Anesthesiology, Article

How much change in pain score does really matter to patients?

a b s t r a c t

Objective: The goal of this study was to determine the minimal change in pain score recognized by patients as meaningful known as minimal clinically important difference (MCID).

Methods: Pain was recorded upon admission, 30 and 60 min later and patients were asked to describe the extent of pain change on a 5-point Likert scale ranging from “much better” to “much worse”. Patients reported their pain by two common pain scales comprising Numeric rating scale (NRS) and visual analog scale . We used receiver operating characteristic curve to assess the accuracy of pain scales. We then calculated the mean change in pain scores among patients who reported their pain change as “a little better” or “a little worse” and also analyzed regression to evaluate the MCID.

Results: A total of 150 patients and 253 pain changes were recruited. The MCID +- SD (95% CI) was 1.65 +- 1.58 (1.32-1.97) for NRS and 16.55 +- 17.53 (12.96-20.15) for VAS. The area under the curve by NRS and VAS were

0.86 and 0.89. For linear regression, the line slope and the y-intercept were 17.56 and 1.88, for VAS; these values were 1.73 and 0.31 for NRS, respectively.

Conclusions: Recognizing the extent of change in pain score that really matters to patients is crucial for the evaluation of treatment effect. Patients perceived a change of 1.65 points on NRS and 16.55 on VAS in their Pain severity as meaningful. This value was not different whether the pain was perceived alleviated or aggravated.

(C) 2019

Introduction

Importance

Pain represents one of the most frequent complaints in the emergency department (ED). It is usually a presentation of a disease process and, at times, an indicator of illness severity. It has been argued that chronic pain may raise the rate of all-cause morbidity and mortality [1], and thus should be assessed and treated appropriately. Pain is an inherently subjective symptom. While it is inaccurate to rely only on the objective interpretation of healthcare providers about the patient’s pain, no direct objective method exists for its assessment from the patient’s perspective. Hence, Pain measurement primarily relies on the patients’ subjective reports [2].

Background

The most widely used pain rating scales in clinical practice are the 100-mm Visual Analog Scale and the 11-point Numeric Rating Scale . These scales are known to be sensitive, reliable and

* Corresponding author.

E-mail address: [email protected] (M. Jalili).

practical for the assessment of pain [3-6]. Apart from the absolute amount of pain, any change in its severity is also important in emergency practice, since it may signal a worsening condition or may indicate a successful treatment. It is crucial to consider the minimally important pain changes as an standard criteria after pain killer for patient satisfaction to continue or abandon a treatment [7]. While the commonly used instruments are able to measure small changes in pain score, the amount of change with clinical significance, also known as the minimal clinically important difference (MCID), is yet to be determined. MCID is the subjective declaration of minimal change in pain score which is recognizable as a meaningful change in pain status by patients [8]. Since pain perception is a complex phenomenon, identical pain scores and changes may be interpreted differently depending on the patient’s race, acuity of pain, personality, as well as social and cultural background. In addition, pain can represent different levels of threat for patients; threat of disease recurrence, loss of opportunity, or employment [9]. Therefore, the results of such studies need to be validated in different settings. Furthermore, VAS has been used to evaluate MCID in several studies, within many of which this assessment was a secondary objective [10-12], but published reports of the MCID for NRS are fewer and anchor-based scales are influenced by recall bias and previous pain scores [13,14]. On the other hand, the methods used for determining MCID varies in different studies; while

https://doi.org/10.1016/j.ajem.2019.158489

0735-6757/(C) 2019

some investigators have considered the mean difference in pain scores of patients who have reported a small change in their pain severity as the MCID, some have used other methods such as the opinion of an expert panel for this purpose [15].

Goal of this investigation

The goal of this study was to assess the accuracy of NRS and VAS pain scales and to determine the MCID in adult patients with acute pain in the ED, using various statistical methods.

Methods

The study was approved by the Institutional Review Board (IRB) of Tehran University of Medical Sciences. Informed consent was obtained from all the participants but written consent was not deemed necessary at the discretion of IRB with the research ethics approval ID- 8911307053. The manuscript adheres to the STROBE guideline.

Study design and setting

To determine the MCID of VAS and NRS, we designed a prospective study performed in two university-affiliated Tertiary care hospitals with more than 40,000 ED visits per year and staffed 24/7 with board- certified emergency medicine specialists and emergency medicine residents.

Selection of participants

Adult patients (18 years or older) presenting to our ED with any type of acute pain were included. There was no upper age limit. Patients were excluded if they were unwilling to give consent, unable to understand explanations given in Persian, unable to cooperate, suffering from altered mental status or Cognitive impairments. In addition, participants with significant visual, auditory and motor disabilities were excluded from the study because of the difficulty in checking the pain-related score.

