Reliability Measures of Subcutaneous Pressure Pain Threshold Measurements: A Proposed Method of Assessing Painful Musculoskeletal Disorders

A randomised cluster sampling technique was employed to establish the sample of the study. A total of 30 (17 males) subjects attending the Kinesiology module in the Technological Educational Institute of Lamia, Greece, were recruited. Sampling was based on a random numbers selection (generated by a computer software application) from the total population of students attending the modules (approximately 120). Only one subject was excluded from the study due to excessive levels of depression and anxiety (see exclusion criteria). Demographics of the sample are presented in Results.

Literature suggests 30-35 subjects for reliability designs in biomedical research when α=0.05 and β=0.20 [22]. On these lines a pragmatic sample of 30 subjects was included fulfilling the criteria. Specifically, the exclusion criteria selected for this study were based on findings concerning PPT rather than sPPT studies, since there are no sPPT related studies in the literature. The specific criteria are detailed below:

Pain complaints and/or treatment at the shoulder/neck area within the last 3 months (as self-reported by the examinees).

Concurrent signs of malaise or fever, involuntary weight loss, rheumatic disease, cancer or psychiatric history (as self-reported by the examinees).

Clinically apparent major depression [23], (As assessed by the HADs scale and the clinical opinion of the examiner). One female was excluded due to this criterion.

Known endocrine or metabolic disorders which are not controlled by the given medical treatment.

Severely hypotensive or hypertensive (blood pressure) individuals.

Previous recent (within 6 months) open wound in the cervical spine or upper extremity (as self-reported by the examinees), only on the basis of severe trauma or surgery

Narcotic intake within one month of the study and analgesic or muscle relaxants within 24 hours of the study (as self-reported by the examinees).

Hypaesthesia or dysesthesia of the skin – permanent or temporary (as self-reported by the examinees in the general questionnaire and assessed by the examiner during initial visit).

Previous experience of subjects with pressure algometry. (to avoid practice effect bias) [24].

The concept of the study was explained and all subjects completed a consent form prior to participation in the study. All parts of the study were developed within the principles and standards of the Declaration of Helsinki and in accordance with the Guidelines on the Practice of Ethics Committees in Medical Research Involving Human Subjects. The study was designed and approved in accordance with the ethical procedures at the TEI of Lamia, Greece.
 

A prospective test-retest within-subject repeated measures design was used. In order to assess short-term reliability, subjects were measured on two consecutive days ("Day 1" and "Day 2"). The assessment of the longer-term reliability required another set of measurements one week after the "Day 1" assessment ("Day 7"). The sPPT measurements were administered by an independent examiner, whereas the initial assessment of the subject was performed by the main investigator (GG). The sPPT measurement involved "pinching" of the skin and subcutaneous tissue without including any muscular tissue.

Protocol: The protocol that was followed for all measurements ("Day 1", "Day 2", "Day 7") is described. Initially, the subjects were given two questionnaires to complete. These two questionnaires (a general one and another to assess psychological distress – see instruments) served to exclude subjects that fulfilled any of the exclusion criteria. The primary investigator recorded blood pressure variables and further examined the subjects for other exclusion criteria (e.g. dysesthesia of the skin). The measuring sites (see later) were then marked on the subjects and the complete procedure was explained. The pressure algometer device was then calibrated according to the procedure described by the manufacturer (by using a pre-determined load and adjusting the algometer accordingly). Practice measurements were then attempted until the subject and the examiner were familiar with the technique. A site different from those included in the main trial was selected for the practice session (e.g. the subject’s thenar muscles).

All sPPT measurements were recorded in the same order to prevent order effect bias on the results [25]. Recordings were repeated three times with a five minutes interval between each [26]. Recordings were taken bilaterally with the left side measured first followed by the right side in half of the subjects with the opposite order applied to the rest of the sample, as determined randomly by a coin flip. The mean average of the last two measures was taken as the best estimate of the sPPT. The first measure was discarded since it is shown to usually provide the values with the greatest variance [27].

