Abstract
Background
The current standard to evaluate the presence of somatosensory dysfunctions is quantitative sensory testing, but its clinical utility remains limited. Low‐cost and time‐efficient clinical sensory testing (CST) batteries have thus been developed. Recent studies show moderate to substantial reliability in populations with neuropathic pain. This study evaluates the inter‐ and intra‐tester reliability of people with spine‐related leg and arm pain, representing mixed pain mechanisms.
Methods
Fifty‐three patients with spine‐related leg (n = 41) and arm pain (n = 12) attended three CST sessions. The CST battery consisted of eleven tests, determining loss and gain of sensory nerve function. CST was performed by the same investigator twice and by an additional investigator to determine inter‐ and intra‐tester reliability. Fleiss' (inter‐tester) and Cohen's (intra‐tester) kappa were calculated for dichotomized and intraclass correlation coefficients (ICC) for continuous outcomes.
Results
Fleiss' kappa varied among modalities from fair to substantial (κ = 0.23–0.66). Cold, warm, and vibration detection thresholds and cold and pressure pain thresholds reached kappa >0.4 (moderate to substantial reliability). Cohen's kappa ranged from moderate to substantial (κ = 0.45–0.66). The reliability of the windup ratio was poor (ICC <0.18).
Conclusion
CST modalities with moderate to substantial inter‐tester reliability could be of benefit as a screening tool. The moderate to substantial intra‐tester reliability for all sensory modalities (except windup ratio) supports their potential use in clinical practice and research to monitor somatosensory changes over time in patients with spine‐related limb pain of mixed pain mechanisms.
Significance
We already know that most modalities of clinical sensory test (CST) batteries achieve moderate to substantial inter‐ and intra‐tester reliability in populations with neuropathic pain.
This study evaluates the reliability of a CST battery in populations with mixed pain mechanisms. We found inter‐tester reliability varied from poor to substantial for sensory modalities, questioning the value of some CST modalities. The CST battery showed moderate to substantial intra‐tester reliability, suggesting its usefulness to monitor sensory changes over time in this cohort.
1. INTRODUCTION
Somatosensory dysfunction is a hallmark feature of neuropathic pain (Colloca et al., 2017), but may also occur in nociceptive pain conditions (Moloney et al., 2015; Tampin et al., 2012). Clinically, these can present as loss (e.g. hypoesthesia or hypoalgesia) or gain (e.g. allodynia or hyperalgesia) of sensory function (Baron et al., 2017). The current reference standard to evaluate the presence and nature of somatosensory dysfunction is quantitative sensory testing (QST). QST is a standardized method to quantify and monitor both loss and gain of nerve function by evaluating somatosensory modalities mediated by primary sensory afferents (Rolke et al., 2006). Promisingly, some QST modalities might have prognostic value for the development of musculoskeletal pain and disability (Georgopoulos et al., 2019; Petersen et al., 2021), highlighting the potential benefits of monitoring specific somatosensory modalities in clinics. While QST is considered a valid measurement to identify somatosensory dysfunction (Rolke et al., 2006), the equipment is expensive, requires training, and is time‐consuming, thus limiting its clinical application.
To overcome these barriers, research effort has gone into the development of clinical sensory test (CST) batteries to detect somatosensory dysfunctions with low‐cost equipment and higher time efficiency (Reimer et al., 2020; Ridehalgh et al., 2018; Zhu et al., 2019). Several studies have examined the concurrent validity of CST (i.e., correlation with QST) (Koulouris et al., 2020; Reimer et al., 2020; Zhu et al., 2019). Although concurrent validity varies among the different sensory modalities, thermal and mechanical detection thresholds, as well as cold and pressure pain thresholds, achieve moderate to relatively strong correlations compared to QST (Zhu et al., 2019). Another aspect adding to the validity of CST batteries is their reliability across different examiners (inter‐tester) and by the same examiner on different occasions (intra‐tester) (Mokkink et al., 2010). The good reliability of CST is imperative for its use as a screening tool and as a clinical tool to measure outcome and monitor treatment response. Recently, moderate to substantial inter‐ and intra‐tester reliability has been reported for different CST batteries in populations with predominant neuropathic pain (Baad‐Hansen et al., 2013; Koulouris et al., 2020; Reimer et al., 2020; Wasan et al., 2020). CST might thus offer benefits compared to a standard neurological examination, especially considering its comprehensive assessment of both loss and gain of function of different fibre populations and the notable variability in inter‐tester reliability of routine bedside neurological examinations (Dyck et al., 2010).
It remains unknown whether CST batteries are also reliable in patient populations with mixed pain mechanisms. Spine‐related limb pain can involve both neuropathic and nociceptive pain mechanisms, with often subtle somatosensory dysfunction (Attal et al., 2011; Tampin et al., 2012). This study therefore aims to assess the inter‐ and intra‐tester reliability of a CST battery in a population with spine‐related leg or arm pain to reflect a wide range of pain mechanisms.
2. METHODS
2.1. Study cohort
Patients aged 18 years and above experiencing spine‐related arm or leg pain were recruited from the Vesalius‐Clinic in Bad Rappenau rehabilitation centre and a private physiotherapy practice in Germany. These sites represent secondary and primary care settings, respectively, and therefore help generalizability of our findings. The clinical diagnosis of unilateral spine‐related leg or arm pain was made by physiotherapists specialized in musculoskeletal disorders. As there is no diagnostic gold standard to identify spinal‐related leg and arm pain, we relied on a pragmatic approach based on a detailed subjective and objective assessment. The following two criteria had to be fulfilled: (1) Pain radiating below the gluteal fold in case of spine‐related leg pain (Lin et al., 2014) and below the acromion in the case of spine‐related arm pain and (2) the limb pain had to be modifiable through spinal movement (Rastogi et al., 2022). Patients included those with painful radiculopathy, radicular pain, or spine‐related somatic referred pain. Patients were excluded if there was evidence of a central nervous system disorder or metabolic conditions (e.g., myelopathy or diabetes), a clinical diagnosis of anxiety or depression (indicated by a HADS‐D total score of >21 (Mitchell et al., 2010; Vodermaier & Millman, 2011)) as well as bilateral spine‐related limb pain. Patients with insufficient command of the German language to complete questionnaires and follow instructions were excluded.
The study protocol was approved by the ethics committee from the department of computer science at the Trier University of Applied Sciences (03–2021). Prior to study entry, all participants gave their informed written consent. The reporting of the study is consistent with the guideline for reporting reliability and agreement studies (Kottner et al., 2011).
