A mixed-methods research approach was employed for this two-phased study. In the first phase, a formative qualitative methodology was used to translate and culturally adapt the ABC scale (WHO, 2016). In the second phase, a descriptive research design was used to determine the reliability of the ABC-S scale and describe the self-perceived balance confidence among first language Sepedi speakers in a rural community in Limpopo.

The second phase of this study was conducted in a rural area of the Greater Sekhukhune District Municipality of the Limpopo province, where Sepedi is the predominant language spoken, accounting for 83% of the population (Integrated Development Plan, 2021).

Using purposive sampling, a total of 32 participants were recruited for the second phase of the study. Participants were required to meet the following criteria: (1) be between the ages of 60 and 88 years (Kamanji, 2016; Marques et al., 2013) to allow for comparison with the isiZulu (Kamanji, 2016) and Brazilian (Marques et al., 2013) versions of the scale; (2) be first language Sepedi speakers; (3) reside in a rural community; (4) be ambulatory and capable of independently performing daily activities inside and outside of the home either with or without walking aids; and (5) be able to understand and follow multistep instructions. Individuals with medical conditions that increase the risk of falling (e.g. neurological disorders, musculoskeletal disorders, orthostatic hypotension, etc.) as well as those with unmanaged hearing and visual impairments were excluded from the study.

The mean age of participants was 70.06 years (±7.26; range: 60–88) and the majority were female (n = 84%). The gender distribution aligns with the distribution of the broader Sekhukhune District Municipality population where 65% of the population in this age group is female (Integrated Development Plan, 2021).

Data collection involved the use of a self-developed case history questionnaire (Appendix 1) and the ABC-S scale. The case history questionnaire consisted of 19 questions organised in four sections: demographic information, balance history, medical history and family history. The development of the case history questionnaire was guided by literature (Kamanji, 2016). The ABC-S scale has 16 items rated on an 11-point subscale ranging from 0% (no confidence) to 100% (completely confident) with increments of 10%. Scores below 50% indicate a low level of physical functioning. Scores above 80% represent a high level of physical functioning (Marques et al., 2013). Scores of <67% have shown to be predictive of future falls (Reelick et al., 2009).

Ethical clearance to conduct this study was obtained before data collection from the University of the Witwatersrand School of Human and Community Development Ethics Committee (No. STA_2021_5). The research adhered to the ethical principles outlined in the World Medical Association Declaration of Helsinki (World Medical Association [WMA], 2013). Permission to conduct the study was obtained from relevant authorities. Participants were fully informed of the nature of the study as well as their rights and assured of anonymity and confidentiality. Participants identified as having a fall risk were referred to their closest audiology department for assessment. Three options were available: the local community-based audiology clinic or two hospitals in the Greater Sekhukhune District Municipality. The two hospitals also offer physiotherapy services.

The translated ABC scale (Appendix 2) was culturally adapted by a bilingual expert panel to suit the needs of the rural South African context. Ten of the original 16 items were modified to ensure semantic, contextual and cultural appropriateness. These modifications are documented in Table 1.

The adjustments were necessary because of the distinct local reality and cultural differences of the South African participants, considering that the original scales development is based on the Canadian population (Powell & Myers, 1995).

For instance, item 3 originally ‘pick up a slipper from the front of the closet floor’ underwent modifications to ‘pick up a shoe from the area where one stores their belongings’ to better align with local terminology and cultural context. Specifically, the terms ‘slipper’ and ‘closet’ were not familiar to the South African participants. Marques et al. (2013) caution that using unfamiliar terminology may hinder comprehension. None of the participants reported any doubts as to the meaning during the administration of the ABC-S scale. The adaptations aimed to ensure comprehension without altering the activity’s intended meaning, as evidenced by comparable scores to the Brazilian-Portuguese and Canadian studies (Table 5).

Similarly, item 4 was also adapted by changing the phrase ‘reach for a small can off a shelf at eye level’ to ‘reach for a small tin that is on top of a shelf that you can see’ to align with the common expression in the Sepedi culture.

Items 8 and 9 were adapted to account for socioeconomic differences, replacing ‘walk outside the house to a car parked in the driveway’ with ‘walk outside to a vehicle in the yard’ emphasising ‘vehicle’ to accommodate varied transportation modes. Most of the residents in Limpopo use taxis (Stats SA, 2020b), as cars are owned by less than 15% of households in rural communities (Stats SA, 2018). Driveways are also uncommon in rural communities.

To align with the contextual realities of the Elias Motsoaledi Local Municipality (EMLM) area and the lifestyle of its residents, the term ‘shop(s)’ was added to items 10, 12 and 13. Most of the people who reside in the EMLM area and other rural areas in South Africa go to shops, which are more accessible to them compared to malls.

During the pilot study, participants expressed unfamiliarity with the term ‘ramp’ in item 11 (‘walk up or down a ramp’), leading to its modification ‘walk up or down a hill’. In item 15, the word ‘parcels’ was replaced with ‘luggage’ to better suit the context of the study population.

Item 16 (‘walk outside on icy sidewalks’) underwent the final modification to account for differences in climate, as snow is extremely rare in South Africa. The term ‘sidewalk’ was also changed. Item 16 was thus changed to ‘walk outside on a surface largely covered by dew’ to reflect context and align with the Sepedi expression.

These meticulous adaptations underscore the significance of tailoring assessment tools for cultural relevance, ensuring the reliability of the ABC-S scale in the local South African setting.

