Table 2.

Evidence table on risk-factors of antipsychotic medication adherence and non-adherence.

Study type	Study	Outcome measures	Sociodemographic factors	Clinical factors	Treatment-related factors	Family involvement and therapeutic relations
RCT	1. Lindenmayer et al. (2009)	PANSS; MADRS; GAF; CGI-S; QLS; Simpson–Angus Scale; BARS and AIMS	Demographics (gender, age, ethnicity), illness characteristics, baseline weight (n.s.)	MADRS scores [baseline total mean (SD), adherent 13.90 (8.80) vs. non-adherent 15.85 (8.50), p = 0.010]; worsening PANSS depressive factor (HR = 1.2, 95% CI 1.06–1.35, p = 0.003); hostility (HR = 1.14, 95% CI 1.02–1.26, p = 0.020); change in PANSS total score and history of substance abuse (n.s.)	adverse events (n.s.) and weight change (n.s.)
	2. Morken et al. (2007)	Expressed emotion assessment based on CFI	male sex (OR = 6.11, 95% CI 1.2–29.74, p = 0.025)	Symptom severity (BPRS) (OR = 1.13, 95% 1.01–1.27, p = 0.034)		Patients living with family with high expressed emotion (OR = 36.43, 95% CI 2.18–608.01, p = 0.012); lower expressed emotion: 1st year (OR = 19.59, 95% CI 1.64–234.22, p = 0.019); both years (OR = 6.04, 95% CI 1.07–34.13, p = 0.042)
	3. Weiden et al. (2012)				Route of administration (n.s.)
	4. Kahn et al. (2008)				FGA vs. SGA (n.s.)
CT—open label, naturalistic, flexible dose	5. Guo et al. (2011)				FGA vs. SGA (n.s.)
CT—observational, longitudinal	6. Winton-Brown et al. (2017)	GAF, PANSS, CDSS, insight rating scale (David et al., 1992), relapse	Non-Caucasian (OR = 3, 95% CI 1.3–7.2, p = 0.01)	Use of illicit substances (OR = 0.3, 95% CI 0.1–0.5, p < 0.001)	Presence of EPS (OR = 8.1, 95% CI 1–65.3, p = 0.050)	Carer involvement (OR = 2.2, 95% CI 1–4.9, p = 0.048);
	7. Coldham et al. (2002)	QLS; ESRS; Bares Akathisia Scale; Premorbid Adjustment Scale	Young age (F = 4.5, p = 0.010); young age of onset (F = 6.7, p = 0.002); younger age (OR = 1.13, 95% CI 1.02–1.24, p = 0.015)	Relapse in first year (F = 4.16, p = 0.020); positive symptoms at 1 year (F = 7.88, p = 0.001); QoL at baseline (F = 3.45, p = 0.030); QoL at 1 y (F = 4.47, p = 0.010); poor premorbid functioning (OR = 0.07, 95% CI 0.00–0.24, p = 0.006); alcohol at baseline (F = 3.31, p = 0.020); alcohol at 1 y (F = 6.21, p = 0.003); cannabis at baseline (F = 3.17, p = 0.040); cannabis at 1 y (F = 3.17, p = 0.001); cannabis use (OR = 0.46, 95% CI 0.25–0.84, p = 0.012); alcohol abuse n.s.; insight at baseline (F = 4.08, p = 0.020); insight at 1 y (F = 4.26, p = 0.02)		lack of family involvement (OR = 0.19, 95%CI 0.05–0.75, p = 0.017)
	8. Mohamed et al. (2009)	GAF; ITAQ; DAI		Baseline illness insight (t = 2.48, p < 0.050); change in insight scores from baseline to follow-up up (ITAQ: 0.078, p < 0.001; DAI: 0.235, p < 0.001); positive attitudes toward medication (r = 0.154, p < 0.001)
	9. Quach et al. (2009)	GAF; SUMD; ROMI	Young age (OR = 1.79, 95% CI 1.16–2.75, p = 0.008)	Comorbid addiction (OR = 2.03, 95% CI 1.17–3.52); high global functioning (GAF) (OR = 1.73, 95% CI 1.07–2.81, p = 0.0300); unawareness of the effect of medication (OR = 2.34, 95% CI 1.44–3.82, p = 0.0010); negative attitude toward medication (OR = 2.13, 95% CI 1.43–3.17, p = 0.0001)		No upbringing by both parents (OR = 1.64 95% CI 1.11–2.42, p = 0.010); no key supporting relative (OR = 1.54, 95% CI 1.05–2.25, p = 0.030)
