Adm Policy Ment Health & Ment Health Serv Res (2007) 34:307314

DOI 10.1007/s10488-006-0108-5

ORIGINAL PAPER

The Relationship of Antipsychotic Medication Class

and Adherence with Treatment Outcomes and Costs
for Florida Medicaid Beneficiaries with Schizophrenia
Marion A. Becker M. Scott Young
Ezra Ochshorn Ronald J. Diamond

Published online: 9 January 2007

Springer Science+Business Media, LLC 2007

Abstract While some studies show a significant adherence for persons with schizophrenia may be as
advantage in adherence rates with use of atypical important to treatment costs and benefits as the class of
versus typical antipsychotic medication, others show no medication used.
advantage or mixed results (Jones et al. (2006).
Archives of General Psychiatry, 63, 10791087; Keywords Antipsychotic medication
Adherence

Rosenheck, (2006). Archives of General Psychiatry, Medicaid
Cost-effectiveness
63, 10741076). This study examined treatment out-
comes and costs associated with adherence rates by
antipsychotic medication class for adult Medicaid Introduction
beneficiaries in Florida diagnosed with schizophrenia.
Outcomes examined include arrests, involuntary com- Although schizophrenia only affects about 1% of the
mitments, and physical and behavioral healthcare population in the United States, this disorder consumes
costs. Study findings demonstrate that medication a disproportionately larger share of total national
healthcare expenditures, approximately $35$65 bil-
lion annually (Rice, 1999; Wyatt, Henter, Leary, &
This paper was previously presented at the Academy Health Taylor, 1995). Presently, the most effective treatment
Research Meeting: Becker, M.A., (June, 2005). Antipsychotic for schizophrenia involves the use of antipsychotic
medication in schizophrenia: What is the impact of adherence on
treatment outcomes. Paper presented at the Academy Health medications (Rittmannsberger, Pachinger, Keppelmul-
Research Meeting, Boston, M.A. ler, & Wancata, 2004). However, growing concerns
over rapidly rising expenditures for these drugs have
M. A. Becker (&)
made benefit versus price a critical issue (Fichtner,
Department of Mental Health Law & PolicyMHC 2735,
Louis de la Parte Florida Mental Health Institute, Hanrahan, & Luchins, 1998; Gardner et al., 2003;
University of South Florida, Tampa, FL, USA Gibson, Dalmer, Jackson, Wilder, & Ramsey, 2004;
e-mail: [email protected] Hogan, 1999). Faced with skyrocketing Medicaid
expenditures, many states are adopting restrictive
M. S. Young
Department of Mental Health Law & PolicyMHC 2603, prescribing policies (e.g., mandatory generic require-
Louis de la Parte Florida Mental Health Institute, ments, prior authorization, medication algorithms,
University of South Florida, Tampa, FL, USA drug formulary restrictions, and preferred drug lists)
to contain these costs (Awad & Voruganti, 2004;
E. Ochshorn
Department of Mental Health Law & PolicyMHC 2625G, Becker & Lemrow, 2004; Boyd, 2003; Geddes,
Louis de la Parte Florida Mental Health Institute, Freemantle, Harrison, & Bebbington, 2000).
University of South Florida, Tampa, FL, USA Since their development in the 1950s, conventional
antipsychotic medications also referred to as first-
R. J. Diamond
Department of Psychiatry, University of Wisconsin- generation or typical antipsychotics (TAs)have been
Madison, Madison, WI, USA a primary component of treatment for schizophrenia

