Insight,psychosis,and depression in Africa: A cross-sectional survey from an inpatient unit in Ghana

Abstract

Few studies of the relationship of insight to psychopathology have been conducted in non-Western populations. This study examined the relationships between insight and depression, anxiety, and positive and negative symptoms on patients with schizophrenia resident in a psychiatric hospital in Ghana. A sample of 49 participants, (37 men and 12 women), with DSM-IV defined schizophrenia took part in semistructured interviews consisting of the Hamilton Rating Scales for Depression (HAM-D) and Anxiety (HAM-A); the Schedule for the Assessment of Insight – Expanded Version (SAI-E) and the Positive and Negative Syndrome Scale (PANSS). Bivariate correlations between variables were examined and those significantly correlated with an insight domain were included in multiple regression models. Variables associated with the total insight score were age, gender, anxiety symptoms, depression symptoms, and treatment compliance. In the final model, HAM-D positively predicted total SAI-E score, whilst PANSS-pos was negatively associated with total SAI-E score. The results are broadly consistent with those found in Western samples regarding insight and depressive symptoms. Implications of these results for competing theories of insight in psychoses are discussed. Patients able to identify themselves as ill may be aware of their affective symptoms.

Keywords

depression Ghana insight psychosis schizophrenia West Africa

Introduction

The distinction between psychosis and neurosis traditionally rested on the preservation of insight in neurosis. This distinction was abandoned in ICD-10 and DSM-III partly because insight proved so stubbornly resistant to definition (Fulford, 2004). However, renewed interest over the past two decades has generated some consensus in relation to assessment and measurement of insight. Evidence suggests that insight is comprised of at least three partially interdependent domains: self-recognition of being ill; the relabeling of psychotic experiences as abnormal; and perceived need for treatment (Amador, Strauss, Yale, &Gorman, 1991; David, 1990).

Three models have been proposed to account for impaired insight in psychosis. A neuro-cognitive deficit may impair self-recognition of symptoms of mental illness. This deficit may reflect a global cognitive impairment or a more specific deficit associated with parietal and/or frontal lobe dysfunction (Flashman &Roth, 2004). A recent well-designed longitudinal study found neuropsychiatric measures of global-attention, memory, and executive function did not correlate with any of the three scales measuring insight used (Cuesta, Peralta, Zarzuela, &Zandio, 2006). However, a systematic review and meta-analysis of 35 studies found that while there is a weak positive association between insight and general cognitive function, there is a moderately strong association with impaired set-shifting and error monitoring during the Wisconsin Card Sorting Test (WCST; Aleman, Agrawal, Morgan, &David, 2006). This suggests a specific impairment in executive functioning.

An alternative model hypothesizes that patients’ prior cultural or idiosyncratic beliefs induce them to misattribute the symptoms of illness to another cause, such as sorcery (Saravanan, Jacob, Prince, Bhugra, &David, 2004). Finally, a psychodynamic model suggests that knowledge of symptoms may be defensively denied to preserve self-esteem from the devastating reality of the illness. The psychodynamic theory is consistent with the finding that greater insight is associated with anxiety and depression. Another meta-analysis of 40 studies found evidence that awareness of symptoms of mental illness is moderately correlated with a negative affect, while other measures such as awareness of the social consequences of illness and the need for treatment are only weakly associated with this (Mintz, Dobson, &Romney, 2003). Both awareness of social consequences and need for treatment, however, may be associated with suicidal ideation (Schwartz, 2000), though it is currently unknown if this translates into suicidal acts. Those with little insight score higher on a scale measuring self-deception (Moore, Cassidy, Carr, &O’Callaghan, 1999), but this measure has yet to be validated in a population with psychosis. Perhaps the mediating factor is cognitive rather than psychodynamic, as those who hold negative beliefs about mental illness are more aware of their own illness, but also are more likely to feel hopeless, self-critical (Lysaker, Roe, &Yanos, 2007), and have lower self-esteem, even in the absence of current low mood (Cooke et al., 2007). Social factors also appear important in insight (Kaiser, Snyder, Corcoran, &Drake, 2006). Subjects who are dissatisfied with their support networks feel low in mood. The satisfied tend to either score well on measures of insight or very poorly.

