- Email: [email protected]
Electronically monitored adherence in outpatients with schizophrenia or schizoaffective disorder: A comparison of first- vs. second-generation antipsychotics
Schizophrenia Research 82 (2006) 107 – 114 www.elsevier.com/locate/schres
Electronically monitored adherence in outpatients with schizophrenia or schizoaffective disorder: A comparison of first- vs. second-generation antipsychotics Paul A. Nakonezny a, Matthew J. Byerly b,* a
Center for Biostatistics and Clinical Science and Department of Psychiatry, The University of Texas Southwestern Medical Center, 6363 Forest Park Rd, Suite 651, Dallas, TX 75235-8828, USA b Department of Psychiatry, The University of Texas Southwestern Medical Center, 6363 Forest Park Rd, Suite 651, Dallas, TX 75235-8828, USA Received 8 August 2005; received in revised form 18 October 2005; accepted 31 October 2005 Available online 22 December 2005
Abstract Objective: To examine the effect of first- vs. second-generation antipsychotics on electronically monitored adherence in outpatients with schizophrenia or schizoaffective disorder. Method: The sample consisted of 61 outpatients with a DSM-IV diagnosis of schizophrenia or schizoaffective disorder who took either a first-generation (N = 25) or second-generation (N = 36) oral antipsychotic during study participation. Treatment group designation (first-vs. second-generation antipsychotic) was based on the particular antipsychotic medication the patient was receiving as part of routine care at study entry (i.e., non-random assignment). Adherence to prescribed antipsychotic medication was assessed monthly over a 6-month study period using electronic monitoring (EM) of medication bottle opening. Various participant characteristics were collected at baseline to test for group differences and for potential associations with prospectively measured adherence. The primary data analysis was a mixed-model analysis of repeated measures. Results: The analysis of EM adherence revealed no significant difference between those taking first-generation (6-month adjusted mean adherence = 64.35%) and second-generation antipsychotics (6-month adjusted mean adherence = 69.17%; Group effect, p = .29) and no significant Group X Period interaction ( p = .13). Conclusion: There was no statistical difference in EM adherence, over a 6-month period, between patients taking first- and second-generation antipsychotics. However, since the patients were not randomized, conclusions must be interpreted within the context of the quasi-experimental design used in the current study. D 2005 Elsevier B.V. All rights reserved. Keywords: Antipsychotic medication adherence; Medication adherence; Schizophrenia
* Corresponding auhtor. Tel.: +1 214 648 5290; fax: +1 214 648 5323. E-mail address: [email protected] (M.J. Byerly). 0920-9964/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2005.10.015
108
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
1. Introduction Non-adherence to antipsychotic medication is a recurrent problem in persons with schizophrenia and schizoaffective disorder (Young et al., 1986; Fenton et al., 1997). Previous research suggests that nonadherence is associated with serious consequences, including exacerbation of psychotic symptoms (Ayuso-Gutierrez and del Rio Vega, 1997), increased aggression toward self and others (Steadman et al., 1998), worse prognosis (Wyatt, 1991; Lieberman et al., 1998), increased hospital and emergency room use (Fenton et al., 1997; Olfson et al., 2000), and high health care costs (Weiden and Olfson, 1995). To address this treatment challenge, studies of medication type and antipsychotic medication adherence in schizophrenia have proliferated in recent years (Dolder et al., 2002). One of the important lines of investigation, based on the supposition that the improved tolerability of second-generation antipsychotics might lead to improved adherence with these agents, has led to 6 published studies comparing adherence associated with first- vs. second-generation agents (Olfson et al., 2000; Dolder et al., 2002; Rosenheck et al., 2000; Cabeza et al., 2000; Grunebaum et al., 2001; Diaz et al., 2004). Of the 3 retrospective studies (Dolder et al., 2002; Cabeza et al., 2000; Grunebaum et al., 2001), 1 study found greater adherence among patients taking secondgeneration antipsychotics (Dolder et al., 2002), whereas 1 study found greater adherence among patients taking first-generation antipsychotics (Grunebaum et al., 2001), and 1 of the studies found no adherence differences (Cabeza et al., 2000). Three recent trials have prospectively evaluated the adherence associated with first- vs. second-generation antipsychotics, with all 3 studies finding a lack of difference in adherence between first- and secondgeneration agents (Olfson et al., 2000; Rosenheck et al., 2000; Diaz et al., 2004). Differences in previous adherence findings are perhaps due in part to methodological variability in the studies, particularly differences in study samples and methods for measuring and analyzing adherence. The purpose of the current study was to build on prior research of antipsychotic medication adherence in schizophrenia. Specifically, the current study used EM to examine the effect of first- vs. second-
generation antipsychotics on prospectively assessed adherence in outpatients with schizophrenia or schizoaffective disorder. In contrast to a recent study (Diaz et al., 2004), which compared the EM adherence of first- and second-generation antipsychotics among newly discharged inpatients, the current study involved outpatients irrespective of hospitalization history.
2. Method 2.1. Study design and participants A quasi-experimental design, and data from 61 participants, was used to prospectively assess adherence to prescribed antipsychotic medication. Data were collected from March of 2003 to April of 2004. During the study period, individual patient participation included a screening, baseline, and then up to 6 consecutive monthly adherence evaluations. Participants were recruited from any one of three Dallas County public mental health outpatient clinics. Eligibility criteria included (a) diagnosis of schizophrenia or schizoaffective disorder as established by the Structured Clinical Interview for DSM-IV, (b) currently taking a single oral antipsychotic, and (c) an age requirement of at least 18 years. Patients were excluded if they received a depot antipsychotic within one treatment cycle or if they used a pillbox. The study protocol was reviewed and approved by the institutional review board of The University of Texas Southwestern Medical Center at Dallas, and written informed consent was obtained from all participants. 2.2. Procedures and measures 2.2.1. Dependent variables 2.2.1.1. Adherence. Adherence to prescribed antipsychotic medication was assessed using electronic monitoring of medication bottle opening. Electronic monitoring was performed with the Medication Event Monitoring System (MEMSR), which is a medication vial cap that has the same appearance as regular medication bottle caps and that electronically recorded the date and time of bottle opening. The MEMSR caps used in this study had no cueing mechanisms.
