Effect of Subjective Reasoning and Neurocognition on Medication Adherence for Persons With Schizophrenia

Understanding the factors underlying adherence to a neuroleptic regimen is important for planning treatment for schizophrenia. Weiden and colleagues ( 1 ) developed the Rating of Medication Influences (ROMI) scale for the assessment of attitudinal and behavioral factors influencing patients' compliance with neuroleptic treatment during the preceding month. Of the 20 ROMI items, seven represent psychological factors related to compliance in persons who have demonstrated adherence in recent months—the subgroup of patients most likely to benefit from rehabilitation programs.

An exploratory factor analysis of ROMI items yielded three subscales: influence of others, including the quality of the relationship with the prescribing doctor or nonprescribing therapist and the belief of family and friends that drug treatment is important; prevention, including the belief that neuroleptic treatment forestalls the illness from returning and fear of rehospitalization; and medication affinity, including perceived daily benefit of neuroleptics and absence of perceived social pressure ( 1 ).

Because neurocognitive impairment is one of the most reliable predictors of functional outcome in schizophrenia ( 2 ), the association between cognitive functioning and adherence should be clarified. The results of previous studies, however, have been mixed. Donohoe and colleagues ( 3 ) reported an association between poor compliance and lower facial recognition. Robinson and colleagues ( 4 ) found that better executive functioning decreased the likelihood of discontinuation of medication. A study specifically using ROMI showed an inverse correlation between influence of others and verbal memory and cognitive flexibility ( 5 ). Accordingly, the goal of this study was to examine the relationship between patients' reasoning concerning adherence as assessed by the ROMI and neurocognitive function of Japanese patients diagnosed as having schizophrenia.

Methods

The participants were recruited from among the outpatients of Tokyo University Hospital attending the psychiatric day treatment unit. They regularly participated in rehabilitation programs for one to four days per week to improve social, coping, and job-related skills; however, they were not given specific group psychoeducation for medication self-management. We asked the 40 regular users of the center to participate in the study, and 16 (11 men and five women; mean±SD age of 28.5±4.2 years) with schizophrenia agreed to participate. Diagnosis of schizophrenia was made through the Structured Clinical Interview for DSM-IV Axis I Disorders by experienced psychiatrists. The illness duration was 8.3±4.8 years. The duration of education was 13.9±1.8 years. No participant had a history of alcohol or other drug abuse or dependence, nor did any have a neurological illness that affected the central nervous system. The Ethical Committee of the Faculty of Medicine, University of Tokyo, approved this study. After a complete explanation of the study to the participants, we obtained their written informed consent. The data were collected from February through May 1999.

Participants' mean intelligence quotient (IQ) on the Wechsler Adult Intelligence Scale-Revised was 97.2± 13.5 on the verbal IQ (VIQ), 81.5± 12.0 on the performance IQ (PIQ), and 89.8±13.1 on the full-scale IQ (FIQ) tests. Their psychiatric symptoms were mild; the mean rating on the positive subscale of the Positive and Negative Syndrome Scale was 11.1±3.1, and the mean rating on the negative subscale was 17.4±5.9; possible scores on both subscales range from 7 to 49, with higher scores indicating more severe symptoms. Also, the general psychopathology scale score was 30.5±7.8; possible scores range from 16 to 112, with higher scores indicating more severe symptoms. The mean score on the Global Assessment of Functioning was 48.9± 9.2; possible scores range from 0 to 100, with higher scores indicating better functioning. All participants were taking neuroleptic medications, with a mean daily chlorpromazine-equivalent dose of 443.8±257.2 mg.

The ROMI ( 1 ) was used to assess subjective reasoning on medication adherence. The ROMI consists of a semistructured interview (containing items on demographic characteristics, general attitudes toward medication, and so on) followed by a structured interview exploring factors that influence medication adherence. According to Weiden and colleagues ( 1 ), the 13 noncompliance items in the ROMI apply only to patients who have not taken their medication for at least one week for any part of the past month; otherwise only the first seven compliance items are administered. The latter situation applied to all participants in this study. Each question of the ROMI is answered on a 4-point scale. Two independent raters rated the ROMI on five randomly selected cases, which yielded intraclass correlation coefficients from .73 to 1.00 for items that assessed compliance.

Executive functioning was assessed by the Wisconsin Card Sorting Test (WCST) ( 6 ). The index was percentage of perseverative errors, calculated as the number of perseverative errors divided by the total number of responses. The result was a mean percentage of 18.7±11.4; possible scores range from 0 to 100, with higher scores indicating worse perseveration. Verbal memory was assessed with the Rey Auditory Verbal Learning Test (RAVLT) ( 7 ), on which the participants were asked to recall a list of 15 words spoken by a tester. The procedure was repeated five times (sessions 1-5), and the sum of recalled words from sessions 1 to 5 was used for the analysis. The mean score was 39.6±9.8; possible scores range from 0 to 75, with higher scores indicating better memory.

