Responsiveness and Minimal Important Change of the NeckPix
in Subjects with Chronic Neck Pain Undergoing RehabilitationThis section was compiled by Frank M. Painter, D.C.
Send all comments or additions to: Frankp@chiro.org
FROM: European Spine Journal 2018 (Jun); 27 (6): 1324–1331 ~ FULL TEXT
Marco Monticone, Luca Frigau, Howard Vernon, Barbara Rocca, Francesco Mola
Department of Medical Sciences and Public Health,
University of Cagliari,
Cittadella Universitaria,
Strada Statale, 554 - Monserrato,
Cagliari, Italy.
marco.monticone@unica.it
PURPOSE: The NeckPix© is a simple and rapid means of measuring the beliefs of subjects with chronic neck pain concerning pain-related fears of a specific set of activities of daily living. The original version showed satisfactory psychometric properties. This observational study is aimed at evaluating its responsiveness and minimal important changes (MICs) in subjects with chronic neck pain.
METHODS: At the beginning, at the end of an 8-week rehabilitation programme as well as at the one-year follow-up, 153 subjects completed the NeckPix. After the programme and at follow-up, subjects and physiotherapists also completed the global perceived effect (GPE) scale, which was divided to produce a dichotomous outcome. Responsiveness was calculated by distribution [effect size (ES); standardised response mean (SRM)] and anchor-based methods [receiver-operating characteristics (ROC) curves; correlations between change scores of the NeckPix and GPEs]. ROC curves were also used to compute MICs.
RESULTS: The ES ranged from 0.95 to 1.26 and the SRM from 0.84 to 0.98 at post-treatment and follow-up based on subjects' and physiotherapists' perspective. The ROC analyses revealed AUCs of 0.89 and 0.97 at post-treatment and follow-up, respectively; MICs (sensitivity; specificity) were of 6 (0.82; 0.88) and 8 (0.80; 0.92) at post-treatment and of 8 (0.95; 0.90 based on subjects and 0.95; 0.92 based on physiotherapists perspective) at follow-up. The correlations between change scores of the NeckPix and global perceived effects (GPEs) ranged from -0.69 to -0.82.
CONCLUSIONS: The NeckPix was sensitive in detecting clinical changes in subjects with chronic neck pain undergoing rehabilitation. We recommend taking the minimal important changes (MICs) provided into account when assessing subjects' improvement or planning studies in this clinical context.
KEYWORDS: Chronic neck pain; Kinesiophobia; Minimal important changes; NeckPix; Responsiveness
From the FULL TEXT Article:
Introduction
Integrating the management of psychological factors, such as fear of movement (i.e. kinesiophobia) to multidisciplinary rehabilitation is recommended in subjects with chronic neck pain (NP) to improve their course of disability, pain and quality of life. [1–4]
Interestingly, it has been suggested that the presentation of images of activities of daily living (ADL’s) that persons might find stressful or consider difficult to perform can allow a more in-depth investigation of the situations important to each individual subject which they are avoiding during everyday activities. [5] However, the number of image-based instruments for assessing fear-avoidance-based activity limitations is limited. [6–8]
Among these tools, the NeckPix© was published in 2015 as a simple and rapid means of measuring the beliefs of subjects with chronic NP concerning pain-related fears of a specific set of ADL’s. [8] This 10-item questionnaire was originally developed in Italian using a process of item generation and reduction/selection. The images did not require translation, while the instructions and captions were adapted also into English to facilitate its widest use: an Italian/English- speaking investigator made the first translation, which was back translated by another English-speaking investigator (see supplementary material). In completing the Neck- Pix©, subjects are asked to rate each picture from 0 (no fear) to 10 (greatest fear) according to the question: How much do you fear doing this activity would hurt your neck?, and the scale total score (0–100) is expected to generalise to a measure of activity-related kinesiophobia. The results of the initial study identified one main cognitive-behavioural component (explained variance: 71.12%; item-factor loadings: 0.786–0.921) and demonstrated satisfactory psychometric properties (internal consistency: 0.954; reproducibility: 0.979; and validity: 0.455–0.759, moderately to highly associated with related measures of fear-avoidance beliefs, pain catastrophising, NP disability and pain intensity). The developers did not find any floor or ceiling effects. [8] Additional analyses involving item response theory might be of interest as they have been already conducted for other cervical tools. [9]
However, it is of importance to investigate additional psychometric properties which make an important contribution to individuals management and research in measuring clinical change, such as responsiveness (i.e. the ability of an instrument to detect changes in the construct to be measured over time) and minimal important change (MIC, i.e. the smallest change in score of the construct to be measured that subjects perceive to be important). [10–12]
The aim of this study was to determine the responsiveness and MICs of the NeckPix© in subjects with chronic NP undergoing multidisciplinary rehabilitation using both distribution-based and anchor-based methods mainly suggested in the current literature and based on the ‘‘COnsensus- based Standards for the selection of health status Measurement INstruments’’ (COSMIN) [13, 14]; influences of subjects and physiotherapists on responsiveness and MICs were assessed.