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
Previous research suggests that MEMSR is a noninvasive, valid, reliable, and sensitive method for assessing the medication adherence of patients with general medical conditions (Farmer, 1999). The feasibility, validity, and sensitivity of electronic monitoring have also been demonstrated in at least four prior studies of persons with schizophrenia and schizoaffective disorder (Diaz et al., 2004, 2001; Byerly et al., 2005a,b). Electronic monitoring, nonetheless, does have shortcomings. Most notably, EM may overestimate adherence, because the events captured by EM (date/ time of bottle opening) cannot ensure ingestion of prescribed medication. In addition, EM technology has not been adapted for pill boxes and is not appropriate for liquid forms of medicine. A period documenting a lack of medication bottle opening, however, most likely represents an episode of nonadherence (Osterberg and Blaschke, 2005). We measured antipsychotic adherence as the proportion of medication vial caps opened in a given month relative to the prescribed doses for that month. Patients were made aware of the purpose and function of the MEMSR cap prior to study initiation, but they did not have access to EM-generated adherence results. If patients with multiple-dosing regimens opened the MEMSR cap at least the number of times prescribed each day, they received full credit for adherence for that day. Excessive bottle opening on a given day, however, did not did not count toward overall adherence. 2.3. Independent variables 2.3.1. Antipsychotic medication Patients were not randomly assigned to antipsychotic treatments for the study. Instead, treatment group designation (first-vs. second-generation antipsychotic) was based on the particular antipsychotic medication the patient was receiving as part of routine care at study entry. No patients switched antipsychotic medication type during the course of the trial. The first-generation antipsychotic medication group consisted of 25 patients with schizophrenia (N = 14) or schizoaffective disorder (N = 11) who took a single oral antipsychotic (fluphenazine, N = 1; haloperidol, N = 7; loxapine, N = 2; molindone, N = 1; perphenazine, N = 9; thiothixene, N = 3; trifluoperazine, N = 2)
109
during study participation. The second-generation antipsychotic medication group consisted of 36 patients with schizophrenia (N = 21) or schizoaffective disorder (N = 15) who took a single oral antipsychotic (aripiprazole, N = 1; clozapine N = 3; olanzapine, N = 13, quetiapine, N = 13, risperidone, N = 4, ziprasidone, N = 2) during study participation. 2.4. Covariates 2.4.1. Baseline characteristics The following information was collected at baseline to test for group differences and for potential associations with prospectively measured adherence over the 6 month study period: age, gender (0=female, 1=male), race (0=nonwhite, 1=white), education (0=less than high school, 1=high school or above), age of illness onset, length of illness, symptom severity [Positive and Negative Syndrome Scale (PANSS) total score] (Kay et al., 1987), insight [Schedule for the Assessment of Insight (SAI)] (David et al., 1992), alcohol abuse (Alcohol Use Scale] (Drake et al., 1990), drug abuse (Drug Use Scale) (Drake et al., 1990), total number of medications, medication side effects (Likert-type severity scale, which ranged from 1 to 10), whether antipsychotic dose was within an acceptable range as defined by the Texas Medication Algorithm Project (TMAP) treatment guidelines (Miller et al., 2005), dosing schedule, and presence of a caregiver (0=no, 1=yes). Although prior research suggests that these factors can potentially moderate medication adherence (Fenton et al., 1997; Lacro et al., 2002), stepwise (multiple) linear regression determined that just three of the 15 baseline characteristics were significantly associated with electronically determined adherence in this study (race, education, and PANSS total score). Therefore, only these three baseline characteristics were included as covariates in our analysis comparing adherence rates of first- vs. second-generation agents. 2.5. Data analysis The primary data analysis was a 2 Treatment Group (first-generation antipsychotics, second-generation antipsychotics) X 6 Period (Months 1–6) mixedmodel analysis of repeated measures, with Treatment Group serving as a between-subjects factor and Period
110
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
serving as a within-subjects factor. Restricted maximum likelihood estimation and Type 3 tests were used, with the Kenward–Roger correction (Kenward and Roger, 1997) applied to the unstructured covariance model. The sample of patients in our mixedmodel analysis was treated as a random effects factor. The baseline PANSS total score, educational level, and race were included as covariates in the mixedmodel analysis. The main effect of Treatment Group and the Treatment Group X Period interaction effect
were examined. Treatment Group X Period least squares means were also compared to assess the simple Treatment Group effects in each period (month). Cohen’s d was calculated and interpreted as the effect size estimator for the between-subjects Group effect. The level of significance for all tests was set at p b .05. The following model checking methods were performed before conducting the mixed-model analyses: (a) checked for outliers in the data and none were found; (b) the normality of the
Table 1 Participant characteristics by medication group and their relationship with EM adherence Baseline/participant characteristic
Age in years, M (SD) Gender, N (%) Male (1) Female (0) Race, N (%) White (1) Nonwhite (0) Education, N (%) bhigh school (0) zhigh school (1) Age at illness onset, years, M (SD) Length of illness, years, M (SD) PANSS total, M (SD) Insight, M (SD) Alcohol abuse, M (SD) Drug abuse, M (SD) Medications total, M (SD) Medication side effects, M (SD) Dose range, N (%) Within range (1) Not within range (0) Dosing schedule, N (%) 1/day 2/day 3/day 4/day Routine plus as needed Presence of caregiver, N (%) Yes (1) No (0)
First generation antipsychotics
Second generation antipsychotics
(N = 25)
(N = 36)
45.5 (9.3)
43.4 (8.9)
15.0 (60.0) 10.0 (40.0)
15.0 (41.7) 21.0 (58.3)
04.0 (16.0) 21.0 (84.0)
15.0 (41.7) 21.0 (58.3)
14.0 11.0 25.5 19.9 79.5 9.8 1.2 1.2 5.7 4.0
(56.0) (44.0) (11.6) (10.8) (8.9) (2.4) (.50) (.41) (3.2) (2.1)
09.0 (36.0) 16.0 (64.0) 17.0 06.0 01.0 00.0 01.0
(68.0) (24.0) (04.0) (00.0) (04.0)
11.0 (44.0) 14.0 (56.0)
10.0 26.0 21.4 21.7 71.7 10.1 1.3 1.1 4.7 3.1
(27.8) (72.2) (8.6) (10.7) (13.4) (2.6) (.52) (.33) (2.5) (2.3)
Regression resultsa
p
6-month EM adherence (%) b b
0.38 0.20c
0.35 3.2
p 0.34 0.63
0.04c
22.7
0.001d
0.03c
27.6
0.0001d
0.12b 0.54b 0.01b 0.79b 0.43b 0.13b 0.21b 0.12b 0.07c
0.65 0.32 0.88 0.58 2.1 2.2 1.3 1.3 2.2
0.05 0.31 0.0009d 0.67 0.74 0.82 0.27 0.37 0.74
0.41c
2.8
0.57
0.05c
1.8
0.81
22.0 (61.1) 14.0 (38.9) 30.0 05.0 01.0 00.0 00.0
(83.3) (13.9) (02.8) (00.0) (00.0)
07.0 (19.4) 29.0 (80.6)
b=unstandardized regression coefficient = the amount of % change in EM adherence due to a one unit change in the predictor (baseline) variable (e.g., for every 1 unit increase in the total PANSS score, EM adherence decreased by .88%; going from bhigh school education (0) to zhigh school education (1) leads to a 27.6% increase in EM adherence). a Each baseline variable (predictor) was used in a separate linear regression. b F statistic was used to test for mean differences between the two antipsychotic medication groups. c Fisher’s exact test (two-sided) was used to test for differences between the two antipsychotic medication groups. d Baseline characteristics retained in the stepwise regression model as significant predictors of the 6-month average EM adherence.