The models of insight summarized above have been developed through research predominantly conducted on Western populations; the validity and generalizability of these models needs to be explored in non-Western populations. The meta-analysis by Mintz et al. (2003) and more recent studies have found that both positive and negative symptoms of schizophrenia are significantly though modestly correlated with poor insight, while depressed mood has a modest positive association with insight (Mutsatsa, Joyce, Hutton, &Barnes, 2006). Age, female gender, and lower educational status have also been found to be associated with reduced insight (Mintz et al., 2003; Ritsner &Blumenkrantz, 2007). This study aimed to explore whether these same factors that were associated with insight also pertain in a non-Western population in Ghana.

Method

Setting and sample

Two of the authors (JC and NP) undertook the research while working for the Ghanaian Health Service. The study was conducted at Pantang Psychiatric Hospital located in the outskirts of Accra, the capital of Ghana. Built during a period of optimism and in the spirit of pan-Africanism that came after independence from the British, the hospital was intended to serve the needs of several West African nations. However, this ambition was never fully realized and the hospital now mainly serves Greater Accra and the Eastern Region. The hospital consists of 10 wards, acute and chronic, each admitting 50 patients and an outpatient department caring for over 100 patients per day, including 10 to 15 new cases.

Patients with a DSM-IV diagnosis of schizophrenia made by a consultant psychiatrist on admission, who were resident in the hospital during a three-week period in June 2007 and who could converse freely in English with the British interviewers (JC and NP) were included in the study. The senior nurse on each ward was asked to identify potential participants who were then approached by one of the researchers. If the participant met the inclusion criteria he or she was invited to participate in the study. A total of 49 patients, 37 men, were recruited. One individual declined participation. The DSM-IV diagnosis was obtained for each subject through a combination of case-note review and clinical interview with either JC or NP, both senior trainees in psychiatry. The clinical interview, which took around 30 minutes, assessed current and past symptoms, medical and psychiatric co-morbidity, and level of function. The case-notes were used to obtain collateral history of recent and past mental health status and symptoms. Ethical approval was granted by the local ethics committee and written consent obtained.

Measures

Insight was measured using the Schedule for the Assessment of Insight – Expanded Version (SAI-E) (David, 1990; Kemp &David, 1996). The SAI-E was chosen because it reliably assesses three domains of current insight: treatment adherence (SAI-TA); illness recognition (SAI-IR); and symptom relabelling (SAI-SR). The SAI-E includes a separate subsection for the primary nurse to rate the patient’s adherence to treatment (SAI-C). SAI-C is not considered part of the total insight score. We particularly wanted to assess treatment adherence because it was the observation of both JC and NP that patients experiencing psychotic illness in Ghana with impaired insight nevertheless complied with medication, even when mental illness was explicitly rejected as the cause of distress. Psychopathology was rated with the Positive and Negative Syndrome Scale (PANSS; Kay, Fiszbeinm, &Opler, 1987), which is composed of a positive (PANSS-pos), negative (PANSS-neg), and general psychopathology subscale (PANSS-gen). Depression and anxiety were rated using the Hamilton Rating Scales for Depression (HAM-D; Hamilton, 1960) and Anxiety (HAM-A; Hamilton, 1959) respectively. Each participant was interviewed by either NP or JC except for 10 randomly selected participants who were interviewed serially by both to assess interrater reliability.

Data analysis

The intra-class correlation coefficient using a two-way random model with measures of consistency was computed to test interrater reliability. Bivariate relationships between each variable and total SAI-E score were examined. Data were tested for normality using the Anderson–Darling Normality Test. Spearman’s rank correlation was performed to look at the associations between numerical variables and the total SAI-E score. A Mann–Whitney test was used to determine any gender differences. Variables to be included in the regression analysis were selected on the basis of research results from previous in the West. Given the relatively small sample size, only the total insight score was used as a dependent variable and only four variables were used: HAM-D, PANSS-pos, PANSS-neg, and SAI-C. These were selected to assess the association with total SAI-E score. Nonparametric variables were log-transformed. This resulted in non-interpretable beta-weights, which are therefore not presented in our results section. All analyses were done using SPSS (SPSS, 2001) and Minitab (Minitab Inc., 2007), with a significance level of 5%.