Methods
This research was part of an observational study approved by the Institutional Review Board of our Hospital (date of approval: 22/12/2014). Subjects gave their written consent to participate.
Subjects
Outpatients admitted to our Rehabilitation Unit were enrolled between January 2015 and April 2015. The inclusion criteria were: a diagnosis of chronic axial NP (i.e. a documented history of pain lasting for more than 12 weeks), a good understanding of Italian, and an adult age. The exclusion criteria were acute (lasting up to 4 weeks) and subacute axial NP (lasting up to 12 weeks), specific causes of NP (e.g. disc herniation, canal stenosis, spinal deformity, fracture, spondylolisthesis, or infections), and central or peripheral neurological signs. Subjects with systemic illness, cognitive impairment (MMSE of < 24), recent myocardial infarctions, cerebrovascular events, or chronic renal diseases were excluded. Case histories, cervical radiographs and, in doubtful cases, Computed Tomography or Nuclear Magnetic Resonance were used to confirm inclusion/exclusion criteria; common degenerative changes, such as disc degeneration or spondyloarthrosis, were not considered as exclusion criteria. [15] Subjects who previously received cognitivebehavioural therapy for their NP were also excluded. The subjects’ sociodemographic and clinical characteristics were investigated using a specific form.
Procedures and outcome measures
All of the participants were provided written information concerning the questionnaires and procedures by two research assistants. Those satisfying the entry criteria underwent an eight-week outpatient rehabilitation programme that included exercises aimed at improving postural control, strengthening and stabilising the neck muscles, and stretching; subjects also received cognitive-behavioural therapy and education in ergonomic principles. This rehabilitation programme was the same for all of the subjects and was already tested for its efficacy. [16] Mild analgesics and nonsteroidal anti-inflammatory drugs (NSAIDs) were permitted during the study and an excessive use of medicines for pain control (> 3 pills of any type per day) was checked.
The NeckPix© was administered to all of the subjects as part of the pre-rehabilitation, post-rehabilitation and followup assessment.
At the end of treatment (8 weeks) and one year before follow-up, subjects’ and physiotherapists’ global perceived effect (pGPE and phGPE, respectively) was evaluated using the question, respectively: “Overall, how much did the treatment you received help your fear of movement due to current neck pain?” and “Overall, how much did the treatment you delivered help your subject’s fear of movement due to her/ his current neck pain?”; the GPE was determined using a five-level Likert scale with two improvement levels, one nochange level and two worsening levels. [17]
The questionnaires were administered by secretarial staff who checked them and returned any uncompleted part for completion to minimise the rate of missing/multiple responses. At follow-up, the subjects returned to the Institute or were contacted by phone by the same secretaries to complete the questionnaires.