123
308 Adm Policy Ment Health & Ment Health Serv Res (2007) 34:307314

(Lehman & Steinwachs, 1998). However, TAs are Surprisingly, many studies of schizophrenia have
associated with diminished tolerability and patient neglected to consider medication adherence in their
compliance due in part to a high incidence of adverse analysis. Even the most effective antipsychotic medi-
extrapyramidal side effects including parkinsonism and cation will fail if not taken. Indeed, research suggests
tardive dyskinesia (Brown, Markowitz, Moore, & that adherence is far more important than medication
Parker, 1999; Dolder et al., 2002). TAs are reported to class in determining treatment outcomes (Valenstein
have only partial efficacy against negative, cognitive, et al., 2004). Partial or nonadherence is strongly
and depressive symptoms of schizophrenia and are associated with hospital readmission, which is a large
associated with patients complaints of reduced quality part of treatment costs (Svarstad, Shireman, & Swee-
of life (Csernansky & Schuchart, 2002; Naber et al., ney, 2001). In prior research showing cost savings for
2001). Since 1989, when the Food and Drug Adminis- AAs, these cost savings have been partially derived
tration (FDA) began approving second-generation or from lower hospitalization rates thought to be due to a
atypical antipsychotic (AA) medications, these newer, higher level of medication compliance associated with
more expensive antipsychotics have gained acceptance the use of AAs (Davis et al., 2003; Foster & Goa, 1999;
as the first-line treatment choice for persons with Glazer, 1998; Hargraves & Shumway, 1996; Lehman &
schizophrenia. Some research suggests that AAs may Steinwachs, 1998; McCombs et al., 2000). Unsuccessful
be more broadly effective than TAs; they have better management of medication adherence is correlated
safety profiles, lower relapse rates, enhanced tolerabil- with suboptimal clinical responses and may be the most
ity and increased adherence rates (Csernansky & common cause of relapses, which increased treatment
Schuchart, 2002; Davis, Chen, & Glick, 2003; Dolder costs (Weiden, Aquila, & Standard, 1996; Weiden
et al., 2002). Other research, however, has questioned et al., 2004).
these advantages (Jones et al., 2006; Rosenheck, 2006). Almost 40% of the national annual cost of rehos-
The CATIE study, for instance, failed to find any pitalization for persons with schizophrenia is attributed
significant differences between medication classes in to adherence problems (Weiden & Olfson, 1995).
time to discontinuation (Lieberman et al., 2005). Non- Based on findings from randomized clinical trials,
adherence is a major problem with both classes of treatment guidelines recommend that antipsychotic
antipsychotics (Eaddy, Grogg, & Locklear, 2005; medication regimens be followed for at least a year
Thieda, Beard, Richter, & Kane, 2003; Weiden, Kozma, after symptom remission (Mojtabai et al., 2002).
Grogg, & Locklear, 2004). Adherence is defined as the Recent data suggest that the majority of patients
extent to which a persons behavior coincides with with schizophrenia considered fully compliant with
medical or health advice (Haynes, 1979, p. 2). their antipsychotic medication actually exhibited
While some studies show a significant superiority in considerable variation when patient self-reports were
adherence rates with AAs, others show little advantage contrasted with blood samples or measures such as
or inconclusive results (Dolder et al., 2002; Lacro, the Medication Event Monitoring System (Javaid,
Dunn, Dolder, Leckband, & Jeste, 2002; Rosenheck Holland, & Janicak, 1991; Olivieri, Matsui, Hermann,
et al., 2003; Vanelli, Burstein, & Crammer, 2001). & Koren, 1991). Further, the prevalence of discontin-
Depending on the study design, measurement meth- uous or interrupted use of antipsychotic medication
ods, and success criteria, reported rates of nonadher- increases over time and is reported to be approxi-
ence to antipsychotic medication range from 11 to mately 50% after 1 year and 75% after 2 years
80%, with an average rate of 50% (Dolder, Lacro, & (Weiden & Zygmunt, 1997).
Jeste, 2003; Lacro et al., 2002; Weiden & Olfson, 1995). In order to better understand the relationship
Factors contributing to this inconclusive clinical and between antipsychotic medication class, consistent
cost-effectiveness picture include studies that have use of a specific medication regime, treatment out-
been too brief (given schizophrenias long-term treat- comes and costs, this retrospective intent-to-treat study
ment prognosis), and outcome variables concentrating used administrative data from non-elderly Florida
on direct medical costs (Rosenheck, 2006). Few studies Medicaid beneficiaries with schizophrenia to compare
have examined the impact of AAs on more distal physical health treatment, behavioral health treatment,
outcomes such as long-term adherence employment, or length of medication adherence, criminal justice
perceived quality of life (Becker & Diamond, 2005; outcomes, and service costs among persons taking
Lehman & Steinwachs, 1998). Further complicating different classes of antipsychotic medication. To our
existing research is the lack of common definitions or knowledge, this is the first study to include incarcer-
standards for clinical improvement, remission, and ation and involuntary examination as variables in
recovery (Andreason et al., 2005). assessing treatment outcome.