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
data was assessed and no significant departure from normality was found; (c) heteroscedasticity was assessed, but it was not significantly present in our mixed-model analyses; and (d) the covariance pattern of the repeated measurements was determined and accounted for in the modeling. 3. Results 3.1. Participant characteristics The sample consisted of 61 participants: 30 males (49.2%) and 31 females (50.8%). The age range was 21 to 59 years (with an average age of 44.3 years, SD = 9.1). The racial distribution included 40 (65.6%) African Americans, 19 (31.2%) Caucasians, and 2 (3.2%) Hispanics. The educational level included 24 (39.3%) participants with less than a high school education and 37 (60.7%) with at least a high school education (the average level of education was high school graduate to some college). Participant characteristics by group are reported in Table 1.
111
3.2. Medication adherence The analysis of EM adherence revealed no significant group effect (first-vs. second-generation agent; F = 1.12, df = 1, 59.2, p = .29) and no significant Group X Period interaction ( F = 1.78, df = 5, 49.1, p = .13). Further, the test of simple Treatment Group effects (in each period) showed no significant group differences on electronically determined adherence in any of the six prospective months (F’s b 3.79, p’s N .05). The results of the adherence analyses, including effect sizes, are shown in Table 2 and Fig. 1.
4. Discussion The current study compared the antipsychotic medication adherence of first- vs. second-generation antipsychotics in 61 outpatients with schizophrenia or schizoaffective disorder over a prospective 6-month study period using electronic monitoring.
Table 2 Medication Group X time period simple effects for mean level of EM adherence Month and group
Adjusted EM adherence (%) M
Month 1 and group 1st-gen AP groupa 2nd-gen AP groupa Month 2 and group 1st-gen AP groupa 2nd-gen AP group Month 3 and group 1st-gen AP groupa 2nd-gen AP group Month 4 and group 1st-gen AP groupb 2nd-gen AP groupa Month 5 and group 1st-gen AP groupc 2nd-gen AP groupd Month 6 and group 1st-gen AP groupc 2nd-gen AP groupc Overall (months 1– 6) 1st-gen AP group 2nd-gen AP group
SD
95% CI
n
75.61 76.25
23.27 22.60
66.09 to 85.13 68.58 to 83.92
24 35
67.46 70.41
27.78 27.66
56.11 to 78.80 61.19 to 79.64
24 36
61.31 70.03
27.09 26.94
50.26 to 72.37 61.04 to 79.02
24 36
62.18 64.65
30.07 30.23
49.66 to 74.71 54.43 to 74.89
23 35
57.36 71.18
26.31 25.36
46.14 to 68.57 61.92 to 80.45
22 30
56.32 66.50
28.24 28.03
44.26 to 68.36 56.71 to 76.30
22 33
64.35 69.17
24.75 24.18
54.43 to 74.27 61.09 to 77.24
25 36
F
p
d
0.01
0.91
0.03
0.17
0.68
0.11
1.57
0.21
0.32
0.10
0.76
0.08
3.79
0.06
0.53
1.80
0.19
0.36
0.61
0.44
0.19
Note. M=Least-Squares Means are adjusted for race, education, and baseline PANSS; SD=Standard Deviation; 95% CI=95% Confidence Interval for the expected mean value; n=sample size; Cohen’s d=Effect Size Estimator. 1st-Gen AP Group=first-generation antipsychotics; 2ndGen AP Group=second-generation antipsychotics. Overall=6-month average. F statistic was used to test for mean adherence differences between the two antipsychotic medication groups. a 1 missing observation. b 2 missing observations. c 3 missing observations. d 6 missing observations.
112
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
Fig. 1. Group X Period simple effects for mean level of EM adherence. Least-squares means for EM adherence are adjusted for race, education, and baseline PANSS. 1st-Gen AP=first-generation antipsychotics; 2nd-Gen AP=second-generation antipsychotics.
The findings indicate that there was no significant difference in the EM adherence of patients taking first-generation antipsychotics (6-month adjusted mean adherence = 64.35%) and those taking secondgeneration antipsychotics (6-month adjusted mean adherence = 69.17%). This finding is consistent with the majority of previous studies comparing antipsychotic adherence rates of first- vs. second-generation agents (Olfson et al., 2000; Rosenheck et al., 2000; Cabeza et al., 2000; Diaz et al., 2004). Of note, however, studies with more rigorous designs and data analytic methods—including those utilizing prospective evaluation (Olfson et al., 2000; Rosenheck et al., 2000; Diaz et al., 2004), treatment randomization (Rosenheck et al., 2000), or EM to assess adherence (Diaz et al., 2004)—all found a lack of adherence difference between first- and second-generation antipsychotic medications. The current study extends the recent findings of Diaz et al. (2004)—which reported the adherence rates of first-vs. second-generation antipsychotics among newly discharged inpatients with schizophrenia and schizoaffective disorder—to a sample of relatively stable outpatients with the same diagnoses. However, the average EM adherence of those taking first-generation antipsychotics (64.35%) and those taking second-generation antipsychotics (69.17%) reported in the current study is higher than the average EM adherence to conventional (29.6%) and atypical (41.6%) antipsychotics reported in Diaz et al. (2004) This is perhaps due, in part, to differences in the patient samples of the two studies. Prior research
suggests that recently hospitalized patients, which were the focus of the Diaz et al. (2004) study, demonstrate particularly poor adherence (Fenton et al., 1997) and, thus, might be expected to exhibit worse adherence than the relatively stable outpatients included in the current study. The use of EM may provide an important advance in assessing antipsychotic medication adherence of persons with schizophrenia and schizoaffective disorder. A recently published New England Journal of Medicine review of adherence assessment stated that EM bprovides the most accurate and valuable data on adherence in difficult clinical situations and in the setting of clinical trials and adherence research.Q (Osterberg and Blaschke, 2005) Recent research suggests that, relative to EM assessment, patients and prescribing clinicians significantly over-estimate adherence and research assistants either over-estimate (Byerly et al., 2005a) or misclassify the adherence status of patients (Byerly et al., 2005b). The results of this study may be tempered by the nature of the study. In particular, a lack of randomization to antipsychotic type (first- or second-generation) makes for at least two potential limitations. First, non-random assignment makes it possible that other factors besides medication type either combined with or independently influenced the antipsychotic medication adherence patterns during the study period. But the obvious demographic/socioeconomic influences as well as participant characteristics were accounted for and evaluated in our analyses. Second, a lack of randomization can lead to selection bias. We
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
note that the participant characteristics results presented earlier (Table 1), in part, reduce concern over selection bias. Another potential limitation of the study is the relatively small sample size of 61 patients, which could have contributed to a false negative finding (Type II error) (Cook, 1979). However, the results of a post hoc power analysis (which suggests a power level of 0.75) at least partially mitigate concern about the threat of a Type II error.
