Cross cultural adaptation of the MFES into Igbo.
Translation of most items on the E-MFES was uncontroversial. All the 14 items on the original MFES were retained but item 8 was a bit modified. The terms “cabinets” and “closets” in item 8 had no Igbo equivalent terms. The expert review committee agreed that the Igbo equivalent of “cupboard” could give the most similar meaning. Thirty older adults (70% female; mean age 72.7±7.63 years) participated in the pretesting and cognitive debriefing interview of this study. The participants had at least secondary education. All the participants indicated clarity of language and ease of understanding of all the items during the cognitive debriefing interview. The participant also agreed that the term used to replace “cabinets” and “closets” was suitable and Igbo friendly description for the terms. At the second expert panel meeting, the consensus was that the cross-cultural adaptation of the MFES into Igbo was good and that participants’ responses at pretest justified the term adapted for “cabinets” and “closets”. No further modification was made.
Validation of the I-MFES
A total of 109 patients (43.1% males) with mean age of 74.45±8.775 participated in the psychometric testing of the I-MFES. 14.7% of the participant were widowed, whereas 67.9% attained at least secondary level of education. Majority (82.6%) of the participants were still occupationally active with farming (37.6%) being the most predominant occupation (table 1).
Validities of the Igbo version of the I-MFES
All the MFES scores (both item and total) measured on both the E-MFES and the I-MFES fell within the range of being “fearful”. The participants scored highest and lowest in items 6 (answering the door or telephone) and 12 (crossing the road) respectively on both the E-MFES and the I-MFES. There was no significant difference between any of the corresponding item and total scores on the E-MFES and the I-MFES (p>0.05) indicating that the scores on the two instruments are statistically similar (table 2).
The convergent validity coefficient (rho=-0.93) of the I-MFES estimated by correlating the total scores on the I-MFES and the FES-I was excellent, indicating that the two instruments measure the same construct (fear of falling). The correlation between the total scores on the I-MFES and the FES-I is pictorially represented on a scatter plot (Figure 1). The concurrent (known-group) validity coefficients (r=0.97-1.00) of the I-MFES obtained by correlating corresponding total and item scores on the I-MFES and the E-MFES were all excellent indicating that the two versions of the MFES are equivalent (table 3). The correlation between the total scores on the I-MFES and the E-MFES is pictorially represented on a scatter plot (Figure 2). The heteroscedasticity of the participants’ scores on the I-MFES and the E-MFES is displayed on Figure 3.
Reliability of the I-MFES
Internal consistency coefficient (alpha=0.97) of the items on the I-MFES estimated by means of Cronbach's alpha using the split half method was excellent. The standard error of mean (SEM) of the item and total scores on the I-MFES ranged from 0.29 to 0.32. The minimal detectable difference (MDD) of the item and total scores on the I-MFES ranged from 0.79 to 0.89. Items 3 and 7 had the highest and lowest SEM and MDD scores respectively (table 4).
Structural validity of the Igbo version of the I-MFES
The data was fit for factorial analysis. Inspection of the correlation matrix revealed that all the coefficients were above 0.3, which suggests reasonable factorability. The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.94 exceeding the recommended value of 0.6, and Barlett’s test of sphericity reached statistical significance (X2(91)=3405.93; p=0.0001 supporting the factorability of the correlation matrix. Finally, the communalities were all above 0.3 (table 5), thus confirming that each item shared some common variance with other items. Given these overall indicators, factor analysis was conducted with all the 14 items of the I-MFES. The PCA revealed the presence of only one component with the initial eigen value exceeding 1, accounting for 88.73% of the variances (table 6). Scree plot also confirmed the presence of one component by showing a clear reflection after the first factor (figure 4).