2.2. Study procedure
Participants attended three examinations: two on the same day (T1a and T1b) and the third after a period of two to 7 days (T2) (Figure 1). During the first examination, demographics and clinical data, including age, pain duration, MRI evidence of nerve root compression if available, and findings of a bedside neurological examination (Appendix S1), were recorded by the main investigator (CB). Based on the information available for each patient, we evaluated the presence of probable and definite neuropathic pain according to the grading system described by Finnerup et al. (2016). Patients were asked to define their area of maximal pain, and this was marked by one of the examiners on a body diagram. Patients rated the present pain intensity on an 11‐point numeric rating scale (NRS, 0 = no pain; 10 = the worst pain imaginable) at the start of the three examinations.
FIGURE 1.
Study protocol. CST, Clinical Sensory Testing; NPSI, Neuropathic Pain Symptom Inventory; NRS, Numeric Rating Scale; MSK‐HQ, Musculoskeletal Health Questionnaire.
The CST battery was conducted by three physiotherapists specialized in musculoskeletal therapy who had at least 5 years of professional experience. Before data collection, the main investigator (CB) was trained online for 1 h in the CST performance, including wording and rating criteria, by a specialist who was involved in the original design of the CST (AS). The main investigator then provided an in‐person training session for the other two examiners for standardization purposes. Four months after the start of data collection, the main investigator organized follow‐up training with the other two examiners to prevent drifts in performance.
To determine inter‐tester reliability, the first CST session (T1a) was performed by one examiner, followed by the second session (T1b) by a second examiner 20–30 min later. The main investigator (CB) was always involved as one examiner in T1a or T1b. The examiners were staff from each of the two institutions. We have not standardized or randomized the order of examiners between T1a and T1b due to logistical issues with therapist availability. In the case of a pain flare up after T1a (defined by an increase in pain intensity ≥2 points on the NRS), the start of T1b was delayed until the pain decreased to levels recorded at T1a.
To determine intra‐tester reliability, participants returned to the study site for a third CST assessment (T2) within two to 7 days after their first session. The main investigator (CB) performed the CST at T2 in all participants.
Examiners were blinded to the CST results of the other examiners, but blinding of the patients and the same examiner between T1 and T2 was not possible. However, the CST battery contains eleven tests, and we considered it highly unlikely that the examiner or patients would remember the outcomes. Throughout all three sessions, examiners were blinded to the outcomes of the questionnaires.
2.3. CST battery
The CST battery was based on previously published testing protocols (Ridehalgh et al., 2018; Scholz et al., 2009; Zhu et al., 2019). It has been validated as a screening tool to detect sensory dysfunction by identifying deviations from normal levels (i.e., beyond one or two standard deviations of QST z‐scores from healthy controls) using dichotomous scaling (normal/abnormal) in individual patients (Zhu et al., 2019). The protocol consisted of eleven readily available and easy‐to‐use devices for determining loss and gain of sensory nerve function (Figure 2). Before performing the CST, the examiner determined skin temperature in the area of maximal pain using an infrared thermometer (TFA® Infrared‐thermometer FLASH PEN). The required skin temperature was set to at least 31°C. Warm packs were used to reach the target temperature if required. The order of the CST battery was standardized. After the completion of one test modality, we moved immediately to the next test with no breaks between the modalities to accommodate time constraints imposed by a clinical setting. The test protocol took about 10–15 min and was based on two time measurements taken during examiner training with patients in both primary and secondary care settings. Participants were tested with their eyes closed first on the symptom‐free contralateral control area, followed by the corresponding area of maximal pain. The area of maximal pain marked on the body diagram was available to the examiners at each session.
FIGURE 2.
CST devices used in the testing protocol. (1) TipTherm© for cold and warm detection threshold, (2) cooling pack for cold pain threshold, (3) glass vial for heat pain threshold, (4) cotton wool for mechanical detection threshold, (5) von Frey filament 256 mN and (6) Neurotip (Neurpen©, Owen Mumford) for mechanical pain threshold, (7) brush (SESELab™‐Brush‐05) for dynamic mechanical allodynia, (8) toothpick for wind‐up ratio, (9) tuning fork 128 Hz for vibration detection threshold, (10) eraser for pressure pain threshold; CST, Clinical Sensory Testing.
2.3.1. Modalities for loss of function
First, a TipTherm© was used to determine the ability to discriminate thermal sensation. For the cold detection threshold (CDT), the metal end of the TipTherm© was used, while the warm detection threshold (WDT) was tested with the polymer end, which is generally perceived as warm or neutral at room temperature. Both modalities were tested for 2 s on the patients' skin.
Mechanical detection threshold (MDT) was tested via a light stroke of 2–3 cm and 1 s duration using a ball of cotton wool.
The vibration detection threshold (VDT) was evaluated with a tuning fork (Rydel‐Seiffer, Arno Barthelmes & Co. GmbH Germany) of 128 Hz frequency, which was placed for 10 s on the skin. The amplitude of the tuning fork was standardized by releasing the metal fork from a fully approximated position.
The mechanical pain threshold (MPT) was evaluated in two ways. First with a standardized von Frey filament weighing 256 mN (MPT VF256 loss of function (LoF)) and second with a pinprick (MPT PP) using a Neurotip mounted on a Neuropen© (Owen Mumford). For the latter, the stimulus was standardized using the integrated spring pressure of 40 g. Both modalities were placed for 1–2 s on the patients' skin.
2.3.2. Modalities for gain of function
For cold pain threshold (CPT), a cooling pack (8 cm x 15 cm) previously stored in a freezer compartment was placed for 10 sec on the patients' skin.
Heat pain threshold (HPT) was evaluated with a glass vial filled with hot water (40°C) and placed for 10 s over the skin. The water was heated to 40°C with a kettle and temperature determined with the laser thermometer.
MPT was performed as above with a standardized von Frey filament weighting of 256 mN but recorded as gain of function (MPT VF256 gain of function (GoF)) if perception was increased on the maximal pain site compared to the contralateral site.
The pressure pain threshold (PPT) was evaluated with an eraser mounted on a pencil. Pressure was applied for 10 s over the closest muscle belly to the area of maximal pain. The pressure was sufficient to indent the soft tissue, leading to skin blanching.
The wind‐up ratio (WUR) was established with a toothpick by applying a single stimulus followed by a train of 10 stimuli at a frequency of 1/sec. Participants rated the pain on an 11‐point NRS from 0 (no pain at all) to 10 (worst pain imaginable). WUR was calculated as the ratio of the single stimulus rating over the average rating for the train of stimuli. In a post hoc analysis, we also calculated the difference between the single and train of stimuli to avoid data loss due to zero ratings on the single stimulus.
Dynamic mechanical allodynia (DMA) was assessed with a brush (SESELab™‐Brush‐05, Somedic Sweden) by gently stroking the skin three times over a length of 2–3 cm with a 1 s duration.
2.3.3. CST interpretation
Immediately after the application of each CST stimulus, patients were asked whether the stimulus applied over the affected site was perceived as increasing, decreasing, or having the same intensity compared to the contralateral control area. In loss of function modalities (detection thresholds, MPT PP and MPT VF256), a perception of decreased sensation was interpreted as loss of function. In the gain of function modalities (pain thresholds), a perception of increased sensation was considered to reflect gain of function. DMA was rated as present (pain provocation) or absent (no pain provocation).