Statistics
Responsiveness was determined using distribution and anchor-based methods [12, 18]: the former included the effect size (ES), also using Guyatt’s approach, and the standardised response mean (SRM). The ES is a standardised measure of change over time calculated on the whole sample by dividing the difference between the pre- and post-test scores by the pre-test standard deviation (SD); in the case of Guyatt’s approach, the change computed on the whole sample is divided by the pre-test SD calculated only for stable subjects whose clinical status remained unchanged (GPE = 3). The ES therefore represents individual change in terms of the number of pre-test SDs, with values of 0.20, 0.50, and 0.80, respectively, representing small, moderate, and large changes. The SRM (also referred to as the responsiveness- treatment coefficient or efficacy index) is the ratio between individual change and the SD of that change. It has been suggested that SRM values of 0.20, 0.50, and 0.80, respectively, represent small, moderate, and large changes.
As an anchor-based method, receiver-operating characteristic (ROC) curves were selected, which are useful indicators of the relationship between a measure and an external indicator of change, such as the GPE. Subjects were dichotomized into two groups based on GPE scores and considered improved when the GPE score was equal to 1 and 2, and stable when the GPE score was equal to 3. Responsiveness is described in terms of sensitivity (the probability that the measure correctly classifies subjects who demonstrate change when an external criterion of clinical change is used) and specificity (the probability that the measure correctly classifies subjects who do not demonstrate change when the external criterion is used). The sensitivity and specificity of each value of change in the measure are calculated and used to plot a ROC curve. The sensitivity values and falsepositive rates (1-specificity) are plotted on the y and the x axis of the curve, and the area under the curve (AUC) represents the probability a measure correctly classifies subjects as improved or unchanged. This area theoretically ranges from 0.5 (no discriminating accuracy) to 1.0 (perfect accuracy), and an AUC of at least 0.70 is considered to be acceptable. [17] The optimal cutoff point was computed using the Youden index and taken as the MIC, which indicates the change score associated with the least misclassification. [19] The sample size required for the ROC analysis is about 50 subjects per dichotomized group. [19]
External responsiveness was also investigated by means of correlation analyses with external criteria (GPE). [17] Correlations between the pre–post treatment change scores and the pre-treatment/follow-up in the NeckPix© and the GPE scores were tested by estimating Spearman’s rank order correlation coefficients (r < 0.30 = low; 0.30 < r < 0.60 = moderate; r > 0.60 = high. P < 0.05 was considered statistically significant). [20]
Results
Subjects
Table 1
Table 2 188 subjects were invited to participate, of whom 17 (9.1%) refused. Of the 171 selected subjects, 18 dropped out before starting the rehabilitation sessions because of logistic problems [12], economic difficulties [4], or personal problems. [3] The final study population consisted of 153 subjects (101 females, 66%, and 52 males, 34%) a mean age of 47.39 ± 15.98 years, a mean pain duration of 20.68 ± 16.69 months and a mean body mass index of 23.05 ± 3.51 kg/m2. All of the sociodemographic characteristics of the subjects are illustrated in Table 1.
Mean values (standard deviation) for the total at pre-treatment, post-treatment and follow-up were of 57.1 (15.7), 42.3 (17.4) and 37.5 (17.0), respectively. Baseline, post-treatment and follow-up scores for improved and stable subjects based on pGPE and phGPE are reported in Table 2.
Procedures
The study procedures were well accepted by all of the subjects, who did not raise any specific questions during the instruction phase or the administration of the questionnaires; no missing or multiple answers were found. None of the procedures led to any problems and all of the subjects completed the rehabilitation programme. No excessive use of medicines to control pain was shown. No specific issues were raised by the physiotherapists.
Psychometric properties
Table 3
Table 4
Figure 1 The dichotomisations based on the pGPE and phGPE are shown in Table 3. Given the number of the subjects in each group, the sample size was considered suitable for calculating responsiveness.