5. Conclusion The findings indicate that there was no statistical difference in adherence, over a 6-month period, between those taking first- vs. second-generation antipsychotics. These findings add to a growing literature suggesting that the reduced side effect liability of oral second-generation antipsychotics is not associated with improved adherence in persons with schizophrenia and schizoaffective disorder. Further research, including large randomized trials utilizing rigorous adherence assessment methods such as electronic monitoring, is needed in schizophrenia and other populations.
Acknowledgements This study was supported by NIMH K-Award grant # 5 K23 MH064930 and by a grant from Janssen Medical Affairs, LLC.
References Ayuso-Gutierrez, J.L., del Rio Vega, J.M., 1997. Factors influencing relapse in the long-term course of schizophrenia. Schizophr. Res. 28, 199 – 206. Byerly, M., Fisher, R., Whatley, K., et al., 2005a. A comparison of electronic monitoring vs. clinician rating of antipsychotic adherence in outpatients with schizophrenia. Psychiatry Res. 133, 129 – 133. Byerly, M.J T.A, Carmody T., Bugno R., Erwin T., Rush A.J. (Eds.), 2005b. Electronic monitoring of antipsychotic adherence of outpatients with schizophrenia and schizoaffective disorder, and association with symptoms, quality of life and cognition. In: New Research Program and Abstracts of the 2005 New Clinical Drug Evaluation Unit Meeting. Boca Raton, FL. Abstract; June 8, 2005.
113
Cabeza, I.G., Amador, M.S., Lopez, C.A., et al., 2000. Subjective response to antipsychotics in schizophrenic patients: clinical implications and related factors. Schizophr. Res. 41, 349 – 355. Cook, T.D.C.D., 1979. Quasi-Experimentation: Design and Analysis Issues for Field Settings. Houghton Mifflin Company, Boston, MA. David, A., Buchanan, A., Reed, A., et al., 1992. The assessment of insight in psychosis. Br. J. Psychiatry 161, 599 – 602. Diaz, E., Levine, H.B., Sullivan, M.C., et al., 2001. Use of the Medication Event Monitoring System to estimate medication compliance in patients with schizophrenia. J. Psychiatry. Neurosci. 26, 325 – 329. Diaz, E., Neuse, E., Sullivan, M.C., et al., 2004. Adherence to conventional and atypical antipsychotics after hospital discharge. J. Clin. Psychiatry 65, 354 – 360. Dolder, C.R., Lacro, J.P., Dunn, L.B., et al., 2002. Antipsychotic medication adherence: is there a difference between typical and atypical agents? Am. J. Psychiatry 159, 103 – 108. Drake, R.E., Osher, F.C., Noordsy, D.L., et al., 1990. Diagnosis of alcohol use disorders in schizophrenia. Schizophr. Bull. 16, 57 – 67. Grunebaum, M.F., Weiden, P.J., Olfson, M., 2001. Medication supervision and adherence of persons with psychotic disorders in residential treatment settings: a pilot study. J. Clin. Psychiatry 62, 394 – 399. (quiz 400-1). Farmer, K.C., 1999. Methods for measuring and monitoring medication regimen adherence in clinical trials and clinical practice. Clin. Ther. 21, 1074 – 1090. (dicussion 1073). Fenton, W.S., Blyler, C.R., Heinssen, R.K., 1997. Determinants of medication compliance in schizophrenia: empirical and clinical findings. Schizophr Bull. 23, 637 – 651. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261 – 276. Kenward, M.G., Roger, J.H., 1997. Small sample inference for fixed effects from restricted maximum likelihood. Biometrics 53, 983 – 997. Lacro, J.P., Dunn, L.B., Dolder, C.R., et al., 2002. Prevalence of and risk factors for medication nonadherence in patients with schizophrenia: a comprehensive review of recent literature. J. Clin. Psychiatry 63, 892 – 909. Lieberman, J.A., Sheitman, B., Chakos, M., et al., 1998. The development of treatment resistance in patients with schizophrenia: a clinical and pathophysiologic perspective. J. Clin. Psychopharmacol. 18, 20S – 24S. Miller, A.L.H.C., Crismon, M.L., Chiles, J.A. TIMA Procedural Manual; Schizophrenia Module, April 7, 2005. http://www.dshs. state.tx.us/mhprograms/timasczman.pdf. Olfson, M., Mechanic, D., Hansell, S., et al., 2000. Predicting medication noncompliance after hospital discharge among patients with schizophrenia. Psychiatr. Serv. 51, 216 – 222. Osterberg, L., Blaschke, T., 2005. Adherence to medication. N. Engl. J. Med. 353, 487 – 497. Rosenheck, R., Chang, S., Choe, Y., et al., 2000. Medication continuation and compliance: a comparison of patients treated with clozapine and haloperidol. J. Clin. Psychiatry 61, 382 – 386.
114
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
Steadman, H.J., Mulvey, E.P., Monahan, J., et al., 1998. Violence by people discharged from acute psychiatric inpatient facilities and by others in the same neighborhoods. Arch. Gen. Psychiatry 55, 393 – 401. Weiden, P.J., Olfson, M., 1995. Cost of relapse in schizophrenia. Schizophr. Bull. 21, 419 – 429.
Wyatt, R.J., 1991. Neuroleptics and the natural course of schizophrenia. Schizophr. Bull. 7, 325 – 351. Young, J.L., Zonana, H.V., Shepler, L., 1986. Medication noncompliance in schizophrenia: codification and update. Bull. Am. Acad. Psych. Law 14, 105 – 122.