2.4. Questionnaires
Participants completed the Musculoskeletal Health Questionnaire (MSK‐HQ) and Neuropathic Pain Symptom Inventory (NPSI) at both T1 and T2. These questionnaires served the primary purpose of evaluating symptom stability, checking that participants' characteristics remained consistent throughout the testing period. This stability assessment was crucial to establishing the intra‐tester reliability of CST modalities. To reduce potential information bias, the baseline questionnaires were completed between T1a and T1b to shift participants' focus off the sensory testing.
2.4.1. Musculoskeletal health questionnaire‐MSK‐HQ
The German translation of the MSK‐HQ (MSK‐HQG) was implemented to assess patient health status on a functional level. With 14 items on pain, disability, emotions, sleep, and self‐confidence to manage the condition, the MSK‐HQ provides a holistic view of the impact of the condition (Hill et al., 2016; Karstens et al., 2021). The MSK‐HQ is scored from 0 to 56, with a higher score indicating better musculoskeletal health status. The minimally important change has been estimated in a range of musculoskeletal conditions at 8.5 points (Karstens et al., 2021). The German translation shows good test‐retest reliability and good construct validity (Karstens et al., 2021).
2.4.2. Neuropathic pain symptom inventory‐NPSI
The German version of the NPSI (NPSI‐G) was used to obtain information about different pain characteristics. The NPSI includes five clusters of items (superficial spontaneous pain, deep spontaneous pain, paroxysmal pain, evoked pain, as well as paraesthesia and dysaesthesia) to discriminate and quantify different dimensions of neuropathic pain. In the NPSI, each sub score can range from 0 to 10 and the total score from 0 to 100, with higher scores representing higher symptom severity (Bouhassira et al., 2004). It has been shown that the NPSI is sensitive to detecting change (Bouhassira et al., 2004; Sommer et al., 2011).
2.5. Statistical analysis and sample size
Before statistical analysis was performed, the CST variables (increased, decreased, or normal) were dichotomized (normal or abnormal), as validated previously by Zhu et al. (2019). The dichotomous scaling offers clinicians the advantage of quickly identifying the presence or absence of somatosensory dysfunction in an individual patient. A perceived decreased response (for loss of function tests) or increased response (for gain of function tests) compared to the control area was defined as a sensory dysfunction (abnormal). An equal response was defined as a normal function (normal).
For dichotomized measures, inter‐tester agreement (T1a and T1b) was estimated utilizing Fleiss' Kappa (accommodates non‐full cross‐over designs), while intra‐tester reliability (T1 and T2) was estimated using Cohen's Kappa statistics. Inter‐tester and intra‐tester agreement estimations were interpreted adopting the method introduced by Landis and Koch ((Landis & Koch, 1977): <0 poor, 0–0.2 slight, 0.21–0.4 fair, 0.41–0.6 moderate, 0.61–0.8 substantial, and 0.81–1.0 almost perfect agreement). We also provide observed agreement and proportion of specific agreement for each test as a measure to assist the assessment of agreement of individuals in clinical practice.
Symptom stability from T1 to T2 was measured based on the change of the NPSI and MSK‐HQ for all dichotomized sensory modalities. We applied a Shapiro‐Wilk test to test for normal distribution, followed by a paired t‐test or Wilcoxon test to determine if the scores of the two questionnaires changed between timepoints. Due to statistically significant changes in the questionnaire scores over time, which may have impacted the CST findings and thus reliability, a two‐stage logistic regression model for analysing the intra‐tester agreement with the change in scores as covariates was performed as a post hoc sensitivity analysis to check if the agreement of the two testers was still better than chance, even when taking both questionnaire scores as covariates into account (Lipsitz et al., 2003). We considered the two‐stage regression as an extension of Kappa that can handle covariates. In the first regression stage, the marginal probabilities for each one of the testers to give an abnormal rating, given the covariates, were calculated. Then an offset term, representing agreement by chance only, was formulated as the logit of agreement between the fitted probabilities for each one of the testers to give an abnormal rating. This offset was then used alongside the covariates (change in MSK‐HQ and NPSI from T1 to T2) in the second logistic regression stage to model the overall, chance corrected, agreement between the two testers. This method first used the random agreement of the two testers (first stage, offset) as well as the chance‐corrected agreement of the covariates and an offset (second stage). We then tested the hypothesis that the overall chance‐corrected agreement considering the covariates was higher than agreement due to chance only using a one‐sided t‐test.
For the continuous measures (WUR and temporal summation of pain (TSP)), intraclass correlation coefficients (ICCs) were calculated to determine the magnitude of inter‐tester and intra‐tester agreement. We computed the inter‐tester ICC based on a one‐way random‐effects model, single measures, and the absolute‐agreement of three observers. Intra‐tester ICC was calculated utilizing a two‐way mixed‐effects model, single measures, and absolute agreement. Due to the high number of zero ratings during the first stimulus of WUR testing, we performed a post hoc ICC analysis for TSP based on the difference between the average rating for the train of stimuli and the single stimulus rating. ICC was interpreted according to Koo and Li: ICC <0.5 poor, ICC >0.5 moderate, ICC >0.75 good, and ICC >0.9 excellent correlation (Koo & Li, 2016).
All statistical analyses were performed using the statistical computing language R (v4.0.0; R Core Team, 2020). Sample size estimation was performed a priori by applying a confidence interval approach (Rotondi & Donner, 2012). The desired bounds for a two‐sided confidence interval were set to +/− 0.2 from the anticipated preliminary kappa value of 0.61, resulting in a sample size estimation of a minimum of 53 participants.
3. RESULTS
Fifty‐three patients were recruited for the study procedure and all but one completed all three sessions (Covid‐related dropout at T2). Patient characteristics are described in Table 1. Forty‐one patients had spine‐related leg pain and 12 spine‐related arm pain. MRI data were available for 29 patients. A neurological deficit upon bedside neurological examination was found in 30 patients. Eleven patients (20.8%) met the criteria for probable neuropathic pain, while nineteen patients (35.9%) were identified to have definite neuropathic pain (Finnerup et al., 2016). We recorded a decrease in present pain NRS from 4 (SD = 2.4) at T1 to 3.3 (SD = 2.4) at T2. A pain flare‐up between T1a and T1b was recorded in three patients. Their second testing session was delayed until the pain returned to baseline values. Details on symptom stability between T1 and T2 according to the MSK‐HQ and NPSI can be found in Table S2.
TABLE 1.
