Initial Medication Adherence by Psychiatric Outpatients in a General Hospital and Relevant Factors

Background: The International Society for Pharmacoeconomics and Outcomes Research proposed two types of medication adherence: initial and long-term. Initial adherence is a predictor of long-term adherence and thus is a crucial metric to explore and support. This study aimed to investigate initial medication adherence by psychiatric outpatients and relevant factors. Methods: The study surveyed psychiatric outpatients using a 30-day timely return visit rate (TRVR) after the rst visit to indicate initial adherence. All participants agreed to engage in the self-designed survey and assessments of the Eysenck Personality Questionnaire (EPQ) and Symptoms Checklist-90 (SCL-90). Clients who missed timely return visits received telephone follow-up to determine the main reasons. Results: The overall TRVR was 59.4%, and 40.6% of clients missed return visits. Logistic regression analysis revealed risk factors for initial adherence were work, tense family atmosphere, negative attitudes towards medication, higher EPQ psychoticism score, and lower SCL-90 phobic anxiety score. The main reasons given for non-timely return visits were improvement suggesting lack of need for a return visit, various limitations, no improvement, and side effects. Conclusion: Psychiatric outpatients had poor initial medication adherence related to multiple dimensional factors, including job, family, personality characteristics, mental status, and thoughts about mental illness and treatments. The present study focused on the IMA of psychiatric outpatient rst-visit clients in a relatively large sample. We used TRVR as the indicator of IMA and found an overall TRVR of 59.4% as well as its multiple relevant factors, such as attitude towards medication, family atmosphere, working status, age, and EPQ psychoticism and SCL-90 phobic anxiety subscale scores. From the clients’ perspective, through a telephone follow-up, we found that the main reasons for the NTRV included considering return visits unnecessary because of improvement and other various barriers.


Background
Adherence is an assessment indicator of a patient's illness behavior, referring to the extent to which the patient adopts behavior consistent with a treatment plan negotiated by the doctor and patient [1]. The International Society for Pharmacoeconomics and Outcomes Research proposed two types of medical adherence: initial (IMA) and long-term (LTA). IMA refers to adherence after being prescribed a new drug, and LTA refers to adherence during continuous medication or long-term treatment [2,3]. Poor IMA can predicate worse LTA [4]. Currently, there is no consistent de nition of the "initial" phase with most studies de ning it as 14 to 28 days after a new prescription or the rst visit but others including periods of months or even a year [2]. This study de ned the initial phase as 30 days after the rst visit and used timely return visit (TRV) behavior as the outcome indicator of IMA.
Poor adherence is a global challenge for psychiatric clinical practices [5]. High proportions of psychiatric patients have shown poor treatment adherence, leading to poor treatment outcomes and resulting in a massive waste of social resources [4], [6,7], [8] [9], [10]. Factors found to impact treatment adherence among psychiatric patients are cognitive responses, patients' trust in doctors, patients' perception of their mental health, and age, unpleasant side effects of drugs, attitudes towards drugs, paranoia symptoms, substance abuse, self-knowledge, and family environment [11] [12].
Most previous studies have focused on long-term medication adherence. The IMA in psychiatric outpatients and its relevant factors remain unclear. Thus, the present study investigated the IMA in a large sample and performed a telephone follow-up in clients with poor adherence. The ndings of this study could provide meaningful evidence for practical efforts to improve IMA in psychiatric outpatients.

Participants
All participants were rst-visit clients of the psychiatric outpatient department of the Third A liated Hospital of Sun Yat-sen University from May to October 2018. The inclusion criteria were as follows: (1) provided informed consent, (2) age between 18 and 65 years, (3) education level above primary school, (4) doctors prescribed medicine at the rst visit, and (5) provided complete data. Patients hospitalized after the rst visit were excluded. This study was in line with the Declaration of Helsinki, Ethical Principles for Medical Research Involving Human Subjects (World Medical Association, 2013) [13].

