Determinants of Medication Compliance in Schizophrenia: Empirical and Clinical Findings

Determinants of Medication Compliance inSchizophrenia: Empirical and Clinical Findings

by Wayne S. Fenton, Crystal R. Blyler, and Robert K. Heinssen

Abstract

Advances in psychopharmacology have producedmedications with substantial efficacy in the treatmentof positive and negative symptoms of schizophreniaand the prevention of relapse or symptom exacerba-tion after an acute episode. In the clinical setting, theindividual patient's acceptance or rejection of pre-scribed pharmacological regimens is often the singlegreatest determinant of these treatments' effectiveness.For this reason, an understanding of factors thatimpede and promote patient collaboration with pre-scribed acute and maintenance treatment shouldinform both pharmacological and psychosocial treat-ment planning. We review the substantive literature onmedication adherence in schizophrenia and describe amodified health belief model within which empiricalfindings can be understood. In addition to factorsintrinsic to schizophrenia psychopathology, medica-tion-related factors, available social support, substanceabuse comorbidity, and the quality of the therapeuticalliance each affect adherence and offer potentialpoints of intervention to improve the likelihood of col-laboration. Because noncompliance as a clinical prob-lem is multidetermined, an individualized approach toassessment and treatment, which is often best devel-oped in the context of an ongoing physician-patientrelationship, is optimal. The differential diagnosis ofnoncompliance should lead to interventions that targetspecific causal factors thought to be operative in theindividual patient

Schizophrenia Bulletin, 23(4):637-651,1997.

Although advances in psychopharmacology have vastlyimproved the range of treatment options for schizophre-nia, outcome variance explained by the choice of medica-tion is likely small compared with that accounted for byhow and if the patient takes what is prescribed.

Compliance is defined as "the extent to which a person'sbehavior coincides with medical or health advice"(Haynes 1979, p. 2). The term has been criticized for over20 years as a reflection of an outmoded and paternalisticconception of the physician-patient relationship (Stimson1974; Holm 1993). Nonetheless, compliance is a wordoften used in clinical settings where for clinicians,patients, and families it remains one of the most vexingchallenges in psychopharmacology.

Compliance is difficult to quantify and study for sev-eral reasons. Clinicians' ability to identify which patientsdo not take medicine is limited (McClellan and Cowan1970; Norell 1981). Other measures of adherence includepatient or relative self-report, prescription renewals andpill counts, saliva and urine screens, or steady-state serumdeterminations. Concordance across different measures ofcompliance is often low, although self-reported noncom-pliance is corroborated more often than is self-reportedadherence (Rickels and Briscoe 1970; Gordis 1976;Boczkowski et al. 1985). Measurement is further compli-cated because compliance is rarely an all-or-none phe-nomenon, but may include errors of omission, mistakes indosage and timing, and taking medications that are notprescribed (Blackwell 1976).

A 1986 review of 26 studies using a variety of defini-tions and detection methods to assess medication useamong outpatients with schizophrenia reported a mediandefault rate of 41 percent (range, 10% to 76%) with oralmedications and 25 percent (range, 14% to 36%) withdepot injections over time periods up to 1 year (Young etal. 1986). Fifteen subsequent studies using varying defini-tions of noncompliance and many mixing patients takingoral and depot medications reported a median 1-month to2-year noncompliance rate of 55 percent (range, 24% to88%) (Hogan et al. 1983; Ayers et al. 1984; Carman et al.

Reprint requests should be sent to Dr. W.S. Fenlon, ChestnutLodge Hospital, 500 West Montgomery Ave., Rockville, MD 20850.

637

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Schizophrenia Bulletin, Vol. 23, No. 4, 1997 W.S. Fenton et al.

1984; Boczkowski et al. 1985; Gaebel and Pietzcker1985; Munetz and Roth 1985; Frank and Gunderson1990; Kelly and Scott 1990; Pristach and Smith 1990;Buchanan 1992; Adams and Howe 1993; Parker andHadzi-Pavlovic 1995; Razali and Yahya 1995;Macpherson et al. 1996a, 19966; Owen et al. 1996).

The belief that noncompliance is a direct result of dis-ease processes in schizophrenia dominates the clinical per-ception of noncompliance for these patients. Reportednoncompliance rates for schizophrenia, however, are in themiddle range of those reported for other common medicaldisorders. Medication noncompliance rates of 55 to 71percent have been reported for patients with arthritis (Berget al. 1993), 54 to 82 percent for patients with seizure dis-orders (Shope 1988), 20 to 57 percent for patients withbipolar affective disorder (Elixhauser et al. 1990), and 19to 80 percent for patients with diabetes (Friedman 1988).Half of patients with hypertension drop out of care within1 year, and only two-thirds of those who remain take ade-quate medication (Eraker et al. 1984).

A review of compliance with maintenance regi-mens—rheumatic fever prophylaxis, glaucoma, isoniazidfor tuberculosis, and self-administered insulin—found amean noncompliance rate for these long-term illnesses of54 percent (Sacket 1976). Compliance is lowest when thecondition is prolonged, treatment is prophylactic or sup-pressive, and the consequences of stopping treatment aredelayed. In disorders sharing these features, adherencedeclines with time (Blackwell 1973).

Through 1994 at least 14,000 English-language arti-cles have addressed compliance-related issues in medicalcare (Donovan 1995). Recent reviews converge in con-cluding that noncompliance is far better documented thanunderstood and that a focus on the patient's decisionmak-ing process is often a key missing ingredient in extantresearch (Trostle 1988; Morris and Schulz 1993; Donovan1995). Health belief (Becker and Maiman 1975; Kirschtand Rosenstock 1979) or health decision (Eraker et al.1984) models that emphasize a patient's subjective assess-ment of the risks and benefits of treatment in the contextof personal values and goals are advanced as best inte-grating available data on compliance research. Althoughthese models may require modification in disorders likeschizophrenia in which cognition and motivation areaffected directly by illness processes (Babiker 1986;Bebbington 1995), they do facilitate a shift in perspective:rather than viewing noncompliance as the patient's prob-lem, it is redefined as an indication that the therapeuticregimen is not assisting the individual patient to achievehis or her goals.

In this article, we review the substantive literature onmedication compliance in schizophrenia with an emphasis

on empirical studies that (1) identify current or predictivecorrelates of adherence and noncompliance and (2) assessinterventions targeted to improve adherence. These find-ings provide an empirical basis for the differential diagno-sis and understanding of noncompliance within a modi-fied health belief or health decision model.

Methods

This review is limited to studies of medication adherencein schizophrenia: studies focused on psychiatric patientswithout specifying diagnosis and those focused on adher-ence to other aspects of treatment, such as aftercareappointments, were not included. Clinical correlates ofcompliance that have been studied include patient socio-demographic features; illness characteristics, includingcomorbidity and insight; medication features, includingside effects and route and frequency of administration;family and social support; and treatment system charac-teristics, including quality of the physician-patient rela-tionship. Interventions studied include reinforcement,education, skills training, and memory enhancement.

Potentially relevant English-language articles wereidentified from the English language psychiatric and psy-chological literature with the aid of computer searchesusing such key words as compliance, adherence, psy-chopharmacology, and schizophrenia. Bibliographiesfrom primary sources and reviews were then reviewed toidentify earlier relevant works. In addition to empiricalstudies (those including some measure of compliance orintention to comply), clinical reports were included forreview if they presented useful perspectives on social orpsychodynamic issues that would generally be inaccessi-ble to empirical study. In selecting studies for review, reli-ability testing and corroboration of patient self-reportwere not required criteria; had they been, very few studieswould be left for review.

Correlates of Adherence andNoncompliance

Patient-Related Factors.Patient demographic characteristics. As in other

medical disorders, demographic variables are not consist-ently associated with compliance in schizophrenia. Elevenstudies assessed the relationship between one or morepatient demographic characteristics and compliance (Leffand Wing 1971; Hoffman et al. 1974; Soskis 1978; Hoganet al. 1983; Pan and Tantam 1989; Buchanan 1992;Draine and Solomon 1994; Sellwood and Tamer 1994;Parker and Hadzi-Pavlovic 1995; Razali and Yahya 1995;

638

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Determinants of Medication Compliance Schizophrenia Bulletin, Vol. 23, No. 4, 1997

Owen et al. 1996). Eight of ten found no association withage, six of nine found no association with gender, four offive found no association with ethnicity, and four of fourno association with education or income. In contrast, twostudies found noncompliance associated with youth, threewith male gender, one with single marital status, and onewith African-Caribbean ethnicity.

Illness Characteristics.Illness history. Studies have failed to reveal an

association between compliance and age at onset andduration of illness (Buchanan 1992), age at first hospital-ization (McEvoy et al. 1984), and premorbid functioning(Adams and Howe 1993). Data relating the number ofprior hospitalizations to compliance are contradictory:Three studies indicated more prior hospitalizations(Nelson et al. 1975; Pan and Tantam 1989; Sellwood andTarrier 1994); two studies, fewer prior hospitalizations(Reilly et al. 1967; McEvoy et al. 1984); and one study,no difference in prior hospitalizations (Hogan et al. 1983)among patients judged to be noncompliant at an indexassessment. Although these data do not strongly supportthe contention that patients learn to adhere to medicationsafter repeated relapses, hospitalization may improve com-pliance in the period immediately after discharge: Twostudies assessing compliance among patients before andafter an index admission showed significant decreases innoncompliance rates at followup (Christensen 1974;Owen et al. 1996).