Patient characteristics.
| n = 53 a | |
|---|---|
| Age [mean ± SD] (range) | 52.4 ± 10.3 (31–80) |
| Sex [n] (%) | |
| Female | 31 (58%) |
| Male | 22 (42%) |
| Area of maximal pain [n] (%) | |
| Upper limb | 12 (22.6)% |
| Upper arm | 1 (1.9%) |
| Forearm | 5 (9.4%) |
| Hand | 6 (11.3%) |
| Lower limb | 41 (77.4%) |
| Thigh | 19 (35.9%) |
| Lower leg | 15 (28.3%) |
| Foot | 7 (13.2%) |
| Pain duration, months [mean ± SD] (range) | 23.0 ± 54.4 (1–360) |
| Present pain intensity NRS [mean ± SD] | |
| T1a | 4.0 (2.4) |
| T1b | 3.9 (2.4) |
| T2 | 3.3 (2.4) |
| MSK‐HQG score [mean ± SD] | |
| T1 | 33.0 ± 11.1 |
| T2 (n = 52) | 34.8 ± 10.2 |
| NPSI‐G score [mean ± SD] | |
| T1 | 28.9 ± 18.2 |
| T2 (n = 52) | 26.1 ± 18.1 |
| MRI evidence of nerve root compression [n] (%) | |
| Cervical spine | 10 (18.9%) |
| C4/5 | 2 (3.8%) |
| C5/6 | 4 (7.6%) |
| C6/7 | 4 (7.6%) |
| Lumbar spine | 21 (39.6%) |
| L3/4 | 3 (5.7%) |
| L4/5 | 10 (18.9%) |
| L5/S1 | 9 (17.0%) |
| Loss of function in bedside neurological examination [n] (%) | |
| Cervical spine | |
| Sensory (light touch) | 6 (11.3%) |
| Muscle strength | 4 (7.6%) |
| Tendon reflex | 4 (7.6%) |
| Lumbar spine | |
| Sensory (light touch) | 15 (28.3%) |
| Muscle strength | 13 (24.5%) |
| Tendon reflex | 7 (13.2%) |
| Neuropathic pain grading [n] (%) | |
| Probable | 11 (20.8%) |
| Definite | 19 (35.9%) |
Abbreviations: MSK‐HQ, musculoskeletal health questionnaire; NPSI, neuropathic pain symptom inventory; NRS, numeric rating scale; SD, standard deviation.
Unless noted differently.
We did not consider DMA in our statistical analysis as it was rated as absent in all patients.
3.1. Inter‐tester reliability
The percentage of observed agreement for dichotomized CST modalities ranged from 66% (MPT PP) to 83% (CDT), with CDT, CPT, and PPT reaching >80% agreement (Table 2). Kappa statistics for dichotomized modalities ranged from fair (0.23 for MPT VF256 (GoF)) to substantial (0.66 for CDT). We identified fair agreement for MDT, MPT (both loss and gain of function), MPT PP, and HPT; moderate agreement for WDT, VDT, and CPT; and substantial agreement for CDT and PPT. We were able to estimate WUR only for seventeen patients, as thirty‐six did not experience pain on the initial stimulus. The ICC for WUR was poor between examiners (0.14) (Table 3).
TABLE 2.
Observed agreement and inter‐tester reliability (T1a and T1b) for dichotomized CST modalities (n = 53).
| Modalities | Observed agreement (%) | Inter‐tester reliability (κ) | Lower and upper limits of CI (95%) for κ |
|---|---|---|---|
| Loss of function | |||
| CDT | 83% | 0.66 | 0.43; 0.85 |
| WDT | 79% | 0.56 | 0.28; 0.78 |
| MDT | 74% | 0.40 | 0.12; 0.65 |
| MPT VF265 (LoF) | 62% | 0.24 | −0.05; 0.49 |
| MPT PP | 66% | 0.32 | 0.06; 0.56 |
| VDT | 79% | 0.46 | 0.13; 0.73 |
| Gain of function | |||
| CPT | 81% | 0.60 | 0.37; 0.80 |
| HPT | 75% | 0.39 | 0.06; 0.66 |
| MPT VF265 (GoF) | 68% | 0.23 | −0.11; 0.52 |
| PPT | 81% | 0.61 | 0.38; 0.81 |
Note: 0.82–1, almost perfect (
); 0.61–0.8, substantial (
); 0.41–0.6, moderate (
); 0.21–0.4, fair (
); 0–0.2, slight (
); <0, poor (grey).
Abbreviations: CDT, cold detection threshold; CPT, cold pain threshold; CST, clinical sensory testing; HPT, heat pain threshold; MDT, mechanical detection threshold; MPT VF265, mechanical pain threshold van Frey hair weighting 265 mN; LoF/GoF, loss/gain of function; MPT PP, mechanical pain threshold pin prick; PPT, pressure pain threshold; VDT, vibration detection threshold; WDT, warm detection threshold.
TABLE 3.
Inter‐tester (n = 17) and intra‐tester reliability (n = 23) for WUR.
| Modalities | Inter‐tester reliability (ICC) | Lower and upper limits of CI (95%) for ICC | Intra‐tester reliability (ICC) | Lower and upper limits of CI (95%) for ICC |
|---|---|---|---|---|
| WUR | 0.14 | 0–0.46 | 0.18 | −0.19; 0.52 |
Note: >0.9, excellent (
); >0.75, good (
); >0.5, moderate (
); < 0.5, poor (
).
Abbreviations: ICC, intraclass correlation coefficient; WUR, wind‐up ratio.
3.2. Intra‐tester reliability
The mean time interval between T1 and T2 was 3.7 days. The percentage of observed agreement for dichotomized CST modalities ranged from 73% (MPT VF256 (GoF)) to 85% (CPT), with CDT, MDT, CPT, and HPT reaching >80% agreement (Table 4). Kappa statistics for dichotomized modalities ranged from moderate (0.45 for VDT and MPT VF256 (GoF)) to substantial (0.66 for CPT). We identified moderate agreement for all modalities, except for CDT and CPT, which reached substantial agreement (Table 4).
TABLE 4.
Observed agreement and intra‐tester reliability (T1 and T2) for dichotomized CST modalities (n = 52).
| Modalities | Observed agreement (%) | Intra‐tester reliability (κ) | Lower and upper limits of CI (95%) for (κ) |
|---|---|---|---|
| Loss of function | |||
| CDT | 81% | 0.61 | 0.39; 0.83 |
| WDT | 77% | 0.51 | 0.27; 0.75 |
| MDT | 83% | 0.59 | 0.35; 0.83 |
| MPT VF265 (LoF) | 75% | 0.48 | 0.24; 0.73 |
| MPT PP | 79% | 0.55 | 0.31; 0.79 |
| VDT | 79% | 0.45 | 0.16; 0.74 |
| Gain of function | |||
| CPT | 85% | 0.66 | 0.45; 0.88 |
| HPT | 81% | 0.51 | 0.24; 0.79 |
| MPT VF265 (GoF) | 73% | 0.45 | 0.20; 0.69 |
| PPT | 75% | 0.50 | 0.27; 0.74 |
Note: 0.82–1, almost perfect (
); 0.61–0.8, substantial (
); 0.41–0.6, moderate (
); 0.21–0.4, fair (
); 0–0.2, slight (
); <0, poor (grey).