Study protocol

Data was collected in three steps by one of the researchers (MB) through convenience sampling of participants with acute pain. Her shifts were spread evenly throughout both days, nights and weekends. Three episodes of pain were recorded upon first ED visit, after 30 and 60 min of their first episode of pain measurement. After 30 min and 60 min of the baseline pain, participants were asked to compare their current pain state with the pain they were experiencing when they expressed their pain score by choosing one of the following descriptors: “much better”, “a little better”, “about the same pain”, “a little worse” or “much worse”. At least 2 measurements were required to report a pain change and patients were excluded if they left the ED prior to declare a second pain episode. Patients were not reminded of their previous pain score or pain change. Treatment measures to control pain (such as pain medications, immobilization, nitroglycerine, etc.) were conducted according to the routine protocols of the ED or at the discretion of the treating physician. We also documented the patients’ age, gender, level of education, type of pain (traumatic or non-traumatic), and pain location (head and neck, torso, limb or generalized). The illogical values were defined as pain changes that perceived in the opposite direction of the reported pain scores (i.e. a patient reported the pain change as “a little less” although the corresponding pain score increased).

Analysis

Data was exported to SPSS, version 21.0 for analysis. Estimates were reported with the precision of 95% confidence intervals (95% CIs),

significance b 0.05 and type II error b 0.02. The MCID was defined as the mean difference between current and preceding scores rated either “a little better” or “a little worse” pain. Furthermore, the percentage of pain changes was calculated and reported separately. The Receiver Operating Characteristic (ROC) Curve was used to assess the accuracy of pain scales to assess pain changes. Moreover, regression was applied to evaluate the minimal clinically important difference and the extent of relationship between clinical change and pain score changes. Also, a subgroup analysis of data belonging to patients who had reported their pain score changes as “a little more” or “a little less” was performed separately.

Results

Characteristics of study subjects

Of 242 patients presenting with acute pain to our ED, 92 cases were excluded. The flow of participants in the study is depicted in Fig. 1.

The mean age of the participants was 32 years and 78.7% were male.

Most patients (68%) presented with pain that was not related to trauma. Limb injuries and abdominal pain were the two most common complaints among study population, accounting for 32 and 31% of the cases respectively. The level of education was reported as secondary school and college education in 38% and 32% of patients, respectively. All subjects were of the same race (Caucasian) and their primary language was Persian. Totally, 253 pain measurements were reported. In the NRS group, the mean difference between two pain scores in patients reporting their pain change as “a little better” and “a little worse” was 1.65 +- 1.58 (1.32-1.97). In the VAS group, the mean difference was 16.55 +- 17.53 (12.96-20.15). Table 1 defines the mean difference in NRS and VAS pain scores of patients based on their reporting of the change. Totally, 15 patients in the NRS group and 14 patients in the VAS group reported their perception of pain change in the opposite direction of the reported pain scores (illogical values).

Area under the curve (AUC)

The AUCs were 0.819 (0.758 – 0.879) for NRS and 0.830 (0.772 –

0.888) for VAS (Fig. 2).

The comparison of mean pain changes

The overall mean differences in pain scores are mentioned in Table 1 according to the extent of pain change. There was no statistically significant difference between a little better or worse pain changes (P

= 0.343 for NRS and 0.587 for VAS).

Regression

We used linear polynomial regression to estimate curves of VAS and NRS changes on pain changes (Fig. 3). For linear regression, the line slope and the y-intercept were 17.56 and 1.88, for VAS; these values were 1.73 and 0.31 for NRS, respectively.

There was moderate goodness of fit (R2: 0.34 for VAS, R2: 0.38 for NRS) in linear regression between VAS or NRS changes by pain changes. Table 2 demonstrates the MCID estimated by regression.

Regression was repeated omitting illogical figures. This time, the slope of the line and the y-intercept were 21.10 and 1.00, for VAS; these values were 2.13 and 0.16 for NRS, respectively. There was moderate goodness of fit (R2: 0.53 for VAS, R2: 0.53 for NRS).

Discussion

In this study, we sought to determine the MCID in pain score of patients in the ED. The MCID values determined in this study for NRS and VAS were mainly compatible with the previous studies. In order

Fig. 1. Flow chart of the study participants.

to evaluate the treatment effect, it is crucial to determine the extent of change in pain that matters to patients, known as MCID.