The association between compliance and future hos-pitalization risk is far less equivocal: Seven studies indi-cated that patients rated as noncompliant have a 6-monthto 2-year risk of relapse that is an average of 3.7 timesgreater than patients rated as compliant (Leff and Wing1971; Linn et al. 1982; Gaebel and Pietzcker 1985;Munetz and Roth 1985; Kashner et al. 1991; McFarlane etal. 1995; Parker and Hadzi-Pavlovic 1995). Two addi-tional studies did not allow calculation of relative risk, butidentified irregular medication adherence as a significantpredictor of relapse (Falloon et al. 1978; Verghese et al.1989). The magnitude of elevated risk associated withnoncompliance seems comparable with that conveyed byrandomization to placebo in maintenance trials(Baldessarini et al. 1990). Consistent with these data isthe finding of recent medication noncompliance in the his-tory of 38 to 68 percent of relapsed patients (Reilly et al.1967; Christensen 1974; Herz and Melville 1980; delCampo et al. 1983; Parker and Hadzi-Pavlovic 1995;Owen et al. 1996). Because relapse typically occursweeks to months after the discontinuation of medication,however, patients only rarely attribute it to noncompliance(Chien 1975; Herz and Melville 1980).

Although the association between noncompliance andrelapse is robust, causality is likely bidirectional:Crawford and Forrest (1974) found erratic pill-takingassociated with worsening symptoms for patients takingplacebo tablets while maintained on depot phenothiazineinjections. Likewise, in a study that randomized schizo-phrenia patients to drug or placebo groups 2 months afterhospital discharge, Hogarty et al. (1973) found that about50 percent of the relapsers from either group were non-compliant compared with only 15 percent of patients fromeither group who had not relapsed within 6 months.

Illness severity and subtype. Both inpatient drugrefusal and outpatient noncompliance are consistently asso-ciated with more severe ratings of psychopathology. Eightstudies assessed the relationship between symptom severityor global functioning and inpatient medication refusal,future outpatient medication compliance, or attitude towardmedication. Marder et al. (1983) found more severe psy-chopathology, including disorganization, hostility, and sus-piciousness, associated with inpatient drug refusal. Fiveinvestigations reported a positive association betweensymptom severity at or after discharge and poor outpatientcompliance (Renton et al. 1963; Van Putten et al. 1976;Kelly et al. 1987; Pan and Tantam 1989) or poor attitudetoward compliance (Draine and Solomon 1994). One studyfound only the Brief Psychiatric Rating Scale (Overall andGorham 1962) grandiosity score to be associated with poorcompliance (Bartko et al. 1988), and one study reported norelationship between symptom severity at discharge andfuture outpatient compliance (Ayers et al. 1984).

Seven studies investigating the relationship betweenparanoid suspiciousness, persecutory delusions, or schizo-phrenia subtype and medication adherence yielded mixedresults. Two studies reported that noncompliance preced-ing a hospitalization is more common among patientswith paranoid schizophrenia subtype (Reilly et al. 1967;Pristach and Smith 1990). An additional investigationreported greater noncompliance as measured with urinescreens among inpatients with paranoid delusions (Wilsonand Enoch 1967). In contrast, one investigation found noassociation between paranoid schizophrenia subtype andthe expressed willingness to take medications (Soskis1978), one study found no association between subtypeand self-reported outpatient compliance (Hoffman et al.1974), and another study reported no association betweenparanoid ratings and missed depot appointments (Bartkoet al. 1988). In a study that may reconcile these discrepantfindings, Van Putten et al. (1976) found no associationbetween compliance and paranoid schizophrenia subtype,but noted that 85 percent of paranoid schizophreniapatients with delusions of persecution or influence habitu-ally complied with medications, whereas 92 percent of

639

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Schizophrenia Bulletin, Vol. 23, No. 4, 1997 W.S. Fenton et al.

paranoid patients with grandiose delusions habituallyrefused medications.

Cognition/memory. Neither overall intelligence(Adams and Howe 1993), discharge Mini-Mental StateExam score (Folstein et al. 1975; Buchanan 1992), norNeurobehavioral Cognitive Status Exam results (NorthernCalifornia Neurobehavioral Group 1988; Cuffel et al.1996) has been associated with compliance. The potentialassociation between specific neuropsychological deficitsand compliance has not been explored. A significant per-centage of outpatients, however, attribute noncomplianceto forgetting (Hoffman et al. 1974) or indicate that theybelieve reminders to take their medicine would be helpful(Serban and Thomas 1974).

Insight. The concept of insight has undergone con-siderable elaboration in recent years, coincident with ashift in explanatory focus from psychological to neurolog-ically based formulations of self-awareness deficits(Amador et al. 1993, 1994). A review emphasized themultidimensional nature of insight and its relative inde-pendence from symptom severity (Amador et al. 1991).Using a variety of self-report measures of illness aware-ness, nine studies assessed the relationship betweeninsight and adherence with prescribed pharmacologicalregimens. Poor insight was consistently associated withnoncompliance. Three studies reported an associationbetween poor insight assessed at admission or during hos-pitalization and medication noncompliance among inpa-tients (Lin et al. 1979; Marder et al. 1983; McEvoy et al.1989). Four studies reported an association between alack of insight at hospital admission, discharge, or post-discharge assessment and poor outpatient compliance(Nelson et al. 1975; Van Putten et al. 1976; Bartko et al.1988; Macpherson et al. 1996a, 1996b). Perhaps reflect-ing the fact that both insight and compliance can fluctuatewith clinical state, one study found that an awareness ofillness and medication compliance were related onlywhen measured concurrently (Cuffel et al. 1996). Onestudy reported an association between involuntary admis-sion status (an indirect index of insight) and poor 2-yearpostdischarge medication compliance, but no associationbetween ratings from attitude questionnaires and compli-ance (Buchanan 1992). Although a statistical relationshipbetween insight and adherence has been replicated in avariety of settings, several investigators noted that a siz-able subgroup of patients who do not believe they are illor require medication nonetheless are regularly compliant.

Other Health Beliefs. Except for the consistent rela-tionship between awareness of the presence of a psychi-atric illness and medication compliance, the associationbetween specific health beliefs and medication compli-ance among patients with schizophrenia is more ambigu-

ous. Kelly et al. (1987) found that greater perceived sus-ceptibility to rehospitalization was associated withincreased compliance among a population made uplargely of patients with schizophrenia. Hogan et al. (1983)also found that schizophrenia outpatients rated by theirtherapists as generally compliant with medication weremore likely than noncompliant patients to believe thatstaying on medication would prevent relapse.Noncompliant patients were more likely to believe thatmedication should only be taken when one feels sick, thatit would harm them physically, and that it is unnatural totake medication. In contrast, Pan and Tantam (1989)found no difference in beliefs about the possibility ofrelapse if maintenance treatment were stopped betweenregular attenders and outpatients who had missed two ormore appointments over 12 months at a depot neurolepticclinic. Buchanan (1992) found no association betweencompliance and self-appraisal of the likelihood of becom-ing ill again. Serban and Thomas (1974) found that mosthospitalized schizophrenia patients who reported that theydid not use prescribed medications between hospitaliza-tions failed to do so despite their expressed belief that reg-ular medication would be helpful. It should be noted thatthis disjunction between health beliefs and behaviors is byno means unique to schizophrenia.

Subjective Weil-Being. Perceived immediate benefitand a subjective sense of well-being derived from medica-tions seem to be associated more consistently with com-pliance than are expressed beliefs concerning susceptibil-ity to relapse. Patients who do not comply are likely tofeel that their medications do not help, are of no benefit,or are ineffective and unnecessary (Nelson et al. 1975;Soskis 1978; Lin et al. 1979; Herz and Melville 1980).Patients who consent to and comply with neuroleptics aremore likely to report feeling better (Marder et al. 1983),getting help (Buchanan 1992), and endorsing a direct(Hogan et al. 1983; Razali and Yahya 1995) or indirect(Adams and Howe 1993) beneficial effect of medicationon their well-being. After resolution of an acute episode,however, some patients stop medications because theyfeel well and therefore no longer in need of treatment(Reilly et al. 1967; Hoffman et al. 1974). As describedbelow in the section on side effects, to the extent that sub-jective well-being is associated with initial and long-termadherence, subjective discomfort is associated with med-ication refusal or noncompliance.

