1749–1756. Svestka, J., & Bitter, I. (2007). Nonadherence to antipsychotic treatment in patients with schizophrenic disorders. Neuro Endocrinology Letters, 28, 95–116. Swartz, M. S., Swanson, J. W., & Hannon, M. J. (2003). Does fear of coercion keep people away from mental health treatment? Evidence from a survey of persons with schizophrenia and mental health professionals. Behavioral Sciences & the Law, 21(4), 459–472. Teferra, S., Hanlon, C., Beyero, T., Jacobsson, L., & Shibre, T. (2013). Perspectives on reasons for non-adherence to medication in persons with schizophrenia in Ethiopia: A qualitative study of patients, caregivers and health workers. BMC Psychiatry, 13(1), 1. Tranulis, C., Goff, D., Henderson, D. C., & Freudenreich, O. (2011). Becoming adherent to antipsychotics: A qualitative study of treatment-experienced schizophrenia patients. Psychiatric Services, 62(8), 888–892. Tunis, S. L., Faries, D. E., Stensland, M. D., Hay, D. P., & Kinon, B. J. (2007). An examination of factors affecting persistence with initial antipsychotic treatment in patients with schizophrenia. Current Medical Research and Opinion, 23(1), 97–104. United Nations Development Programme (UNDP) (2011). UNDP in Africa. http://www. undp.org/africa Vauth, R., Löschmann, C., Rüsch, N., & Corrigan, P. W. (2004). Understanding adherence to neuroleptic treatment in schizophreniaı́. Psychiatry Research, 126(1), 43–49. Velligan, D. I., Weiden, P. J., Sajatovic, M., Scott, J., Carpenter, D., Ross, R., & Docherty, J. P. (2009). Expert consensus panel on adherence problems in serious and persistent mental illness. The expert consensus guideline series: Adherence problems in patients with serious and persistent mental illness. Journal of Clinical Psychiatry, 70(4), 1–46. Weiden, P. J., Mackell, J. A., & McDonnell, D. D. (2004). Obesity as a risk factor for antipsychotic noncompliance. Schizophrenia Research, 66(1), 51–57. Weiden, P. J., & Olfson, M. (1995). Cost of relapse in schizophrenia. Schizophrenia Bulletin, 21(3), 419–429. Weiden, P., Rapkin, B., Mott, T., Zygmunt, A., Goldman, D., Horvitz-Lennon, M., & Frances, A. (1994). Rating of medication influences (ROMI) scale in schizophrenia. Schizophrenia Bulletin, 20(2), 297–310. Weiden, P. J., & Ross, R. (2002). Why do patients stop their antipsychotic medications? A guide for families and friends. Journal of Psychiatric Practice, 8(6), 413–416. Weiden, P., & Zygmunt, A. (1997). Medication noncompliance in schizophrenia. Part I. Assessment. Journal of Practical Psychiatry and Behavioral Health, 3, 106–110. World Health Organisation (2003). Adherence to long-term therapies. Evidence for action. Geneva: World Health Organisation. World Health Organisation (2011). Schizophrenia. Geneva: World Health Organisation. World Health Organisation (2016). Webpage on schizophrenia. Accessed from http:// www.who.int/mental_health/management/schizophrenia/en/ (on 14.04.2016) Yamada, K., Watanabe, K., Nemoto, N., Fujita, H., Chikaraishi, C., Yamauchi, K., ... Kanba, S. (2006). Prediction of medication noncompliance in outpatients with schizophrenia: 2-year follow-up study. Psychiatry Research, 141(1), 61–69. Yamada, K., Watanabe, K., Yagi, G., & Asai, M. (1999). Factors and assessment of noncompliance of medication. Japanese Journal of Clinical Psychiatry, 28, 633–647 (in Japanese). Zygmunt, A., Olfson, M., Boyer, C. A., & Mechanic, D. (2002). Interventions to improve medication adherence in schizophrenia. American Journal of Psychiatry, 159(10), 1653–1664.