Insight impairment and associated factors among patients with bipolar disorder: a cross-sectional study at Muhimbili National Hospital, Dar es Salaam, Tanzania, December 2020 to January 2021

¹Department of Psychiatry and Mental Health, Temeke Regional Referral Hospital, Dar es Salaam, Tanzania, ²Department of Psychiatry and Mental Health, Muhimbili University of Health and Allied Sciences, Dar es Salaam, Tanzania, ³Department of Public Health, Muhimbili University of Health and Allied Sciences, Dar es Salaam, Tanzania, ⁴Department of Research, Tanzania Health Promotion Support, Dar es Salaam, Tanzania, ⁵Department of Family Medicine and Population Health, University of Antwerp, Antwerp, Belgium

^&Corresponding author
Fridah Tobias Mtui, Department of Psychiatry and Mental Health, Temeke Regional Referral Hospital, Dar es Salaam, Tanzania

Introduction    

Insight is a multidimensional concept which involves awareness of having a mental illness and symptoms of the illness, understanding the need for treatment, together with awareness of the social and medical consequences of mental illness [1,2]. Insight impairment is one of the fundamental characteristics in patients with mental disorders, and it is more pronounced during the acute phase of the illness.

Bipolar disorder (BD) is a mood disorder characterized by both depressive and manic phases, with about 40 million people globally living with BD in 2019 [3,4]. There is limited data about insight impairment among patients with BD, especially in Africa. Available literature shows the prevalence of impaired insight among patients with BD type 1 in Norway to be 47% in remitted patients and 94% in symptomatic patients [5]. Additionally, a study conducted among patients with BD in Portugal showed that 64.3% of patients had impaired insight [6].

Insight impairment in BD varies according to type and phase of illness; BD type 1 and mixed type experience more insight impairment compared to patients with type 2 [7]. Factors such as higher education, better cognitive function, higher self-reported quality of life, fewer episodes at the time of assessment, more hospitalizations, and better medication adherence were positively correlated with BD insight [6,8-11]. Older age, being female, and shorter duration of illness were negatively associated with BD insight [6,12,13].

Poor insight in BD is a major predictor of non-adherence to medication, increased relapse rates, frequent hospitalizations, and increased risk of suicidality, ultimately contributing to a higher burden of illness and reduced quality of life [14]. On the other hand, evidence shows that interventions for impaired insight among patients with BD are promising targets that can result in better clinical outcomes [15]. To optimize clinical outcomes, pharmacological interventions are frequently augmented with various psychotherapies. These include group psychoeducation [16], compliance therapy, cognitive- behavioral therapy, interpersonal and social rhythm therapy [17] and cognitive rehabilitation targeted at addressing neurocognitive deficits [18].

Most studies on insight impairment have focused on schizophrenia, providing valuable contributions to improving patient care and management [12] . There is limited research on insight among patients with BD, especially in lower- and middle-income countries like Tanzania. This study aimed to assess insight impairment and determine associated factors among adult outpatients with BD at Muhimbili National Hospital (MNH).

Methods     

Study design: we carried out a cross-sectional study among adult patients attending the outpatient psychiatric clinic at MNH over one month (from December 2020 to January 2021).

Study setting, participants and selection of participants: Muhimbili National Hospital is located in Dar es Salaam, Tanzania, and serves as the national referral and teaching hospital with a 1500-bed capacity. It has a psychiatric unit attending about 140 adult outpatients with BD per month from the 2019 MNH Database, serving patients within the city and other regions. We included all patients with BD aged 18 years and above attending the outpatient psychiatric clinic at MNH. Patients with comorbid neurological illness and patients in the acute phase of mental or medical illness, thus unable to consent or comprehend research tools, were excluded. Sample size estimation of 154 participants based on the continuous data formula.

Where; Z= confidence interval 95%=1.96, δ = standard deviation =0.57 mean score [19]; E= margin of error of 5% to a mean score of 1.89= 0.0945 mean score [19]; N= (1.96x0.57)/ 0.0945)2 =139.76 ≈ 140. Adjusting for non-response, we add 10% =140 + 14; when oversampled by 10%, the sample size becomes 154. Purposive sampling was employed to screen all attending outpatients and identify individuals with BD presenting for routine follow-up clinic. Following informed consent, participants were screened using the bipolar module of the Mini-International Neuropsychiatric Interview (M.I.N.I.) version 7.0.2; those who met the criteria for bipolar disorder were formally enrolled in the study. Out of the 170 patients recruited, only 154 agreed to participate in the study and met the inclusion criteria, which were patients aged 18 years and older diagnosed with BD attending the outpatient psychiatric clinic at MNH. However, 5 patients were excluded due to having psychotic disorders as assessed by MINI, 8 patients did not consent, and 3 patients rushed for consultation in other departments.