	10. Baloush-Kleinman et al. (2011)	CGI, SAPS; SANS; Cognitive Appraisal of Health Scale; Scale to Assess Unawareness of Mental Disorder; MacArthur Competence Assessment Tool; ESRS; Liverpool University Neuroleptic Side Effect Rating Scale; patient-rated Trust in Physician Scale; DAI; Visual Analog Scale (perception of family involvement)	Mode of admission, diagnosis of schizoaffective disorder, duration of illness (all n.s.)	Higher levels of insight into illness (t = 0.13, p = 0.009), awareness of the need for treatment (t = 3.82, p < 0.001), awareness of the social consequences of illness (n.s.)	Side-effects in adherent group (t = 2, p = 0.036); medication class (n.s.)	Perceptions of doctor–patient trust in the therapeutic alliance (t = 3, p = 0.012), perceived family involvement and attitudes toward medication in the family (t = 5, p < 0.001)
	11. Janssen et al. (2006)	GAF; DOTES; PANSS	Number of previous psychiatric hospitalizations (p < 0.010); involuntary admission (OR = 0.60, 95% CI 0.41–0.89, p < 0.050); no school graduation (OR = 0.59, 95% CI 0.41–0.86, p < 0.010); gender, primary diagnosis, first or multiple episode admission, duration of illness (all n.s.)	History of aggressive behavior (OR = 0.57, 95% CI 0.38–0.85), PANSS negative subscore above 25 (admission) (OR = 0.61, 95% CI 0.43–0.85, p < 0.01), PANSS paranoid/belligerence subscore above 9 (admission) (OR = 0.69, 95% CI 0.48–0.99, p < 0.01); substance disorder (OR = 0.52, 95% CI 0.32–0.85, p < 0.01)	Neurological side effects (n.s.); SGA monopharmacy at discharge > FGA mono or FGA + SGA (p < 0.005, χ2 = 17.6); FGA monotherapy switch to SGA vs. continue to take FGA (p < 0.001, χ2 = 12.6); mean dosage of initial antipsychotic treatment (n.s.); route of admin at admission (n.s.); depot vs. oral AP at discharge (p < 0.05, χ2 = 6.3)
	12. Acosta et al. (2009)	Amador Insight scale, PANSS	Age, sex, marital status, education level, living alone or with someone, length of illness, number of prior hospitalizations, time since last hospitalization (all n.s.)	PANSS conceptual disorganization (OR = 1.74, CI 0.96–3.17, p = 0.068); present and past substance use or abuse (n.s.); poor insight (OR = 1.22, 95% CI 1.01–1.48, p = 0.040)	Medication class and dosage (n.s.)
	13. Yang et al. (2012)	CDSS; CGI; PANSS; LUNSERS; DAI; SWN; Revised Insight Scale for Psychosis; WIS	All n.s.	CDSS (n.s.); CGI-S at baseline (r = −0.301, p < 0.050); CGI-S at 4 w (r = −0.403, p < 0.010); CGI-S at 8 w (r = −0.426, p < 0.010); PANSS score excitement [mean (SD), adherent 1.23 (0.43) vs. non-adherent 1.63 (0.83), p = 0.032], poor impulse control [mean (SD), adherent 1.23 (0.43) vs. non-adherent 1.58 (0.77), p = 0.049], and preoccupation [mean (SD), adherent 1.27 (0.58) vs. non-adherent 1.74(0.93), p = 0.035]; neurocognitive functions and insight (n.s.); attitudes toward medication (r = 0.49, p < 0.010)	Side effects (n.s.); polypharmacy (r = 0.358, p < 0.050);	Lower perceived support from significant other (only significant in parts of analysis; mean (SD), adherent 3.49 (1.54) vs. non-adherent 4.59 (1.62), p = 0.017);
CT—cross-sectional	14. Klingberg et al. (2008)	PANSS, GAF, SCL-GSI; UKU; EPS; AIMS	All n.s.	PANSS, GAF, SCL-GSI, global functioning and neurocognitive function (all n.s); lack of insight (OR = 0.41, 95% CI 0.183–0.915, p = 0.030); positive attitude toward medication (r = 0.382; p < 0.001)	Medication class and dosage (n.s.)	Frequency social contact, patient has a close friend, contact to relatives >10 h per week, influence family criticism, resignation and overprotection (all n.s.)