Electronically monitored adherence in outpatients with schizophrenia or schizoaffective disorder: A comparison of first- vs. second-generation antipsychotics Paul A. Nakonezny a, Matthew J. Byerly b,* a
Center for Biostatistics and Clinical Science and Department of Psychiatry, The University of Texas Southwestern Medical Center, 6363 Forest Park Rd, Suite 651, Dallas, TX 75235-8828, USA b Department of Psychiatry, The University of Texas Southwestern Medical Center, 6363 Forest Park Rd, Suite 651, Dallas, TX 75235-8828, USA Received 8 August 2005; received in revised form 18 October 2005; accepted 31 October 2005 Available online 22 December 2005
Abstract Objective: To examine the effect of first- vs. second-generation antipsychotics on electronically monitored adherence in outpatients with schizophrenia or schizoaffective disorder. Method: The sample consisted of 61 outpatients with a DSM-IV diagnosis of schizophrenia or schizoaffective disorder who took either a first-generation (N = 25) or second-generation (N = 36) oral antipsychotic during study participation. Treatment group designation (first-vs. second-generation antipsychotic) was based on the particular antipsychotic medication the patient was receiving as part of routine care at study entry (i.e., non-random assignment). Adherence to prescribed antipsychotic medication was assessed monthly over a 6-month study period using electronic monitoring (EM) of medication bottle opening. Various participant characteristics were collected at baseline to test for group differences and for potential associations with prospectively measured adherence. The primary data analysis was a mixed-model analysis of repeated measures. Results: The analysis of EM adherence revealed no significant difference between those taking first-generation (6-month adjusted mean adherence = 64.35%) and second-generation antipsychotics (6-month adjusted mean adherence = 69.17%; Group effect, p = .29) and no significant Group X Period interaction ( p = .13). Conclusion: There was no statistical difference in EM adherence, over a 6-month period, between patients taking first- and second-generation antipsychotics. However, since the patients were not randomized, conclusions must be interpreted within the context of the quasi-experimental design used in the current study. D 2005 Elsevier B.V. All rights reserved. Keywords: Antipsychotic medication adherence; Medication adherence; Schizophrenia
* Corresponding auhtor. Tel.: +1 214 648 5290; fax: +1 214 648 5323. E-mail address: [email protected] (M.J. Byerly). 0920-9964/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.schres.2005.10.015
108
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
1. Introduction Non-adherence to antipsychotic medication is a recurrent problem in persons with schizophrenia and schizoaffective disorder (Young et al., 1986; Fenton et al., 1997). Previous research suggests that nonadherence is associated with serious consequences, including exacerbation of psychotic symptoms (Ayuso-Gutierrez and del Rio Vega, 1997), increased aggression toward self and others (Steadman et al., 1998), worse prognosis (Wyatt, 1991; Lieberman et al., 1998), increased hospital and emergency room use (Fenton et al., 1997; Olfson et al., 2000), and high health care costs (Weiden and Olfson, 1995). To address this treatment challenge, studies of medication type and antipsychotic medication adherence in schizophrenia have proliferated in recent years (Dolder et al., 2002). One of the important lines of investigation, based on the supposition that the improved tolerability of second-generation antipsychotics might lead to improved adherence with these agents, has led to 6 published studies comparing adherence associated with first- vs. second-generation agents (Olfson et al., 2000; Dolder et al., 2002; Rosenheck et al., 2000; Cabeza et al., 2000; Grunebaum et al., 2001; Diaz et al., 2004). Of the 3 retrospective studies (Dolder et al., 2002; Cabeza et al., 2000; Grunebaum et al., 2001), 1 study found greater adherence among patients taking secondgeneration antipsychotics (Dolder et al., 2002), whereas 1 study found greater adherence among patients taking first-generation antipsychotics (Grunebaum et al., 2001), and 1 of the studies found no adherence differences (Cabeza et al., 2000). Three recent trials have prospectively evaluated the adherence associated with first- vs. second-generation antipsychotics, with all 3 studies finding a lack of difference in adherence between first- and secondgeneration agents (Olfson et al., 2000; Rosenheck et al., 2000; Diaz et al., 2004). Differences in previous adherence findings are perhaps due in part to methodological variability in the studies, particularly differences in study samples and methods for measuring and analyzing adherence. The purpose of the current study was to build on prior research of antipsychotic medication adherence in schizophrenia. Specifically, the current study used EM to examine the effect of first- vs. second-
generation antipsychotics on prospectively assessed adherence in outpatients with schizophrenia or schizoaffective disorder. In contrast to a recent study (Diaz et al., 2004), which compared the EM adherence of first- and second-generation antipsychotics among newly discharged inpatients, the current study involved outpatients irrespective of hospitalization history.
2. Method 2.1. Study design and participants A quasi-experimental design, and data from 61 participants, was used to prospectively assess adherence to prescribed antipsychotic medication. Data were collected from March of 2003 to April of 2004. During the study period, individual patient participation included a screening, baseline, and then up to 6 consecutive monthly adherence evaluations. Participants were recruited from any one of three Dallas County public mental health outpatient clinics. Eligibility criteria included (a) diagnosis of schizophrenia or schizoaffective disorder as established by the Structured Clinical Interview for DSM-IV, (b) currently taking a single oral antipsychotic, and (c) an age requirement of at least 18 years. Patients were excluded if they received a depot antipsychotic within one treatment cycle or if they used a pillbox. The study protocol was reviewed and approved by the institutional review board of The University of Texas Southwestern Medical Center at Dallas, and written informed consent was obtained from all participants. 2.2. Procedures and measures 2.2.1. Dependent variables 2.2.1.1. Adherence. Adherence to prescribed antipsychotic medication was assessed using electronic monitoring of medication bottle opening. Electronic monitoring was performed with the Medication Event Monitoring System (MEMSR), which is a medication vial cap that has the same appearance as regular medication bottle caps and that electronically recorded the date and time of bottle opening. The MEMSR caps used in this study had no cueing mechanisms.