123
Adm Policy Ment Health & Ment Health Serv Res (2007) 34:307314 309

Methods Procedure

Participants After obtaining university IRB approval, we identified

Medicaid beneficiaries diagnosed with schizophrenia
The study included all community dwelling Florida (as indicated by ICD-9 codes of 295.0295.9) who had
Medicaid beneficiaries: (1) between the ages of 18 used the same antipsychotic medication for at least four
and 61 during FY 99-00, (2) continuously enrolled in continuous months. (Persons switching medications
Medicaid for 36 months, including 12 months prior to during this initial 4-month period were not analyzed
and 20 months following continuous use of an anti- because the study timeframe could not accommodate
psychotic drug for at least 4 months, (3) diagnosed adequate follow-up.) Adherence was operationalzed as
with schizophrenia in FY 99-00, as indicated by at months on one medication, and was calculated as a
least one inpatient or two outpatient claims contain- proportion of the possible Medicaid prescriptions filled
ing that diagnosis, (4) not enrolled in a Medicaid over the 2-year follow-up period. For example, if
managed care plan, and (5) not living in a nursing Medicaid claims data showed an individual received
home. Persons diagnosed with schizophrenia who the same antipsychotic medication 18 out of 24 months,
were not receiving antipsychotic medications were that person was rated as having 75% adherence to the
deleted from the analysis. It is not known whether medication regime. Both inpatient and outpatient
excluding persons enrolled in a Medicaid managed pharmacy records were used to assure complete data.
care plan introduced bias, but previous research All antipsychotic medications paid for by Medicaid,
comparing fee-for-service and managed care partici- including clozapine, were included in the analyses.
pants within the Florida Medicaid program suggests Beneficiaries were grouped into four categories based
these populations are comparable (Shern et al., on time on their specific medication: (1) maximal
2005). adherence (75100% use over the 2-year study), (2)
moderate adherence (5074.9%), (3) minimal adher-
Measures ence (2549.9%), and (4) negligible adherence (<25%).
We also examined three types of antipsychotic medica-
Florida Medicaid paid claims data were used to tion users within the study sample: (A) patients who
identify all study subjects and to collect information received TAs only, (B) patients who received AAs only,
on the following services: inpatient and outpatient and (C) patients who received a combination of AAs and
mental health, crisis unit or emergency room, TAs (i.e., on both types of medication at the same time).
substance abuse, pharmacy, case management, and The use and cost of behavioral and physical health,
physical health. Medicaid data were also used to pharmacy, criminal justice, and involuntary commit-
identify pre-existing patient characteristics: age, sex, ment services received were compared across the three
race, Department of Children and Families (DCF) medication class groups and four levels of adherence.
health planning district, and the existence of other
mental health comorbidities. The study utilized the Analyses
Florida Alcohol, Drug Abuse, and Mental Health Data
Warehouse (ADMDW) service event data to capture We used simple descriptive statistics to describe
publicly funded behavioral health services used by the medication usage, and a multivariate logistic regression
study sample not covered by Floridas Medicaid analysis to predict maximal adherence. Service costs
program. Additionally, we used the Florida Baker were calculated thusly: Number of units of services
Act Data File to identify those who had an involuntary patients used X price per service unit. Previously
psychiatric (or Baker Act) examination. Arrest developed standard unit costs were used for specific
data were provided by the Florida Department of services (Jones et al., 2006). These cost standards were
Law Enforcement (FDLE). Data on quarters of applied to ADMDW service events, Baker Act (invol-
paid employment were obtained from the Florida untary) examinations, and arrests. For Medicaid
Unemployment Insurance Employment Wage File services, including pharmacy, the cost was the amount
(FUIEWF), a database maintained by the Florida Medicaid reimbursed. To compare statistically the
Department of Education. We operationalized the individual and total service costs associated with the
levels of adherence to antipsychotic medications by different medication classes, a series of univariate
the frequency with which patients refilled their one-way analyses of variance were performed to
monthly prescriptions over the studys 24-month compare the four adherence levels on each of the
follow-up period. service cost variables.

123
310 Adm Policy Ment Health & Ment Health Serv Res (2007) 34:307314

Table 1 Sample characteristics by medication class

Characteristic Typical only Atypical only (AA) Both TA & AA Total
(TA)(N = 3,277) (N = 6,023) (N = 1,030) (N = 10,330)

Mean # months on drug (SD) 17.44 (7.16) 19.75 (6.54) 15.75 (7.44) 18.61 (6.98)
Adherence pattern
Maximal (75100%) 1,827 (55.8%) 4,319 (71.7%) 463 (45.0%) 6,609 (64.0%)
Moderate (5075.9%) 518 (15.8%) 581 (9.6%) 177 (17.2%) 1,276 (12.4%)
Limited (2549.9%) 755 (23.0%) 892 (14.8%) 293 (28.4%) 1,940 (18.8%
Negligible (024.9%) 177 (5.4%) 231 (3.8%) 97 (9.4%) 505 (4.9%)
Age at study entry
1830 282 (8.6%) 915 (15.2%) 123 (11.9%) 1,320 (12.8%)
3144 1,481 (45.2%) 2,827 (46.9%) 523 (50.8%) 4,831 (46.8%)
4564 1,514 (46.2%) 2,281 (37.9%) 384 (37.3%) 4,179 (40.4%)
Mean age (SD) 43.54 (9.10) 41.17 (9.84) 41.48 (9.28) 41.95 (9.61)
Gender
Male 1,790 (54.6%) 3,256 (54.1%) 630 (61.2%) 5,676 (54.9%)
Race
White 1,485 (45.3%) 3,340 (55.5%) 540 (53.4%) 5,365 (51.9%)
Non-white 1,792 (54.7%) 2,683 (44.5%) 490 (47.6%) 4,965 (48.1%)
Comorbidities
Psychotic disorder otherthan 423 (12.9%) 771 (12.8%) 183 (17.8%) 1,377 (13.3%)
schizophrenia
Mood disorder 784 (23.9%) 1,871 (31.1%) 282 (27.4%) 2,937 (28.4%)
Substance abuse diagnosis 414 (12.6%) 820 (13.6%) 179 (17.4%) 1,413 (13.7%)
Note: Cells include raw frequencies and column percents or means followed by standard deviations in parentheses