Co-Occurring Alcohol and Drug Use. Comorbid alco-hol or other substance abuse is common among individu-als with schizophrenia (Regier et al. 1990) and is a strongpredictor of neuroleptic noncompliance. Drake et al.(1989) studied 115 outpatients with schizophrenia and

640

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Determinants of Medication Compliance Schizophrenia Bulletin, Vol. 23, No. 4, 1997

found that 45 percent were occasional and 23 percentheavy alcohol users. More severe alcohol use and abusewere associated with medication noncompliance, psy-chosocial problems (including homelessness), disorga-nized and hostile behavior, medical problems, and fre-quent rehospitalizations over a 1-year followup. In asecond group of outpatients, Kashner et al. (1991) foundthat substance-abusing patients with schizophrenia were13 times more likely than non-substance-abusing patientsto be noncompliant with antipsychotic medication. In agroup of schizophrenia inpatients with a rate of self-reported noncompliance (72%) before admission that wastoo high to find an overall association with their substanceabuse history, Pristach and Smith (1990) reported that 62percent stopped taking medications while drinking.Patients often cited their physicians' advice not to mixmedicine and alcohol as a reason for intermittent noncom-pliance. Among inpatients assessed for substance abuseand followed after discharge for 6 months, Owen et al.(1996) reported that substance abuse in the 30 days beforeindex assessment was the strongest predictor of noncom-pliance at followup assessment. Substance abuseincreased the risk of noncompliance eightfold and seemedto interact with decreased outpatient contact to result inpoor clinical outcome. Of potential relevance to compli-ance are reports that tardive dyskinesia and akathisia maybe more prevalent among patients who abuse alcohol(Lutz 1976; Olivera et al. 1990; Dixon et al. 1992).

Medication-Related Factors.Side effects. Neuroleptic side effects that may be

particularly unpleasant include sedation, anticholinergiceffects, cognitive blunting, depression, sexual dysfunc-tion, and extrapyramidal syndromes—dystonia, akinesia,Parkinsonian effects, akathisia, and tardive dyskinesia(Weiden et al. 1986). Between one-quarter and two-thirdsof patients who unilaterally discontinue prescribed neu-roleptic medicines cite side effects as their primary reasonfor noncompliance (Renton et al. 1963; Reilly et al. 1967;Hoffman et al. 1974; del Campo et al. 1983). Among out-patients, both self (Falloon et al. 1978; Hogan et al. 1983;Kelly et al. 1987) and physician (Nelson et al. 1975; Panand Tantam 1989; Buchanan 1992) ratings of side effectsare associated with or predictive of noncompliance.

Although side effects are consistently associated withpoor maintenance adherence among outpatients, inpa-tients may not reliably report a history of side effects.Pristach and Smith (1990) did not find self-reported his-tory of neuroleptic side effects to be related to noncompli-ance among inpatients before admission. Marder et al.(1983) found no overall association between the self-reported history of side effects and inpatient drug refusal:

Inpatient drug refusers were described as including a sub-group of patients with severe side effects and a secondgroup in whom illness-related factors, such as denial, hos-tility, and grandiosity, were major determinants of refusal.Fleischhacker et al. (1994) attributed their failure to findan association between side effects during the first 4weeks of treatment and subsequent compliance amongpatients on haloperidol and clozapine to an aggressiveapproach to detect and treat adverse effects by changingmedications, lowering dosages, and prescribing concomi-tant medications.

In an important series of studies based on the obser-vation that normal volunteers differed in their reactions toa test dose of phenothiazines (Heninger et al. 1965), VanPutten et al. (1974) described a subgroup of schizophreniapatients who experienced a dysphoric response to a vari-ety of conventional neuroleptics available at that time.These patients felt miserable on neuroleptics, complainedabout drug effects, and pleaded to have their medicationsstopped or dosages reduced. In contrast, nondysphoricresponders reported that they "liked" medications and"felt better" on them. Akathisia was found to be the mostsignificant factor underlying a dysphoric neurolepticresponse and was highly associated with medicationrefusal or outpatient noncompliance or both. In somepatients, akathisia was experienced as a catastrophic senseof terror and impending annihilation that was phenomeno-logically similar to an exacerbation of psychosis (VanPutten 1974). More frequently, patients reported a subtleinner restlessness, anxiety, and inability to feel comfort-able in any position, a phenomenon that was incompatiblewith any productive activities and could not be toleratedfor any period of time. However, much of the akathisiaresponsible for outpatient noncompliance was describedas sufficiently mild as to remain undetected by anobserver who lacked a close and continuous relationshipwith the patient (Van Putten 1974). Two subsequent stud-ies replicated the finding of a significant associationbetween an initial dysphoric response to a test dose ofthiothixene or haloperidol, akathisia, and subsequent med-ication noncompliance (Van Putten et al. 1981, 1984). In athird independent sample, a greater proportion of patientswith a dysphoric than a syntonic response to a chlor-promazine test dose were noncompliant over a 9-monthfollowup period (Ayers et al. 1984). Together, theseresults point to a strong association among a dysphoricresponse to medication, akathisia, and medication refusalor noncompliance or both.

Dosage and agent Higher (Pan and Tantam 1989),lower (Nelson et al. 1975), and no different (Hogan et al.1983) neuroleptic dosages have been reported among out-patients rated as less compliant with maintenance treat-

641

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Schizophrenia Bulletin, Vol. 23, No. 4, 1997 W.S. Fenton et al.

ment. A curvilinear relationship between dosage and com-pliance, with very low doses associated with lack of effi-cacy and very high doses with excessive side effects,seems likely. However, few data are available that assessdifferential compliance rates to different agents. Carmanet al. (1984) found noncompliance rates as measured byserum and urine assays to be significantly higher amongpatients taking high-potency compared with low-potencyagents (65% vs. 13%). Among outpatients receiving phe-nothiazine injections, Carney and Sheffield (1976)reported a higher noncompliance rate (43%) among thosereceiving fluphenazine enanthate than among patientsreceiving fluphenazine decanoate or flupenthixoldecanoate (23%). The difference was attributed to ahigher rate of extrapyramidal side effects among thoseprescribed fluphenazine enanthate. We were unable tolocate any randomized controlled study assessing compli-ance with different agents.

The correlation between greater psychopathology atindex assessment and noncompliance likely reflects anassociation between the efficacy of prescribed treatmentand compliance. The finding that noncompliance rates forchlorpromazine were substantially higher than those forimipramine among depressed (70% vs. 44%) but notschizophrenia (32% vs. 25%) outpatients provides addi-tional support for a relationship between efficacy andcompliance (Willcox et al. 1965).

Route. A 1986 review of 26 studies reporting non-compliance rates indicated a lower mean default rate(25%) in studies of depot compared with oral (41%) neu-roleptics (Young et al. 1986). Based on the hypothesizedadvantage of depot preparations in improving compliance,six controlled studies comparing relapse rates amongpatients randomized to oral versus depot neurolepticswere reviewed more recently (Davis et al. 1993). Thesestudies suggest a modest advantage for the depot route inreducing relapse rates that may be greater in nonresearchsamples (Dixon et al. 1995). Changing patients to depotpreparations does not, however, seem to be an effectiveglobal strategy to eliminate noncompliance: Van Putten etal. (1976) found that 83 percent of habitually noncompli-ant schizophrenia outpatients who were switched todecanoate did not return with any regularity for bimonthlyinjections. Likewise, Falloon et al. (1978) reported that 73percent of schizophrenia patients returned to the commu-nity after hospital treatment who were irregular in theirtablet taking also missed at least one injection in 12months. Buchanan (1992) found no difference in compli-ance rates over 2 years postdischarge for patients takingoral and depot neuroleptics.

Although long-acting phenothiazine injections do notensure medication compliance because they must be

administered by a treatment provider, noncompliance withthis type of treatment can be detected quickly and withcertainty. Such noncompliance allows an assessment ofclinical impact for the individual patient and may triggerassertive interventions. For this reason, the major advan-tage of depot neuroleptics may be the ability to eliminatecovert noncompliance as a cause of clinical decompensa-tion (Schooler and Keith 1993).

Complexity of regimen. Although the complexityof a medication regimen is associated with complianceacross a broad range of medical disorders (Haynes 1976),only one (Razali and Yahya 1995) of four empirical stud-ies that focused exclusively on schizophrenia identified astatistically significant association between complexity ofregimen and compliance. Hoffman et al. (1974), Hogan etal. (1983), and Buchanan (1992) found no such associa-tion.

Environmental Factors.Family and social support. Social support, in gen-

eral, and the availability of family or friends to assist orsupervise medications, in particular, are consistently asso-ciated with outpatient adherence. Eight studies indicatedthat patients living with relatives or whose medicationsare supervised by relatives are more likely than thoselacking such support to maintain adherence to prescribedantipsychotic medication (Parkes et al. 1962; Renton et al.1963; Reilly et al. 1967; Hoffman et al. 1974; Nelson etal. 1975; Van Putten et al. 1976; Buchanan 1992; Razaliand Yahya 1995). An additional study found a nonsignifi-cant association between stability of living arrangementsand compliance (Owen et al. 1996). Causality is likelybidirectional in determining the association between fam-ily or social support and compliance. In a group of mostlyschizophrenia outpatients, Draine and Solomon (1994)found that better social functioning and more extensivesocial networks were related to positive attitudes towardmedication compliance. In addition, negative or stressfulsocial interactions may counteract the positive effect oncompliance of living with others (Reilly et al. 1967).

