- Email: [email protected]
Utility of the Oswestry Disability Index for studies of back pain related disability in nurses: Evaluation of psychometric and measurement properties
International Journal of Nursing Studies 47 (2010) 604–607
Contents lists available at ScienceDirect
International Journal of Nursing Studies journal homepage: www.elsevier.com/ijns
Utility of the Oswestry Disability Index for studies of back pain related disability in nurses: Evaluation of psychometric and measurement properties Anna P. Dawson a,*, Emily J. Steele b, Paul W. Hodges a, Simon Stewart c a b c
Centre for Clinical Research Excellence in Spinal Pain Injury and Health, The University of Queensland, Brisbane, Australia Discipline of Public Health, University of Adelaide, Adelaide, Australia Baker IDI Heart and Diabetes Research Institute, Melbourne, Australia
A R T I C L E I N F O
A B S T R A C T
Article history: Received 2 January 2009 Received in revised form 25 July 2009 Accepted 23 October 2009
Background: Disability due to back pain in nurses results in reduced productivity, work absenteeism and attrition from the nursing workforce internationally. Consistent use of outcome measures is needed in intervention studies to enable meta-analyses that determine efficacy of back pain preventive programs. Objective: This study investigated the psychometric and measurement properties of the Oswestry Disability Index (ODI) in nursing students to determine its suitability for assessing back pain related disability in intervention studies. Methods: Bachelor of Nursing students were recruited. Test–retest reliability and the ability of the ODI to discriminate between individuals with serious and non-serious back pain were investigated. The measurement error of the ODI was examined with the minimal detectable change at the 90% confidence level (MDC90). Results: Student nurses (n = 214) had a low mean ODI score of 8.8 7.4%. Participants with serious back pain recorded higher scores than the rest of the cohort (p < 0.05). Test–retest reliability examined in 33 individuals was ICC = 0.88 (95%CI 0.77–0.94). The MDC90 = 6%, and 36% of nursing students scored below the MDC90 indicating the tool had limited ability to detect longitudinal change in disability in this population. Conclusion: Data from this and previous studies demonstrate that the measurement properties of the ODI are inappropriate for studying back pain related disability in nurses. The ODI is not recommended for back pain intervention studies in the nursing population and an alternative tool that is sensitive to lower levels of disability must be determined. ß 2009 Elsevier Ltd. All rights reserved.
Keywords: Back pain Disability evaluation Questionnaire Nursing Reliability (epidemiology) Minimal detectable change
What is already known about the topic? Consistent use of outcome measures is needed in studies of back pain prevention in nurses in order to enable meta-analyses of intervention efficacy.
The Oswestry Disability Index (ODI) measures disability due to back pain and has been used in studies with nurse cohorts. No studies have investigated the psychometric and measurement properties of the ODI in a nursing sample. What this paper adds
* Corresponding author at: Centre for Clinical Research Excellence in Spinal Pain Injury and Health, School of Health and Rehabilitation Sciences, The University of Queensland, St. Lucia, QLD 4072, Brisbane, Australia. Tel.: +61 7 3346 7467; fax: +61 7 3365 2775. E-mail address: [email protected] (A.P. Dawson). 0020-7489/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijnurstu.2009.10.013
Nursing students have low mean ODI scores. Data from this and other studies demonstrate that meaningful reduction in disability beyond measurement
A.P. Dawson et al. / International Journal of Nursing Studies 47 (2010) 604–607
error cannot be detected in a considerable proportion of nurses and nursing students. The ODI is not appropriate for studies of back pain in the nursing population. 1. Introduction Back pain (BP) is frequently reported by nurses and nursing students. Disability due to BP (BP-D) has a major impact on the productivity of the nursing workforce due to restricted duties, work absenteeism and attrition from the profession. The nursing profession is facing a critical shortage of workers, hence spinal pain and resultant disability must be prevented. Standardised use of measurement tools is needed in the international research community to enable data pooling in meta-analyses that determine intervention efficacy. A review of interventions to prevent BP and back injury in nurses identified inconsistent application of outcome instruments (Dawson et al., 2007). The Oswestry Disability Index (ODI) (Roland and Fairbank, 2000) measures the degree to which back or leg pain impacts functional activities, with a score range of 0–100%. It has been used to assess BP-D in a descriptive study (Carta et al., 2007) and intervention trials with nurse cohorts (Cooper et al., 1996, 1998; Yassi et al., 2001). The psychometric and measurement properties of the ODI have not been examined in nurses and it is uncertain whether the instrument is suitable for assessing BP-D in this population. The aims of this study were to determine in a nursing student sample: first, whether the tool is able to discriminate different classifications of BP; second, test– retest reliability; and third, whether the tool is able to detect longitudinal changes in BP-D. 2. Methods 2.1. Sample and data collection Bachelor of Nursing students in all three years of training who attended lectures on campus at an Australian university were provided with information and invited to participate during class time. A compulsory Nursing course from each year of training was attended for recruitment. Of 460 student nurses who attended university during recruitment, 373 volunteered (81% participation). All volunteers completed a demographic questionnaire and the extended Nordic Musculoskeletal Questionnaire (NMQ-E) (Dawson et al., 2009), and individuals reporting annual BP (upper and/or lower back region) completed the ODI version 2.0 (Roland and Fairbank, 2000). We also invited individuals to participate in a test–retest reliability study. In this analysis, we examine ODI data provided by n = 214 participants who reported annual BP (Cohort 1) and a subset of these volunteers (n = 33) who completed the ODI on two occasions (Cohort 2). 2.2. Data analysis Annual BP severity was classified into serious (that which required sick leave and/or treatment as recorded by
605
the NMQ-E), and non-serious (that which occurred without sick leave or treatment). The ability of the ODI to detect cross-sectional differences between participants with serious and non-serious BP (discriminative function) was assessed with Mann–Whitney U-tests. The test–retest reliability of the ODI was assessed using the proportion of observed agreement (po), and the intraclass correlation coefficient (ICC[2,1]). ODI ability to detect longitudinal within-subject changes in disability (evaluative function) was assessed with the minimal detectable change at the 90% confidence level (MDC90). This is the amount considered with 90% confidence to be a real change exceeding that due to measurement error. First the standard error of measurement (SEM) was determined using the standard deviation (s) from Cohort 1: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SEM ¼ s ð1 ICCÞ pffiffiffi. The MDC90 was then calculated: MDC90 ¼ 1:64 2 SEM, and the proportion of participants who scored
3. Results 3.1. Sample The personal characteristics of participants are presented in Table 1. 3.2. Discriminative function of ODI in nurses in training Mean (range) ODI scores for Cohort 1 include 8.8 7.4 (0–38)% for the entire sample, 7.3 6.2 (0–24)% for nonserious BP (n = 80) and 9.7 7.9 (0–38)% for serious BP (n = 134). Zero scores were recorded for 13% of the sample. Students with serious BP had higher scores than those with non-serious symptoms (p = 0.031). 3.3. Test–retest reliability Cohort 2 administered the ODI mean (range) 1.15 0.36 (1–2) days apart. ICC estimates for this cohort are presented in Table 2 and suggest the ODI total score was more reliable than individual questions.
Table 1 Sample characteristics.
Sample size Age (years) (mean SD) Weight (kg) (mean SD) Height (cm) (mean SD) Female (%) Annual upper back pain (%) Annual lower back pain (%) a
Mean of two testing occasions.
Cohort 1
Cohort 2
214 25.9 9.0 68.1 14.6 167.1 8.6 88.3 49.5 92.1
33 23.2 6.5 68.0 12.2 166.6 8.7 90.9 52.3a 94.0a
A.P. Dawson et al. / International Journal of Nursing Studies 47 (2010) 604–607
606
Table 2 Test–retest reliability of individual sections and total score of the Oswestry Disability Index in Bachelor of Nursing students (n = 33). Section
Pain/activity
ICC[2,1] (95%CI)
po
1 2 3 4 5 6 7 8 9 10
Pain intensity Personal care Lifting Walking Sitting Standing Sleep Sex Social life Travel
0.65 (0.40–0.81)
0.67 1.00 0.74 0.94 0.84 0.72 0.91 0.94 0.85 0.82
Total ODI score
a
0.74 0.78 0.71 0.59 0.82 0.25 0.52 0.51
(0.53–0.87) (0.60–0.88) (0.48–0.85) (0.31–0.78) (0.67–0.91) (0.11–0.55) (0.22–0.73) (0.21–0.73)
0.88 (0.77–0.94)
NA
ICC[2,1] = intraclass correlation coefficient (two-way model, single measures); 95%CI = 95% confidence interval; po = proportion of observed agreement; NA = not applicable. a ICC could not be calculated as all respondents scored zero on both testing occasions.