Correlational analyses using Spearman's rho were performed for scores on each cluster of the ROMI and neurocognitive and other clinical indices. The level for significance was set at .05 (two-tailed).

Results

The cluster for prevention showed a significantly positive association with the percentage of perseverative errors for the WCST ( ρ =.504, p=.047). For other clinical indices, influence of others was significantly associated with education ( ρ =-.627, p=.009), medication dosage ( ρ =.625, p=.01), PIQ ( ρ =-.623, p=.01), and FIQ ( ρ = -.550, p=.027). Medication affinity was significantly associated with education ( ρ =.554, p=.026). Prevention was significantly associated with age ( ρ =.574, p=.02).

Discussion

This study examined the relationship between subjective reasoning on adherence to neuroleptic treatment and neurocognitive and other clinical variables in a treatment-compliant sample of outpatients with schizophrenia. We found a significant association between higher prevention cluster scores and lower executive functioning. This finding replicated the previous finding of a correlation between higher influence of others and lower executive functioning—that is, higher WCST perseverative error scores ( 5 ). Other studies, however, showed a positive correlation between medication compliance and neurocognitive functioning ( 3 , 4 ). This discrepancy may be partly a result of the differential characteristics of the sample—our sample comprised only treatment-compliant outpatients, whereas other studies included treatment-noncompliant participants. Another possibility is that medication continuation or discontinuation was the indicator for outcome in other studies, whereas in our study the ROMI Scale evaluated subjective reasoning regarding adherence, not adherent behavior per se. The prevention cluster was also positively associated with participants' age, which may indicate that older participants had more experience with subjective feelings of anticipation of relapse and readmission.

The findings that influence of others showed negative associations with education and general IQ and positive association with medication dosage suggest that a positive attitude of the patient's relatives toward neuroleptic treatment and a positive psychiatrist-patient relationship may be important factors for medication adherence for patients with lower education and IQ. These findings imply that individuals who realize the limits of their knowledge may become more reliant on the influence of others. On the other hand, the medication affinity cluster showed a positive association with education, which may suggest that patients with higher education tend to be compliant because of perceived everyday benefits of neuroleptic treatment by symptom control and not because of any social pressure. Thus the routes to compliance may be very different for persons according to their cognitive capacity.

Taken together, these results suggest that consideration of executive functioning, education, and general IQ may be important to assess individual motivation for medication adherence. This pattern of correlation may have important implications for the design of psychoeducational approaches and compliance therapy, such as a need to provide different services for individuals on the basis of individual clinical characteristics and different neurocognitive profiles.

That neither negative symptoms nor positive symptoms were associated with compliance behavior in contrast to the previously reported findings ( 8 ) may reflect the low level of psychopathological symptoms in our sample and the lack of a noncompliant subsample.

Our study has limitations. First, the sample was small and had a restricted range of clinical characteristics. Only outpatients with mild symptoms and high compliance were included. There may have been a selection bias. Those who agreed to participate may have had better relationships with staff, may have felt greater benefit from the treatment, and may have been more compliant than those who did not. Second, the design of this study was cross-sectional. Therefore, assessment of patients with varying degrees of severity in various settings in a longitudinal design may be necessary for future investigations.

In summary, we observed relationships between clusters of compliance and neuropsychological indices of prefrontal functioning and clinical characteristics among treatment-compliant outpatients with schizophrenia. The findings support the notion that assessment of individual levels of neurocognitive functioning and clinical characteristics is important to plan efficient psychoeducation about medication adherence for persons with schizophrenia.

Acknowledgments

This work was supported in part by grants-in-aid for scientific research (16790675 and 17025015 to KK; 17-5234 to MAR) from the Japan Society for the Promotion of Science and the Ministry of Education, Culture, Sports, Science, and Technology, Japan. Dr. Rogers was supported by postdoctoral fellowship grant 237027 from the National Health and Medical Research Council of Australia. The authors thank Kazuo Yamada, M.D., Ph.D., for providing the Japanese version of the Rating of Medication Influences scale, which was made possible with the permission of Peter Weiden, M.D. The authors also thank Takahiro Ohno, M.D., and Akinobu Hata, M.D., Ph.D., for data collection, and Masato Fukuda, M.D., Ph.D., for his contribution to the intellectual content of the study.