Standardised instructions were given before each measurement on all occasions. Subjects were instructed to "report as soon as the sensation of pressure changes to pain by pressing the special button attached to the algometer". Usually all procedures were completed within 30–40 min minimising the effect of fatigue [28] and preventing subjects’ loss of concentration [29]. Subjects were kept uninformed of their scores throughout the study to prevent previous scores from influencing the results [26]. All measurements took place in the same location (research laboratory), by the same examiner, between 10:00 – 14:00 hrs in order to avoid diurnal variation [30]. The environment (temperature, draughts, cold, dampness) was kept as steady as possible throughout the study in order to control for potential effects on the pain threshold as suggested by Hildebrandt et al. [31]. The examiner remained blind to the previous sPPT values when performing the measurements in order to eliminate the possible effects of examiner's expectancy [32] and to follow the guidelines of successful blinding in reliability studies [33].

The order of sPPT measurements and its topographical location follows:

The T0 point, T2 point and T3 point (Figure 1). The upper trapezius points (T0, T2, T3). Subjects were seated with their arms hanging freely by their sides.

the Ulna point (Figure 2)

the Biceps point (Figure 2)

the Triceps point (Figure 2)

the Gastrocnemius point (Figure 2)

Instruments (Device - Questionnaires): The Pressure Algometer: The pressure algometer used in this study was an electronic Somedic type II device (Figure 3). The stimulation unit of the Somedic is gun-shaped with the stimulation tip situated at the end of the barrel, connected to a pressure transducer built into the handle. The measurement units are in kPa/cm². The accuracy of measurement is 3% and the range used in the current study was from 0-1000 kPa/cm² (the capacity of the device is up to 2000 kPa/cm²). The device is constructed with a visual display (built into the handle in the form of horizontal flashing light bars), ensuring a smooth and controlled rate of pressure application. The Somedic® algometer is equipped with a push-button that “freezes” the digital display value for 10 sec, when the PPT or sPPT is reached i.e. when the subject pushes the button. This innovation allows for accurate pressure recording of the PPT and sPPT measures [34,35].

The sPPT measurements require that a special attachment is attached to the algometer to allow pinching of the skin and the subcutaneous tissue (Figure 3) [36].

The recording of the sPPT requires that the examiner “pinches” the skin and subcutaneous tissue using this device whilst avoiding any muscular tissue and gradually increasing the pressure with a steady rhythm. When the feeling of pressure changes to discomfort (very early pain – "first pain") the subject presses the button, indicating that the pain threshold has been reached. The absolute value of the measurement depends on several technical characteristics of the algometer and the conditions of the procedure.

Fischer [37] has suggested measuring sPPT using the same parameters as for measuring PPT. However, from unpublished pilot work, practical experience and feedback from pilot subjects the following parameters were indicated for measuring sPPT:

A constant application rate of approximately 10 kPa/sec in order to allow enough time for the pain sensation to build up steadily and the subject to respond accurately. This rate of application allowed approximately the same response time (3-5 sec) as the typical 50 kPa/sec does for the deeper PPT, which is considered an appropriate time interval.

Selection of the smaller of the available tip-sizes (diameter=0.8 cm). This diameter corresponds to a surface of 0.50 cm². List et al. showed that for PPT measurements, the smaller sizes rendered larger PPT measurements possibly due to a spatial summation phenomenon. It is highly likely that this will apply for sPPT measurement as well. Using a smaller surface also permits more precise localisation of the subcutaneous tissue of interest. It would be interesting however, if the manufacturer had provided a smaller surface for the tipsize, to assess Fischer’s assertion of a 0.2 cm² appropriateness to record skin tenderness [34].

Questionnaires: Two self-reported questionnaires were used in this study: a general and a psychometric one. Both questionnaires were self-reported and administered to the participants prior to the initiation of the study. Each questionnaire took approximately five minutes to complete.

The general questionnaire addressed demographic information and questions regarding the exclusion criteria. The subjects were also asked questions regarding systematic diseases, hypertension and hypotension, and other relevant features. The purpose of this general questionnaire was to identify the appropriate subjects for inclusion in the study.