Results health service data indicate that, regardless of adher-

ence level, individuals receiving typical antipsychotic
As shown in Table 1, a total of 10,330 individuals medications had fewer behavioral-health related
met study inclusion criteria. Over half the partici- hospitalizations, fewer CSU or ER visits, and fewer
pants (58.3%) received AA medication only, slightly residential treatment days. Patients on combination
less than one third (31.7%) received TA medication therapy had the highest rate of hospitalizations, CSU
only, and 10%, received combination AA & TA or ER visits, and residential treatment days. Com-
treatment. In keeping with study expectations, pared to persons on AAs or combination therapy,
persons on AAs demonstrated greater adherence, persons on TAs had fewer involuntary psychiatric
remaining on their medication 2.3 months longer than (Baker Act) examinations. As expected, persons on
those on TAs or combination treatment. Further TAs were more likely to take medication for side
data inspection indicated that maximal adherence effects than those on AAs (73.8% vs. 41.3%). Persons
was the most common pattern across all three taking AAs had fewer arrests than persons taking
medication groups, with negligible adherence the TAs. All variables in Table 2 were used to compute
least common. total costs for each medication class and adherence
There were modest differences in age, gender, group.
diagnosis and race across the medication groups (see Figure 1 presents the total mean per-user-per-month
Table 1). With a mean age of 43.5 years, TA users (PUPM) service costs for each level of adherence
were on average 2 years older than individuals receiv- within each medication class. These total costs relate to
ing AAs. The combination drug group had a higher outcomes listed in Table 2, including expenditures for
proportion of males than the other groups. Persons in antipsychotic and side effect medications. Total costs
the combination drug group were also more likely to were lowest for persons taking only typical antipsy-
have a co-occurring substance abuse or psychiatric chotics and were highest for persons receiving com-
diagnosis other than schizophrenia. In addition, per- bined therapy. A series of one-way analyses of variance
sons with minority status were less likely to receive indicated these results were significant across all four
AAs. These demographic differences were consistent adherence levels (all Ps < 0.001). Within each med-
across levels of adherence. ication class, higher adherence levels were associated
Table 2 presents the service utilization outcome with significantly lower total costs, though the corre-
variables by medication class. Physical and behavioral lation was weaker for persons taking only atypical

123
Adm Policy Ment Health & Ment Health Serv Res (2007) 34:307314 311

Table 2 Outcomes by antipsychotic medication class (N = 10,330)

Outcome Typical only (TA) Atypical only (AA) Both TA & AA Total F or v2
(N = 3,277) (N = 6,023) (N = 1,030) (N = 10,330)

Mean months on drug 17.4 19.7 15.7 18.6 <0.0001

Persons hospitalized for 552 (16.8%) 1,035 (17.2%) 236 (22.9%) 1,823 (17.7%) <0.0001
behavioral health
Persons with state psychiatric 42 (1.3%) 100 (1.7%) 15 (1.5%) 157 (1.5%) 0.3563
hospital stays
Persons with CSU or ER visits 640 (19.5%) 1,426 (23.7%) 276 (26.8%) 2,342 (22.7%) <0.0001
Persons with residential treatment 177 (5.4%) 623 (10.3%) 136 (13.2%) 936 (9.1%) <0.0001
Persons on other psychotropic 2,331 (71.1%) 4,901 (81.4%) 826 (80.2%) 8,058 (78.0%) <0.0001
meds
Persons on side effect meds 2,417 (73.8%) 2,489 (41.3%) 798 (77.5%) 5,704 (55.2%) <0.0001
Persons with outpatient MH 3,226 (98.4%) 5,942 (98.7%) 1,018 (98.8%) 10,186 (98.6%) 0.5693
servicesa
Persons with one or more Baker 560 (17.1%) 1,186 (19.7%) 274 (26.6%) 2,020 (19.6%) <0.0001
Act Exams
Persons with multiple Baker Act 227 (6.9%) 501 (8.3%) 136 (13.2%) 864 (8.4%) 0.0378
Exams
Persons with one or more arrests 266 (8.1%) 471 (7.8%) 74 (7.2%) 811 (7.9%) 0.6183
Persons with multiple arrests 117 (3.6%) 205 (3.4%) 42 (4.1%) 364 (3.5%) 0.5471
Persons with addictions services 375 (11.4%) 720 (12.0%) 155 (15.1%) 1,250 (12.1%) 0.0072
Persons with a physical health 3,122 (95.3%) 5,766 (95.7%) 1,001 (97.2%) 9,889 (95.7%) 0.0298
Medicaid claim
Note: Percentages represent column percentages
a
Includes all Medicaid outpatient behavioral health services, including case management