Both at post-treatment and follow-up, subjects and physiotherapists agreed on improvements, whereas the differences in stable and worsen subjects were shown. The subjects improved at post-treatment were the same retrieved at follow-up, whereas stable and worsen subjects showed minor changes based on pGPE and phGPE.
The results of the distribution-based and anchor-based methods are reported in Table 4. The ES suggested that the intervention needed large changes at post-treatment, both on the pGPE and phGPE (0.95). A further improvement was required at follow-up (1.26); and it was similar when the Guyatt’s approach was adopted (0.87–1.20 and 0.84–1.23, respectively). No differences in SRM were found at posttreatment and follow-up (0.84 and 0.98, respectively) based on pGPE and phGPE.
ROC analyses showed acceptable values of AUC, demonstrating good abilities to discriminate between improved and stable subjects at both post-treatment and follow-up (see Fig. 1). However, sensitivity and specificity differed: a higher specificity both at post-treatment and at follow-up was found on phGPE, suggesting a slightly better capability of identifying those who are actually stable; on the contrary, a higher sensitivity at post-treatment was found on pGPE, demonstrating a slightly better ability of identifying those who actually improved.
At post-treatment, the best cutoff points were 6 based on the pGPE and 8 based on the phGPE, and this means that a pre–post treatment change of > 6 and > 8, respectively, would have been considered as a clinically important change. At follow-up, the MICs were 8 in both cases and this means that a pre to follow-up change of > 8 would have been considered a clinically important change.
To statistically compare the AUC of the ROC curves based on pGPE and phGPE, we performed a DeLong’s test. [21] We could not reject the hypothesis that AUCs were equal both at post-treatment and at follow-up, and consequently they had equivalent responsiveness properties. On the other hand, we could reject the hypothesis that the AUC at post-treatment and follow-up were equal. Based on the high number of females enrolled, we additionally performed a number of ROC analyses adjusted for gender for each of the possible circumstances (i.e. pGPE at post-treatment, pGPE at follow-up, phGPE at post-treatment, and phGPE at follow-up). Given the estimates found (p values ranging from 0.46 to 0.88), we could not reject the null hypothesis that gender had no effect on the ROC analysis.
Figure 2 shows the relative frequency distributions of the change scores based on GPEs. The plots show the distributions of stable and improved subjects, and the MICs for pGPE and phGPE both at post-treatment and follow-up. At post-treatment, both for pGPE and phGPE, it arises that the distributions of stable and improved are considerably separated in spite of some acceptable overlapping due to the wide range of the scale. At follow-up, the overlapping is reduced, allowing an even better distinction.
The correlations between change scores of the NeckPix© and GPEs were high (0.69–0.82, see Table 4): their values improved moving from post-treatment to follow-up, with estimates based on pGPE higher than phGPE, both at posttreatment and at follow-up.
Discussion
This paper describes the estimation of responsiveness and the MICs of the NeckPix© questionnaire in a population of subjects with chronic NP undergoing multidisciplinary rehabilitation. Analysing the responsiveness and MIC of an outcome measure is an ongoing process and is strongly recommended to strengthen its psychometric properties and expand its applicability. [14, 22] Different approaches have been used to calculate responsiveness, but as yet, there is still no consensus as to which method is the best. [18] Hence, in this study, we used both distribution-based (ES, Guyatt ES, and SRM) and anchor-based methods (ROC analysis).
Distribution-based methods showed large responsiveness to the multidisciplinary rehabilitation programme at both post-treatment and one-year follow-up, as well as based on subjects’ or physiotherapists’ perspective. Other findings were not available in previous researches and, therefore, comparisons cannot be performed.
It has been recommended that distribution methods should be used cautiously as they tend to measure the magnitude of change scores rather than their validity. [22] When a general measure of change in patient-reported outcomes (PRO) such as the GPE is available and can be dichotomized into representative groups of improved and stable subjects, an anchor-based method such as ROC analysis is preferred as the AUC measures the ability of an instrument to discriminate between improved and stable subjects. [14]
The findings of this study showed that AUCs are always ≥ 0.89 at both post-treatment and at the one-year follow-up as well as based on subjects’ or physiotherapists’ perspective for the questionnaire under investigation. Again, no similar findings were retrieved in the literature available and, therefore, no comparisons can be made.