The test-retest reliability of the ABC-S scale is shown in Table 2. The ICC for the total score exhibited a high level of reliability, measuring at 0.85, with a 95% confidence interval. This is consistent with the results obtained for the isiZulu ABC scale, which demonstrated a score of 0.82 (Kamanji, 2016) and balance confidence in individuals with stroke at 0.85 (Botner et al., 2005). Other versions of the ABC scale have demonstrated excellent test-retest reliability, such as the Brazilian-Portuguese (Marques et al., 2013), Hindi (Moiz et al., 2017) and Chinese (Hsu & Miller, 2006) versions with scores of 0.94, 0.97 and 0.99, respectively.

Table 3 shows the inter-rater reliability of the ABC-S scale. The ICC for the total score exhibits a high level of reliability, measuring at 0.81 with a confidence interval of 95%. This is consistent with other versions of the ABC scale such as the Brazilian-Portuguese translation, with a score of 0.80 (Marques et al., 2013). These findings suggest that the scale produces consistent results over time and between different raters and can be used in clinical and research settings.

Table 4 shows the internal consistency of the ABC-S scale, which was determined using the Cronbach’s alpha coefficient and item-total correlations. The scale showed a high internal consistency of α = 0.92. This was reinforced by moderate-to-strong item-total correlations for 13 of the items, ranging from 0.41 to 0.85 and exceeding the cutoff value of 0.70 (Ferketich, 1991).

An α ≥ 0.9 is considered excellent (Bland & Altman, 1997). This indicates that the items of the ABC-S scale measure a single, unidimensional construct (Tavakol & Dennick, 2011) and is thus appropriate for use in clinical and research settings. The internal consistency of the ABC-S scale is consistent with findings reported in previous studies (see Guan et al., 2012; Kamanji, 2016; Mak et al., 2007; Powell & Myers, 1995; Salbach et al., 2006; Schott, 2008; Van Heuvelen et al., 2005). Subsequent analysis indicated that the majority of the items are suitable for retention, leading to a decrease in the alpha coefficient if they are excluded. Nevertheless, there were four exceptions to this trend, specifically items 1, 3, 10 and 16, for which the removal would result in an increased alpha coefficient of 0.92. Therefore, it is advisable to consider removing these items. This aligns with the observed item-total correlations for items 1, 3 and 16, which demonstrated a weak correlation. According to Powell and Myers (1995), the internal consistency of the scale does not exhibit a significant improvement when a few items are removed. Thus, the deletion of item 16, which pertains to ‘walking outside on icy sidewalks’, should not pose any challenges for administering the scale in warmer climates, such as South Africa (Myers et al., 1998).

The mean confidence level of first language Sepedi speakers in a rural community in the Limpopo province was 81.3 (±10.3; range: 58.1–95.9) (Table 5). This is suggestive of a high level of self-perceived confidence in their ability to maintain balance during various daily activities (Myers et al., 1998).

Subjective fear of falling was assessed through the unvalidated single-item question: ‘Do you have a fear of falling?’ asked during case history. Of the 12 participants who indicated fear of falling, 6 (50%) obtained a total score of <67, with scores ranging from 23.13 to 65. Only two participants (6.25%) who did not report a fear of falling obtained scores below 67. The single-item question was significantly correlated to the ABC scores in the sample (r = −0.567).

The two items with the lowest degrees of confidence in this study were identical to the findings in the Brazilian (Marques et al., 2013), isiZulu (Kamanji, 2016) and Canadian populations (Powell & Myers, 1995). These items generally involve activities that pose greater challenges or perceived risks for the participants. These include item 6 (‘stand on a chair and reach for something’) and item 15 (‘step onto or off an escalator while holding onto parcels such that you cannot hold onto the railing’). These items had values ranging from 58.1 to 64.4, consistent with findings from other studies (Botner et al., 2005; Mak et al., 2007; Salbach et al., 2006; Schepens et al., 2010). Item 16 in this study, however, had the highest score (95.9), in contrast to the Canadian and Brazilian population at 20.7 and 61.2, respectively. This difference may be explained by the perceived level of difficulty walking on a surface covered by dew (this study) in comparison to walking on ice.

A score lower than 67% has been identified as an indicator of fall risk (Lockhart et al., 2019; Reelick et al., 2009). When comparing the mean scores across different cultural and linguistic contexts, it is evident that the Sepedi (81.3) and Brazilian-Portuguese (81.7) studies yielded considerably higher mean scores compared to the isiZulu (51) and Canadian (59.6) studies, as shown in Table 5. In both the Sepedi and Brazilian-Portuguese studies, only three items obtained scores lower than 67%, indicating a relatively low fall risk among the participants. This finding is in stark contrast to the isiZulu (Kamanji, 2016) and Canadian (Powell & Myers, 1995) versions of the ABC scale, where 13 and 12 items, respectively, received scores below the fall risk threshold. These results suggest that elderly individuals in the Sepedi and Brazilian-Portuguese-speaking populations demonstrated a higher level of self-perceived balance confidence when compared to their Canadian counterparts, and a lower likelihood of falling when compared to their isiZulu and Canadian counterparts.

The disparity on fall risk among these populations may be attributed to differences in the selection criteria employed in the original and isiZulu studies. In a study conducted by Lajoie and Gallagher (2004), individuals who had experienced a stroke or had preexisting decreased postural control were included as participants. This selection criterion likely contributed to lower balance confidence scores in the Canadian study population. Similarly, the isiZulu study by Kamanji (2016) included participants with physical impairments. These impairments may have negatively impacted balance confidence levels, resulting in the comparatively lower mean scores observed in the isiZulu study. In contrast, the current and Brazilian-Portuguese studies (Marques et al., 2013) excluded individuals with medical conditions that increase the risk of falling and those with unmanaged hearing and visual impairments. Consequently, the participants in these studies may have had a higher baseline level of balance confidence, leading to the elevated mean scores observed.