Assessing pain intensity with the visual analog scale: A plea for uniformity

"Those who do not feel pain seldom think that it is felt." The sensation of pain is undeniably subjective. Personality, recollections of painful events, emotional state, age, culture, context, and other factors influence an individual's responses to, and description of, pain.1-5 Linguistically, the word "pain" encompasses multidimensional attributes well beyond a single sensation varying only in intensity.2, 3 Every patient's "pain vocabulary" contains nuanced connotations: aching, pulling, burning, stinging, stabbing, or radiating, to name a few.6 Quantifying "how much it hurts" eludes visual documentation (X-ray, MRI) or objective physiologic measurement (heart rate, blood pressure, hormone response, or other empirical information). No matter the clinical setting or circumstance, however, adequate pain management requires precise, thorough pain assessment—an individualized approach.4, 7-9 Assessing, treating, and monitoring pain is germane to every discipline—from general clinical practice through all subspecialties.10 Likewise, research assessing novel therapies and treatments pose additional challenges, particularly when changes in pain represent the primary outcome variable. The most credible expression of pain comes via each patient's self-report using their unique pain vocabulary.4, 6, 8 When cognitively and physically capable, the patient is the best source for accurately communicating the extent to which pain moderates, worsens, or varies over time.4, 6, 8 The European Palliative Care Research Collaborative, and scores of others charged with pain measurement, agree the most clinically relevant dimension to pain assessment, regardless of disease or condition, is pain intensity (PI).5, 7 A preponderance of evidence demonstrates that the 100 mm visual analog scale (VAS) is by far the most frequently used assessment instrument5, 8-11 to evaluate analgesic effects of various therapies and detect minute pain changes during analgesic administration. Structurally, we find the VAS a simple tool to use by anyone cognitively capable of understanding the parameters and responding to clinician instructions. Indeed its popularity is frequently attributed to the ease and convenience of the VAS in a fast-paced clinical setting.4 PI may be the most convenient dimension to assess quickly, but interpreting a VAS reading is no simple matter.12 Context plays an important role when translating a VAS score into an evidence-based treatment. One specialty treats nociceptive pain; another focuses on relieving patients suffering neuropathic pain.13 Hodgins2 and other authorities8 suggest that identical scores on a pain scale may be treated differently based on the specialty and individual patient.2 High pain scores may be termed acceptable immediately following traumatic injury, yet unacceptable if they persist over time.2, 3 The underlying cause of pain (surgery vs. migraine headache), and clinical setting (acute care vs. hospice) illustrates that no "one size fits all." Because VAS is an extremely common means of estimating and monitoring PI, it raises this conundrum: is VAS really the most effective assessment tool? Could flaws within the tool itself contaminate results? Or, might clinical interpretation of pain scores rob the VAS of consistency and reliability? We posit the tool and its clinical application combine to produce undependable, sometimes capricious results. Moreover, this variability and imprecision inherent in the standard VAS confounds extrapolation of findings from differing research investigations when applied to a global patient population. Ours is not an especially new or novel perspective. Authors representing diverse clinical and research domains have reported similar concerns over the decades.1, 2, 4, 7-16 Many clinicians and investigators consider the 100 mm continuum of the VAS (Figure 1A and B) superior in capturing subtle differences (sensitivity) compared to the few discrete points of categorical pain scales that limit responses to 4, 5, 7, or 11 choices8 (Figure 2A and B). Ironically, these investigators cannot reach consensus on which precise tick-mark on the 100 mm continuum corresponds to clinically relevant "moderate pain." 3, 5, 9, 14 Studies differ on what is considered a moderate pain threshold, beginning with >30 mm, up to >60 and even >75 mm.14, 17 Similarly, Stinson et al15 question whether a difference of 10 mm represents a meaningful change in PI. Williams et al8 report the meaning of identical tick-marks on a scale vary by patient. One individual's number expresses "average, everyday pain." A second patient considers the same VAS location as "barely tolerable."18 Several studies describe a "ceiling and floor" effect—a clustering of results around lower and higher endpoint descriptors.1, 8, 18 Our own recent use of the VAS19 supports the interpretive