The optimal cutoff point estimated on the basis of ROC analysis at the end of treatment was about 6 and 8 based on subjects’ or physiotherapists’ perspective, respectively, suggesting the need for a greater improvement in kinesiophobia when an external judgment is advocated. Otherwise, similar estimates (i.e. MIC = 8) were achieved at follow-up, suggesting the need for a greater improvement in the fear of movement as assessed by the NeckPix© to achieve the clinical relevance and a perfect consistency at the end of the study by subjects and physiotherapists.
The external responsiveness was also investigated by means of correlation analyses with GPE, which reflect the extent to which changes in a PRO measure over a specific time relate to corresponding changes in an external standard, defined as an accepted indication of change in the condition of a subject. [13] We found that the pre–post treatment and pre-follow-up changes in the measure under investigation were –0.78 and –0.82 based on pGPE, and –0.69 and –0.77 on phGPE, correlated to the change in perceived effect, suggesting that they are responsive to GPE score, being able to predict changes in perceived treatment effect. Again, similar findings were not available in previous researches and comparisons cannot be conducted.
This study does have some limitations. First of all, the NeckPix© might not have been responsive to worsening outcomes as the subjects who were a “little worse” or “worse” were excluded from the analyses. Second, GPE was assessed using a five-point Likert scale, and clinically important changes would probably have been more discriminating if a seven-point scale had been used. Third, the applicability of this study is limited to an Italian population and similar studies are recommended in other countries.
In future, it would be interesting to evaluate responsiveness and MICs values of other scales measuring the same construct in the same population, to allow a comparison in terms of psychometrics performances and thus providing clinical indications. Despite the validity of clinical test commonly used in subjects with neck pain has not been firmly established, it would be also interesting to compare our findings with similar outputs deriving from tests to evaluate their actual impact on the management of kinesiophobia. [23] As recently reported [24], other psychological factors, such as outcome expectancy, have shown acceptable capability in predicting treatment success and it could be attractive to further investigate its contribution in addition to an imagebased instrument for fear-avoidance, such as the NeckPix©.
In conclusion, the findings of this study show that the NeckPix© questionnaire is a responsive measure in subjects with chronic NP undergoing multidisciplinary rehabilitation. It is recommended taking these MIC estimates into account when assessing improvement or planning clinical studies on a similar sample.
Acknowledgements
The authors would like to thank all the patients and the physiotherapists who took part in the study.
Conflict of interest
The authors declare that they have no conflict of interest.
References:
Jensen IB, Bergström G, Ljungquist T, Bodin L (2005)
A 3-year follow-up of a multidisciplinary rehabilitation programme for back and neck pain.
Pain 115:273–283Johansen JB, Røe C, Bakke ES et al (2013)
The determinants of function and disability in neck patients referred to a specialized outpatient clinic.
Clin J Pain 29:1029–1035Monticone M, Ambrosini E, Rocca B, et al (2016)
Group-based multimodal exercises integrated with cognitive-behavioural therapy improve disability, pain and quality of life of subjects with chronic neck pain: A randomized controlled trial with oneyear follow-up.
Clin Rehabil 269215516651979Monticone M, Cedraschi C, Ambrosini E et al (2015)
Cognitive? behavioural treatment for subacute and chronic neck pain.