Variables: we specified the socio-demographic characteristics of the participants as age, sex, marital status, education level, and occupation status. We then specified the clinical characteristics into age at first BD episode, number of admissions, substance use, use of medications, illness duration in years, type of medications, psychotherapy service, type of bipolar disorder, and bipolar disorder phase. Our dependent variable was insight level by considering ISAD mean score.

Data resource and measurement: a structured questionnaire was originally developed in English and subsequently translated into Swahili to gather data on participants' socio-demographic and clinical characteristics. Two research assistants trained on ethical protocols, sampling procedures, questionnaire adaptation, and administration of study tools were recruited for data collection. The questionnaire included the ISAD, a tool developed to assess insight impairment in patients with affective disorders [20]. The ISAD consists of 17 items: items 1-3 assess insight into the overall condition, while items 4-17 focus on insight into specific symptoms of bipolar disorder. Each item is rated on a 0-5 Likert scale, where 0 indicates no symptoms or an inapplicable situation, 1 reflects full awareness of psychiatric morbidity, 2 reflects awareness, 3 represents moderate awareness, 4 represents slight awareness, and 5 indicates a complete lack of awareness. The tool was translated into Swahili, and some modifications were made to the Likert scale for clarity during interviews: 1 represented the absence of awareness, 2 represented slight awareness, 3 indicated moderate awareness, 4 indicated awareness, and 5 signified full awareness of psychiatric morbidity. However, during data entry and analysis, responses were converted back to the original Likert scale. During validation in Brazil, the tool demonstrated strong psychometric properties. Intraclass correlation coefficients (ICCs) across the items ranged from 0.80 to .99, confirming high inter-rater reliability. The scale also exhibited robust internal consistency, with an overall Cronbach's alpha of 0.90. Subscale analysis revealed acceptable internal consistency for items 1-3 alpha = 0.77 and excellent consistency for items 4-17 (alpha = 0.89) [12]. No evidence of being used in Tanzania or in a similar sociocultural context. Apart from confirming BD diagnosis, the M.I.N.I. version 7.0.2 bipolar module was also used to pick manic, depressive, and euthymic phases of BD. To determine insight impairment among patients with BD, we used a cut-off point of 3 on the ISAD total mean score. Those with a mean score above 3 were classified as having impaired insight, while patients with an ISAD total mean score of 3 or below were categorized as having preserved insight.

Data analysis: statistical analysis was realized by the software Statistical Package for the Social Sciences (SPSS) version 23.0. Descriptive statistics for demographic and clinical characteristics were summarized using frequencies, percentages, mean, and corresponding standard deviation. Independent t-tests and one-way ANOVA were utilized for bivariate analysis to examine how socio-demographic and clinical factors were associated with the ISAD total mean score. All the variables associated with the ISAD total mean score with a p-value less than 0.2 in bivariate analysis were further analyzed using linear regression analysis to control for confounding effects. A p< 0.05 was considered significant.

Ethical issues and protection of human subjects: ethical approval to conduct the study was granted by the Muhimbili University of Health and Allied Sciences (MUHAS) Senate Research and Publications Committee, IRB reference number MUHAS-REC-10-2020-410, and permission to conduct the study was granted by MNH (letter reference number MNH/TRCU/Perm/2020/0014. Oral informed consent was obtained from all participants before they enrolled in the study. No personal identifying information was obtained from the participants. Data were accessible only to authorized study personnel. The study adhered to ethical standards of the Declaration of Helsinki.

Results     

Social demographic and clinical characteristics of the outpatients with BD at Muhimbili National Hospital: a total of 154 eligible and consented patients with BD participated in the study. Females accounted for half of the sample; the mean (SD) age was 36.32 (11.5) years, with an age range of 18- 67 years. The majority, 54.6% (84/154) of the participants were single; 29.2% (45/154) were either married or cohabiting, while 16.2% (25/154) were either divorced/ separated or widowed. In education level 37.7% (58/154) of participants reported having secondary education, while 31.8% (49/154) reported having primary education and college or university education 30.5% (47/154). Half (77/154) of participants reported being self-employed in either business or agricultural activities.