difficulties Stinson and colleagues cite.17, 18 This is further supported by investigators who point out that pain is not a linear phenomenon.15 For this reason, one may not assume a score of 50 is precisely two times that of 25.17, 18 Granted that at the bedside, the patient's report of increasing or decreasing PI trend should provide data on which to base individual clinical pain management decisions. Nevertheless, the advantages of more consistent, universal VAS score comparisons in clinical and research settings would broaden this assessment tool's accuracy and reliability. Equally important, greater consistency among VAS values enhances the scale's comparative applicability across pain treatment study publications. Many have attempted to enhance the precision and broader applicability of the standard VAS (Figure 2). Design and formatting variables among VAS pain scale versions include: (1) anchor terminology, (2) presence/absence of line markers, (3) interpreting the baseline and successive measures, and (4) the units of measurement and length of the scale, and (5) horizontal or vertical orientation.11 Some of these "modified VAS" instruments are shown in Figure 1C and 1D. We believe some modifications can and do enhance the value of the data derived. However, for extrapolation across differing patient populations and accounting for the tremendous heterogeneity in patient sensations and descriptions of PI, incorporating a few basic changes should result in a universally accepted VAS. One especially constructive modification would result from adopting identical and clinically universal anchoring terminology—the descriptive terms that provide user reference points. A review of published VAS tools reveals a spectrum of anchor descriptor terms. The differing, seemingly arbitrary, anchor point label verbiage interjects an important variation for patient and clinician. Universally accepted definitions or numerals to quantify PI beginning with "no pain" and escalating to "worst pain ever" (or other nomenclature) do not exist. In reality "worst pain ever" is infinite. It is always possible for a patient to experience more pain than they deemed the "worst ever" just 60 min ago.1 By forcing patients into selecting "worst pain ever" a second or third time, their response must be interpreted as "no change" even if that person's pain steadily increased.1 In fact, Williams quotes one study subject as saying, "having to put 100 would make me feel worse."8 Several less finite anchor labels are already in use: "unbearable pain," "excruciating pain," "very intense pain," and others. Even "no pain," although finite, is affected by each patient's pain threshold and subjective distinction between what they deem simply an unpleasant sensation versus clinically relevant pain.1 This lack of uniformity, precision, and consensus in the format and application of VAS within institutions, between medical practitioners, and even among two clinicians working side-by-side in the same department, renders relative clinical comparison extremely difficult. With the increasing demand for provision of evidence-based medicine, collaborative interdisciplinary care models, and trend toward comparative effectiveness research—we can, and must, do better in assessing PI. We posit that clinicians and researchers would be better served with a universal VAS construct and more consistent interpretive practices within the context of each specialty. We believe such consistency will bring clarity to published findings and yield more clinically relevant pain assessment data for investigators and clinicians alike and, most importantly, foster more accurate, and thus, clinically relevant data extrapolation across studies for applicability to the individual patient. Given The Joint Commission20 mandates pain assessment for all patients and the Institute of Medicine's Committee on Advancing Pain Research, Care, and Education21 identifies pain management as a "moral imperative," perhaps the time is ripe for a comprehensive multicenter cross-disciplinary study with significant power to establish consistency in how this instrument is administered and interpreted. The ultimate goal should be to accumulate sufficient evidence to support "gold standard" guidelines which are then "tweaked" to meet specific needs of each medical specialty, clinical setting, and research study design. Once accomplished, frequent and effective training in how to administer and interpret VAS scores according to the guidelines should be mandated to improve rater and interpretive consistency within and among varied clinical settings. Enhanced uniformity of pain assessments and interventions across disciplines can only benefit the science and, above all, the patients under our care. Although a substantial and time-consuming undertaking, it could prove well worth the pain! The authors have no financial interests to disclose.