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
Previous research suggests that MEMSR is a noninvasive, valid, reliable, and sensitive method for assessing the medication adherence of patients with general medical conditions (Farmer, 1999). The feasibility, validity, and sensitivity of electronic monitoring have also been demonstrated in at least four prior studies of persons with schizophrenia and schizoaffective disorder (Diaz et al., 2004, 2001; Byerly et al., 2005a,b). Electronic monitoring, nonetheless, does have shortcomings. Most notably, EM may overestimate adherence, because the events captured by EM (date/ time of bottle opening) cannot ensure ingestion of prescribed medication. In addition, EM technology has not been adapted for pill boxes and is not appropriate for liquid forms of medicine. A period documenting a lack of medication bottle opening, however, most likely represents an episode of nonadherence (Osterberg and Blaschke, 2005). We measured antipsychotic adherence as the proportion of medication vial caps opened in a given month relative to the prescribed doses for that month. Patients were made aware of the purpose and function of the MEMSR cap prior to study initiation, but they did not have access to EM-generated adherence results. If patients with multiple-dosing regimens opened the MEMSR cap at least the number of times prescribed each day, they received full credit for adherence for that day. Excessive bottle opening on a given day, however, did not did not count toward overall adherence. 2.3. Independent variables 2.3.1. Antipsychotic medication Patients were not randomly assigned to antipsychotic treatments for the study. Instead, treatment group designation (first-vs. second-generation antipsychotic) was based on the particular antipsychotic medication the patient was receiving as part of routine care at study entry. No patients switched antipsychotic medication type during the course of the trial. The first-generation antipsychotic medication group consisted of 25 patients with schizophrenia (N = 14) or schizoaffective disorder (N = 11) who took a single oral antipsychotic (fluphenazine, N = 1; haloperidol, N = 7; loxapine, N = 2; molindone, N = 1; perphenazine, N = 9; thiothixene, N = 3; trifluoperazine, N = 2)
109
during study participation. The second-generation antipsychotic medication group consisted of 36 patients with schizophrenia (N = 21) or schizoaffective disorder (N = 15) who took a single oral antipsychotic (aripiprazole, N = 1; clozapine N = 3; olanzapine, N = 13, quetiapine, N = 13, risperidone, N = 4, ziprasidone, N = 2) during study participation. 2.4. Covariates 2.4.1. Baseline characteristics The following information was collected at baseline to test for group differences and for potential associations with prospectively measured adherence over the 6 month study period: age, gender (0=female, 1=male), race (0=nonwhite, 1=white), education (0=less than high school, 1=high school or above), age of illness onset, length of illness, symptom severity [Positive and Negative Syndrome Scale (PANSS) total score] (Kay et al., 1987), insight [Schedule for the Assessment of Insight (SAI)] (David et al., 1992), alcohol abuse (Alcohol Use Scale] (Drake et al., 1990), drug abuse (Drug Use Scale) (Drake et al., 1990), total number of medications, medication side effects (Likert-type severity scale, which ranged from 1 to 10), whether antipsychotic dose was within an acceptable range as defined by the Texas Medication Algorithm Project (TMAP) treatment guidelines (Miller et al., 2005), dosing schedule, and presence of a caregiver (0=no, 1=yes). Although prior research suggests that these factors can potentially moderate medication adherence (Fenton et al., 1997; Lacro et al., 2002), stepwise (multiple) linear regression determined that just three of the 15 baseline characteristics were significantly associated with electronically determined adherence in this study (race, education, and PANSS total score). Therefore, only these three baseline characteristics were included as covariates in our analysis comparing adherence rates of first- vs. second-generation agents. 2.5. Data analysis The primary data analysis was a 2 Treatment Group (first-generation antipsychotics, second-generation antipsychotics) X 6 Period (Months 1–6) mixedmodel analysis of repeated measures, with Treatment Group serving as a between-subjects factor and Period
110
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
serving as a within-subjects factor. Restricted maximum likelihood estimation and Type 3 tests were used, with the Kenward–Roger correction (Kenward and Roger, 1997) applied to the unstructured covariance model. The sample of patients in our mixedmodel analysis was treated as a random effects factor. The baseline PANSS total score, educational level, and race were included as covariates in the mixedmodel analysis. The main effect of Treatment Group and the Treatment Group X Period interaction effect
were examined. Treatment Group X Period least squares means were also compared to assess the simple Treatment Group effects in each period (month). Cohen’s d was calculated and interpreted as the effect size estimator for the between-subjects Group effect. The level of significance for all tests was set at p b .05. The following model checking methods were performed before conducting the mixed-model analyses: (a) checked for outliers in the data and none were found; (b) the normality of the
Table 1 Participant characteristics by medication group and their relationship with EM adherence Baseline/participant characteristic
Age in years, M (SD) Gender, N (%) Male (1) Female (0) Race, N (%) White (1) Nonwhite (0) Education, N (%) bhigh school (0) zhigh school (1) Age at illness onset, years, M (SD) Length of illness, years, M (SD) PANSS total, M (SD) Insight, M (SD) Alcohol abuse, M (SD) Drug abuse, M (SD) Medications total, M (SD) Medication side effects, M (SD) Dose range, N (%) Within range (1) Not within range (0) Dosing schedule, N (%) 1/day 2/day 3/day 4/day Routine plus as needed Presence of caregiver, N (%) Yes (1) No (0)
First generation antipsychotics
Second generation antipsychotics
(N = 25)
(N = 36)
45.5 (9.3)
43.4 (8.9)
15.0 (60.0) 10.0 (40.0)
15.0 (41.7) 21.0 (58.3)
04.0 (16.0) 21.0 (84.0)
15.0 (41.7) 21.0 (58.3)
14.0 11.0 25.5 19.9 79.5 9.8 1.2 1.2 5.7 4.0
(56.0) (44.0) (11.6) (10.8) (8.9) (2.4) (.50) (.41) (3.2) (2.1)
09.0 (36.0) 16.0 (64.0) 17.0 06.0 01.0 00.0 01.0
(68.0) (24.0) (04.0) (00.0) (04.0)
11.0 (44.0) 14.0 (56.0)
10.0 26.0 21.4 21.7 71.7 10.1 1.3 1.1 4.7 3.1
(27.8) (72.2) (8.6) (10.7) (13.4) (2.6) (.52) (.33) (2.5) (2.3)
Regression resultsa
p
6-month EM adherence (%) b b
0.38 0.20c
0.35 3.2
p 0.34 0.63
0.04c
22.7
0.001d
0.03c
27.6
0.0001d
0.12b 0.54b 0.01b 0.79b 0.43b 0.13b 0.21b 0.12b 0.07c
0.65 0.32 0.88 0.58 2.1 2.2 1.3 1.3 2.2
0.05 0.31 0.0009d 0.67 0.74 0.82 0.27 0.37 0.74
0.41c
2.8
0.57
0.05c
1.8
0.81
22.0 (61.1) 14.0 (38.9) 30.0 05.0 01.0 00.0 00.0
(83.3) (13.9) (02.8) (00.0) (00.0)
07.0 (19.4) 29.0 (80.6)
b=unstandardized regression coefficient = the amount of % change in EM adherence due to a one unit change in the predictor (baseline) variable (e.g., for every 1 unit increase in the total PANSS score, EM adherence decreased by .88%; going from bhigh school education (0) to zhigh school education (1) leads to a 27.6% increase in EM adherence). a Each baseline variable (predictor) was used in a separate linear regression. b F statistic was used to test for mean differences between the two antipsychotic medication groups. c Fisher’s exact test (two-sided) was used to test for differences between the two antipsychotic medication groups. d Baseline characteristics retained in the stepwise regression model as significant predictors of the 6-month average EM adherence.