Table 3 Evaluative function of the Oswestry Disability Index in Bachelor of Nursing students. Back pain prevalence
SEM (%)
MDC90 (%)
Proportion with ODI score < MDC90 (%)
BP previous year (n = 214) BP on day of study (n = 82)
2.56 2.77
5.94 6.43
36.4 25.6
SEM = standard error of measurement; MDC90 = minimal detectable change at the 90% confidence level.
3.4. Evaluative function Table 3 presents the SEM and MDC90 for the ODI in student nurses, and the proportion of students who scored
calculated from BP patient data that range from 8 to 15 points (Davidson and Keating, 2002; Fritz and Irrgang, 2001; Grotle et al., 2003; Mannion et al., 2006) (Note: MDC90 was calculated from MDC95 data where necessary). Given that 39.5% of treatment group nurses and 54.4% of control group nurses scored ODI 10 points (Cooper et al., 1996) it is probable that greater than 15% scored below MDC90 at baseline. Hence, these data demonstrate that the ODI is also limited in its ability to detect longitudinal reduction in BP-D in nurses with back injury. Even though low ODI scores were detected in Cooper and colleagues’ trial (1996), the same investigating team applied the ODI in a subsequent trial of nurses. They found baseline scores in working nurses and unit assistants ranged from 5.2 to 7.6 points across treatment groups (Yassi et al., 2001). In light of our findings, these scores are too low to detect meaningful BP-D reduction. Hence the international research community must acknowledge the limitations of the ODI in nurses and cease its application in nursing samples. The suitability of an alternative instrument that is sensitive to lower levels of disability needs to be examined in a nursing cohort. An occupation-specific tool that assesses disability in relation to nursing tasks could be developed, however would preclude comparisons with studies comprising non-nursing samples. The RolandMorris Disability Questionnaire (RDQ) exhibits proportionally higher scores than the ODI when simultaneously applied in patient populations, and may identify changes in disability when ODI scores are at a minimum (Roland and Fairbank, 2000). Mean RDQ scores in nurses with BP reporting modified duties or work absence are 7.6 4.1 (Denis et al., 2007), equating to 32% of the maximum score. Hence, the RDQ may be appropriate for application in studies of nurses with BP and its measurement properties in this population warrant further investigation. Data in this study was sourced from nursing students who attended university, and does not represent enrolled students who do not attend lectures on campus. We used a methodological estimation for MDC90 calculation whereby interval-scale data was approximated by ordinal-scaled ODI data. This is consistent with all previous studies of the ODI (Davidson and Keating, 2002; Fritz and Irrgang, 2001; Grotle et al., 2003; Mannion et al., 2006). As ordinal data is ranked but not equidistant, there is likely to be greater distance between very low or high ODI scores compared with middle-range scores. Consequently, our MDC90 estimate (6%) may be slightly inflated. However, a strong correction of our estimate to account for marginal differences in MDC90 does not alter our conclusions. In our sample, 25.7% of individuals scored 2%. Therefore, even with an MDC90 estimate as low as 3% – which would likely exceed any correction provided by transformation of data to the interval scale – we could confidently state that the measurement properties of the ODI are inappropriate in this sample. In summary, the ODI is well utilised and endorsed in back pain research. In nurses and nursing students, however, ODI scores cluster at the lower range of the scale and the ODI has limited ability to detect longitudinal reduction in disability. The ODI is not recommended for studies of BP related disability in the nursing population.