The sample's mean (SD) age at the first bipolar disorder (BD) episode was 26.7 (8.5) years (range: 15-52 years), with approximately half of the cohort (50.6%,78/154) experiencing their first episode between ages 25 and 50. The mean duration of illness was 9.8 (9.7) years, and the mean number of hospital admissions was 2.5 (2.96). Regarding treatment, 98.7% (152/154) of participants were actively taking medication at the time of the interview. Among those medicated, 49.3% (75/152) were on combination medication, 45.4% (69/152) were taking antipsychotics, and 5.3% (8/152) were utilizing mood medications (mood stabilizers or antidepressants). Most participants (72.1%,111/154) had received psychoeducation exclusively, while 27.9% (43/154) received a combination of psychological interventions (psychoeducation plus other). Clinically, the vast majority (87.7%,135/154) were diagnosed with Bipolar I disorder. At the time of evaluation, 72.7% (112/154) of the sample were euthymic, 15.6% (24/154) were manic, and 11.7% (18/154) were in a depressed phase. Additionally, a large majority (85.7%,132/154) reported no substance use within the past 12 months (Table 1).

Prevalence of insight impairment among outpatients with BD attending at Muhimbili National Hospital: the ISAD total mean score/ high score was 2.48/5. About a third (28.6%, 44/154) of participants had insight impairment with a mean score >3.

Association of social demographic, clinical characteristics and ISAD total mean score among outpatients with BD at Muhimbili National Hospital: in bivariate analysis, we found a significant difference in the ISAD total mean scores among the employed 1.96 (1.03), unemployed 2.67 (0.89), and self-employed 2.62 (0.95) participants (p=0.001). We found no statistically significant difference in the ISAD total mean score by sex, age, marital status, or education level. In clinical characteristics, the total mean ISAD score differed significantly across categories of illness duration (p = 0.019). Participants with a disease duration of 3 years or less exhibited the highest impairment, with a mean (SD) score of 2.78 (0.99). This was followed by those with an illness duration between 3.1 and 12 years 2.34 (0.94), while the lowest impairment was observed in individuals with illness duration exceeding 12 years 2.29 (0.97). Type of BD was significantly associated with variation in ISAD total mean score p = 0.016. Specifically, participants with BD I had higher mean (SD) score of 2.49 (1.02) compared to those with BD II 2.40 (0.73). The current clinical phase of illness was significantly associated with differences in insight levels p = 0.001. Participants in a manic phase demonstrated the highest insight impairment, with a total mean ISAD score of 3.14 (1.08), followed by those in a depressive phase 2.41 (0.75), while the lowest impairment was observed during the euthymic phase 2.36 (0.95). Conversely, no statistically significant differences in mean total ISAD scores were found across other clinical characteristics, including age at first bipolar disorder episode, history of psychiatric admissions, substance use, current medication use, specific type of medication prescribed, or history of psychotherapy intervention (Table 2).

In the multivariable linear regression analysis, employment status, illness duration, and clinical phase emerged as independent predictors of total mean ISAD scores. Compared to employed participants, lower total mean ISAD scores were independently predicted by being self-employed (beta = 0.302; 95% CI: 0.193, 0.994; p = 0.004) and being unemployed (beta = 0.302; 95% CI: 0.235, 1.102; p = 0.003). Additionally, an illness duration of ≤ 3 years significantly predicted higher insight impairment (beta = 0.306; 95% CI: 0.246, 1.006; p = 0.001) relative to a duration exceeding 12 years. Presenting in a manic phase also remained a strong, significant predictor of higher impairment scores (beta = 0.286; 95% CI: 0.359, 1.191; p < 0.001). In contrast, bipolar disorder subtypes demonstrated no independent association with total mean ISAD scores in the multivariable model (Table 2.1).

Discussion     

In our study, BD had similar prevalence between females and males, which has also been found in other studies [21]. Almost all participants 98.7% were on medications, and the majority were in a euthymic state; this emphasizes the importance of proper pharmacological management of patients with BD to help them attain a stable state. Our findings revealed that 28.6% of patients with BD exhibited impaired insight. This prevalence is lower compared to studies done in Norway and Portugal 70.2% and 6>4.3%, respectively [6,22]. The difference in insight assessment tools and participants can explain this finding. Participants in our study were outpatients, and the majority were in a euthymic state, which is associated with better insight compared to other phases of BD [9,12,23]. Unemployed and self-employed status showed association with level of insight. Patients without employment (27.3%) had a lower level of insight compared to patients who were self- employed. This outcome could stem from the socioeconomic reality that patients lacking employment typically present with challenges sustaining proper management of BD, as studies have shown an association of symptomatic BD with lower levels of insight [12].