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
data was assessed and no significant departure from normality was found; (c) heteroscedasticity was assessed, but it was not significantly present in our mixed-model analyses; and (d) the covariance pattern of the repeated measurements was determined and accounted for in the modeling. 3. Results 3.1. Participant characteristics The sample consisted of 61 participants: 30 males (49.2%) and 31 females (50.8%). The age range was 21 to 59 years (with an average age of 44.3 years, SD = 9.1). The racial distribution included 40 (65.6%) African Americans, 19 (31.2%) Caucasians, and 2 (3.2%) Hispanics. The educational level included 24 (39.3%) participants with less than a high school education and 37 (60.7%) with at least a high school education (the average level of education was high school graduate to some college). Participant characteristics by group are reported in Table 1.
111
3.2. Medication adherence The analysis of EM adherence revealed no significant group effect (first-vs. second-generation agent; F = 1.12, df = 1, 59.2, p = .29) and no significant Group X Period interaction ( F = 1.78, df = 5, 49.1, p = .13). Further, the test of simple Treatment Group effects (in each period) showed no significant group differences on electronically determined adherence in any of the six prospective months (F’s b 3.79, p’s N .05). The results of the adherence analyses, including effect sizes, are shown in Table 2 and Fig. 1.
4. Discussion The current study compared the antipsychotic medication adherence of first- vs. second-generation antipsychotics in 61 outpatients with schizophrenia or schizoaffective disorder over a prospective 6-month study period using electronic monitoring.
Table 2 Medication Group X time period simple effects for mean level of EM adherence Month and group
Adjusted EM adherence (%) M
Month 1 and group 1st-gen AP groupa 2nd-gen AP groupa Month 2 and group 1st-gen AP groupa 2nd-gen AP group Month 3 and group 1st-gen AP groupa 2nd-gen AP group Month 4 and group 1st-gen AP groupb 2nd-gen AP groupa Month 5 and group 1st-gen AP groupc 2nd-gen AP groupd Month 6 and group 1st-gen AP groupc 2nd-gen AP groupc Overall (months 1– 6) 1st-gen AP group 2nd-gen AP group
SD
95% CI
n
75.61 76.25
23.27 22.60
66.09 to 85.13 68.58 to 83.92
24 35
67.46 70.41
27.78 27.66
56.11 to 78.80 61.19 to 79.64
24 36
61.31 70.03
27.09 26.94
50.26 to 72.37 61.04 to 79.02
24 36
62.18 64.65
30.07 30.23
49.66 to 74.71 54.43 to 74.89
23 35
57.36 71.18
26.31 25.36
46.14 to 68.57 61.92 to 80.45
22 30
56.32 66.50
28.24 28.03
44.26 to 68.36 56.71 to 76.30
22 33
64.35 69.17
24.75 24.18
54.43 to 74.27 61.09 to 77.24
25 36
F
p
d
0.01
0.91
0.03
0.17
0.68
0.11
1.57
0.21
0.32
0.10
0.76
0.08
3.79
0.06
0.53
1.80
0.19
0.36
0.61
0.44
0.19
Note. M=Least-Squares Means are adjusted for race, education, and baseline PANSS; SD=Standard Deviation; 95% CI=95% Confidence Interval for the expected mean value; n=sample size; Cohen’s d=Effect Size Estimator. 1st-Gen AP Group=first-generation antipsychotics; 2ndGen AP Group=second-generation antipsychotics. Overall=6-month average. F statistic was used to test for mean adherence differences between the two antipsychotic medication groups. a 1 missing observation. b 2 missing observations. c 3 missing observations. d 6 missing observations.
112
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
Fig. 1. Group X Period simple effects for mean level of EM adherence. Least-squares means for EM adherence are adjusted for race, education, and baseline PANSS. 1st-Gen AP=first-generation antipsychotics; 2nd-Gen AP=second-generation antipsychotics.
The findings indicate that there was no significant difference in the EM adherence of patients taking first-generation antipsychotics (6-month adjusted mean adherence = 64.35%) and those taking secondgeneration antipsychotics (6-month adjusted mean adherence = 69.17%). This finding is consistent with the majority of previous studies comparing antipsychotic adherence rates of first- vs. second-generation agents (Olfson et al., 2000; Rosenheck et al., 2000; Cabeza et al., 2000; Diaz et al., 2004). Of note, however, studies with more rigorous designs and data analytic methods—including those utilizing prospective evaluation (Olfson et al., 2000; Rosenheck et al., 2000; Diaz et al., 2004), treatment randomization (Rosenheck et al., 2000), or EM to assess adherence (Diaz et al., 2004)—all found a lack of adherence difference between first- and second-generation antipsychotic medications. The current study extends the recent findings of Diaz et al. (2004)—which reported the adherence rates of first-vs. second-generation antipsychotics among newly discharged inpatients with schizophrenia and schizoaffective disorder—to a sample of relatively stable outpatients with the same diagnoses. However, the average EM adherence of those taking first-generation antipsychotics (64.35%) and those taking second-generation antipsychotics (69.17%) reported in the current study is higher than the average EM adherence to conventional (29.6%) and atypical (41.6%) antipsychotics reported in Diaz et al. (2004) This is perhaps due, in part, to differences in the patient samples of the two studies. Prior research
suggests that recently hospitalized patients, which were the focus of the Diaz et al. (2004) study, demonstrate particularly poor adherence (Fenton et al., 1997) and, thus, might be expected to exhibit worse adherence than the relatively stable outpatients included in the current study. The use of EM may provide an important advance in assessing antipsychotic medication adherence of persons with schizophrenia and schizoaffective disorder. A recently published New England Journal of Medicine review of adherence assessment stated that EM bprovides the most accurate and valuable data on adherence in difficult clinical situations and in the setting of clinical trials and adherence research.Q (Osterberg and Blaschke, 2005) Recent research suggests that, relative to EM assessment, patients and prescribing clinicians significantly over-estimate adherence and research assistants either over-estimate (Byerly et al., 2005a) or misclassify the adherence status of patients (Byerly et al., 2005b). The results of this study may be tempered by the nature of the study. In particular, a lack of randomization to antipsychotic type (first- or second-generation) makes for at least two potential limitations. First, non-random assignment makes it possible that other factors besides medication type either combined with or independently influenced the antipsychotic medication adherence patterns during the study period. But the obvious demographic/socioeconomic influences as well as participant characteristics were accounted for and evaluated in our analyses. Second, a lack of randomization can lead to selection bias. We
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
note that the participant characteristics results presented earlier (Table 1), in part, reduce concern over selection bias. Another potential limitation of the study is the relatively small sample size of 61 patients, which could have contributed to a false negative finding (Type II error) (Cook, 1979). However, the results of a post hoc power analysis (which suggests a power level of 0.75) at least partially mitigate concern about the threat of a Type II error.