Abbreviations: CDT, cold detection threshold; CPT, cold pain threshold; CST, clinical sensory testing; HPT, heat pain threshold; MDT, mechanical detection threshold; MPT VF265, mechanical pain threshold van Frey hair weighting 265 mN; LoF/GoF, loss/gain of function; MPT PP, mechanical pain threshold pinprick; PPT, pressure pain threshold; VDT, vibration detection threshold; WDT, warm detection threshold.
The changes in the MSK‐HQ and NPSI were found to be statistically significant from T1 to T2 (Table S2). Therefore, a two‐stage logistic regression analysis for intra‐tester agreement adjusted for the two covariates NPSI and MSK‐HQ, followed by a one‐sided t‐test, revealed that the agreement of the testers could not be explained by chance only, that is agreement between the fitted probabilities of each tester modelled after the covariates alone (Table 5). The ICC for WUR was poor for intra‐tester reliability (0.18) (Table 3). The post hoc ICC analysis for TSP revealed comparably poor inter‐tester and intra‐tester reliability (Table S3).
TABLE 5.
Two stage logistic regression analysis of intra‐ tester agreement (T1 and T2) adjusted for covariates (MSK‐HQ, NPSI).
| Modalities | Estimate | t | p value |
|---|---|---|---|
| Loss of function | |||
| CDT | 1.52 | 12.01 | <0.005 |
| WDT | 1.04 | 8.33 | <0.005 |
| MDT | 1.85 | 6.56 | <0.005 |
| MPT VF256 (Lof) | 1.02 | 18.10 | <0.005 |
| MPT PP | 1.10 | 24.39 | <0.005 |
| VDT | 1.47 | 7.80 | <0.005 |
| Gain of function | |||
| CPT | 1.44 | 40.77 | <0.005 |
| HPT | 1.23 | 7.43 | <0.005 |
| MPT VF256 (GoF) | 0.98 | 20.29 | <0.005 |
| PPT | 1.50 | 9.32 | <0.005 |
Note: Results show the hypothesis testing that the chance corrected agreement is equal to the agreement by chance (the agreement between the marginal probabilities for each tester to give an abnormal rating by only taking into account the covariates), that is, their true difference is 0. The column “Estimate” shows the estimate of their true difference, “t” holds the t‐statistic and the “p value” shows the significance level of the one‐sided t‐test p value.
Abbreviations: CDT, cold detection threshold; CPT, cold pain threshold; HPT, heat pain threshold; MDT, mechanical detection threshold; MPT VF265, mechanical pain threshold van Frey hair weighting 265 mN; LoF/GoF, loss/gain of function; MPT PP, mechanical pain threshold pinprick; MSK‐HQ, musculoskeletal health questionnaire; NPSI, neuropathic pain symptom inventory, PPT, pressure pain threshold; VDT, vibration detection threshold; WDT, warm detection threshold.
Further details on proportion of specific agreement relating to whether sensory modalities were rated as abnormal or normal can be found in Table S1.
4. DISCUSSION
This study investigated the inter‐tester and intra‐tester reliability of a comprehensive, time‐efficient, and low‐cost CST battery in a sample of patients with spine‐related leg and arm pain. We found that most CST modalities achieved moderate reliability (inter‐tester: WDT, VDT, and CPT; intra‐tester: WDT, MDT, MPT VF256 (LoF), MPT PP, VDT, HPT, MPT VF 256 (GoF), and PPT), while four modalities reached substantial agreement (inter‐tester: CDT and PPT; intra‐tester: CDT and CPT). We found poor inter‐ and intra‐tester agreement for WUR as well as fair inter‐tester agreement for five CST modalities (MDT, MPT (LoF), MPT PP, HPT, and MPT (GoF)), which challenges their value in this patient population.
4.1. Inter‐tester reliability
A clinical test must have sufficient inter‐tester reliability to be valid as a screening tool in clinical practice. Overall, inter‐tester reliability ranged substantially between sensory modalities. We found that three sensory modalities (WDT, VDT, and CPT) reached comparable levels and two (CDT and PPT) achieved better inter‐tester reliability than previously reported in similar CST batteries among patients with neuropathic pain (Reimer et al., 2020; Wasan et al., 2020). Notably, this is the first study to investigate the reliability of thermal thresholds using the TipTherm® device. Arguably, the TipTherm® might be more time‐efficient than other devices to screen for thermal detection thresholds as it does not need time to adjust the target temperature. Five sensory modalities (MDT, MPT PP, MPT VF256 (LoF and GoF), HPT, and WUR) were found to have worse inter‐tester reliability than previously reported (Schmid et al., 2009; Vroomen et al., 2000; Wasan et al., 2020). Several reasons may have led to this divergence. First, Schmid et al. (2019) examined MDT exclusively in the upper extremity, which has increased tactile sensitivity compared to the lower limb (Ackerley et al., 2014). Second, more rigorous training, including a certification exam, might add to the higher inter‐tester reliability found in Wasan et al. (Wasan et al., 2020). Third, we used lower temperatures to detect HPT compared to other studies (Reimer et al., 2020; Wasan et al., 2020). Potentially, slightly higher temperatures of 45°C to 47°C create stronger heat hyperalgesia and therefore achieve higher inter‐tester reliability (0.62 and 0.55, respectively) than the more subtle heat (40°C) used here. Promisingly, we found moderate to substantial inter‐tester reliability for several sensory modalities (WDT, VDT, CPT, CDT, and PPT), which further supports their usefulness as screening devices to detect somatosensory dysfunction in populations with mixed pain mechanisms. However, poor to fair inter‐tester reliability was found for MDT, MPT PP, MPT 256VF (LoF and GoF), HPT, and WUR, which questions their clinical value.
4.2. Intra‐tester reliability
Good intra‐tester reliability is important to monitor changes over time. Except for WUR, all CST modalities showed moderate to substantial intra‐tester reliability. These levels of agreement are in accordance with previous research (Koulouris et al., 2020; Wasan et al., 2020). However, we found poorer reliability for WUR than previously reported (Nothnagel et al., 2017; Schmid et al., 2019). One possible explanation is that we included a patient population whereas previous research was performed in healthy participants which are expected to show more homogenous sensory function. Of note, our analysis for WUR was underpowered as we had to exclude a high number of participants from WUR analysis (n = 36 of 53 for inter‐tester and n = 29 of 52 for intra‐tester reliability) due to zero ratings on the single stimulus. Alternative analysis of TSP using all participants by calculating differences of ratings did not improve reliability. The toothpick we used for WUR was not sharp enough to elicit a painful response in every participant. In line with this, our previous work suggested that a toothpick is inferior to a Neurotip in identifying small‐fibre degeneration (Ridehalgh et al., 2018).