A.P. Dawson et al. / International Journal of Nursing Studies 47 (2010) 604–607
Acknowledgements Anna Dawson is supported by The Sir Robert Menzies Memorial Foundation Limited. Emily Steele, Paul Hodges and Simon Stewart are supported by the National Health and Medical Research Council of Australia. We thank the University of South Australia for partially funding this study. Conflict of interest. None. Funding. The University of South Australia contributed partial funding to this study. Ethical approval. Ethical approval was granted by the Human Research Ethics Committee at the University of South Australia, reference P057/04. References Carta, A., Parmigiani, F., Campagna, M., Parrinello, G., Porru, S., 2007. LBP and disability in nursing personnel performing manual handing of patients in a large Italian hospital [Article in Italian]. G Ital. Med. Lav. Ergon. 29 (3 Suppl.), 581–583. Cooper, J.E., Tate, R.B., Yassi, A., 1998. Components of initial and residual disability after back injury in nurses. Spine 23 (19), 2118–2122. Cooper, J.E., Tate, R.B., Yassi, A., Khokhar, J., 1996. Effect of an early intervention program on the relationship between subjective pain
607
and disability measures in nurses with low back injury. Spine 21 (20), 2329–2336. Davidson, M., Keating, J., 2002. A comparison of five low back disability questionnaires: reliability and responsiveness. Phys. Ther. 82 (1), 8–24. Dawson, A., McLennan, S., Schiller, S., Jull, G., Hodges, P., Stewart, S., 2007. Interventions to prevent back pain and back injury in nurses: a systematic review. Occup. Environ. Med. 64 (10), 642–650. Dawson, A., Steele, E., Hodges, P., Stewart, S., 2009. Development and test– retest reliability of an extended version of the Nordic Musculoskeletal Questionnaire (NMQ-E): a screening instrument for musculoskeletal pain. J. Pain 10 (5), 517–526. Denis, S., Shannon, H., Wessel, J., Stratford, P., Weller, I., 2007. Association of low back pain, impairment, disability & work limitations in nurses. J. Occup. Rehabil. 17 (2), 213–226. Fritz, J., Irrgang, J., 2001. A comparison of a modified oswestry low back pain disability questionnaire and the quebec back pain disability scale. Phys. Ther. 81 (2), 776–788. Grotle, M., Brox, J., Vøllestad, N., 2003. Cross-cultural adaptation of the Norwegian versions of the Roland-Morris Disability Questionnaire and the Oswestry Disability Index. J. Rehabil. Med. 35 (5), 241–247. Mannion, A., Junge, A., Fairbank, J., Dvorak, J., Grob, D., 2006. Development of a German version of the Oswestry Disability Index. Part 1. Cross-cultural adaptation, reliability, and validity. Eur. Spine J. 15 (1), 55–65. Roland, M., Fairbank, J., 2000. The Roland-Morris Disability Questionnaire and the Oswestry Disability Questionnaire. Spine 25 (24), 3115–3124. Yassi, A., Cooper, J.E., Tate, R.B., Gerlach, S., Muir, M., Trottier, J., Massey, K., 2001. A randomized controlled trial to prevent patient lift and transfer injuries of health care workers. Spine 26 (16), 1739–1746.
Contents lists available at ScienceDirect
International Journal of Nursing Studies journal homepage: www.elsevier.com/ijns
Utility of the Oswestry Disability Index for studies of back pain related disability in nurses: Evaluation of psychometric and measurement properties Anna P. Dawson a,*, Emily J. Steele b, Paul W. Hodges a, Simon Stewart c a b c
Centre for Clinical Research Excellence in Spinal Pain Injury and Health, The University of Queensland, Brisbane, Australia Discipline of Public Health, University of Adelaide, Adelaide, Australia Baker IDI Heart and Diabetes Research Institute, Melbourne, Australia
A R T I C L E I N F O
A B S T R A C T
Article history: Received 2 January 2009 Received in revised form 25 July 2009 Accepted 23 October 2009
Background: Disability due to back pain in nurses results in reduced productivity, work absenteeism and attrition from the nursing workforce internationally. Consistent use of outcome measures is needed in intervention studies to enable meta-analyses that determine efficacy of back pain preventive programs. Objective: This study investigated the psychometric and measurement properties of the Oswestry Disability Index (ODI) in nursing students to determine its suitability for assessing back pain related disability in intervention studies. Methods: Bachelor of Nursing students were recruited. Test–retest reliability and the ability of the ODI to discriminate between individuals with serious and non-serious back pain were investigated. The measurement error of the ODI was examined with the minimal detectable change at the 90% confidence level (MDC90). Results: Student nurses (n = 214) had a low mean ODI score of 8.8 7.4%. Participants with serious back pain recorded higher scores than the rest of the cohort (p < 0.05). Test–retest reliability examined in 33 individuals was ICC = 0.88 (95%CI 0.77–0.94). The MDC90 = 6%, and 36% of nursing students scored below the MDC90 indicating the tool had limited ability to detect longitudinal change in disability in this population. Conclusion: Data from this and previous studies demonstrate that the measurement properties of the ODI are inappropriate for studying back pain related disability in nurses. The ODI is not recommended for back pain intervention studies in the nursing population and an alternative tool that is sensitive to lower levels of disability must be determined. ß 2009 Elsevier Ltd. All rights reserved.