5. Conclusion The findings indicate that there was no statistical difference in adherence, over a 6-month period, between those taking first- vs. second-generation antipsychotics. These findings add to a growing literature suggesting that the reduced side effect liability of oral second-generation antipsychotics is not associated with improved adherence in persons with schizophrenia and schizoaffective disorder. Further research, including large randomized trials utilizing rigorous adherence assessment methods such as electronic monitoring, is needed in schizophrenia and other populations.
Acknowledgements This study was supported by NIMH K-Award grant # 5 K23 MH064930 and by a grant from Janssen Medical Affairs, LLC.
References Ayuso-Gutierrez, J.L., del Rio Vega, J.M., 1997. Factors influencing relapse in the long-term course of schizophrenia. Schizophr. Res. 28, 199 – 206. Byerly, M., Fisher, R., Whatley, K., et al., 2005a. A comparison of electronic monitoring vs. clinician rating of antipsychotic adherence in outpatients with schizophrenia. Psychiatry Res. 133, 129 – 133. Byerly, M.J T.A, Carmody T., Bugno R., Erwin T., Rush A.J. (Eds.), 2005b. Electronic monitoring of antipsychotic adherence of outpatients with schizophrenia and schizoaffective disorder, and association with symptoms, quality of life and cognition. In: New Research Program and Abstracts of the 2005 New Clinical Drug Evaluation Unit Meeting. Boca Raton, FL. Abstract; June 8, 2005.
113
Cabeza, I.G., Amador, M.S., Lopez, C.A., et al., 2000. Subjective response to antipsychotics in schizophrenic patients: clinical implications and related factors. Schizophr. Res. 41, 349 – 355. Cook, T.D.C.D., 1979. Quasi-Experimentation: Design and Analysis Issues for Field Settings. Houghton Mifflin Company, Boston, MA. David, A., Buchanan, A., Reed, A., et al., 1992. The assessment of insight in psychosis. Br. J. Psychiatry 161, 599 – 602. Diaz, E., Levine, H.B., Sullivan, M.C., et al., 2001. Use of the Medication Event Monitoring System to estimate medication compliance in patients with schizophrenia. J. Psychiatry. Neurosci. 26, 325 – 329. Diaz, E., Neuse, E., Sullivan, M.C., et al., 2004. Adherence to conventional and atypical antipsychotics after hospital discharge. J. Clin. Psychiatry 65, 354 – 360. Dolder, C.R., Lacro, J.P., Dunn, L.B., et al., 2002. Antipsychotic medication adherence: is there a difference between typical and atypical agents? Am. J. Psychiatry 159, 103 – 108. Drake, R.E., Osher, F.C., Noordsy, D.L., et al., 1990. Diagnosis of alcohol use disorders in schizophrenia. Schizophr. Bull. 16, 57 – 67. Grunebaum, M.F., Weiden, P.J., Olfson, M., 2001. Medication supervision and adherence of persons with psychotic disorders in residential treatment settings: a pilot study. J. Clin. Psychiatry 62, 394 – 399. (quiz 400-1). Farmer, K.C., 1999. Methods for measuring and monitoring medication regimen adherence in clinical trials and clinical practice. Clin. Ther. 21, 1074 – 1090. (dicussion 1073). Fenton, W.S., Blyler, C.R., Heinssen, R.K., 1997. Determinants of medication compliance in schizophrenia: empirical and clinical findings. Schizophr Bull. 23, 637 – 651. Kay, S.R., Fiszbein, A., Opler, L.A., 1987. The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr. Bull. 13, 261 – 276. Kenward, M.G., Roger, J.H., 1997. Small sample inference for fixed effects from restricted maximum likelihood. Biometrics 53, 983 – 997. Lacro, J.P., Dunn, L.B., Dolder, C.R., et al., 2002. Prevalence of and risk factors for medication nonadherence in patients with schizophrenia: a comprehensive review of recent literature. J. Clin. Psychiatry 63, 892 – 909. Lieberman, J.A., Sheitman, B., Chakos, M., et al., 1998. The development of treatment resistance in patients with schizophrenia: a clinical and pathophysiologic perspective. J. Clin. Psychopharmacol. 18, 20S – 24S. Miller, A.L.H.C., Crismon, M.L., Chiles, J.A. TIMA Procedural Manual; Schizophrenia Module, April 7, 2005. http://www.dshs. state.tx.us/mhprograms/timasczman.pdf. Olfson, M., Mechanic, D., Hansell, S., et al., 2000. Predicting medication noncompliance after hospital discharge among patients with schizophrenia. Psychiatr. Serv. 51, 216 – 222. Osterberg, L., Blaschke, T., 2005. Adherence to medication. N. Engl. J. Med. 353, 487 – 497. Rosenheck, R., Chang, S., Choe, Y., et al., 2000. Medication continuation and compliance: a comparison of patients treated with clozapine and haloperidol. J. Clin. Psychiatry 61, 382 – 386.
114
P.A. Nakonezny, M.J. Byerly / Schizophrenia Research 82 (2006) 107–114
Steadman, H.J., Mulvey, E.P., Monahan, J., et al., 1998. Violence by people discharged from acute psychiatric inpatient facilities and by others in the same neighborhoods. Arch. Gen. Psychiatry 55, 393 – 401. Weiden, P.J., Olfson, M., 1995. Cost of relapse in schizophrenia. Schizophr. Bull. 21, 419 – 429.
Wyatt, R.J., 1991. Neuroleptics and the natural course of schizophrenia. Schizophr. Bull. 7, 325 – 351. Young, J.L., Zonana, H.V., Shepler, L., 1986. Medication noncompliance in schizophrenia: codification and update. Bull. Am. Acad. Psych. Law 14, 105 – 122.