While we did identify a statistically significant improvement of NPSI and MSK‐HQ scores from T1 to T2, clinically, NPSI change was very small (2.88 out of 100) and MSK‐HQ change (2.57 out 56) was below the minimal important change (8.5) (Karstens et al., 2021) and the agreement of the testers could not be explained by these two covariates alone. This sensitivity analysis further lends confidence in our intra‐tester reliability results.
Taken together, the results of this study show moderate to substantial intra‐tester reliability except for WUR, supporting the usage of most CST modalities to monitor change of somatosensory function in clinical practice and research in populations with spine‐related limb pain of mixed pain mechanisms.
4.3. Limitations and further directions
Several limitations need to be considered. Although we included the pre‐specified sample size of 53 participants, one participant missed the third testing session, and we found spine‐related arm pain underrepresented (n = 12) compared to spine‐related leg pain (n = 41). This limits the generalizability of our results in this population and also increases the degree of uncertainty regarding the true reliability values of the measurements for spine‐related arm pain. Therefore, it may be necessary to conduct further research with a larger sample size to obtain more robust reliability estimates specifically for spine‐related arm pain.
In addition, we have opted to dichotomize the outcomes for all CST parameters except WUR. While this is useful as a screening tool, it limits the CST's ability to quantify the degree of sensory dysfunction, which would be useful for comparison between patients or within patients over time. Optimally, CST modalities would need to allow quantification (e.g. NRS for pain to quantify gain of function measurements or utilizing a tuning fork with a scale), as has been implemented in other CST batteries (Koulouris et al., 2020; Reimer et al., 2020).
Blinding of the participants and examiners from T1 to T2 could not be achieved, possibly introducing recall bias and impacting the results of intra‐tester reliability. Given the high number of tests, we consider it unlikely that participants or investigators could remember each rating. Future studies could include checks for blinding, for instance, by asking the examiner before T2 to guess the outcome of the sensory tests combined with the certainty of the guess (Kolahi et al., 2011). In addition, recall bias could be reduced using a longer time interval between T1 and T2. However, given that our patient population received intensive physiotherapy management and indeed improved over the short testing period, we did not consider it appropriate to extend the time interval between the testing sessions.
Further studies are required to investigate whether CST modalities are sensitive to change as well as their possible prognostic ability. Notably, it has been shown that QST modalities like temporal summation and thermal pain detection thresholds have moderate correlation to pain and disability outcomes respectively (Georgopoulos et al., 2019). In future research, it will be valuable to determine whether CST modalities have prognostic abilities comparable to those of QST. This is particularly relevant as CST is more clinically feasible. Several studies have investigated different aspects of validity and reliability of different CST batteries and synthesizing them (e.g., meta‐analysis) will help to determine the optimal CTS battery and estimate their value as a cost‐effective alternative for QST in different populations.
4.4. Conclusion
This study shows that the inter‐tester reliability of a low‐cost and time‐efficient CST battery varies from poor to substantial depending on the sensory modality. This variability in inter‐tester reliability suggests that only some modalities are likely to be useful as screening tools. In contrast, the CST battery has moderate to substantial intra‐tester reliability, supporting its potential application in clinical practice and research to monitor different sensory modalities in this patient population.
AUTHOR CONTRIBUTIONS
Cedric Bender is the Principal Investigator of this study. Cedric Bender conceived and designed the planning, implementation, and analysis of the study. Cedric Bender collected the data as an examiner. Cedric Bender performed the analysis and wrote the manuscript. Annina B. Schmid is the Supervisor of this study. Annina B. Schmid was involved in designing the study, assisted in the implementation, and analysis. Sven Karstens was involved in designing the study, assisted in data storage, management, and analysis, and had the role of data protection officer. Fabian Muth collected the data as an examiner. Georgios Baskozos contributed a data analysis tool (two stage logistic regression model for analysing the intra‐tester agreement). All listed authors discussed the results and commented on the manuscript.
FUNDING INFORMATION
A.B. Schmid is supported by a Wellcome Trust Clinical Career Development Fellowship (222,101/Z/20/Z) and the Medical Research Foundation (Emerging Leaders Prize in Pain Research). Georgios Baskozos is supported by Diabetes UK under grant agreement 19/0005984 and by the Wellcome Trust under grant reference 223,149/Z/21/Z. The research was supported by the National Institute for Health Research (NIHR) and the Oxford Health Biomedical Research Centre (BRC). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.
DISCLAIMER
This research was funded in whole, or in part, by the Wellcome Trust [222,101/Z/20/Z, 202747/Z/16/Z]. For the purpose of Open Access, the author has applied a CC BY public copyright licence to any Author Accepted Manuscript version arising from this submission.
Supporting information
Data S1.
ACKNOWLEDGEMENTS
We would like to thank Adrian Keller for his involvement in data collection as an examiner as well as Lucy Ridgway for her contribution to the final version of the manuscript.