Practical barriers. In one investigation, 28 percentof patients who had reduced or stopped taking medica-tions before an inpatient admission cited financial burdenas the principal reason for discontinuation (Reilly et al.1967). Sullivan et al. (1995) found that family informantsreported that 7 percent of previously hospitalized patientslacked money for medication and 19 percent had missedmedication because of a lack of transportation to the phar-macy. Practical barriers or lack of access to care may beparticularly salient for homeless individuals, who areoften viewed as noncompliant (Interagency Council onthe Homeless 1992).

642

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Determinants of Medication Compliance Schizophrenia Bulletin, Vol. 23, No. 4, 1997

Physician-patient relationship. The clinical sup-position that a positive therapeutic alliance facilitatesmedication compliance finds empirical support in threestudies. Nelson et al. (1975) found that the single best pre-dictor of medication compliance among discharged schiz-ophrenia patients was the patient's perception of thephysician's interest in him or her as a person. Marder etal. (1983) found that, compared with patients who refusedmedications, schizophrenia inpatients who consented toneuroleptic treatment rated themselves as more satisfiedwith ward staff and their own physicians and felt that theirphysicians understood them, had their best interests inmind, and had explained the reasons for taking medica-tions and their potential side effects. Illness severity ortreatment response may partially explain these associa-tions. Frank and Gunderson (1990) found that 74 percentof patients with fair or poor therapeutic alliances (rated at6 months) failed to comply fully with prescribed medica-tion regimens during the next year and a half. In contrast,only 26 percent of schizophrenia patients with a goodalliance with their therapist (rated at 6 months) were non-compliant. In this study, the association between therapeu-tic alliance and medication compliance was independentof the patient's severity of psychopathology, type ordosage of medication, or inpatient/outpatient status.

Attitude of staff. Irwin et al. (1971) reported anonsignificantly higher noncompliance rate as determinedby urine screen (39% vs. 25%) among outpatients treatedby physicians who viewed medication as having question-able value, compared with patients of physicians whoviewed medication as an essential aspect of treatment.

Interventions. Psychosocial treatments for schizophre-nia often include promotion of medication compliance asan implicit or explicit goal. Data bearing on the efficacy ofindividual psychotherapy, social skills training, case man-agement, family psychoeducation, and assertive commu-nity treatment programs have been reviewed recently(Lehman et al. 1995). Here we review a more narrow setof interventions that specifically target medication compli-ance: reinforcement, education, and memory enhancement.

Reinforcement. When characterized by institu-tional surroundings, long waits, and impersonal orbureaucratic treatment, mental health clinics can beuninviting in a way that discourages attendance and com-pliance (Talbott et al. 1986; Chen 1991; Dencker andLiberman 1995). Making the setting more appealing byproviding reinforcement has improved adherence.Liberman and Davis (1975) designed a program to rein-force compliance by serving lunch at a monthly medica-tion clinic and allowing patients who tested positive forneuroleptics to select among several rewards, including

toiletries and personal items. Compared with patients ran-domly assigned to a control medication group, the experi-mental group showed better attendance, higher compli-ance levels, and more positive attitudes towardmedication. Cassino et al. (1987) successfully increasedattendance among schizophrenia patients at a decanoateclinic from 58 to 76 percent over a 17-week period byoffering brunch at morning sessions of the clinic. Offeringlunch-type food at an afternoon session, however, had lit-tle effect on attendance.

Education. Circumscribed educational interven-tions aimed at providing information about schizophreniaand its treatment have been ineffective at increasing com-pliance. Boczkowski et al. (1985) found no difference incompliance between an experimental group of outpatientsprovided a 30- to 50-minute information session and acontrol group for whom no specific attempt to focus onmedication or diagnosis was made. Macpherson et al.(1996a) randomly assigned patients to one of threegroups: one receiving three educational sessions atweekly intervals, one receiving a single educational ses-sion, and one having no educational intervention.Although participants in the group receiving the three ses-sions did have fewer knowledge errors at 1-month fol-lowup, their scores on a medication compliance scale didnot change. Similarly, Brown et al. (1987) documented anincrease in knowledge among schizophrenia outpatientswho received two instructional sessions 1 month apart,but noted that instruction did not affect independentlyrated compliance.

Skills training in areas related to medication seems tobe more effective than providing factual information.Eckman et al. (1990) designed a medication managementmodule that trained patients in four skill areas: obtaininginformation about medications, administering medicationand evaluating its benefits, identifying side effects, andnegotiating medication with health care providers. Themodule was delivered to patients in a variety of settingsfor 3 hours per week over 15 to 20 weeks. Upon comple-tion of the module and over a 3-month followup, knowl-edge about medication, skill utilization, and complianceimproved over baseline. Compliance assessed by thepatients' psychiatrists increased from 67 percent beforetraining to 82 percent after training, and complianceassessed by designated caregivers increased from 60 to 79percent.

Based on a randomized trial of individual and familyeducation, Kelly and Scott (1990) described two circum-scribed interventions that each reduced noncompliance at6-month followup. The individual intervention was deliv-ered by a health educator before the first two postdis-charge aftercare appointments and focused on increasing

643

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Schizophrenia Bulletin, Vol. 23, No. 4, 1997 W.S. Fenton et al.

the patient's ability to communicate with providers byexpressing concerns and asking questions. The familyintervention included up to three home visits that focusedon the development of an individualized behaviorally ori-ented compliance plan that, if necessary, included familyinvolvement in aftercare. The authors believed that thecritical ingredients in each of these effective interventionswere frequent repetition and behavioral modeling, ratherthan appealing to attitudes and beliefs. Skills training ofthis sort was described as least effective for patients withcomorbid alcoholism, who frequently dropped out oftreatment. Nelson et al. (1975) reported that a basic formof skills training—allowing schizophrenia patients to self-administer medications while hospitalized—increasedoutpatient compliance over a 6- to 24-week followup. Theintervention was effective, however, only for patientswho, based on Rorschach test data, accepted that theywere psychiatrically disturbed.

In a study of the impact of psychoeducationally ori-ented family therapy on medication compliance, Strang etal. (1981) randomly assigned to individual supportive orfamily therapy recently discharged schizophrenia patientsliving with a relative who exhibited high expressed emo-tion. Patients receiving family therapy that included spe-cific behavioral compliance strategies worked outbetween patient and family (Falloon et al. 1984) weremore likely to take their prescribed tablets, less likely torequire a change to depot neuroleptics, and showed higherand more stable neuroleptic plasma levels, despite identi-cal mean daily doses for the two groups.

Memory enhancement. Boczkowski et al. (1985)described a "behavioral tailoring" intervention thatincluded identifying a highly visible location for storingmedication, pairing medication intake with specific rou-tine behaviors, and prescribing a self-monitoring calendarwith tear-off slips. In a randomized trial, behavioral train-ing participants were more compliant at 1- and 3-monthfollowup than patients receiving a didactic educationalsession or control intervention.

Psychodynamic Considerations. Data concerning therole of psychodynamic factors in medication compliancederive from the observations of clinicians prescribingmedication to patients with schizophrenia over time.Three areas are consistently identified as pertinent: thepsychological meaning of medication to the individualpatient, the role of psychotic symptoms in maintainingself-esteem or personality organization, and issues relatedto transference and countertransference.

Psychological meaning. Clinicians have reported awide range of psychological meanings ascribed to med-ications. Patients who are preoccupied with issues ofauthority and control may be particularly prone to engage

in struggles over medication (Amdur 1979). In view ofthe significant loss of personal control associated withpsychosis, Diamond (1984) described noncompliance asan effort to regain control over one's life and feel better.Gutheil (1977) noted that individual patients may con-cretely equate medication with sickness ("/ / / need drugsI must be sick. The higher the dose the sicker I am. I'llstop being sick if I stop taking drugs"). Book (1987)described several dynamic issues affecting compliance:paranoid patients' experience of being controlled, poi-soned, or invaded by medication; the painful reminder ofa defective, "about to fall apart" self among patients whomake extensive use of denial; and the possibility thatpatients' attitudes toward medication are influenced byidentification with relatives who received similar medica-tion and experienced poor outcomes, such as suicide.

While viewing compliance primarily as a learnedbehavioral response, Falloon (1984) noted that somepatients fear that prolonged medication may lead todependence and addiction or equate the need for medica-tion with having a weak character. Taking medication mayalso be equated with physical or psychological weaknessso that the recovering patient who feels strong enoughwill stop taking medicine (Amdur 1979). Noncompliancein this context may be a test or gamble designed to deter-mine if the illness is still present (Morris and Schulz1993).

Medication may be be an area around which familyor interpersonal conflicts are enacted, so patients stopmedication to express anger toward a relative or mentalhealth professional (Kane 1983). Similarly, patients maydiscontinue medications in the face of increased pressureto improve functioning or on the verge of hospital dis-charge or beginning a new job, school, or rehabilitationprogram. In these circumstances noncompliance can beunderstood as an unconscious expression of the fear ofautonomy or as a communication that expectations havebeen set too high (Fenton and McGlashan 1995).