Fig. 1 Total cost by <25% 25-49% 50-74% 75-100%

antipsychotic medication class $1,800
$1,694
and adherence level $1,608
$1,600
$1,473
Total Per-User Per-Month Cost

$1,400 $1,322
$1,280 $1,256
$1,238
$1,189
$1,200
$1,102
$1,023
$1,000
$907

$800
$729

$600

$400

$200

$0
Atypical Typical Both
Antipsychotic Medication Class

antipsychotics. For persons receiving combination without a co-occurring substance abuse disorder, and
therapy, the lowest costs were associated with the being on AA medication all increased adherence rates
two highest levels of adherence. over the two-year study. Persons on AA medication
were more than twice as likely to have maximal
Multivariate Analysis adherence than persons on TA medication. Overall,
comorbid substance abuse more than doubled the rate
As shown in Table 3, logistic regression analysis of nonadherence. These study findings support previ-
revealed that being middle aged or older, male, white, ous research that found a relationship between

123
312 Adm Policy Ment Health & Ment Health Serv Res (2007) 34:307314

Table 3 Logistic regression analysis predicting maximal adherence (N = 10,330)

2
Predictor variable Values Odds ratio 95% CI Wald v p Value

Age category 3144 vs. 1830 1.197 1.0501.364 7.249 0.0071

4564 vs. 1830 1.396 1.2201.597 23.578 <0.0001
Gender Female versus male 0.822 0.7550.895 20.317 <0.0001
Ethnicity Non-white versus white 0.737 0.6770.801 50.519 <0.0001
Comorbid substance abuse SA versus No SA 0.459 0.4040.521 144.986 <0.0001
Dual eligibility Dual versus non-dual 0.955 0.8751.043 1.033 0.3096
Type of antipsychotic Atypical versus typical 2.050 1.8722.246 238.669 <0.0001
Both versus typical 0.650 0.5640.751 34.669 <0.0001

medication adherence and age, gender and ethnicity Banks, 2003). Medicaid datasets offer detailed infor-
(Fenton, Blyler, & Heinssen, 1997; Ziguras, Klimidis, mation on beneficiaries, minimize attrition due to lost
Lambert, & Jackson, 2001). contact over time, and can readily be used at minimal
cost because the data are already in place.

Study Limitations
Discussion
This study has a number of limitations. First, admin-
istrative data record-keeping is imperfect and these Nonadherence with antipsychotic medication remains
data lacked measures of illness severity, health status, a well-recognized but poorly understood problem that
or role functioning. Second, this study does not include threatens the treatment success for schizophrenia and
service costs incurred outside of the Medicaid and increases costs. When evaluating the cost-effectiveness
public behavioral health service system. This includes of antipsychotic medications, clinicians, and research-
care provided within the criminal justice system and ers should consider the important role of medication
state hospitals, along with out-of-pocket costs or adherence. In this study, individuals receiving the more
private pay services. Therefore true behavioral and costly AAs were more than twice as likely to achieve
physical health service costs are underestimated. A maximal or excellent adherence than those receiving
third limitation is that we only included beneficiaries TAs. Subjects with co-morbid substance abuse were
continuously enrolled in Medicaid for the 36-month less than half as likely to achieve maximal adherence
follow-up period. Thus, outcomes for Medicaid drop- than those with no substance use problems. While
outs were not examined in this study. Fourth, there are other variables were associated with increased adher-
apparent systematic biases in the decision to place ence, AA medication and absence of substance use
recipients on AA or TA medications. People who are disorders were the strongest predictors. The more
white, have a substance abuse diagnosis or have a tolerable side effects profile of AAs likely contribute to
mood disorder are more likely to be placed on AA the increased adherence found in this and other
medications (Daumit et al., 2003; Herbeck et al., 2004; studies. This finding is important as prior research
Ren et al., 2002). Fifth, our outcome measures were suggests that inconsistent use of antipsychotic medica-
limited. For instance, any improvement in quality of tion may negatively impact long-term costs and patient
life associated with decreased criminal justice involve- outcomes, particularly if it occurs early in the treat-
ment and increased paid employment resulting from ment course (Tandon, 1998; Weiden & Olfson, 1995).
increased medication adherence is not reflected in Adherence in this intent-to-treat study was defined
these cost analyses. Finally, this study used pharmacy by the length of time patients continued refilling
refill methodology that records only medications dis- prescriptions for their initially prescribed antipsychotic
pensed. It is certainly possible that some subjects medication regimen. We felt assigning patients to a
received but did not take their medication, while others medication class group based on the initial four-month
may have received medication from sources outside of period was justified; four months provides ample time
Medicaid. for pharmacological stabilization and demonstrates a
Despite the limitations noted above, large adminis- clear intention-to-treat with the identified medication.
trative databases are inclusive, cost-effective, and allow Our data indicates that most patients were maxi-
investigations of treatment administered in real world mally adherent across drug classes.
settings with diverse populations, including subpopu- Additionally, given that interrupted antipsychotic
lations rarely included in clinical trials (Pandiana & medication use and polypharmacy have been shown to