Results

The sample consisted of 49 independent observations. There were 12 women (24.5%) and the average age was 37.35 (SD 11.73), ranging from 19 to 63 years. The median age at onset was 22.5 years (range 14 to 47). Fifteen of the participants completed junior high school, 13 finished high school, six finished college while 15 had had no formal education. Age and other demographic variables can be viewed in Table 1.

Table 1.

Demographic characteristics of sample

Variable	Value	Range
Age, years, mean (SD)	37.6 (11.7)	19–63
Age at onset, years, mean (SD)	25.0 (7.9)	14–47
Length of stay in months, n (%)	29 (60.42)	1–418
1–6 months	3 (6.25)
6–12 months	4 (8.33)
12–24 month	3 (6.25)
24–48 months	9 (18.75)
> 48 months
Medicated, n (%)	49 (100)

Descriptive statistics for each of the rating scale measures are provided in Table 2. The intraclass correlation coefficient as a measure of inter-rater agreement is presented for each scale in Table 2. Agreement is shown to be excellent for most scales and good for the PANSS-neg and SAI-SR subscales. Cronbach’s alpha was acceptable for all measures. Several measures consisted of single scores (SAI-TA, SAI-IR, SAI-SR, SAI-C) therefore Cronbach’s alpha was not calculated for these.

Table 2.

Descriptive statistics and interrater reliability scores for rating scales

Score	Mean	SD	Median	Range	Intraclass coefficient	Cronbach’s alpha
SAI total	8.32	6.22	6.0	2.0, 26.0	.94	0.86
SAI-TA	2.92	1.06	3.0	1.0,6.0	.92	n/a
SAI-IR	2.92	1.06	0.0	0.0,8.0	.93	n/a
SAI-SR	3.25	3.48	2.0	0.0,14.0	.77	n/a
SAI-C	5.37	0.78	5.0	4.0,7.0	.94	n/a
HAM-A	7.53	6.70	7.0	0.0,25.0	.90	0.86
HAM-D	11.22	7.31	9.0	2.0,33.0	.86	0.86
PANNS general	25.67	10.30	25.0	8.0,60.0	.95	0.79
PANNS positive	11.47	6.0	12.0	1.0,23.0	.92	0.71
PANNS negative	14.49	8.7	13.0	1.0,32.0	.71	0.9

Note. SAI: Schedule for the Assessment of Insight; -TA: treatment adherence; -IR: illness recognition; -SR: symptom relabeling; -C: primary nurse’s rate of the patient’s adherence to treatment; HAM-A: Hamilton Rating Scale for Anxiety; HAM-D: Hamilton Rating Scale for Depression; PANSS: Positive and Negative Syndrome Scale.

The data were analysed with multiple linear regression, using as predictors HAM-D, PANSS-pos, PANSS-neg, and SAI-C. The regression was a rather poor fit (R²adj = .34). HAM-D was positively associated with insight (p = .002), while PANSS-pos was negatively associated (p = .029). Neither PANSS-neg nor SAI-C contributed significantly to the model.

Discussion

This study is the first to explore the relationship between insight and psychopathology in a West African population with schizophrenia. Although all participants were inpatients the mean score on measures of psychopathology were quite low. Despite this, the mean SAI-E scores indicated a poor degree of insight, a reflection, perhaps, of a cultural tendency to attribute abnormal psychological experiences to supernatural causes. Alternatively, patients may actively deny mental illness because it is highly stigmatized in many African communities (Crabb et al., 2012)

Statistical analysis revealed that insight was positively associated with depressive symptoms and inversely associated with positive symptoms of psychosis. However, only a small modest portion of the variance in insight was explained by this model, consistent with studies in Western settings (Amador &Kronengold, 2004). In a similar study in Malawi, Crumlish et al. (2007) also found that positive symptoms of psychosis were inversely correlated with the SAI-E total score but they also found negative symptoms were inversely correlated with the SAI-TA subscale. This difference may reflect the different scales used to assess symptoms. Crumlish and colleagues used the Scale for the Assessment of Positive Symptoms and Scale for the Assessment of Negative Symptoms (SAPS and SANS, respectively), which correlate well with the PANSS, but differ from it in subtle ways (Norman, Malla, Cortese, &Diaz, 1996). For example, inattention is a negative symptom recorded in the SANS but is not designated as either positive or negative in the PANSS. The sample studied by Crumlish and colleagues was similar age in age and chronicity (9 years mean duration of illness against 12 years in ours), insight scores, and level of psychopathology. However, their sample consisted of outpatients and included many more women (45%). Our sample had been hospitalized for many years. The moderately high mean score on the PANSS negative in our inpatient sample could reflect secondary negative symptoms, which may be less associated with loss of insight.