Table 1: Socio-demographic profiles of the participants
Variable Class Frequency Percentage (%)
Gender Male 47 43.1
Female 62 56.9
Marital status Single 1 0.9
Married 92 84.4
Widowed 16 14.7
Highest Educational Primary 35 32.1
Attainment Secondary 62 56.9
Tertiary 12 11.0
Occupational Unemployed 6 5.5
Status Retiree 13 11.9
Farming 41 37.6
Trading 26 23.9
Civil/Public service 18 16.5
Artisan 5 4.6
Table 2: Mann-Whitney U test comparing fear of falling values of the participants on both the English and Igbo versions of the Modified Falls Efficacy Scale
Scores Mean±SD Mean rank U P
Igbo English Igbo English
Item 1 7.76±3.17 7.76±3.17 109.50 109.50 5940.50 1.00
Item 2 7.42±3.10 7.46±3.09 109.20 109.80 5908.00 0.94
Item 3 6.71±3.36 7.20±6.03 108.80 110.20 5864.50 0.87
Item 4 7.50±3.21 7.50±3.21 109.61 109.39 5929.00 0.98
Item 5 7.43±3.17 7.45±3.18 109.09 109.91 5896.00 0.92
Item 6 7.99±3.02 7.99±3.02 109.50 109.50 5940.50 1.00
Item 7 7.98±2.98 7.99±2.96 109.48 10952 5938.50 0.99
Item 8 6.83±3.32 6.83±3.32 109.50 109.50 5940.50 1.00
Item 9 7.22±3.28 7.23±3.29 109.22 109.78 5910.00 0.95
Item 10 7.33±3.16 7.33±3.13 109.72 109.28 5916.00 0.96
Item 11 5.75±3.34 5.81±3.30 109.01 109.99 5887.00 0.91
Item 12 5.49±3.32 5.50±3.28 109.53 109.47 5937.50 0.99
Item 13 6.83±3.26 6.83±3.26 109.50 109.50 5940.50 1.00
Item 14 6.70±3.26 6.68±3.30 109.60 109.40 5930.00 0.98
Total score 7.07±3.02 7.11±109 109.04 109.96 5890.50 0.91
Table 3: Spearman rank order correlation showing the relationships between equivalent scores of the participants on the English and Igbo versions of the Modified Falls Efficacy Scale
Item 1 1.00 <0.01*
Item 2 1.00 <0.01*
Item 3 1.00 <0.01*
Item 4 1.00 <0.01*
Item 5 0.99 <0.01*
Item 6 1.00 <0.01*
Item 7 1.00 <0.01*
Item 8 1.00 <0.01*
Item 9 1.00 <0.01*
Item 10 1.00 <0.01*
Item 11 0.99 <0.01*
Item 12 0.97 <0.01*
Item 13 1.00 <0.01*
Item 14 1.00 <0.01*
KEY:
*= Significant at p<0.05
Table 4: Standard error of mean and minimal detectable difference of the item and domain scores on the Igbo version of the Modified Fall Efficacy Scale.
Item 1 0.30 0.84
Item 2 0.30 0.82
Item 3 0.32 0.89
Item 4 0.31 0.85
Item 5 0.30 0.84
Item 6 0.29 0.80
Item 7 0.29 0.79
Item 8 0.32 0.88
Item 9 0.32 0.87
Item 10 0.30 0.84
Item 11 0.32 0.87
Item 12 0.32 0.88
Item 13 0.31 0.86
Item 14 0.31 0.87
KEY:
I-MFES= Igbo version of the Modified Fall Efficacy Scale
SEM= Standard error of mean
MDD= Minimal detectable difference
Table 5: Communalities of the items on the I-MFES
I_MFES1 1.00 0.91
I_MFES2 1.00 0.88
I_MFES3 1.00 0.86
I_MFES4 1.00 0.92
I_MFES5 1.00 0.92
I_MFES6 1.00 0.90
I_MFES7 1.00 0.88
I_MFES8 1.00 0.89
I_MFES9 1.00 0.90
I_MFES10 1.00 0.91
I_MFES11 1.00 0.83
I_MFES12 1.00 0.79
I_MFES13 1.00 0.93
I_MFES14 1.00 0.91
______________________________________________________________________________
Key:
Extraction Method: Principal Component Analysis.
I-MFES: Igbo version of the MFES.
Table 6: Factor analysis and Monte Carlo principal component analysis for parallel analysis of Igbo Version of the MFES.
Factor Random Decision %variance %cumulative
Eigen values
Factor1 12.422 Accept 88.73 88.73
Factor 2 0.620 Reject 4.43 93.16
Factor 3 0.26 Reject 1.82 94.98
Factor 4 0.154 Reject 1.10 96.08
Factor 5 0.140 Reject 1.00 97.08
Factor 6 0.104 Reject 0.74 97.82
Factor 7 0.071 Reject 0.51 98.32
Factor 8 0.061 Reject 0.44 98.76
Factor 9 0.051 Reject 0.34 99.12
Factor 10 0.043 Reject 0.31 99.43
Factor 11 0.031 Reject 0.22 99.65
Factor 12 0.026 Reject 0.19 99.84
Factor 13 0.015 Reject 0.11 99.95
F Factor14 0.007 Reject 0.05 100.00