123
Adm Policy Ment Health & Ment Health Serv Res (2007) 34:307314 313

contribute to poor treatment outcomes and higher Boyd, D. L. (2003). The bursting State Fiscal Bubble and State
costs for persons with schizophrenia, these variables Medical Budgets. Health Affairs, 22(1), 4661.
Brown, C., Markowitz, J., Moore, T., & Parker, N. (1999).
should be included in future research on the relative Atypical antipsychotics, part II: Adverse effects, drug
merit of specific antipsychotic medications and medi- interactions, and costs. Annals of Pharmacotherapy, 33,
cation classes (Becker & Lemrow, 2004; Geddes et al., 210217.
2000). Csernansky, J. G., & Schuchart, E. K. (2002). Relapse and
rehospitalisation rates in patients with schizophrenia:
Findings indicate that persons with minority status Effects of second generation antipsychotics. CNS Drugs,
were considerably less likely to receive AAs. These 16, 473484.
differences are of great concern and are consistent with Daumit, G. L., Crum, R. M., Guallar, E., et al. (2003). Outpa-
prior research documenting that racial and ethnic tient prescriptions for atypical antipsychotics for African
Americans, Hispanics, and Whites in the United States.
disparities exist in all aspects of psychiatric treatment Archives of General Psychiatry, 60, 121128.
(Baker & Bell, 1999; Becker, Jang, & Kane, 2006; Moy, Davis, J. M., Chen, N., & Glick, I. D. (2003). A meta-analysis of
Dayton, & Clancy, 2005; Nicole, 2005; Schneider, the efficacy of second generation antipsychotics. Archive of
Zaslavsky, & Epstein, 2002). General Psychiatry, 60, 553564.
Dolder, C. R., Lacro, J. P., Dunn, L. B., et al. (2002). Antipsy-
To conclude, medication adherence is a critical link chotic medication adherence: Is there a difference between
between a physicians drug treatment plan and a typical and atypical agents? American Journal of Psychiatry,
patients outcome. Our study suggests adherence may 159, 103108.
be as important to treatment costs and benefits as the Dolder, C. R., Lacro, J. P., & Jeste, D. V. (2003). Adherence to
antipsychotic and nonpsychiatric medications in middle-
medication type. Additional studies should combine aged and older patients with psychotic disorders. Psychoso-
administrative data with quality of life measures to matic Medicine, 65, 156162.
assess the true value of AAs. Eaddy, M., Grogg, A., & Locklear, J. (2005). Assessment of
Future research would benefit from a design that compliance with antipsychotic treatment and resource
utilization in a Medicaid population. Clinical Therapeutic,
included extended follow-up periods in order to 27(2), 263272.
examine long-term outcomes. Improved understanding Fenton, W. S., Blyler, C. R., & Heinssen, R. K. (1997).
of factors that influence patient behavior could facil- Determinants of medication compliance in schizophrenia:
itate the development of strategies to increase medi- Empirical and clinical findings. Schizophrenia Bulletin,
23(4), 637651.
cation adherence and thereby improve treatment Fichtner, C. G., Hanrahan, P., & Luchins, D. J. (1998).
effectiveness. Pharmacoeconomic studies of atypical antipsychotics: Re-
view and perspective. Psychiatric Annals, 28(7), 381396.
Foster, R. H., & Goa, K. L. (1999). Olanzapine: A pharmaco-
economic review of its use in schizophrenia. Pharmacoeco-
References nomics, 15, 611640.
Gardner, J. C., Luo, Z., Bradley, C. J., Polverrjan, E., Holmes-
Andreasen, N. C., Carpenter, W. T., Kane, J. M., Lasser, R. A., Rovnes, M., & Rovner, D. (2003). Longitudinal assessment
Marder, S. R., & Weinberger, D. R. (2005). Remission in of cost in health care intervention. Health Services Research
schizophrenia: Proposed criteria and rationale for consen- Methodology, 3, 149168.
sus. American Journal of Psychiatry, 162, 441449. Geddes, J., Freemantle, N., Harrison, P., & Bebbington, P.
Awad, A. G., & Voruganti, L. N. (2004). New antipsychotics, (2000). Atypical antipsychotics in treatment of schizophre-
compliance, quality of life, and subjective tolerabilityAre nia: Systematic overview and meta-regression analysis.
patients better off? Canadian Journal of Psychiatry, 49(5), British Nursing Journal, 321, 13711376.
297302. Gibson, J., Damler, R., Jackson, A. E., Wilder, T., & Ramsey, J.
Baker, F. M., & Bell, C. (1999). Issues in the psychiatric L. (2004). The impact of olanzapine, risperidone, or haldol
treatment of African Americans. Psychiatric Services, 50(3), on the cost of schizophrenia care in a Medicaid population.
362350. Value in Health, 7(1), 2235.
Becker, M., & Diamond, R. (2005). Quality of life measurement Glazer, W. M. (1998). Formulary decisions and health econom-
in persons with schizophrenia: Are we measuring whats ics. Journal of Clinical Psychiatry, 59(Suppl. 19), 2329.
important? In H. Katschnig & H. Freeman (Eds.), Quality Hargraves, W. A., & Shumway, M. (1996). Pharmacoeconomics
of life in mental disorders (2nd ed.). Chichester, England: of antipsychotic drug therapy. Journal of Clinical Psychiatry,
Wiley. 57(Suppl. 9), 6676.
Becker, M., Jang, Y., & Kane, M. (2006). Evaluation of Florida Haynes, R. B. (1979). Introduction. In R. B. Haynes, D. L.
Medicaid behavioral health pharmacy practice by racial/ Sackett, & D. W. Taylor (Eds.), Compliance in health care
ethnic minorities across the lifespan (pp. 127). Tampa, FL: (pp. 110). Baltimore, MD: John Hopkins University
Louis de la Parte Florida Mental Health Institute, Univer- Press.
sity of South Florida. Herbeck, D. M., West, J. C., Ruditis, I., et al. (2004). Variations
Becker, M., & Lemrow, N. (2004). Behavioral health care services in use of second-generation antipsychotic medication by
associated with switching atypical antipsychotic medication race among adult psychiatric patients. Psychiatric Services,
for persons with SMI (pp. 176). Tampa, FL: Louis de la 55, 677684.
Parte Florida Mental Health Institute, University of South Hogan, M. F. (1999). Public-sector mental health care: New
Florida. challenges. Health Affairs, 18, 106111.