Psychological homeostasis. Psychotic symptomsmay be syntonic or serve to support an individual againstfurther personality disintegration or the collapse of self-esteem. Grandiose delusions cast the self as powerful, andpersecutory delusions mark the sufferer as worthy of spe-cial persecution. When psychosis provides a more posi-tive self-image than can be provided by reality, patientswill cling tenaciously to delusions and resist efforts toameliorate them (Van Putten et al. 1976; Corrigan et al.1990). Under these circumstances a frontal attack on psy-chotic symptoms is rarely effective and may precipitate acatastrophic collapse in self-esteem that leads to self-destructive behavior (Drake and Sederer 1986). Somepatients, particularly young men, who organize a sense ofself-cohesion around body well-being and activity may

644

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Determinants of Medication Compliance Schizophrenia Bulletin, Vol. 23, No. 4,1997

experience the physical effect of neuroleptics as a threatto self-organization (Heninger et al. 1965). Other patientsmay adapt to impending personality disorganization byglobally organizing in opposition to the will of others andmay resist medications as part of an overall effort tomaintain a tenuous sense of effectiveness or control.

Transference/countertransference. Clinicianshave noted that patients' views of medications arise inrelation to their attitude toward the prescriber and may bedistorted by these attitudes. In the context of a relation-ship perceived as authoritarian, the physical effects ofmedication may be perceived as rejecting, hostile, orthreatening (Sarwer-Foner 1960) or as a bodily attack orinvasion (Gutheil 1977). Other interpretations may castthe prescription as a negative dismissal of the patient, orthe patient may fear that the reduction of symptoms willbe accompanied by a parallel reduction in the physician'sinterest and attention. Positive transference toward theprescriber may spill over to the medication, which can beviewed as a "relationship equivalent" or gift (Gutheil1978).

Countertransference, or feelings and attitudes evokedin the physician by the patient, has also been described aspotentially undermining pharmacotherapy. Hopelessnessand frustration in the face of patient noncompliance and adesire to see the patient taught a lesson by suffering arelapse have been described as common countertransfer-ence reactions (Weiden et al. 1986; Book 1987). The urgeto abandon or humiliate the noncompliant patient mayalso be felt. In this respect, allowing the noncompliantpatient who leaves treatment against medical advice to doso with dignity can at least set the stage for more collabo-rative interactions should the patient return in the future(Diamond 1983; Frances and Weiden 1987).

Discussion

Major difficulties in empirical studies of noncomplianceinclude both ascertainment and the fact that the mostseverely noncompliant individuals leave treatment alto-gether. Those who remain are the "curiously ambivalent"individuals who continue in care, but do not adhere toprescriptions (Blackwell 1972). In view of these limita-tions it is perhaps surprising that available empirical stud-ies yield a relatively consistent set of correlates of non-compliance in schizophrenia, as outlined in table 1.

In evaluating these findings, it is important to notethat data suggest multiple possible causes of noncompli-ance. Because noncompliance can have many causes, itsstatistical association with any single factor is diluted bythe presence of patients in the sample for whom other fac-tors are causal. Similarly, the potential impact of interven-

Table 1. Empirical correlates of noncompliancein schizophrenia

Patient-related factorsGreater illness symptom severity or grandiosity or bothLack of insightSubstance abuse comorbidity

Medication-related factorsDysphoric medication side effectsSubtherapeutic or excessively high dosages

Environmental factorsInadequate support or supervisionPractical barriers, such as lack of money or transportation

Clinician-related factorsPoor therapeutic alliance

tions that focus on a single cause of noncompliance—inadequate knowledge or skills, for example—is limitedto those in the sample whose noncompliance derives fromthat cause.

Based on the empirical and clinical literaturereviewed here, a general set of treatment recommenda-tions to maximize the likelihood of compliance are out-lined in table 2. Given noncompliance in an individualpatient, however, a process of differential diagnosisshould lead to specific hypotheses regarding underlyingcauses that allow the design of focal and targeted clinicalinterventions.

Health Belief Model and Differential Diagnosis ofNoncompliance. The health belief model posits thathealth behavior is a product of an implicit and subjectiveassessment of the relative costs and benefits of compli-ance in relation to personal goals and the constraints ofeveryday life. Elements of this model include (1) individ-ual goals and priorities; (2) an evaluation of the perceivedadverse effects of illness and the personal risk of sufferingthese effects; (3) the individual's perception of the advo-cated health behavior's likely effectiveness and feasibility(the patient's subjective assessment of benefits weighedagainst the costs of treatment, including physical, psycho-logical, and practical disadvantages and barriers toaction); and (4) the availability of internal or external cuesto action that trigger health behavior (Becker and Maiman1975; Bebbington 1995). Because schizophrenia may dis-rupt illness perception and the capacity to plan and act,consideration of the cognitive and motivational resourcesavailable to assess risk and formulate action should be anadditional element of a health belief model applicable toschizophrenia.

Elements of this modified health belief model outlineareas of assessment pertinent to the differential diagnosisof noncompliance. A structured interview that explores

645

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Schizophrenia Bulletin, Vol. 23, No. 4, 1997 W.S. Fenton et al.

Table 2. General interventions to maximize thelikelihood of compliance

• Conduct an assessment of compliance history andrisk factors, including substance abuse and financial orother practical barriers, as part of the evaluation of everypatient.

• Allow sufficient time to know the patient as a personand to understand his or her personal goals, concerns,and psychodynamic issues. Place assisting the patient inmeeting self-defined goals at the center of treatment.

• Use a negotiated approach to medication. Create atherapeutic environment where deviation from recommen-dations can be discussed openly, rather than concealed.Show an interest in medication by asking in a nonauthori-tarian manner how much is being taken and the effects.Involve the patient in medication treatment by allowingself-regulation of dosage, if possible.

• Maximize efficacy and minimize side effects inchoosing agents and dosages. Attend seriously to all sideeffects and actively elicit and respond to concerns.

• Educate patient and family regarding the biologicalunderpinnings of illness, relapse prevention, and medica-tion side effects.

• Enlist support in the community, including family,friends, and employers. If needed, arrange for supervisedmedication administration.

• Employ cognitive and memory-enhancing strategiesif disorganization or forgetfulness is a problem.

• When the patient is rendered incompetent becauseof illness, be prepared to recommend judicial intervention.

• If the patient will not comply and is competent, man-age countertransference to allow for a continued relation-ship and the possibility of future treatment.

• Promote the patient's participation in activities thatcan compete with psychosis as sources of gratificationand serf-esteem.

many of these areas has been developed to facilitatepatient evaluation (Weiden et al. 1994).

Noncompliance may signal that patient and physiciangoals and priorities differ (Weiden et al. 1986). Main-taining sexual functioning, avoiding obesity, or not miss-ing work for a doctor's appointment, for example, may beof primary importance to the patient, whereas relapsereduction is the physician's priority. An assessment ofpatient goals forms the basis of a negotiated approach toprescription that is likely to enhance compliance(Eisenthal et al. 1979; Wilson 1995).

Since pursuing strategies designed to remedy inad-vertent noncompliance will fail when noncompliance ispurposeful, the differential diagnosis should attempt toseparate intentional from inadvertent noncompliance.Empirical studies suggest that from the patient's perspec-tive an immediate subjective dysphoria or other side

effects are the most significant costs of compliance.Psychological reports suggest that stigma, loss of the sickrole, or disturbed psychic homeostasis may be less dis-cernible costs. These costs may weigh heavily, particu-larly when accompanied by a low perceived benefit ofmedication that derives from a lack of knowledge, poorinsight, denial, or grandiosity. Interventions under suchcircumstances should target potentially modifiable ele-ments thought to be operative for the individual patient.Neuroleptic dosage reductions, use of adjunctive agents,or a change to an atypical neuroleptic with fewer sideeffects may reduce the cost side of the equation.Promoting participation in self-esteem-enhancing activi-ties may diminish the need to maintain delusional sourcesof self-esteem. Change to an agent with greater efficacymay reduce denial and grandiosity and create the potentialfor greater insight and perceived medication benefits.

Inadvertent noncompliance is associated whh severepsychopathology, including cognitive disorganization,memory impairment, or motivational deficits. The chaoticlife circumstances associated with substance abusecomorbidity, as well as such practical considerations asfinances and transportation, may also be operative. Herebehavioral skill and memory-enhancing interventions,assertive outreach efforts, and recruiting the assistance offamily or other supports to supervise medication aremajor treatment considerations.

New Neuroleptics Versus Depot Preparations. Be-cause of their reduced extrapyramidal side effects andgreater efficacy against positive and negative symptoms,new neuroleptics, such as clozapine, risperidone, olanza-pine, and sertindole, should provide greater patient bene-fits at a reduced perceived cost. Depot preparations havethe advantage of eliminating covert noncompliance andmaximizing the likelihood of steady-state neurolepticblood levels in patients with cognitive disorganization,memory disturbance, or motivational deficits. Althoughan empirical basis for choosing among these two pharma-cological interventions is not available, the full range offactors associated with noncompliance might be consid-ered clinically relevant. Patients with good insight and agood therapeutic alliance but who report intolerable sideeffects are likely the best candidates for a trial of a newagent. Patients with poor insight, grandiosity, or otherpsychotic symptoms or those with memory, motivational,or cognitive deficits might also benefit from a trial of anew agent in the absence of comorbid substance abuseand the presence of either a good therapeutic alliance oradequate family or other supervision to ensure regularadherence. Poor insight and severe psychopathology inthe absence of sufficient supervision favor the use ofdepot agents. Weiden (1995) has suggested that family

646

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Determinants of Medication Compliance Schizophrenia Bulletin, Vol. 23, No. 4,1997

factors may also have a bearing on the decision betweenatypical depot agents: Family concern over akinesia orother side effects favors a trial of a new agent, whereaschronic family conflict over taking oral medicationsfavors depot preparations.