Survey and Instruments
(1) First-visit questionnaire: This self-designed questionnaire was used to investigate age, gender, education level, marital status, employment status, monthly income, residence, family atmosphere, living status, and attitudes towards pharmacotherapy. The respondents completed this survey on their rst visit according to their actual circumstances.
(2) Timely return visit (TRV): Based on the hospital information system records, TRV refers to a client completing at least one return visit within 30 days after the rst visit; otherwise, the patient was labeled "non-timely return visit" (NTRV). TRV rate (TRVR)=TRV cases number / total sample number × 100%. This study used TRVR as the indicator of initial adherence.
(3) Eysenck Personality Questionnaire (EPQ) Chinese version: The EPQ has been widely used to evaluate personality characteristics in China since it was translated into Chinese and standardized in the 1980s [14]. It includes 88 items and four subscales: Psychoticism (P), Extraversion (E), Neuroticism (N), and Lie (L) [15].
(5) Telephone follow-up questionnaire: This questionnaire was self-designed by the researchers. It listed various presupposed reasons for NTRVs: patients' refusal, unsatisfactory effects, unsatisfactory outpatient environment, improvement leading to believe there is no further need for a return visit, fear of side effects, signi cant side effects, no time, far distance, failure to make an appointment, economic problems, and others. According to their actual conditions, the respondents could choose one presupposed reason or supply a speci c other reason on the telephone.

Statistic analysis
The dataset was built via SPSS 25.0. Descriptive statistics were used in distributions of the TRVR and reasons for NTRVs. Group comparisons of numerical data such as EPQ and SCL-90 subscale scores were analyzed via an independent t-test. Group comparisons for category data such as age group, marital status, monthly income, and residence were made via χ² test. The relevant factors for the NTRV were analyzed using Wald forward binary logistic regression. The signi cance level was set at α=0.05.

Investigation into the TRVR and its relevant factors
Overall TRVR In total, 2843 cases meeting inclusion criteria entered statistical analysis, and 1689 completed at least one return visit 30 days after the rst visit. The overall TRVR was 59.4%, and 1154 cases (40.6%) missed the TRV (34.2% in patients with psychotic disorders). Thus, participants were divided into two groups: TRV (n=1689) and NTRV (n=1154).
Demographic data Table 1 shows the participants' demographic data and the TRVR distribution among subgroups. There were statistically signi cant TRVR differences among age, job status, and residence subgroups.

Self-reported reasons for NTRVs
According to respondents' answers, the main reasons for NTRVs were as follows: 383 (55.3%) considered the return visit unnecessary because they felt better, 87 (12.6%) felt no improvement, 60 (8.7%) felt noticeable side effects of the medicine, 20 (2.9%) were afraid of side effects of the medicine, 51 (7.4%) reported inconvenience of traveling to the hospital, 61 (8.8%) had no time, 24 (3.5%) failed to make appointments, 4 (0.6%) reported rejecting the return visit, 1 (0.1%) acknowledged di cult economic circumstances, and 2 (0.3%) were unsatis ed with the hospital environment. These reasons were merged and divided into four groups to meet the statistical conditions: improvement and consideration of no need for a return visit (55.3%), no improvement (12.6%), side effects of medicines (11.5%), and various limitations (20.6%). The distribution of self-reported reasons for the NTRV differed signi cantly among 4 conditions after the rst interview (χ 2 = 221.989, P<0.001). See Table 4.