123
314 Adm Policy Ment Health & Ment Health Serv Res (2007) 34:307314

Javaid, J. I., Holland, D., & Janicak, P. G. (1991). Blood level patients before admission to inpatient treatment. Psychiatric
monitoring of antipsychotics and antidepressants. Psychiat- Services, 55(2), 174179.
ric Medicine, 9(1), 163187. Rosenheck, R. (2006). Outcomes, costs, and policy caution: A
Jones, K., Chen, H. J., Jordan, N., Boothroyd, R. A., Ramoni- commentary on the cost utility of the latest antipsychotic
Perazzi, J., & Shern, D. L. (2006). Examination of formal drugs in schizophrenia study (CUtLASS 1). Archives of
treatment costs in the provision of managed care for a General Psychiatry, 63, 10741076.
Medicaid population with psychiatric disabilities. Medical Rosenheck, R., Perlick, D., Bingham, S., et al. (2003). Effective-
Care, 44, 320327. ness and cost of olanzapine and haloperidol in the treatment
Jones, P., Barnew, T., Davies, L., Dunn, G., Lloyd, H., Hayhurst, of schizophrenia. Journal of the American Medical Associ-
K., Murray, R., Markwick, A., & Lewis, S. (2006). Ran- ation, 290(20), 26932702.
domized controlled trial of the effect on quality of life of Schneider, E. C., Zaslavsky, A. M., & Epstein, A. M. (2002).
second- vs first-generation antipsychotic drugs in schizo- Racial disparities in the quality of care for enrollees in
phrenia. Archives of General Psychiatry, 63, 10791087. medicare managed care. Journal of the American Medical
Lacro, J. P., Dunn, L. B., Dolder, C. R., Leckband, S. G., & Jeste, Association, 287(10), 12881294.
D. V. (2002). Prevalence of and risk factors for medication Shern, D., Robinson, P., Giard, J., & Vargo, A., et al. (2005).
nonadherence in patients with schizophrenia: A compre- Evaluation of Floridas Medicaid Managed Mental Health
hensive review of recent literature. Journal of Clinical Plans: Year 8 Report Tampa, FL: Louis de la Parte Florida
Psychiatry, 63(10), 892909. Mental Health Institute, University of South Florida.
Lehman, A. F., & Steinwachs, D. M. (1998). Translating research Svarstad, B. L., Shireman, T. I., & Sweeney, J. K. (2001). Using
into practice: the schizophrenia patient outcomes research drug claims data to assess the relationship of medication
team (PORT) treatment recommendations. Schizophrenia adherence with hospitalization and costs. Psychiatric Ser-
Bulletin, 24, 110. vices, 52, 805811.
Lieberman, J. A., Stroup, T. S., McEvoy, J. P., Swartz, M. S., Tandon, R. (1998). In conclusion, does antipsychotic treatment
Rosenheck, R., Perkins, D., Keefe, R., Davis, S., Davis, C., modify the long-term course of schizophrenic illness?
Lebowitz, B., Severe, J., & Hsiano, J. (2005). Clinical Journal of Psychiatric Research, 32, 251253.
antipsychotic trials of intervention effectiveness (CATIE) Thieda, P., Beard, S., Richter, A., & Kane, J. (2003). An
investigators. Effectiveness of antipsychotic drugs in economic review of compliance with medication therapy in
patients with schizophrenia. New England Journal of Med- the treatment of schizophrenia. Psychiatric Services, 54(4),