It is useful to reassess the decision between depot andnew neuroleptics periodically. Some patients, for example,may require a considerable period of depot treatment toattain a level of clinical stability, therapeutic alliance, andinsight sufficient to render a trial of a new agent feasible.

Conclusion

The prevention and treatment of noncompliance are ofmajor importance in the care of patients with schizophre-nia. Although noncompliance has multiple causes, theempirical literature identifies a circumscribed set of fac-tors that alone or in varying combinations are likely to beoperative in individual cases. Exploring each of these fac-tors within a modified health belief model should allowfor differential diagnosis and an individualized approachto reducing noncompliance. A comprehensive understand-ing and integration of patient, illness, treatment, and envi-ronmental factors are needed to manage noncompliance(Kane 1986). This integrated approach can best be accom-plished within an ongoing physician-patient relationshipthat allows sufficient time for doctor and patient to knoweach other and maintain a collaborative therapeutic asso-ciation over time (Fenton and McGlashan 1995).

References

Adams, S.G., Jr., and Howe, J.T. Predicting medicationcompliance in a psychotic population. Journal of Nervousand Mental Disease, 181(9):558-560,1993.

Amador, X.F.; Andreasen, N.C.; Flaum, M.; Strauss,D.H.; Yale, S.A.; Clark, S.; and Gorman, J.M. Awarenessof illness in schizophrenia, schizoaffective and mood dis-orders. Archives of General Psychiatry, 51:826—836,1994.

Amador, X.F.; Strauss, D.H.; Yale, S.A.; Flaum, M.M.;Endicott, J.; and Gorman, J.M. Assessment of insight inpsychosis. American Journal of Psychiatry, 150:873-879,1993.

Amador, X.F.; Strauss, D.H.; Yale, S.A.; and Gorman,J.M. Awareness of illness in schizophrenia. SchizophreniaBulletin, 17(1):113-132, 1991.

Amdur, M.A. Medication compliance in outpatient psy-chiatry. Comprehensive Psychiatry, 20:339-346,1979.

Ayers, T; Liberman, R.P.; and Wallace, C.J. Subjective

response to antipsychotic drugs: Failure to replicate pre-dictions of outcome. Journal of Clinical Psychopharma-cology, 4<2):89-93, 1984.

Babiker, I.E. Noncompliance in schizophrenia.Psychiatric Developments, 4:329-337, 1986.

Baldessarini, R.J.; Cohen, B.M.; and Teicher, M.Pharmacological treatment. In: Levy, S.T., and Ninan,P.T., eds. Schizophrenia: Treatment of Acute Episodes.Washington, DC: American Psychiatric Press, 1990.pp. 61-118.

Bartko, G.; Herczeg, I.; and Zador, G. Clinical symptoma-tology and drug compliance in schizophrenic patients.Ada Psychiatrica Scandinavica, 77:74-76, 1988.

Bebbington, P.E. The content and context of compliance.International Clinical Psychopharmacology, 9(Suppl.5):41-50,1995.

Becker, M.H., and Maiman, L.A. Sociobehavioral deter-minants of compliance with health and medical care rec-ommendations. Medical Care, 13:10-24, 1975.

Berg, J.S.; Dischler, J.; Wagner, D.J.; Raia, J.J.; andPalmer-Shevlin, N. Medication compliance: A healthcareproblem. Annals of Pharmacotherapy, 27(Suppl.):5-24,1993.

Blackwell, B. The drug defaulter. Clinical Pharmacologyand Therapeutics, 13(6):841-848,1972.

Blackwell, B. Patient compliance. New England Journalof Medicine, 289:249-252,1973.

Blackwell, B. Treatment adherence. British Journal ofPsychiatry, 129:513-531, 1976.

Boczkowski, J.A.; Zeichner, A.; and DeSanto, N. Neuro-leptic compliance among chronic schizophrenic outpa-tients: An intervention outcome report. Journal of Con-sulting and Clinical Psychiatry, 53(5):666-671,1985.

Book, H.E. Some psychodynamics of non-compliance.Canadian Journal of Psychiatry, 32:115-117,1987.

Brown, C.S.; Wright, R.G.; and Christensen, D.B.Association between type of medication instruction andpatients' knowledge, side effects, and compliance.Hospital and Community Psychiatry, 38:55-60, 1987.

Buchanan, A. A two-year prospective study of treatmentcompliance in patients with schizophrenia. PsychologicalMedicine, 22:787-797,1992.

Carman, J.S.; Wyatt, E.S.; Fleck, R.; Martin, D.; andGold, M. Neuroleptic compliance in schizophrenic outpa-tients. Psychiatric Hospital, 15(4): 173-178, 1984.

Carney, M.W.P., and Sheffield, B.F. Comparison ofantipsychotic depot injections in the maintenance treat-ment of schizophrenia. British Journal of Psychiatry,129:476-481,1976.

647

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Schizophrenia Bulletin, Vol. 23, No. 4, 1997 W.S. Fenton et al.

Cassino, T.; Spellman, N.; Heiman, J.; Shupe, J.; andSklebar, H.T. Invitation to compliance: The Prolixinbrunch. Journal of Psychosocial Nursing, 25(10):15-16,1987.

Chen, A. Noncompliance in community psychiatry: Areview of clinical interventions. Hospital and CommunityPsychiatry, 42:282-287, 1991.

Chien, C.-P. Drugs and rehabilitation in schizophrenia. In:Greenblatt, M., ed. Drugs in Combination With OtherTherapies. New York, NY: Grune & Stratton, 1975.pp. 13-34.

Christensen, J.K. A 5-year follow-up study of male schiz-ophrenics: Evaluation of factors influencing success andfailure in the community. Ada Psychiatrica Scandinavica,50:60-72, 1974.

Coirigan, P.W.; Liberman, R.P.; and Engle, J.D. Fromnoncompliance to collaboration in the treatment of schiz-ophrenia. Hospital and Community Psychiatry, 41:1203—1211, 1990.

Crawford, R., and Forrest, A. Controlled trial of depotfluphenazine in out-patient schizophrenics. BritishJournal of Psychiatry, 124:385-391, 1974.

Cuffel, B.J.; Alford, J.; Fischer, E.P.; and Owen, R.R.Awareness of illness in schizophrenia and outpatient treat-ment adherence. Journal of Nervous and Mental Disease,184:653-659, 1996.

Davis, J.M.; Janicak, P.G.; Singla, A.; and Sharma, R.P.Maintenance antipsychotic medication. In: Bames, T.R.E.,ed. Antipsychotic Drugs and Their Side Effects. NewYork, NY: Academic Press, 1993. pp. 183-203.

del Campo, E.J.; Carr, C.F.; and Correa, E. Rehospitalizedschizophrenics: What they report about illness, treatmentand compliance. Journal of Psychosocial Nursing andMental Health Services, 21(6):29-33,1983.

Dencker, S.J., and Liberman, R.P. From compliance tocollaboration in the treatment of schizophrenia. Inter-national Clinical Psychopharmacology, 9(Suppl.5):75-78, 1995.

Diamond, R.J. Enhancing medication use in schizo-phrenic patients. Journal of Clinical Psychiatry, 44(Sect.2):7-14, 1983.

Diamond, R.J. Increasing medication compliance inyoung adult chronic psychiatric patients. In: Pepper, B.,and Ryglewicz, R., eds. Advances in Treating the YoungAdult Chronic Patient (New Directions for Mental HealthServices. No. 21). San Francisco, CA: Jossey-Bass, 1984.pp. 59-69.

Dixon, L.B.; Lehman, A.F.; and Levine, J. Conventionalantipsychotic medications for schizophrenia. Schizo-phrenia Bulletin, 21(4):567-577,1995.

Dixon, L.B.; Weiden, P., Haas, G.; Sweeney, J.; andFrances, A. Increased tardive dyskinesia in alcohol abus-ing schizophrenic patients. Comprehensive Psychiatry,33:121-122,1992.

Donovan, J.L. Patient decision making: The missingingredient in compliance research. International Journalof Technology Assessment in Health Care, 11:443—455,1995.

Draine, J., and Solomon, P. Explaining attitudes towardmedication compliance among a seriously mentally illpopulation. Journal of Nervous and Mental Disease,182(l):50-54,1994.

Drake, R.E.; Osher, F.C.; and Wallach, M.A. Alcohol useand abuse in schizophrenia. Journal of Nervous andMental Disease, 177:408-414,1989.

Drake, R.E., and Sederer, L.I. The adverse effects ofintensive treatment of chronic schizophrenia. Comprehen-sive Psychiatry, 27:313-326,1986.