Discussion
The present study focused on the IMA of psychiatric outpatient rst-visit clients in a relatively large sample. We used TRVR as the indicator of IMA and found an overall TRVR of 59.4% as well as its multiple relevant factors, such as attitude towards medication, family atmosphere, working status, age, and EPQ psychoticism and SCL-90 phobic anxiety subscale scores. From the clients' perspective, through a telephone follow-up, we found that the main reasons for the NTRV included considering return visits unnecessary because of improvement and other various barriers.
Poor IMA could decrease the probability of receiving normative and professional treatment, impair outcomes, and even lead to severe consequences [5]. However, poor IMA remains a global challenge for psychiatric practices [3,19]. Ibironke et al. reported a prevalence of 30% of missed rst appointments 2-3 weeks after the rst outpatient visit in patients with schizophrenia[8]. According to National Comorbidity Survey Replication data, the mental health treatment dropout rate from the general medication sector at 12 months was 31.6%, and over 70% of all dropouts occurred after the rst or second visit [20]. Our ndings of high nonadherence prevalence in psychiatric outpatients and patients with psychotic disorders 30 days after the rst visit support the determination that it is time to pay more attention to and effectively improve IMA in psychiatric outpatients.
Generally, IMA was among the illness behaviors critically in uenced by cognitive appraisal and decisionmaking by patients based on their perceived information from illness and appropriate treatment acknowledge, feelings related to self-distress, doctor interview, effects, side effects, and balancing bene ts against costs [21,22].
Attitude towards medication markedly in uenced IMA. People with negative attitudes towards pharmacotherapy had the worst TRVR of 44.2% (Table 1), and refusal of medicine was a key predictor with the highest OR of 2.265 to the poor IMA (Table 3). At the same time, regarding the self-reported reasons for NTRVs, the present study found that 30.1% of patients abandoned treatments because of medicine side effects (Table 4). In line with the ndings of this study, reviews of Kampman [12] and Mitchell [21] also showed that the patient's choice of treatments was an essential factor in uencing adherence. Given this nding and the substantial in uence of side effects, we suggest a regular survey before the rst visit about attitudes towards treatments, particularly pharmacotherapy, might be constructive for doctors to perform patient education.
Regarding the appraisal of illness severity, patients with higher SCL-90 subscale scores of somatization, anxiety, and phobic anxiety were more likely to complete TRVs ( Table 2). The phobic anxiety subscale score was a protective factor for the NTRV (Table 3). A nonadherence study in patients with rst-episode psychosis found that less severe positive symptoms at baseline could strongly predict poor medication adherence in the next one and two years [10]. This was similar in patients with depression [7]. These ndings suggest that the perception and expectation of signi cant distress in the body or mind might push patients to carry out suitable IMA with less hesitation. Thus, more health education about normative medical intervention should be provided to mild-to-moderate-severity patients and their essential relatives.
This study also represented patients' balance between bene ts and costs of return visits by analyzing the self-reported reasons for NTRVs. For example, patients transferred to other hospitals had higher ratios of "no improvement" (46.6%) and "various limitations" (39.7%) than other subgroups (Table 4). Similarly, being on the job was a risk factor for NTRVs (Table 3), perhaps because a return visit appointment would lead to loss of income. Subsequently, a study on treatment preferences in patients with depression showed that employed patients preferred psychotherapy over antidepressants [23], suggesting that patients on the job might be more afraid of medicine side effects in uencing their functional status.
Many factors could in uence the decision-making regarding a return visit. Social (including family atmosphere, living status, and residence) and personal factors (including age and personality) were relevant to IMA in the present study. A supportive family environment-for example, a happy family atmosphere or living with family members-did help improve IMA in this study, which was in line with many previous studies [12,24,25]. The in uence of age on IMA was complex in the present study; there was a tendency of increasing TRVR with age while patients aged 26-35 years old had the worst IMA with lower TRVR than those younger or older (Table 1). Consistent with our nding, a structural equation modeling study on adherence to psychopharmacological medications in psychiatric patients reported a weakly positive correlation between adherence and age [11]; in adult outpatients with depression, younger age was related to dropout at 12-week follow-up[25]. Usually, age was a background factor related to various psychological phenomena such as health belief, cognition pattern, social role con ict, and treatment experience and expectation. Therefore, we suspect there might be differing mediating factors between adherence and age in different study participants. Overall, younger adult psychiatric outpatients need more health education about IMA.
There were some limitations in this study. As the study was carried out in routine clinical practices, there could be uncontrollable, potentially confounding factors and a relatively high rate of follow-up failure.
Additionally, factors related to doctor-patient communications could also in uence IMA and should be determined in future research.

Conclusions
Poor IMA remained in psychiatric outpatients of the general hospital. Health education to patients, particularly younger and mild-severity patients, on the course of illness development, treatment methods, and the necessity of integrated case management are essential for working with psychiatric outpatients.