Keywords: Back pain Disability evaluation Questionnaire Nursing Reliability (epidemiology) Minimal detectable change
What is already known about the topic? Consistent use of outcome measures is needed in studies of back pain prevention in nurses in order to enable meta-analyses of intervention efficacy.
The Oswestry Disability Index (ODI) measures disability due to back pain and has been used in studies with nurse cohorts. No studies have investigated the psychometric and measurement properties of the ODI in a nursing sample. What this paper adds
* Corresponding author at: Centre for Clinical Research Excellence in Spinal Pain Injury and Health, School of Health and Rehabilitation Sciences, The University of Queensland, St. Lucia, QLD 4072, Brisbane, Australia. Tel.: +61 7 3346 7467; fax: +61 7 3365 2775. E-mail address: [email protected] (A.P. Dawson). 0020-7489/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijnurstu.2009.10.013
Nursing students have low mean ODI scores. Data from this and other studies demonstrate that meaningful reduction in disability beyond measurement
A.P. Dawson et al. / International Journal of Nursing Studies 47 (2010) 604–607
error cannot be detected in a considerable proportion of nurses and nursing students. The ODI is not appropriate for studies of back pain in the nursing population. 1. Introduction Back pain (BP) is frequently reported by nurses and nursing students. Disability due to BP (BP-D) has a major impact on the productivity of the nursing workforce due to restricted duties, work absenteeism and attrition from the profession. The nursing profession is facing a critical shortage of workers, hence spinal pain and resultant disability must be prevented. Standardised use of measurement tools is needed in the international research community to enable data pooling in meta-analyses that determine intervention efficacy. A review of interventions to prevent BP and back injury in nurses identified inconsistent application of outcome instruments (Dawson et al., 2007). The Oswestry Disability Index (ODI) (Roland and Fairbank, 2000) measures the degree to which back or leg pain impacts functional activities, with a score range of 0–100%. It has been used to assess BP-D in a descriptive study (Carta et al., 2007) and intervention trials with nurse cohorts (Cooper et al., 1996, 1998; Yassi et al., 2001). The psychometric and measurement properties of the ODI have not been examined in nurses and it is uncertain whether the instrument is suitable for assessing BP-D in this population. The aims of this study were to determine in a nursing student sample: first, whether the tool is able to discriminate different classifications of BP; second, test– retest reliability; and third, whether the tool is able to detect longitudinal changes in BP-D. 2. Methods 2.1. Sample and data collection Bachelor of Nursing students in all three years of training who attended lectures on campus at an Australian university were provided with information and invited to participate during class time. A compulsory Nursing course from each year of training was attended for recruitment. Of 460 student nurses who attended university during recruitment, 373 volunteered (81% participation). All volunteers completed a demographic questionnaire and the extended Nordic Musculoskeletal Questionnaire (NMQ-E) (Dawson et al., 2009), and individuals reporting annual BP (upper and/or lower back region) completed the ODI version 2.0 (Roland and Fairbank, 2000). We also invited individuals to participate in a test–retest reliability study. In this analysis, we examine ODI data provided by n = 214 participants who reported annual BP (Cohort 1) and a subset of these volunteers (n = 33) who completed the ODI on two occasions (Cohort 2). 2.2. Data analysis Annual BP severity was classified into serious (that which required sick leave and/or treatment as recorded by
605
the NMQ-E), and non-serious (that which occurred without sick leave or treatment). The ability of the ODI to detect cross-sectional differences between participants with serious and non-serious BP (discriminative function) was assessed with Mann–Whitney U-tests. The test–retest reliability of the ODI was assessed using the proportion of observed agreement (po), and the intraclass correlation coefficient (ICC[2,1]). ODI ability to detect longitudinal within-subject changes in disability (evaluative function) was assessed with the minimal detectable change at the 90% confidence level (MDC90). This is the amount considered with 90% confidence to be a real change exceeding that due to measurement error. First the standard error of measurement (SEM) was determined using the standard deviation (s) from Cohort 1: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SEM ¼ s ð1 ICCÞ pffiffiffi. The MDC90 was then calculated: MDC90 ¼ 1:64 2 SEM, and the proportion of participants who scored
3. Results 3.1. Sample The personal characteristics of participants are presented in Table 1. 3.2. Discriminative function of ODI in nurses in training Mean (range) ODI scores for Cohort 1 include 8.8 7.4 (0–38)% for the entire sample, 7.3 6.2 (0–24)% for nonserious BP (n = 80) and 9.7 7.9 (0–38)% for serious BP (n = 134). Zero scores were recorded for 13% of the sample. Students with serious BP had higher scores than those with non-serious symptoms (p = 0.031). 3.3. Test–retest reliability Cohort 2 administered the ODI mean (range) 1.15 0.36 (1–2) days apart. ICC estimates for this cohort are presented in Table 2 and suggest the ODI total score was more reliable than individual questions.