Eckman, T.A.; Liberman, R.P.; Phipps, C.C.; and Blair,K.E. Teaching medication management skills to schizo-phrenic patients. Journal of Clinical Psychopharma-cology, 10(l):33-38, 1990.

Eisenthal, S.; Emery, R.; Lazare, A.; and Udin, H."Adherence" and the negotiated approach to patienthood.Archives of General Psychiatry, 36:393-398,1979.

Elixhauser, A.; Eisen, S.A.; Romeis, J.C.; and Homan,S.M. The effects of monitoring and feedback on compli-ance. Medical Care, 28:883-893, 1990.

Eraker, S.A.; Kirscht, J.P.; and Becker, M.H. Under-standing and improving patient compliance. Annals ofInternal Medicine, 100:258-268, 1984.

Falloon, I.R.H. Developing and maintaining adherence tolong-term drug-taking regimens. Schizophrenia Bulletin,10(3):412-417, 1984.

Falloon, I.R.H.; Boyd, J.L.; and McGill, C.W. FamilyCare of Schizophrenia: A Problem-Solving Approach tothe Treatment of Mental Illness. New York, NY: GuilfordPress, 1984.

Falloon, I.R.H.; Watt, D.C.; and Shepherd, M. A compara-tive controlled trial of pimozide and fluphenazinedecanoate in the continuation therapy of schizophrenia.Psychological Medicine, 8:59-70,1978.

Fenton, W.S., and McGlashan, T.H. Schizophrenia:Individual psychotherapy. In: Kaplan, H.I., and Sadock,B.J., eds. Comprehensive Textbook of Psychiatry. Vol. 1,6th ed. Baltimore, MD: Williams & Wilkins Company,1995. pp. 1007-1018.

Fleischhacker, W.W.; Meise, U.; Gunther, V.; and Kurz,M. Compliance with antipsychotic drug treatment:

648

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Determinants of Medication Compliance Schizophrenia Bulletin, Vol. 23, No. 4, 1997

Influence of side effects. Acta Psychiatrica Scandinavica,89:11-15, 1994.

Folstein, M.; Folstein, S.; and McHugh, P. "Mini-mentalstate." A practical method for grading the cognitive stateof patients for the clinician. Journal of PsychiatricResearch, 12:189-198,1975.

Frances, A., and Weiden, P. Promoting compliance withoutpatient drug treatment. Hospital and CommunityPsychiatry, 38:1158-1160, 1987.

Frank, A.F., and Gunderson, J.G. The role of the therapeu-tic alliance in the treatment of schizophrenia: Relationshipto course and outcome. Archives of General Psychiatry,47:228-236,1990.

Friedman, M. Compliance with chronic disease regimens:Diabetes. Journal of Diabetic Complications, 2:140-144,1988.

Gaebel, W., and Pietzcker, A. One-year outcome of schizo-phrenic patients: The interaction of chronicity and neu-roleptic treatment. Pharmacopsychiatry, 18:235-239,1985.

Gordis, L. Methodologic issues in measurement of patientcompliance. In: Sacket, D.L., and Haynes, R.B., eds.Compliance With Therapeutic Regimens. Baltimore, MD:Johns Hopkins University Press, 1976. pp. 51—66.

Gutheil, T.G. Improving patient compliance: Psycho-dynamics in drug prescribing. Drug Therapy, 7:82-95,1977.

Gutheil, T.G. Drug therapy: Alliance and compliance.Psychosomatics, 19:219-225, 1978.

Haynes, R.B. A critical review of the "determinants" ofpatient compliance with therapeutic regimens. In: Sacket,D.L., and Haynes, R.B., eds. Compliance With Thera-peutic Regimens. Baltimore, MD: Johns HopkinsUniversity Press, 1976. pp. 26-39.

Haynes, R.B. Introduction. In: Haynes, R.B.; Sackett,D.L.; and Taylor, D.W., eds. Compliance in Health Care.Baltimore, MD: Johns Hopkins University Press, 1979.pp. 1-10.

Heninger, G.; Dimascio, A.; and Klerman, G.L. Per-sonality factors in variability of response to pheno-thiazines. American Journal of Psychiatry, 121:1091-1094,1965.

Herz, M.I., and Melville, C. Relapse in schizophrenia.American Journal of Psychiatry, 137(7):801-805, 1980.

Hoffman, R.P.; Moore, W.E.; and O'Dea, L.F. A potentialrole for the pharmacist: Medication problems confrontedby the schizophrenic outpatient. Journal of the AmericanPharmaceutical Association, NS 14(5): 25 2-265,1974.

Hogan, T.P.; Awad, A.G.; and Eastwood, R. A self-reportscale predictive of drug compliance in schizophrenics:

Reliability and discriminative validity. PsychologicalMedicine, 13:177-183,1983.

Hogarty, G.E.; Goldberg, S.C.; and the CollaborativeStudy Group. Drug and sociotherapy in the aftercare ofschizophrenic patients. Archives of General Psychiatry,28:54-64,1973.

Holm, S. What is wrong with compliance? Journal ofMedical Ethics, 19:108-110,1993.

Interagency Council on the Homeless. Outcasts on MainStreet: Report of the Federal Task Force on Homelessnessand Severe Mental Illness. Washington, DC: The Council,1992.

Irwin, D.S.; Weitzel, W.D.; and Morgan, D.W. Pheno-thiazine intake and staff attitudes. American Journal ofPsychiatry, 127(12):67-71,1971.

Kane, J.M. Problems of compliance in outpatient treat-ment of schizophrenia. Journal of Clinical Psychiatry,44(Sect. 2):3-6,1983.

Kane, J.M. Prevention and treatment of neuroleptic non-compliance. Psychiatric Annals, 16:576-579, 1986.

Kashner, T.M.; Rader, L.E.; Rodell, D.E.; Beck, CM.;Rodell, L.R.; and Muller, K. Family characteristics, sub-stance abuse, and hospitalization patterns of patients withschizophrenia. Hospital and Community Psychiatry,42(2): 195-197, 1991.

Kelly, G.R.; Mamon, J.A.; and Scott, J.E. Utility of thehealth belief model in examining medication complianceamong psychiatric outpatients. Social Science Medicine,25(ll):1205-1211, 1987.

Kelly, G.R., and Scott, J.E. Medication compliance andhealth education among outpatients with chronic mentaldisorders. Medical Care, 28(12):1181-1197,1990.

Kirscht, J.P., and Rosenstock, I.M. Patients' problems infollowing recommendations of health experts. In: Stone,G.C.; Adler, N.E.; and Cohen, F., eds. Health PsychologyHandbook: Theories, Applications, and Challenges of aPsychological Approach to the Health Care System. SanFrancisco, CA: Jossey-Bass, 1979. pp. 189-215.

Leff, J.P., and Wing, J.K. Trial of maintenance therapy inschizophrenia. British Medical Journal, 3:599-604, 1971.

Lehman, A.F.; Carpenter, W.T., Jr.; Goldman, H.H.; andSteinwachs, D.M. Treatment outcomes in schizophrenia:Implications for practice, policy, and research. Schizo-phrenia Bulletin, 21(4):669-675, 1995.

Liberman, R.P., and Davis, J. Drugs and behavior analy-sis. Progress in Behavior Modification, 1:307-330,1975.

Lin, I.F.; Spiga, R.; and Fortsch, W. Insight and adherenceto medication in chronic schizophrenics. Journal ofClinical Psychiatry, 40:430-432, 1979.

649

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Schizophrenia Bulletin, Vol. 23, No. 4, 1997 W.S. Fenton et al.

Linn, M.W.; Klett, C.J.; and Caffey, E.M. Relapse of psy-chiatric patients in foster care. American Journal ofPsychiatry, 139:778-783, 1982.

Lutz, E.G. Neuroleptic-induced akathisia and dystoniatriggered by alcohol. Journal of the American MedicalAssociation, 236(21):2422-2423, 1976.

Macpherson, R.; Jerrom, B.; and Hughes, A. A controlledstudy of education about drug treatment in schizophrenia.British Journal of Psychiatry, 168:709-717, 1996a.

Macpherson, R.; Jerrom, B.; and Hughes, A. Relationshipbetween insight, educational background, and cognition inschizophrenia. British Journal of Psychiatry, 168:718—722, 19966.

Marder, S.R.; Mebane, A.; Chien, C.-R; Winslade, W.J.;Swann, E.; and Van Putten, T. A comparison of patientswho refuse and consent to neuroleptic treatment.American Journal of Psychiatry, 140(4):47CM72, 1983.

McClellan, T.A., and Cowan, G. Use of antipsychotic andantidepressant drugs by chronically ill patients. AmericanJournal of Psychiatry, 126(12):113-115, 1970.

McEvoy, J.R; Apperson, L.J.; Apelbaum, P.S.; Ortlip, P.;Brecosky, J.; and Hammill, K. Insight in schizophrenia:Its relation to acute psychopathology. Journal of Nervousand Mental Disease, 177:43-47, 1989.