The psychometric questionnaire: The role of depression and anxiety is relatively clarified in the literature in the determination of PPT and sPPT readings, in the sense that subjects with high levels of psychological distress may respond inappropriately to the examination of PPT and sPPT. In order to identify these individuals, the self-reported Hospital Anxiety and Depression scale (HAD) was administered to all subjects [37]. This questionnaire is considered an adequate alternative to assessing depression when a formal psychological interview is not available [38]. The Greek version of the HAD [39] has evidenced the similar properties of the original version and has been used extensively since its validation (HAD-GR). The HAD-GR consists of 14 questions, seven of which are designed to assess the state of anxiety (HAD-A subscale) while the others provide an insight into the depression levels of the individual (HAD-D subscale). Only the depression sub-scale was analyzed for the purpose of exclusion from the study. Each question according to the answer provides a score from 0-3. Thus, the possible score for depression (7 questions) can range from 0-21. Zigmond and Snaith [37] have suggested a threshold level of eight plus in order to identify potentially pathological cases due to depression. This suggestion has been confirmed by Bjelland et al. [38] in an extensive literature review. In this study a cut-off value of 8 plus was adopted in order to exclude potentially pathological cases. Using this cut-off point, one suspicious case of anxiety and depressive symptomatology was identified and the person was excluded and given advice to seek help to a specialist.
 

The statistical package SPSS® 10.1 was used for all statistical analyses. The normal distribution of data was assessed using the Shapiro – Wilk test. The Shapiro-Wilk test was selected over the Kolmogorov-Smirnov in order to control for the small sample size (N<50). The descriptive analysis of the variables included the arithmetic mean (mean value), the standard error of the mean (S.E.M.), the standard deviation (S.D.) and the range of measurement (minimum and maximum values).

Intraclass Correlation Co-efficient (ICC) was selected as the appropriate procedure to assess the intra-rater reliability of the sPPT measurements across time (on consecutive days, and after a week) [40]. Specifically, the type ICC (2,1) was used for all intra-rater calculations as suggested by Sim and Wright [40] (p335), since the assessor was the same for all ratings and therefore was considered as a fixed effect (intra-rater reliability) and the mean value of each day measurements was employed in the calculations. Two ICC values were calculated for each measuring site in this study. The first ICC (Day 1 – 2) expresses the reproducibility of measurements for consecutive days (short-term reliability), while the second ICC (Day 1 – 7) describes the longer-term reliability (one week apart). The short-term and longer-term reliability for each measuring site were further visually inspected with scatterplots [41].

Delitto and Strube [42] believe that the ICC's take into account "level" differences, but are not true measures of concordance. Thus, Domholdt [41] suggested that: "…researchers who report reliability on the basis of an ICC should still report the results of an absolute reliability indicator such as the standard error of measurement or the repeated measures ANOVA" (p.287). Thus, an ANOVA model (General Linear Model – Repeated Measures) was further applied to control for potential differences as part of the ICC estimate model.

In order to enhance the rigor of the applied statistic model, two further statistical analyses were performed: the standard error of measurement (SEM) [43] and the smallest detectable difference (SDD) [44]. The SEM derives from the ANOVA error components, is based on the within-subject error/variability, is a measure of the "precision" of measurement, is expressed in the same units as the original measurement and therefore can be directly compared against subjects’ own values [43]. The standard error of measurement (SEM) is a standard deviation of measurement errors and the most frequently calculated statistic for this purpose [41]. It is often estimated as follows:



where σ: the standard deviation of the repeated measurements

r: the correlation coefficient (in this study the ICC value)

The SDD is a derivative of the SEM according to the formula:



which can be expressed either in the units of the measurement or as a percentage of the parameter's grand mean. The SDD is a “clinical index” which indicates the level of change in a parameter attributed with 95% certainty to a true change in a subject's condition, instead of being caused by test-retest errors [44]. In a repeated measurements design, the smaller the SDD the more responsive to change a measurement is rendered.