Bender, C. , Karstens, S. , Muth, F. , Baskozos, G. , & Schmid, A. B. (2024). Reliability of a clinical sensory test battery in patients with spine‐related leg and arm pain. European Journal of Pain, 28, 1366–1377. 10.1002/ejp.2267
REFERENCES
- Ackerley, R. , Carlsson, I. , Wester, H. , Olausson, H. , & Backlund Wasling, H. (2014). Touch perceptions across skin sites: Differences between sensitivity, direction discrimination and pleasantness. Frontiers in Behavioral Neuroscience, 8, 54. 10.3389/fnbeh.2014.00054 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Attal, N. , Perrot, S. , Fermanian, J. , & Bouhassira, D. (2011). The neuropathic components of chronic Low Back pain: A prospective Multicenter study using the DN4 questionnaire. The Journal of Pain, 12, 1080–1087. 10.1016/j.jpain.2011.05.006 [DOI] [PubMed] [Google Scholar]
- Baad‐Hansen, L. , Pigg, M. , Ivanovic, S. E. , Faris, H. , List, T. , Drangsholt, M. , & Svensson, P. (2013). Chairside intraoral qualitative somatosensory testing: Reliability and comparison between patients with atypical Odontalgia and healthy controls. Journal of Orofacial Pain, 27, 165–170. 10.11607/jop.1062 [DOI] [PubMed] [Google Scholar]
- Baron, R. , Maier, C. , Attal, N. , Binder, A. , Bouhassira, D. , Cruccu, G. , Finnerup, N. B. , Haanpää, M. , Hansson, P. , Hüllemann, P. , Jensen, T. S. , Freynhagen, R. , Kennedy, J. D. , Magerl, W. , Mainka, T. , Reimer, M. , Rice, A. S. C. , Segerdahl, M. , Serra, J. , … Treede, R.‐D. (2017). Peripheral neuropathic pain: A mechanism‐related organizing principle based on sensory profiles. Pain, 158, 261–272. 10.1097/j.pain.0000000000000753 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bouhassira, D. , Attal, N. , Fermanian, J. , Alchaar, H. , Gautron, M. , Masquelier, E. , Rostaing, S. , Lanteri‐Minet, M. , Collin, E. , Grisart, J. , & Boureau, F. (2004). Development and validation of the neuropathic pain symptom inventory. Pain, 108, 248–257. 10.1016/j.pain.2003.12.024 [DOI] [PubMed] [Google Scholar]
- Colloca, L. , Ludman, T. , Bouhassira, D. , Baron, R. , Dickenson, A. H. , Yarnitsky, D. , Freeman, R. , Truini, A. , Attal, N. , Finnerup, N. B. , Eccleston, C. , Kalso, E. , Bennett, D. L. , Dworkin, R. H. , & Raja, S. N. (2017). Neuropathic pain. Nature Reviews. Disease Primers, 3, 17002. 10.1038/nrdp.2017.2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Dyck, P. J. , Overland, C. J. , Low, P. A. , Litchy, W. J. , Davies, J. L. , Dyck, P. J. , O'Brien, P. C. , Cl vs. NPhys Trial Investigators , Albers, J. W. , Andersen, H. , Bolton, C. F. , England, J. D. , Klein, C. J. , Llewelyn, J. G. , Mauermann, M. L. , Russell, J. W. , Singer, W. , Smith, A. G. , Tesfaye, S. , & Vella, A. (2010). Signs and symptoms versus nerve conduction studies to diagnose diabetic sensorimotor polyneuropathy: Cl vs. NPhys trial. Muscle & Nerve, 42, 157–164. 10.1002/mus.21661 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Finnerup, N. B. , Haroutounian, S. , Kamerman, P. , Baron, R. , Bennett, D. L. H. , Bouhassira, D. , Cruccu, G. , Freeman, R. , Hansson, P. , Nurmikko, T. , Raja, S. N. , Rice, A. S. C. , Serra, J. , Smith, B. H. , Treede, R. D. , & Jensen, T. S. (2016). Neuropathic pain: An updated grading system for research and clinical practice. Pain, 157, 1599–1606. 10.1097/j.pain.0000000000000492 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Georgopoulos, V. , Akin‐Akinyosoye, K. , Zhang, W. , McWilliams, D. F. , Hendrick, P. , & Walsh, D. A. (2019). Quantitative sensory testing and predicting outcomes for musculoskeletal pain, disability, and negative affect: A systematic review and meta‐analysis. Pain, 160, 1920–1932. 10.1097/j.pain.0000000000001590 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hill, J. C. , Kang, S. , Benedetto, E. , Myers, H. , Blackburn, S. , Smith, S. , Dunn, K. M. , Hay, E. , Rees, J. , Beard, D. , Glyn‐Jones, S. , Barker, K. , Ellis, B. , Fitzpatrick, R. , & Price, A. (2016). Development and initial cohort validation of the Arthritis Research UK Musculoskeletal Health Questionnaire (MSK‐HQ) for use across musculoskeletal care pathways. BMJ Open, 5, 6. https://10.1136/bmjopen‐2016‐012331 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Karstens, S. , Christiansen, D. H. , Brinkmann, M. , Hahm, M. , McCray, G. , Hill, J. C. , & Joos, S. (2021). German translation, cross‐cultural adaptation and validation of the musculoskeletal health questionnaire: A cohort study. European Journal of Physical and Rehabilitation Medicine, 56, 771–779. doi: 10.23736/S1973-9087.20.06054-2 [DOI] [PubMed] [Google Scholar]
- Kolahi, J. , Bang, H. , & Park, J. (2011). Towards a proposal for assessment of blinding success in clinical trials: Up‐to‐date review. Community Dentistry and Oral Epidemiology, 37, 447–484. 10.1111/j.1600-0528.2009.00494.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- Koo, T. K. , & Li, M. Y. (2016). A guideline of selecting and reporting Intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine, 15, 155–163. 10.1016/j.jcm.2016.02.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kottner, J. , Audigé, L. , Brorson, S. , Donner, A. , Gajewski, B. J. , Hróbjartsson, A. , Roberts, C. , Shoukri, M. , & Streiner, D. L. (2011). Guidelines for reporting reliability and agreement studies (GRRAS) were proposed. Journal of Clinical Epidemiology, 64, 96–106. 10.1016/j.jclinepi.2010.03.002 [DOI] [PubMed] [Google Scholar]
- Koulouris, A. E. , Edwards, R. R. , Dorado, K. , Schreiber, K. L. , Lazaridou, A. , Rajan, S. , White, J. , Garcia, J. , Gibbons, C. , & Freeman, R. (2020). Reliability and validity of the Boston bedside quantitative sensory testing battery for neuropathic pain. Pain Medicine, 21, 2336–2347. 10.1093/pm/pnaa192 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Landis, J. R. , & Koch, G. G. (1977). The measurement of observer agreement for categorical data. Biometrics, 33, 159. 10.2307/2529310 [DOI] [PubMed] [Google Scholar]
- Lin, C. W. , Verwoerd, A. J. , Maher, C. G. , Verhagen, A. P. , Pinto, R. Z. , Luijsterburg, P. A. , & Hancock, M. J. (2014). How is radiating leg pain defined in randomized controlled trials of conservative treatments in primary care? A systematic review. European Journal of, 18, 455–464. 10.1002/j.1532-2149.2013.00384.x [DOI] [PubMed] [Google Scholar]
- Lipsitz, S. R. , Parzen, M. , Fitzmaurice, G. M. , & Klar, N. (2003). A two‐stage logistic regression model for analyzing inter‐rater agreement. Psychometrika, 68, 289–298. 10.1007/BF02294802 [DOI] [Google Scholar]