McEvoy, J.R; Howe, A.C.; and Hogarty, G.E. Differencesin the nature of relapse and subsequent inpatient coursebetween medication-compliant and noncompliant schizo-phrenic patients. Journal of Nervous and Mental Disease,172(7):412^16, 1984.

McFarlane, W.R.; Lukens, E.; Link, B.; Dushay, R.;Deakins, S.A.; Newmark, M.; Dunne, E.J.; Horen, B.; andToran, J. Multiple-family groups and psychoeducation inthe treatment of schizophrenia. Archives of GeneralPsychiatry, 52:679-687, 1995.

Morris, L.S., and Schulz, R.M. Medication compliance:The patient's perspective. Clinical Therapeutics, 15:593-606,1993.

Munetz, M.R., and Roth, L.H. Informing patients abouttardive dyskinesia. Archives of General Psychiatry,42:866-871,1985.

Nelson, A.A.; Gold, B.H.; Huchinson, R.A.; and Benezra,E. Drug default among schizophrenic patients. AmericanJournal of Hospital Pharmacy, 32:1237-1242,1975.

Norell, S.E. Accuracy of patient interviews and estimatesby clinical staff in determining medication compliance.Social Science and Medicine, 15:57-61,1981.

Northern California Neurobehavioral Group, Inc. Manualfor the Neurobehavioral Cognitive Status Examination.Fairfax, CA: The Group, 1988.

Olivera, A.A.; Kiefer, M.W.; and Manley, N.K. Tardivedyskinesia in patients with substance use disorders.American Journal of Drug and Alcohol Abuse, 16:57-66,1990.

Overall, J.E., and Gorham, D.R. The Brief PsychiatricRating Scale. Psychological Reports, 10:799-812,1962.

Owen, R.R.; Fischer, E.P.; Booth, B.M.; and Cuffel, B.J.Medication noncompliance and substance abuse amongpatients with schizophrenia. Psychiatric Services,47(8):853-858, 1996.

Pan, P.-C, and Tantam, D. Clinical characteristics, healthbeliefs and compliance with maintenance treatment: Acomparison between regular and irregular attenders at adepot clinic. Ada Psychiatrica Scandinavica, 79:564—570,1989.

Parker, G., and Hadzi-Pavlovic, D. The capacity of a mea-sure of disability (the LSP) to predict hospital readmissionin those with schizophrenia. Psychological Medicine,25:157-163, 1995.

Parkes, CM.; Brown, G.W.; and Monck, E.M. The gen-eral practitioner and the schizophrenic patient. BritishMedical Journal, 1:972-976,1962.

Pristach, C.A., and Smith, CM. Medication complianceand substance abuse among schizophrenic patients.Hospital and Community Psychiatry, 41(12):1345-1348,1990.

Razali, M.S., and Yahya, H. Compliance with treatment inschizophrenia: A drug intervention program in a develop-ing program. Ada Psychiatrica Scandinavica, 91:331—335,1995.

Regier, D.A.; Farmer, M.E.; and Rae, D.S. Comorbidityof mental disorders with alcohol and other drug abuse.Journal of the American Medical Association, 264:2511-2518,1990.

Reilly, E.L.; Wilson, W.P.; and McClinton, H.K. Clinicalcharacteristics and medication history of schizophrenicsreadmitted to the hospital. International Journal ofNeuropsychiatry, 39:85-90,1967.

Renton, C.A.; Affleck, J.W.; Carstairs, G.M.; and Forrest,A.D. A follow-up of schizophrenic patients in Edinburgh.Ada Psychiatrica Scandinavica, 39:548-600, 1963.

Rickels, K., and Briscoe, E. Assessment of dosage devia-tion in outpatient drug research. Journal of ClinicalPsychopharmacology and the Journal of New Drugs,10(3): 153-160,1970.

Sacket, D.L. The magnitude of compliance and noncom-pliance. In: Sacket, D.L., and Haynes, R.B., eds.Compliance With Therapeutic Regimens. Baltimore, MD:Johns Hopkins University Press, 1976. pp. 9-25.

650

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from

Determinants of Medication Compliance Schizophrenia Bulletin, Vol. 23, No. 4, 1997

Sarwer-Foner, G.J. The role of rleuroleptic medication inpsychotherapeutic interaction. Comprehensive Psychiatry,1:291-300, 1960.

Schooler, N.R., and Keith, S.J. The clinical research basefor the treatment of schizophrenia. PsychopharmacologyBulletin, 29(4):431-446, 1993.

Sellwood, W., and Tarrier, N. Demographic factors associ-ated with extreme non-compliance in schizophrenia.Social Psychiatry and Psychiatric Epidemiology,29:172-177, 1994.

Serban, G., and Thomas, A. Attitudes and behaviors ofacute and chronic schizophrenic patients regarding ambu-latory treatment. American Journal of Psychiatry,131(9):991-995, 1974.

Shope, J.T. Compliance in children and adults: A reviewof studies. In: Schmidt, D., and Leppik, I.E., eds. Com-pliance in Epilepsy (Epilepsy Research Supplement 1).New York, NY: Elsevier Science Publishers, 1988.pp. 23-47.

Soskis, D.A. Schizophrenic and medical inpatients asinformed drug consumers. Archives of GeneralPsychiatry, 35:645-647, 1978.

Stimson, G.V. Obeying doctors orders: A view from theother side. Social Science and Medicine, 8:97-194, 1974.

Strang, J.S.; Falloon, I.R.H.; Moss, H.B.; Razani, J.; andBoyd, J.L. The effects of family therapy on treatmentcompliance in schizophrenia. PsychopharmacologyBulletin, 17(3):87-88, 1981.

Sullivan, G.; Wells, K.B.; Morgenstern, H.; and Leake, B.Identifying modifiable risk factors for rehospitalization: Acase-control study of seriously mentally ill persons inMississippi. American Journal of Psychiatry,152(12):1749-1756,1995.

Talbott, J.A.; Bachrach, L.; and Ross, L. Noncomplianceand mental health systems. Psychiatric Annals,16:586-599,1986.

Trostle, J.A. Medical compliance as an ideology. SocialScience and Medicine, 27:1299-1308, 198S.

Van Putten, T. Why do schizophrenic patients refuse totake their drugs? Archives of General Psychiatry, 31:67—72,1974.

Van Putten, T; Crumpton, E.; and Yale, C. Drug refusal inschizophrenia and the wish to be crazy. Archives ofGeneral Psychiatry, 33:1443-1446,1976.

Van Putten, T; May, P.R.A.; and Marder, S.R. Responseto antipsychotic medication: The doctor's and the con-

sumer's view. American Journal of Psychiatry,141(1):16-19, 1984.

Van Putten, T; May, P.R.A.; Marder, S.R.; and Wittmann,L.A. Subjective response to antipsychotic drugs. Archivesof General Psychiatry, 38:187-190, 1981.

Van Putten, T.; Mutalipassi, L.R.; and Malkin, M.D.Phenothiazine-induced decompensation. Archives ofGeneral Psychiatry, 30:102-105, 1974.

Verghese, A.; John, J.K.; Rajkumar, S.; Richard, J.; Sethi,B.B.; and Trivedi, J.K. Factors associated with the courseand outcome of schizophrenia in India: Results of a two-year multicentre follow-up study. British Journal ofPsychiatry, 154:499-503,1989.

Weiden, P.J. Antipsychotic therapy: Patient preference andcompliance. Current Approaches to Psychosis, 4:1-7,1995.

Weiden, P.J.; Rapkin, B.; Mott, T ; Zygmunt, A.; Gold-man, D.; Horvitz-Lennon, M.; and Frances, A. Rating ofmedication influences (ROMI) scale in schizophrenia.Schizophrenia Bulletin, 20(2):297-310, 1994.

Weiden, P.J.; Shaw, E.; and Mann, J. Causes of neurolep-tic noncompliance. Psychiatric Annals, 16:571-575,1986.

Willcox, D.R.C.; Gillan, R.; and Hare, E.H. Do psychi-atric out-patients take their drugs? British MedicalJournal, 2:790-792, 1965.

Wilson, B.M. Promoting compliance: The patient-provider partnership. Advances in Renal ReplacementTherapy, 3:199-206,1995.

Wilson, J.D., and Enoch, M.D. Estimation of drug rejec-tion by schizophrenic in-patients, with analysis of clinicalfactors. British Journal of Psychiatry, 113:209-211, 1967.

Young, J.L.; Zonana, H.V.; and Shepler, L. Medicationnoncompliance in schizophrenia: Codification and update.Bulletin of the American Academy of Psychiatry and theLaw, 14:105-122, 1986.

The Authors

Wayne S. Fenton, M.D., is Medical Director, ChestnutLodge Hospital, and Director, Chestnut Lodge ResearchInstitute; Crystal R. Blyler, Ph.D., is ResearchPsychologist, Chestnut Lodge Research Institute; andRobert K. Heinssen, Ph.D., is Director, PartialHospitalization and Psychiatric Rehabilitation forSchizophrenia, Chestnut Lodge Hospital, Rockville, MD.

651

by guest on August 27, 2013

http://schizophreniabulletin.oxfordjournals.org/D

ownloaded from