	15. Mutsatsa et al. (2003)	LUNSERS; ROMI; SAI; SWN; PANSS		Negative symptoms (t = −1.98, p = 0.050); disorganization (t = −2.01, p = 0.050); alcohol or non-alcohol substance misuse (n.s.); poor insight (t = 5.71, p < 0.001); negative attitudes toward medication (t = 3.01, p = 0.003)	Akathisia, parkinsonism, non-neurological side effects and subjective well-being (all n.s.)
	16. Bayle et al. (2015)	CGI; PANSS	Age <40 years (OR = 1.566, 95% CI 1.313–1.869, p < 0.001); diagnosis of schizophrenia (p = 0.008, χ2 test, adherent 43.7% vs. non-adherent 56.3%); sex, marital status, and living arrangements or occupation (all n.s.)	CGI-S ≥4 (OR = 1.986, 95% CI 1.518–2.598, p < 0.0001); lower insight (PANSS-G12) (OR = 1.459, 95% CI 1.225–1.738, p < 0.001)
	17. Molteni et al. (2014)	SE using DAI-30		Positive subjective experience with medication (DAI-30) (OR = 1.10, p = 0.002)
	18. Day et al. (2005)	PANSS; LUNSERS; attitude (DAI, Van Putten, Morisky); BIS; relationship with staff; admission experience		Attitude toward medication (r = 0.26, p = 0.001)		PEESSS (r = 0.73, p < 0.001); PEESSC (r = 0.79, p < 0.001); PEESSI (r = 0.16, p < 0.001)
	19. Meier et al. (2010)	Illness history (CSSRI); BPRS; GAF; MHS; LUNSERS; DAI	Age, marital status, and living arrangements or occupation gender (all n.s.)	Symptom scales (all n.s.); positive attitude to psychotropic medication (for clinician-rated adherence; T = 3.46; p < 0.001)	Side effects (n.s.); medication class (n.s.)
	20. Borras et al. (2007)	PANSS; CGI; “Multidimensional Measurement of Religiousness/Spirituality for Use in Health Research,” the “Religious Coping Index,” and a questionnaire on spiritual and religious adjustment to life events		PANSS positive symptoms (OR = 0.91, 95% CI 0.84–0.98, p < 0.001); substance abuse (OR = 4.0, 95% CI 1.5–10.6, p < 0.001)		Positively influenced by spiritual beliefs (31%); negatively influenced by spiritual beliefs (26%);
	21. Aldebot and de Mamani (2009)	BPRS; denial coping from COPE inventory	Gender, ethnicity, years of education (n.s.)	BPRS (n.s.); acceptance (n.s.); denial coping (t = −2.83, p = 0.008)
	22. McCabe et al. (2012)	PANSS; therapeutic alliance (Helping Alliance Scale)		PANSS total score (OR = 0.984, 95% CI 0.971–0.996, p = 0.014)		Therapeutic relationship (clinician-rated OR = 1.51, 95% CI 1.01–2.25, p = 0.042; patient-rated OR = 1.35, 95% CI 0.95–1.90, n.s.)
	23. (Jonsdottir et al., 2013)	PANSS; IDS; YMRS; BIS; UKU; NART; WIS; WASI; Bergen n-back test; DKEFS; WMS; CVLT	Age, gender, marital status, education (all n.s.); BMI full adherence > partial adherence (p = 0.012)	PANSS n.s.; IDS n.s.; YMRS n.s.; insight: BIS no adherence < full adherence (p = 0.013); neurocognition: WAIS n.s.; NART n.s.; WASI no adherence > full adherence p < 0.05; WMS and CVLT no adherence > full and partial adherence p < 0.05;executive functioning: DKEFS no adherence > full adherence p < 0.05; lifetime diagnosis of addiction or abuse of illicit drugs and alcohol partial adherence > full adherence (p = 0.000)	SE: UKU poor adherence significant for diarrhea, nausea, and orthostatism (p-value NR)