Table 1 Sample characteristics.
Sample size Age (years) (mean SD) Weight (kg) (mean SD) Height (cm) (mean SD) Female (%) Annual upper back pain (%) Annual lower back pain (%) a
Mean of two testing occasions.
Cohort 1
Cohort 2
214 25.9 9.0 68.1 14.6 167.1 8.6 88.3 49.5 92.1
33 23.2 6.5 68.0 12.2 166.6 8.7 90.9 52.3a 94.0a
A.P. Dawson et al. / International Journal of Nursing Studies 47 (2010) 604–607
606
Table 2 Test–retest reliability of individual sections and total score of the Oswestry Disability Index in Bachelor of Nursing students (n = 33). Section
Pain/activity
ICC[2,1] (95%CI)
po
1 2 3 4 5 6 7 8 9 10
Pain intensity Personal care Lifting Walking Sitting Standing Sleep Sex Social life Travel
0.65 (0.40–0.81)
0.67 1.00 0.74 0.94 0.84 0.72 0.91 0.94 0.85 0.82
Total ODI score
a
0.74 0.78 0.71 0.59 0.82 0.25 0.52 0.51
(0.53–0.87) (0.60–0.88) (0.48–0.85) (0.31–0.78) (0.67–0.91) (0.11–0.55) (0.22–0.73) (0.21–0.73)
0.88 (0.77–0.94)
NA
ICC[2,1] = intraclass correlation coefficient (two-way model, single measures); 95%CI = 95% confidence interval; po = proportion of observed agreement; NA = not applicable. a ICC could not be calculated as all respondents scored zero on both testing occasions.
Table 3 Evaluative function of the Oswestry Disability Index in Bachelor of Nursing students. Back pain prevalence
SEM (%)
MDC90 (%)
Proportion with ODI score < MDC90 (%)
BP previous year (n = 214) BP on day of study (n = 82)
2.56 2.77
5.94 6.43
36.4 25.6
SEM = standard error of measurement; MDC90 = minimal detectable change at the 90% confidence level.
3.4. Evaluative function Table 3 presents the SEM and MDC90 for the ODI in student nurses, and the proportion of students who scored
calculated from BP patient data that range from 8 to 15 points (Davidson and Keating, 2002; Fritz and Irrgang, 2001; Grotle et al., 2003; Mannion et al., 2006) (Note: MDC90 was calculated from MDC95 data where necessary). Given that 39.5% of treatment group nurses and 54.4% of control group nurses scored ODI 10 points (Cooper et al., 1996) it is probable that greater than 15% scored below MDC90 at baseline. Hence, these data demonstrate that the ODI is also limited in its ability to detect longitudinal reduction in BP-D in nurses with back injury. Even though low ODI scores were detected in Cooper and colleagues’ trial (1996), the same investigating team applied the ODI in a subsequent trial of nurses. They found baseline scores in working nurses and unit assistants ranged from 5.2 to 7.6 points across treatment groups (Yassi et al., 2001). In light of our findings, these scores are too low to detect meaningful BP-D reduction. Hence the international research community must acknowledge the limitations of the ODI in nurses and cease its application in nursing samples. The suitability of an alternative instrument that is sensitive to lower levels of disability needs to be examined in a nursing cohort. An occupation-specific tool that assesses disability in relation to nursing tasks could be developed, however would preclude comparisons with studies comprising non-nursing samples. The RolandMorris Disability Questionnaire (RDQ) exhibits proportionally higher scores than the ODI when simultaneously applied in patient populations, and may identify changes in disability when ODI scores are at a minimum (Roland and Fairbank, 2000). Mean RDQ scores in nurses with BP reporting modified duties or work absence are 7.6 4.1 (Denis et al., 2007), equating to 32% of the maximum score. Hence, the RDQ may be appropriate for application in studies of nurses with BP and its measurement properties in this population warrant further investigation. Data in this study was sourced from nursing students who attended university, and does not represent enrolled students who do not attend lectures on campus. We used a methodological estimation for MDC90 calculation whereby interval-scale data was approximated by ordinal-scaled ODI data. This is consistent with all previous studies of the ODI (Davidson and Keating, 2002; Fritz and Irrgang, 2001; Grotle et al., 2003; Mannion et al., 2006). As ordinal data is ranked but not equidistant, there is likely to be greater distance between very low or high ODI scores compared with middle-range scores. Consequently, our MDC90 estimate (6%) may be slightly inflated. However, a strong correction of our estimate to account for marginal differences in MDC90 does not alter our conclusions. In our sample, 25.7% of individuals scored 2%. Therefore, even with an MDC90 estimate as low as 3% – which would likely exceed any correction provided by transformation of data to the interval scale – we could confidently state that the measurement properties of the ODI are inappropriate in this sample. In summary, the ODI is well utilised and endorsed in back pain research. In nurses and nursing students, however, ODI scores cluster at the lower range of the scale and the ODI has limited ability to detect longitudinal reduction in disability. The ODI is not recommended for studies of BP related disability in the nursing population.