icine, 353, 12091223. 508516.
McCombs, J. S., Nichol, M. B., Johnstone, B. M., Stimmel, G. L., Valenstein, M., Blow, F. C., Copeland, L. A., Mcarthy, J. F.,
Shi, J., & Smith, R. (2000). Antipsychotic drug use patterns et al. (2004). Poor antipsychotic adherence among patients
and the cost of treating schizophrenia. Psychiatric Services, with schizophrenia: Medication and patient factors. Schizo-
51(4), 525527. phrenia Bulletin, 30(2), 255264.
Mojtabai, R., Lavelle, J., Gibson, P.J., et al. (2002). Gaps in use Vanelli, M., Burstein, P., & Crammer, J. (2001). Refill patterns of
of antipsychotics after discharge by first-admission patients atypical and conventional antipsychotic medication at a
with schizophrenia, 19891996. Psychiatric Services, 53(3), national retail pharmacy chain. Psychiatric Services, 52,
337339. 12481250.
Moy, E., Dayton, E., & Clancy, C. M. (2005). Compiling the Weiden, P., Aquila, R., & Standard, J. (1996). Atypical antipsy-
evidence: The national healthcare disparities reports. Public chotic drugs and long-term outcome in schizophrenia.
Response, March/April, pp. 376386. Journal of Clinical Psychiatry, 57(Suppl. ll), 5360.
Naber, D., Moritz, S., Lambert, M., et al. (2001). Improvement Weiden, P. J., Kozma, C., Grogg, A., & Locklear, J. (2004).
of schizophrenic patients subjective well-being under atyp- Partial compliance and risk of rehospitalization among
ical antipsychotic drugs. Schizophrenia Research, 50, 7988. California Medicaid patients with schizophrenia. Psychiatric
Nicole, L. (2005). Health disparitiesLess talk more action. New Services, 55(8), 886891.
England Journal of Medicine, 35(7), 727729. Weiden, P. J., & Olfson, M. (1995). Cost of relapse in
Olivieri, N., Matsui, D., Hermann, C., & Koren, G. (1991). schizophrenia. Schizophrenia Bulletin, 21, 419429.
Compliance assessed by the medication event monitoring Weiden, P., & Zygmunt, A. (1997). Medication noncompliance
system. Archives of Disease in Childhood, 66, 13991402. in schizophrenia. Part I. Assessment. Journal of Practical
Pandiana, J. A., & Banks, S. M. (2003). Large data sets are Psychiatry and Behavioral Health, 3, 106110.
powerful. Psychiatric Services, 54, 745. Wyatt, R. J., Henter, I., Leary, M. C., & Taylor, E. (1995). An
Ren, X. S., Kazia, L. E., Lee, A. F., et al. (2002). Patient economic evaluation of schizophrenia. Social Psychiatry and
characteristics and prescription patterns of atypical antipsy- Psychiatric Epidemiology, 30, 196205.
chotics among patients with schizophrenia. Journal of Ziguras, S. J., Klimidis, S., Lambert, T. J. R., & Jackson, A. C.
Pharmacy and Therapeutics, 27, 441451. (2001). Determinants of anti-psychotic medication compli-
Rice, D. P. (1999). The economic impact of schizophrenia. ance in a multicultural population. Community Mental
Journal of Clinical Psychiatry, 60(Suppl. 1), 46. Health Journal, 37(3), 273283.
Rittmannsberger, H., Pachinger, T., Keppelmuller, P., & Wan-
cata, J. (2004). Medication adherence among psychotic

123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.