JOURNAL OF THE LOUISIANA STATE MEDICAL SOCIETY
becauseof the lone factor, the solution rotationwas unnecessary. This factor, which was identified assumptively as resilience , held an eigenvalue of 2.8 and explained almost 70% of the variance in BRCS responses. Each of the items loaded strongly with this factor. Table 2 details these findings. BRCS reliability. Reliability analysis on the BRCS revealed a Cronbach’s alpha reliability coefficient of 0.854. Guttman’s coefficient was similar at 0.884. This coefficient, λ-4, was the highest of the lower bound lambdas and thus noted as Guttman’s reliability estimate. Inter-item correlations among all pairings of the four BRCS items were moderate-to-strong, 16 as shown in Table 3. BRCS validity. Convergent validity was examined via zero- order correlations and their specific directions between the BRCS and measures of theoretically linked constructs of coping and back pain/functioning. The BRCS significantly, positively correlated with proactive coping and reflective coping; the BRCS negatively correlated, though not significantly, with avoidance coping and back functioning. These associations showed appropriate directions, per previously mentioned literature. 13,14 Table 4 displays these correlation coefficients relevant to BRCS convergent validity.
Scale. Previous literature has shown resilience to be negatively linked with physical pain. 14 Thus, the BRCS was expected to indicate further convergent validity via a negative correlation with the ODI. Significance ( p ) was observed at the standard 0.05 threshold. Missing Data . Missing values (cases x variables) were less than 1% for the original dataset. Since data absence per variable was less than 20%, single imputation was computed in SPSS. 15 Following calculation of root mean square error, we observed nominal differences in variance between original and imputed datasets. Accordingly, the analyses were conducted using the imputed dataset.
RESULTS
Descriptive Statistics
Sample characteristics . The mean age of sample participants was almost 65 years. The majority of the sample self-identified as male (64%), Caucasian/White (85%) and married (77%). The most common method of payment for scheduled PLIF was private pay (60%). In terms of common health risks to surgery, the majority of patients reported no histories of diabetes (79%), heart disease (93%), or smoking (71%). Table 1 details these sample characteristics. Standardized measures . The aggregate global scores on each of the standardized measures are as follows, beginning with the measure under psychometric focus: • BRCS (resilience): M = 12.6 (SD = 2.74, range: 4-16); • Proactive Coping Scale: M = 44.2 (SD = 6.33, range: 26-55); • Reflective Coping Scale: M = 33.9 (SD = 6.37, range: 11-44); • Avoidance Coping Scale: M = 6.8 (SD = 2.24, range: 3-12); and • ODI (back functioning): M = 26.0 (SD = 6.32, range: 10-43).
BRCS Reliability: Inter-Item Correlations
1
2
Item a. Creativity in addressing adversity b. Locus of control in reacting to adversity c. Personal growth post-adversity d. Replacement of loss
4
3
0.472* a 0.653* b 0.494* a --
0.658* b 0.692* b --
0.596* b --
--
* p < .01 Strength of association (Evans, 1996) a=moderate; b=strong
BRCS Factor Analysis Findings
Table 3: BRCS: Brief Resilient Coping Scale
Factor Loading
BRCS Item a. Creativity in addressing adversity b. Locus of control in reacting to adversity c. Personal growth post-adversity d. Replacement of loss Eigenvalue Variance explained
2.788 0.816* 0.887* 0.858* 0.776* 69.712%
DISCUSSION
Patients in this study reported a high level of psychological resilience, especially when considering their LBP prior to surgery. Regarding scale analysis, factor and reliability observations evidenced support for an internally consistent, single-factor scale for measuring resilience in pre-operative PLIF patients. As anticipated, the BRCS was shown to measure a single factor— resilience—with this factor accounting for a large proportion of the variance in BRCS scores. Relatively strong reliability coefficients across all analyses provided further support for the construct validity of the BRCS. In contrast, convergent validity findings were more mixed. As expected, the BRCS positively correlated with two adaptive coping strategies, proactive and reflective coping. Unexpectedly,
Table 2: BRCS: Brief Resilience Coping Scale * Loading on the factor valued as excellent (Tabachnick & Fidell, 2007) BRCS: Brief Resilience Coping Scale
Psychometric Findings
BRCS factor structure . Principal components analysis extracted a single underlying factor with an eigenvalue greater than 1.0;
4 J La State Med Soc VOL 170 JANUARY/FEBRUARY 2018
Made with FlippingBook Digital Publishing Software