Cochrane Database Syst Rev (5):CD010664Vlaeyen JW, Kole-Snijders AM, Boeren RG, Van Eek H (1995)
Fear of movement/(re) injury in chronic low back pain and its relation to behavioral performance. Pain 62:363–372Vlaeyen JW, de Jong J, Geilen M et al (2002) The treatment of fear of movement/(re) injury in chronic low back pain: further evidence on the effectiveness of exposure in vivo. Clin J Pain 18:251–261
Leeuw M, Goossens ME, van Breukelen GJ et al (2007) Measuring perceived harmfulness of physical activities in patients with chronic low back pain: the photograph series of daily activities— short electronic version. J Pain 8:840–849
Monticone M, Vernon H, Brunati R et al (2015) The NeckPix©: development of an evaluation tool for assessing kinesiophobia in subjects with chronic neck pain. Eur Spine J 24:72–79
Saltychev M, Mattie R, McCormick Z, Laimi K (2017) Psychometric properties of the neck disability index amongst patients with chronic neck pain using item response theory. Disabil Rehabil 1–6
Terwee CB, Bot SD, de Boer MR et al (2007) Quality criteria were proposed for measurement properties of health status questionnaires. J Clin Epidemiol 60:34–42
de Vet HC, Terwee CB, Ostelo RW et al (2006) Minimal changes in health status questionnaires: distinction between minimally detectable change and minimally important change. Health Qual Life Outcomes 4:54
Prinsen CA, Vohra S, Rose MR et al (2016) How to select outcome measurement instruments for outcomes included in a “Core Outcome Set”–a practical guideline.
Trials 17:449Husted JA, Cook RJ, Farewell VT, Gladman DD (2000)
Methods for assessing responsiveness: a critical review and recommendations.
J Clin Epidemiol 53:459–468Mokkink LB, Terwee CB, Knol DL et al (2010)
The COSMIN checklist for evaluating the methodological quality of studies on measurement properties: a clarification of its content.
BMC Med Res Methodol 10:22Nordin M, Carragee EJ, Hogg-Johnson S, Weiner SS, Hurwitz EL, Peloso PM, et al.
Assessment of Neck Pain and Its Associated Disorders:
Results of the Bone and Joint Decade 2000–2010 Task Force on
Neck Pain and Its Associated Disorders
Spine (Phila Pa 1976). 2008 (Feb 15); 33 (4 Suppl): S101–S122Monticone M, Baiardi P, Vanti C et al (2012)
Chronic neck pain and treatment of cognitive and behavioural factors: results of a randomised controlled clinical trial.
Eur Spine J 21:1558–1566Kamper SJ, Ostelo RW, Knol DL et al (2010)
Global Perceived Effect scales provided reliable assessments of health transition in people with musculoskeletal disorders, but ratings are strongly influenced by current status.
J Clin Epidemiol 63(760–766):e1Revicki D, Hays RD, Cella D, Sloan J (2008)
Recommended methods for determining responsiveness and minimally important differences for patient-reported outcomes.
J Clin Epidemiol 61:102–109Zweig MH, Campbell G (1993)
Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine.
Clin Chem 39:561–577Atkinson G, Nevill A (1997)
Comment on the use of concordance correlation to assess the agreement between two variables.
JSTORDeLong ER, DeLong DM, Clarke-Pearson DL (1988)
Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach.
Biometrics 837–845DeVet HC, Terwee CB, Mokkink LB, Knol DL (2011)
Measurement in medicine: a practical guide.
Cambridge University Press, CambridgeLemeunier N, da Silva-Oolup S, Chow N, et al (2017)
Reliability and Validity of Clinical Tests to Assess the Anatomical Integrity of the Cervical Spine
in Adults with Neck Pain and its Associated Disorders: Part 1- A Systematic Review from the
Cervical Assessment and Diagnosis Research Evaluation (CADRE) Collaboration
European Spine Journal 2017 (Sep); 26 (9): 2225–2241Groeneweg R, Haanstra T, Bolman CA et al (2017)
Treatment success in neck pain: the added predictive value of psychosocial variables in addition to clinical variables.
Scand J Pain 14:44–52
Return to CHRONIC NECK PAIN
Return to OUTCOME ASSESSMENT
Since 10-27-2017
| Home Page | Visit Our Sponsors | Become a Sponsor |
Please read our DISCLAIMER |