Illness duration of ≤ 3 years is associated with lower insight compared to illness duration above 3 years among patients with BD. Findings from other studies support that shorter duration of BD is associated with lower insight among patients with BD [9,23]. Patients with short illness duration lack awareness of changes in their experience as symptoms of BD, but also, they are not exposed to different and more intensive BD management, giving them a chance to get biopsychosocial interventions which in the end help them attain remission phase. In our study, all patients received psychotherapy services, whereby 72.1% received psychoeducation while 27.9% received psychoeducation plus other psychotherapies helping in management of BD. However, some studies have shown that longer BD duration is associated with insight impairment due to long duration of residual symptoms of BD and cognitive impairment [24]. Among the three phases of BD, the manic phase was a significant predictor of lower insight level among participants with ISAD mean score of above 3. Findings from studies have shown that the manic phase of BD is associated with lower insight compared with the depressive, mixed, or euthymic phase [23,25-27]. This is since most of the patients in this state experience impaired executive functioning, which accounts for observed insight impairment [6].

In our study we used the ISAD, which is a specific insight assessment tool for affective disorders and has a multidimensional concept of insight to ensure reliable evaluation of insight level. While many previous studies have focused on the level of insight and its associations with clinical and demographic factors, our study contributes to the prevalence of insight impairment, an area with relatively limited research. Limitations of the study include there is possibility of recall bias to some variables like illness duration, age of the first onset and number of admissions. Our study had limitations in sampling, as there were no patients with mixed, which is a quite common phase in BD, and also dealing with only outpatients and not including inpatients has led to assessing much more stable patients in manic and depressive phases, which may have influenced a large number of participants to have preserved insight. Using cross-sectional study did not give us room to follow up the same patient in different phases of BD and assess level of insight thus, use the patient as control of himself reducing confounding factors.

Conclusion     

Findings from this study highlight the importance of routine assessment of insight among patients with BD, with factors such as duration of illness, manic phase, and employment status requiring special attention. Given its implications, we recommend having a validated and culturally accepted tool to measure insight impairment within routine mental health services. The study calls for further research to investigate insight using a larger sample size or a different set-up (e.g inpatient) among adult patients with BD since there are limited data in our context. Emphasis on proper psychopharmacological treatment to patients with BD to improve treatment outcomes and improvement in insight impairment.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Conception and study design: Fridah Tobias Mtui and Samuel Lazarus Likindikoki. Data collection: Fridah Tobias Mtui and Samuel Lazarus Likindikoki. Data analysis and interpretation. Fridah Tobias Mtui, Samuel Lazarus Likindikoki and Anna Louis Jazza. Manuscript drafting: Fridah Tobias Mtui, Samuel Lazarus Likindikoki, Harrieth Peter Ndumwa, Sifaely Abiud Mtaita, Mariam Erick Ngaula, and Anna Louis Jazza. Manuscript revision: Fridah Tobias Mtui, Samuel Lazarus Likindikoki, Harriet Peter Ndumwa, Sifaely Abiud Mtaita, Mariam Erick Ngaula and Anna Louis Jazza. All the authors have read and agreed to the final version of this manuscript.

Acknowledgments     

The authors of the article would like to thank the leadership of Psychiatry and Mental health Department at MNH in Dar es Salaam for their permission to conduct the study and collaboration during data collection. Also, to patients who took part in this study and provided data that describe findings on this study.

Tables     

Table 1: socio-demographic and clinical characteristics of participants with bipolar disorder recruited at Muhimbili National Hospital outpatient clinic, Dar es Salaam, between December 2020 and January 2021(N=154)

Table 2: association of socio-demographic and clinical characteristics and insight scale for affective disorder mean score among participants with bipolar disorder recruited at Muhimbili National Hospital outpatient clinic, Dar es Salaam, between December 2020 and January 2021(bivariate analysis)

Table 2.1: association of socio-demographic and clinical characteristics and insight scale for affective disorder mean score among participants with bipolar disorder recruited at Muhimbili National Hospital outpatient clinic, Dar es Salaam, between December 2020 and January 2021(multivariate analysis)

References