- Mitchell, A. J. , Meader, N. , & Symonds, P. (2010). Diagnostic validity of the hospital anxiety and depression scale (HADS) in cancer and palliative settings: A meta‐analysis. Journal of Affective Disorders, 126, 335–348. 10.1016/j.jad.2010.01.067 [DOI] [PubMed] [Google Scholar]
- Mokkink, L. B. , Terwee, C. B. , Patrick, D. L. , Alonso, J. , Stratford, P. W. , Knol, D. L. , Bouter, L. M. , & de Vet, H. C. W. (2010). The COSMIN study reached international consensus on taxonomy, terminology, and definitions of measurement properties for health‐related patient‐reported outcomes. Journal of Clinical Epidemiology, 63, 737–745. 10.1016/j.jclinepi.2010.02.006 [DOI] [PubMed] [Google Scholar]
- Moloney, N. , Hall, T. , & Doody, C. (2015). Divergent sensory phenotypes in nonspecific arm pain: Comparisons with cervical radiculopathy. Archives of Physical Medicine and Rehabilitation, 96, 269–275. 10.1016/j.apmr.2014.09.015 [DOI] [PubMed] [Google Scholar]
- Nothnagel, H. , Puta, C. , Lehmann, T. , Baumbach, P. , Menard, M. B. , Gabriel, B. , Gabriel, H. H. W. , Weiss, T. , & Musial, F. (2017). How stable are quantitative sensory testing measurements over time? Report on 10‐week reliability and agreement of results in healthy volunteers. Journal of Pain Research, 10, 2067–2078. 10.2147/JPR.S137391 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Petersen, K. K. , Vaegter, H. B. , Stubhaug, A. , Wolff, A. , Scammell, B. E. , Arendt‐Nielsen, L. , & Larsen, D. B. (2021). The predictive value of quantitative sensory testing: A systematic review on chronic postoperative pain and the analgesic effect of pharmacological therapies in patients with chronic pain. Pain, 162, 31–44. 10.1097/j.pain.0000000000002019 [DOI] [PubMed] [Google Scholar]
- R Core Team . (2020). R: A language and environment for statistical computing. R Foundation for Statistical Computing. https://www.R‐project.org/ [Google Scholar]
- Rastogi, R. , Rosedale, R. , Kidd, J. , Lynch, G. , Supp, G. , & Robbins, S. M. (2022). Exploring indicators of extremity pain of spinal source as identified by mechanical diagnosis and therapy (MDT): A secondary analysis of a prospective cohort study. The Journal of Manual & Manipulative Therapy, 30, 172–179. 10.1080/10669817.2022.2030625 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Reimer, M. , Forstenpointner, J. , Hartmann, A. , Otto, J. C. , Vollert, J. , & Klein, T. (2020). Sensory bedside testing: A simple stratification approach for sensory phenotyping. Pain Reports, 5, e820. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ridehalgh, C. , Sandy‐Hindmarch, O. P. , & Schmid, A. B. (2018). Validity of clinical small–Fiber sensory testing to detect small–nerve Fiber degeneration. Journal of Orthopaedic & Sports Physical Therapy, 48, 767–774. 10.2519/jospt.2018.8230 [DOI] [PubMed] [Google Scholar]
- Rolke, R. , Baron, R. , Maier, C. , Tölle, T. R. , Treede, D. R. , Beyer, A. , Binder, A. , Birbaumer, N. , Birklein, F. , Bötefür, I. C. , Braune, S. , Flor, H. , Huge, V. , Klug, R. , Landwehrmeyer, G. B. , Magerl, W. , Maihöfner, C. , Rolko, C. , Schaub, C. , … Wasserka, B. (2006). Quantitative sensory testing in the German research network on neuropathic pain (DFNS): Standardized protocol and reference values. Pain, 123, 231–243. 10.1016/j.pain.2006.01.041 [DOI] [PubMed] [Google Scholar]
- Rotondi, M. A. , & Donner, A. (2012). A confidence interval approach to sample size estimation for interobserver agreement studies with multiple raters and outcomes. Journal of Clinical Epidemiology, 65, 778–784. 10.1016/j.jclinepi.2011.10.019 [DOI] [PubMed] [Google Scholar]
- Schmid, A. B. , Adhikari, K. , Ramirez‐Aristeguieta, L. M. , Chacón‐Duque, J. C. , Poletti, G. , Gallo, C. , Rothhammer, F. , Bedoya, G. , Ruiz‐Linares, A. , & Bennett, D. L. (2019). Genetic components of human pain sensitivity: A protocol for a genome‐wide association study of experimental pain in healthy volunteers. BMJ Open, 9, e025530. 10.1136/bmjopen-2018-025530 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Schmid, A. B. , Brunner, F. , Luomajoki, H. , Held, U. , Bachmann, L. M. , Künzer, S. , & Coppieters, M. W. (2009). Reliability of clinical tests to evaluate nerve function and mechanosensitivity of the upper limb peripheral nervous system. BMC Musculoskeletal Disorders, 10, 11. 10.1186/1471-2474-10-11 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Scholz, J. , Mannion, R. J. , Hord, D. E. , Griffin, R. S. , Rawal, B. , Zheng, H. , Scoffings, D. , Phillips, A. , Guo, J. , Laing, R. J. , Abdi, S. , Decosterd, I. , & Woolf, C. J. (2009). A novel tool for the assessment of pain: Validation in Low Back pain. PLoS Medicine, 6, e1000047. 10.1371/journal.pmed.1000047 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Sommer, C. , Richter, H. , Rogausch, J. P. , Frettlöh, J. , Lungenhausen, M. , & Maier, C. (2011). A modified score to identify and discriminate neuropathic pain: A study on the German version of the neuropathic pain symptom inventory (NPSI). BMC Neurology, 11, 104. 10.1186/1471-2377-11-104 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tampin, B. , Slater, H. , Hall, T. , Lee, G. , & Briffa, N. K. (2012). Quantitative sensory testing somatosensory profiles in patients with cervical radiculopathy are distinct from those in patients with nonspecific neck–arm pain. Pain, 153, 2403–2414. 10.1016/j.pain.2012.08.007 [DOI] [PubMed] [Google Scholar]
- Vodermaier, A. , & Millman, R. D. (2011). Accuracy of the hospital anxiety and depression scale as a screening tool in cancer patients: A systematic review and meta‐analysis. Supportive Care in Cancer, 19, 1899–1908. 10.1007/s00520-011-1251-4 [DOI] [PubMed] [Google Scholar]
- Vroomen, P. C. A. J. , de Krom, M. C. T. F. M. , & Knottnerus, J. A. (2000). Consistency of history taking and physical examination in patients with suspected lumbar nerve root involvement. Spine, 25, 91. 10.1097/00007632-200001010-00016 [DOI] [PubMed] [Google Scholar]
- Wasan, A. D. , Alter, B. J. , Edwards, R. R. , Argoff, C. E. , Sehgal, N. , Walk, D. , Moeller‐Bertram, T. , Wallace, M. S. , & Backonja, M. (2020). Test‐retest and inter‐examiner reliability of a novel bedside quantitative sensory testing battery in Postherpetic neuralgia patients. The Journal of Pain, 21, 858–868. 10.1016/j.jpain.2019.11.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zhu, G. C. , Böttger, K. , Slater, H. , Cook, C. , Farrell, S. F. , Hailey, L. , Tampin, B. , & Schmid, A. B. (2019). Concurrent validity of a low‐cost and time‐efficient clinical sensory test battery to evaluate somatosensory dysfunction. European Journal of Pain, 23, 1826–1838. 10.1002/ejp.1456 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1.