A.P. Dawson et al. / International Journal of Nursing Studies 47 (2010) 604–607
Acknowledgements Anna Dawson is supported by The Sir Robert Menzies Memorial Foundation Limited. Emily Steele, Paul Hodges and Simon Stewart are supported by the National Health and Medical Research Council of Australia. We thank the University of South Australia for partially funding this study. Conflict of interest. None. Funding. The University of South Australia contributed partial funding to this study. Ethical approval. Ethical approval was granted by the Human Research Ethics Committee at the University of South Australia, reference P057/04. References Carta, A., Parmigiani, F., Campagna, M., Parrinello, G., Porru, S., 2007. LBP and disability in nursing personnel performing manual handing of patients in a large Italian hospital [Article in Italian]. G Ital. Med. Lav. Ergon. 29 (3 Suppl.), 581–583. Cooper, J.E., Tate, R.B., Yassi, A., 1998. Components of initial and residual disability after back injury in nurses. Spine 23 (19), 2118–2122. Cooper, J.E., Tate, R.B., Yassi, A., Khokhar, J., 1996. Effect of an early intervention program on the relationship between subjective pain
607
and disability measures in nurses with low back injury. Spine 21 (20), 2329–2336. Davidson, M., Keating, J., 2002. A comparison of five low back disability questionnaires: reliability and responsiveness. Phys. Ther. 82 (1), 8–24. Dawson, A., McLennan, S., Schiller, S., Jull, G., Hodges, P., Stewart, S., 2007. Interventions to prevent back pain and back injury in nurses: a systematic review. Occup. Environ. Med. 64 (10), 642–650. Dawson, A., Steele, E., Hodges, P., Stewart, S., 2009. Development and test– retest reliability of an extended version of the Nordic Musculoskeletal Questionnaire (NMQ-E): a screening instrument for musculoskeletal pain. J. Pain 10 (5), 517–526. Denis, S., Shannon, H., Wessel, J., Stratford, P., Weller, I., 2007. Association of low back pain, impairment, disability & work limitations in nurses. J. Occup. Rehabil. 17 (2), 213–226. Fritz, J., Irrgang, J., 2001. A comparison of a modified oswestry low back pain disability questionnaire and the quebec back pain disability scale. Phys. Ther. 81 (2), 776–788. Grotle, M., Brox, J., Vøllestad, N., 2003. Cross-cultural adaptation of the Norwegian versions of the Roland-Morris Disability Questionnaire and the Oswestry Disability Index. J. Rehabil. Med. 35 (5), 241–247. Mannion, A., Junge, A., Fairbank, J., Dvorak, J., Grob, D., 2006. Development of a German version of the Oswestry Disability Index. Part 1. Cross-cultural adaptation, reliability, and validity. Eur. Spine J. 15 (1), 55–65. Roland, M., Fairbank, J., 2000. The Roland-Morris Disability Questionnaire and the Oswestry Disability Questionnaire. Spine 25 (24), 3115–3124. Yassi, A., Cooper, J.E., Tate, R.B., Gerlach, S., Muir, M., Trottier, J., Massey, K., 2001. A randomized controlled trial to prevent patient lift and transfer injuries of health care workers. Spine 26 (16), 1739–1746.