There is growing consensus that insight is an independent psychopathological variable that is only moderately associated with positive symptoms in severe illness (Amador &Kronengold, 2004). The association between insight and depressive symptoms in psychosis may reflect the impact of awareness of illness on mood. Depressed mood may occur in response to cognitions of hopelessness and lowered self-esteem, as a result of a correct appraisal of the situation. It is also possible that those who identified themselves as ill were actually aware of the depression and anxiety rather than the psychosis per se.

This study has several limitations. The assessment of insight can be biased by the degree of psychopathology elicited (Amador et al., 1991). Ideally assessments of insight and psychopathology should be conducted independently by two different interviewers blind to each other’s findings. The most significant limitation is the generalizability of these results. Only English-speaking Ghanaians were recruited and they are more likely to be educated to high school or college level, originate from an urban area, and be familiar with Western biomedical views on illness. Additionally, we did not formally evaluate literacy in English but included only those able to converse freely. Therefore, English speakers experiencing severe thought disorder have been excluded, biasing the population towards the less ill. Also, the cases were identified by a senior nurse on the ward, this could have introduced a bias towards more compliant or co-operative patients, who might be less severely ill. Finally, the use of a structured clinical interview would have ensured greater reliability for the diagnosis of schizophrenia.

In conclusion, our study demonstrates that in a sample of Ghanaian inpatients with schizophrenia, insight is partially predicted by depressive symptomatology and lower severity of positive symptoms of psychosis, replicating findings in the West. Further research combining psychopathological and sociological methodologies is required to unravel why people with schizophrenia so often disregard their own illness.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Norman Poole, MBChB, MRCPsych, MSc, is a consultant Liaison Psychiatrist at the Royal London Hospital, Whitechapel, London. His interests include cognitive disorders, treatment methods in neuropsychiatry, insight in psychiatric and neurological disorders, and the role of culture in phenomenology. Address: Older Adults Liaison Psychiatry Service, 3rd Floor Out-Patients Department, Royal London Hospital, Whitechapel, London E1 1BB, UK, East London NHS Foundation Trust. Email: norman.poole@eastlondon.nhs.uk

Jim Crabb, MRCPsych, is a consultant in General Adult Psychiatry at Forth Valley Royal Hospital. His interests include post traumatic stress disorder, transcultural psychiatry, medical education, and the development of mental health services in low-income settings. He has international experience in Ghana and Malawi.

Akwasi Osei is the Chief Psychiatrist for Ghana.

Peter Hughes, FRCPsych, is Consultant Psychiatrist at Springfield University Hospital London; member of the International Advisory Committee of the Royal College of Psychiatrists. Acting Mental Health subdean. Trained in Ireland, Australia, and the UK. He has international experience in Africa and Haiti.

David Young is Senior Lecturer in the Department of Mathematics and Statistics at the University of Strathclyde and Senior Consultant Statistician for NHS Greater Glasgow and Clyde. Sixteen years experience as a statistician involved in medical research within the pharmaceutical industry, the National Health Service and with the Health Protection Agency.

References

Aleman

Agrawal

Morgan

K. D.

David

A. S.

(2006) Insight in psychosis and neuropsychological function: Meta-analysis. British Journal of Psychiatry 189: 204–212.

Amador

X. F.

Kronengold

(2004) Understanding and assessing insight. In: Amador

David

(eds). Insight and psychosis: Awareness of illness in schizophrenia and related disorders. Oxford, UK: Oxford University Press, pp. 3–30.

Amador

X. F.

Strauss

D. H.

Yale

S. A.

Gorman

J. M.

(1991) Awareness of illness in schizophrenia. Schizophrenia Bulletin 17: 113–132.

