Do Visual Analogue Scales (VAS) and Numerical Rating Scales (NRS) have interval and ratio properties when used to assess pain intensity? A systematic review (protocol)

Emma Fletcher

Pain is a protective feeling that can only be assessed subjectively. Numerical rating scales and visual analogue scales that require respondents to rate their pain on a continuum from ‘no pain’ through to ‘the worst pain imaginable’ are widely used in experimental pain research.

Pain scales used in experimental research should provide statistically useful data (Hartrick, Kovan & Shapiro 2003). Ratio and interval properties are particularly relevant (Hartrick, Kovan & Shapiro 2003; Price et al. 1994) because they allow a scale to capture the magnitude or percentage of change, and allow for direct comparison across groups, within tests, across time, or between studies (Ferreira-Valente, Pais-Ribeiro & Jensen 2011; Price et al. 1994; Price et al. 1983). Price et al. (1983; 1994) demonstrated ratio properties in a mechanical VAS scale (but not a simple numerical rating scale), and that this is only true when specific instructions and endpoints are used.

It is often an assumption in experimental pain research that VAS and NRS scales have ratio properties, and analyse data accordingly. This systematic review aims to determine whether VAS and NRS have interval and ratio properties when used to assess pain intensity in experimental pain research and to establish the strength of evidence supporting the VAS and NRS scale for use in experimental pain research.

Methods:

At all stages of the review, two independent reviewers will be used. A third reviewer will be used for resolving disputes where discussion has been unsuccessful.

Electronic search:

A systematic search of the peer-reviewed scientific literature will be undertaken to identify studies that have investigated the psychometric properties of self-report scales used to measure pain in experimental research. A computer search of online bibliographic databases (Medline, CINAHL, Embase, Scopus, Web of Science) will be performed using the following search string: pain AND scale AND (ratio OR interval OR linear). Where possible, searches will be limited to English and human studies but no limits will be placed on date.

The reference lists of included studies will be searched, and notable authors will be contacted for expert opinion and to ensure key studies are not missed.

Studies identified following screening will be tracked via the afore mentioned databases and google scholar to identify relevant citations.

Screening:

There will be two screening phases (title/abstract and full-text). Both phases will be performed using Covidence software and records of the screening results will be kept in accordance with the Prisma Statement.

The first phase will examine the title and abstract and will include any experimental study that meets the following:
A) Examines pain intensity measured in response to any stimulus modality (e.g. Thermal, electrical stimulation, pressure), and;
B) Examines if the scale used has interval or ratio properties.

The second phase will examine the full-text and exclude studies that do not meet the following inclusion criteria:
A) English language, full text available
B) Primary experimental research study (Not systematic/literature/scoping review)
C) Assessing self-reported pain intensity using a visual analogue scale (VAS) or numerical rating scale (NRS).

Risk of Bias:

Included studies will undergo critical appraisal to determine their risk of bias and methodological quality. A critical appraisal tool developed by Bialocerkowski, A, Klupp, N, Bragge, P (2010) will be used. It will be piloted to ensure suitability and rigour. The key findings of interest are: reporting bias, rigour of methodology, generalizability, criterion and content validity, measurement bias, analytical bias

Data extraction:

Data will be extracted using a predetermined format. Data to be extracted will be:
A) Demographics – age, sex, population setting (clinical, sociodemographic), location, recruitment, inclusion/exclusion.
B) Sample – size, sample size statement, sampling strategy, generalizability, attrition
C) Modality – i.e. thermal, electrical stimulation, pressure, protocol, methodology, equipment.
D) Scale type – VAS or NRS.
E) Body region used in testing
F) Range of stimuli received.
G) Ratings given by participants and description on how rating was provided
H) Analysis – statistical testing, software,
I) Findings

References:

Bialocerkowski, A, Klupp, N, Bragge, P 2010, ‘How to read and critically appraise a reliability article’, International Journal of Therapy and Rehabilitation, vol. 17, no. 3, pp. 114-120.

Ferreira-Valente, M, Pais-Ribeiro, J & Jensen, M 2011, ‘Validity of four pain intensity rating scales’, Pain, vol. 152, no. 10, pp. 2399-2404.

Hartrick, C, Kovan, J & Shapiro, S 2003, ‘The numeric rating scale for clinical pain measurement: A ratio measure?’, Pain Practice, vol. 3, no. 4, pp. 310-316.

Price, D, McGrath, P, Rafii, A & Buckingham, B 1983, ‘The validation of visual analogue scales as ratio scale measures for chronic and experimental pain’, Pain, vol. 17, no. 1, pp. 45-56.

Price, D, Lowast, A, Bush, F, Long, S & Harkins, S 1994, ‘A comparison of pain measurement characteristics of mechanical visual analogue and simple numerical rating scales’, Pain, vol. 56, no. 2, pp. 217-226.