Demonstrable significant increases in the prevalence of mental health problems throughout the lifespan across the globe mean the probability of most individuals developing a mental health problem is high [3]. Increasing levels of mental health literacy contributes to the promotion of mental health and may play an important role in early identification and intervention when a psychological/mental health problem develop. Unfortunately, there is still very little literature about MHL from developing countries like Malawi [12]. There is even less research around how to appropriately assess MHL in developing country contexts, with most studies using tools that have been developed using cohorts from Western countries or regions [10]. Beyond translation into local language, we also need to deeply consider and prioritise psychometric analyses, validation and relevance of such tools to local cultures, as a pre-requisite to ensure we are intend measuring what is intended [11].
This current study aimed to translate the English version of a previously developed MHL questionnaire [11, 15] into Chichewa for use in the Malawi young adult population and to evaluate psychometric properties of this Chichewa version.
For the CFA outcomes, there was good evidence for Factor 1 (Knowledge of mental health problems) and Factor 3 (First aid skills and help seeking behaviour) of the Chichewa version when comparing with the outcomes from the Portuguese paper [11] but low evidence for factor 2 (Erroneous beliefs or stereotypes). Overall, three factors out of the four-factor solution were a reasonably good fit to the data. Three out of the eight items in factor 2 did not fit into the factor structure, suggesting some stereotypes and erroneous beliefs are not the same as the English version, specifically those regarding judgement or criticism of significant others with a mental health problem (MHLq11), the belief that mental disorders don’t affect people’s feelings (MHLq13r) and one’s belief in their ability to assist someone close to them if they had a mental disorder (MHLq21r). Moreover, one item in Factor 4 regarding “Physical exercise contributes to good mental health” did not fit well with factor 4 in the paper with the Portuguese cohort. In fact, this item (MHLq1) was almost negatively correlated with the factor inferring that most survey respondents believe physical exercise is not a self-help strategy. Again, only one item did not fit well with correlating factors from the Portuguese cohort for both factor 1 and factor 3, namely, MHLq22 “The symptom’s length is one of the important criteria for the diagnosis of a mental disorder”, and MHLq17 “If someone close to me had a mental disorder, I would encourage her/him to see a psychiatrist.”
We suggest reasons for the discrepancies observed in the CFA outcomes. These reasons mainly reflect potential lack of cultural relevance and environmental context of some of the MHLq items. Item 1 of factor 4 regarding physical exercise is a good example. There is strong evidence for physical exercise as a good health behaviour, especially its benefits for preventing and managing non-communicable diseases like mental illness globally. Despite this existing evidence, physical activity is a largely neglected modality in most mental health care policies and systems in Africa [16]. This lack of awareness seems to have been reflected in the responses of our Malawian rural community sample. Linking physical exercise specifically to mental health, especially as a self-help strategy is typically a Western or urban concept where populations are likely challenged by the issue of sedentary lifestyles. For someone based in rural Malawi where everyday tasks are typically physically strenuous i.e., collecting water from distant rivers, farming with manual tools, physical exercise may not be perceived or experienced in the same way to those typically confined to office desk environments or easy access to transport. The poor fit for item MHLq22 of factor 1 is expected considering that many people in Malawi do not know much or talk about mental health, especially types of mental illness and details on symptoms presentation [17]. The low loading on factor 3’s item MHLq17 likely illustrates the very limited mental health workforce in the country. With only 4 psychiatrists in the whole country, most Malawians rarely hear of or have access to a psychiatrist [18]. As a result, a significant proportion seek help from traditional healers, spiritual advisers or religious leaders instead [18–20].
For factor 2, the three items low loading may be highlighting complexity around country specific and cultural differences in mental health beliefs and stereotypes [21]. Specifically, that there are differences in sources of prejudice and beliefs around aetiology or causes of mental illness between young adults in rural Malawi [20] and Portugal [22, 23] that result in responses to specific items in Factor 2 not correlating between the two MHLq versions. Most research evidence on stigma related to mental illness is from Western countries and there is limited evidence from an African context [17, 24, 25]. More research focusing on beliefs, stereotypes and stigma related to mental illness in Africa should be conducted to better inform content and relevance of existing and future MHL questionnaires. This also highlights a need for validated tools that measure mental health stigma in Malawi [26] and Africa more broadly.
CFA outcomes complement findings from our work exploring the cultural applicability of this MHLq to Malawi which will be published in a separate paper. We conducted a mixed methods exploration of the MHLq responses, comparing the rural survey responses to focus group discussions around how accurate the Chichewa translation was and how culturally relevant the MHLq is for Malawian context. Qualitative interrogation of the quantitative data highlighted a discrepancy in the factors, especially around stereotypes and beliefs.
There were high ceiling effects indicating that in most of the data people reported in the highest category. This may have affected the variability in the data and may explain the scattered results seen in the exploratory factor analysis. Indeed, focus group participants from our exploratory work on the MHLq’s cultural relevance also questioned the high category scoring, particularly from the ‘knowledge’ subscale items. They felt the items do not sufficiently explore whether survey respondents in rural Malawi understood terms like psychiatrist, psychologist, schizophrenia. They also felt that there may have been a bit of guess work by respondents when completing some of the survey items. They consequently suggested adapting the MHLq by adding ‘I don’t know’ to the Likert scale to better or more explicitly capture respondents’ understanding of questionnaire items.
Overall score for internal consistency was good. However, extremes were noted in the Cronbach alpha scores for the subscales with factor 1 (α = 0.61) and factor 3 (α = 0.76) scores being acceptable whilst factors 2 (α = 0.20) and 4 (α = 0.15) were unacceptable. The very low and unacceptable alpha score for the ‘Self-help strategies’ subscale could be explained by the scale’s low number of items coupled with the small sample size [27]. The internal consistency increased when MHLq items that were not aligning with the factor structure were taken out. In the case of factor 2, the low alpha may reflect the fact that stereotypes and beliefs differ across countries and cultures [21]. This is especially the case in the context of mental health, where public stigma is not only dependent on the overall understanding of mental health and illness, but also based on specific types of mental illness [28]. Subsequent research should be conducted using larger samples in both rural and urban community settings to reassess internal consistency scores of these subscales. Revision of factors within the Chichewa MHLq version should be considered if the subscales’ alpha scores remain low. In terms of convergent validity, the data was nearly there but the sample size was a bit too small to test for education.
Strengths of the study
This is the first published study in Malawi that has investigated translating and validating a tool that can assess mental health literacy in rural community settings. More generally it is the first study assessing MHL among young adults in these settings. Clear and well-established guidelines for translating, adapting and validating instruments for cross cultural use brought rigor to the translation process [13]. Input from our local stakeholders improved semantics within translated items.
Study limitations
Participating sample in the evaluation of this Chichewa MHLq were young people from rural settings, which may limit the generalisability of the questionnaire’s properties to an older population or young people from urban settings in Malawi. We targeted rural settings as people living here are predominantly literate in Chichewa compared to English language. Secondly, we only carried out the questionnaire once so we cannot carry out test/re-test reliability statistics. We plan to use Chichewa responses from the larger dataset of the national survey and compare with this cohort to explore reliability of the questionnaire.