Cooke

M. A.

Peters

E. R.

Greenwood

K. E.

Fisher

P. L.

Kumari

Kuipers

(2007) Insight in psychosis: Influence of cognitive ability and self-esteem. British Journal of Psychiatry 191: 234–237.

Crabb

Stewart

R. C.

Kokota

Masson

Chabunya

Krishnadas

(2012) Attitudes towards mental illness in Malawi: A cross-sectional survey. BMC Public Health 12: 541–541.

Crumlish

Samalani

Sefasi

Kinsella

O’Callaghan

Chilale

(2007) Insight, psychopathology and global functioning in schizophrenia in urban Malawi. British Journal of Psychiatry 191: 262–263.

Cuesta

M. J.

Peralta

Zarzuela

Zandio

(2006) Insight dimensions and cognitive function in psychosis: A longitudinal study. BMC Psychiatry 6(1): 26–26.

David

A. S.

(1990) Insight and psychosis. British Journal of Psychiatry 156: 798–808.

Flashman

L. A.

Roth

R. M.

(2004) Neural correlates of unawareness of illness in psychosis. In: Amador

David

(eds). Insight and psychosis: Awareness of illness in schizophrenia and related disorders. Oxford, UK: Oxford University Press, pp. 157–176.

10.

Fulford

K. W. M.

(2004) Insight and delusion: From Jaspers to Kraepelin and back again via Austin. In: Amador

David

(eds). Insight and psychosis: Awareness of illness in schizophrenia and related disorders. Oxford, UK: Oxford University Press, pp. 51–78.

11.

Hamilton

(1959) The assessment of anxiety states by rating. British Journal of Medical Psychology 32: 50–55.

12.

Hamilton

(1960) A rating scale for depression. Journal of Neurology Neurosurgery and Psychiatry 23: 56–62.

13.

Kaiser

S. L.

Snyder

J. A.

Corcoran

Drake

R. J.

(2006) The relationships among insight, social support, and depression in psychosis. Journal of Nervous and Mental Disease 194(12): 905–908.

14.

Kay

S. R.

Fiszbeinm

Opler

L. A.

(1987) The Positive and Negative Syndrome Scale (PANSS) for schizophrenia. Schizophrenia Bulletin 13: 261–276.

15.

Kemp

David

(1996) Psychological predictors of insight and compliance in psychotic patients. British Journal of Psychiatry 169(4): 444–450.

16.

Lysaker

P. H.

Roe

Yanos

P. T.

(2007) Toward understanding the insight paradox: Internalized stigma moderates the association between insight and social functioning, hope, and self-esteem among people with schizophrenia spectrum disorders. Schizophrenia Bulletin 33(1): 192–199.

17.

Minitab Inc. (2007). Meet Minitab 15 (Version 16). Coventry, UK: Author.

18.

Mintz

A. R.

Dobson

K. S.

Romney

D. M.

(2003) Insight in schizophrenia: A meta-analysis. Schizophrenia Research 61: 75–88.

19.

Moore

Cassidy

Carr

O’Callaghan

(1999) Unawareness of illness and its relationship with depression and self-deception in schizophrenia. European Psychiatry 14(5): 264–269.

20.

Mutsatsa

S. H.

Joyce

E. M.

Hutton

S. B.

Barnes

T. R. E.

(2006) Relationship between insight, cognitive function, social function and symptomatology in schizophrenia: The West London first episode study. European Archives of Psychiatry and Clinical Neuroscience 256(6): 356–363.

21.

Norman

Malla

A. K.

Cortese

Diaz

(1996) A study of the interrelationship between and comparative interrater reliability of the SAPS, SANS and PANSS. Schizophrenia Research 19: 73–85.

22.

Ritsner

M. S.

Blumenkrantz

(2007) Predicting domain-specific insight of schizophrenia patients from symptomatology, multiple neurocognitive functions, and personality related traits. Psychiatry Research 149: 59–69.

23.

Saravanan

Jacob

K. S.

Prince

Bhugra

David

A. S.

(2004) Culture and insight revisited. British Journal of Psychiatry 184: 107–109.

24.

Schwartz

R. C.

(2000) Insight and suicidality in schizophrenia: A replication study. Journal of Nervous and Mental Disease 188(4): 235–237.

25.

SPSS Inc (2001) SPSS for Windows (Version 16) . Chicago, IL: Author.