Analysis by ROC curve

Lauridson et al. used ROC curve to assess the sensitivity and specificity of pain scales to categorize the pain changes the patients declared. They reported higher responsiveness of low back pain rating scale compared to NRS for low back pain [16]. Angst et al.

reported the point of maximal sensitivity and specificity to be 15 mm on VAS using ROC curve [17]. A study by Kovacs et al. found this figure to be 0.5 for subacute and 1.5 for chronic neck pain by NRS [18].

Krebs et al. reported the ROC-AUC 0.76 for NRS scale considering pain that interferes with function [19]. ROC curve showed an acceptable sensitivity and specificity to determine differences in pain changes with NRS and VAS which suggests them as proper pain scales to find out little pain changes in our study.

Table 1

Mean difference in NRS and VAS pain scores of patients based on their reporting of the change in pain severity.

NRS

VAS

a ? 1 (mm)

b ? 2 (%)

a ? 1 (mm)

b ? 2 (%)

Mean

SD

95% CI

Mean

SD

95% CI

Mean

SD

95% CI

Mean

SD

95% CI

Much better (n = 48)

-4.60c

3.05

-5.49 – -3.72

71.57

43.38

58.77 – 84.37

-44.40

29.80

-53.05 – -35.74

74.42

40.64

62.35 – 86.49

A little better (n = 77)

-1.45

1.95

-1.90 – -1.01

40.34

45.20

30.08 – 50.60

-13.38

21.17

-18.18 – -8.57

39.04

58.11

25.85 – 52.23

No change (n = 107)

-0.16

1.10

-0.37 – 0.04

3.23

22.08

1.53 – 8.00

-1.50

11.39

-3.69 – 0.68

3.43

24.01

1.75 – 8.61

A little worse (n = 17)

0.88

1.16

0.28 – 1.48

17.90

27.72

3.13 -32.68

11.88

16.00

3.66 – 20.11

35.55

61.86

2.59 – 68.51

Much worse (n = 4) -1.00 3.00 -8.45 – 6.45 66.67 55.66 _ d 3.67

47.24

-113.70 – 121.04

62.50

71.14

20.14 – 92.79

NRS: Numeric Rating Scale; VAS: Visual Analog Scale; mm: millimeter; %: percent; SD: Standard Deviation; CI: Confidence Interval. n: number of patients in each group.

a ? 1: Absolute Change in Pain Score.

b ? 2: Relative Change in Pain Score.

c Negative values represent pain relief.

d Since there was only four patient in this group, CI is not reported.

Fig. 2. Receiver Operator Characteristic curves of “a little pain changes” measured with Numeric rating scale and Visual analog scale.

Fig. 3. (left) Linear estimation for VAS changes by clinical pain changes Linear regression formula: y = 1.884 + 17.562 x P value b 0.001 for all estimations (right) Linear estimation for NRS changes by clinical pain changes Linear regression formula: y = 0.315 + 1.73 x P value b 0.001 for all estimations.

Minimal clinically important difference

We found that a change in pain score equal to 1.65 for the NRS and

16.55 for the VAS is sufficient for our patients to report a change worth mentioning. This information may be helpful for Emergency practitioners to more accurately assess pain and show the effect of various treatments aimed at decreasing pain in the ED. Bijur et al. found the MCID for the NRS and VAS to be 1.3 and 1.4, respectively [10]. Todd concluded the MCID in VAS score to be 13 mm [20]. Myles

et al. reported a MCID of 10 for VAS and also assessed the patient acceptable symptom state (PASS) defined as the 25th centile of the VAS scores, based on the acceptable pain control in 33 patients after surgery [21]. Lee et al estimated the mean decrease in VAS for patients with adequate pain control at discharge to be 30 mm (95% CI 36.4 to 23.6) versus 5.7 (95% CI 11.2 to 0.3) for patients with inadequate pain control [22].The MCID was reported to be 13 mm with VAS (95% CI 10 mm to 17 mm) in acute trauma patients [20]. In a systematic review and meta-analysis, the MCID was derived from the studies of acute pain,

Table 2

Minimal Clinically significant difference analyzed by linear polynomial regression.

Equation

Illogic data

A little better

A little worse

A little change

VAS

Included

19.46

-15.69

17.57

Eliminated

22.10

-20.10

21.10

NRS

Included

2.05

-1.42

1.73

Eliminated

2.30

-1.97

2.14

NRS: Numeric Rating Scale; VAS: Visual Analog Scale.

post-operative pain and pain of chronic mechanisms such as rheumatic conditions, etc. This study reported a MCID ranging from 8 to 40 mm [23].

Effect of baseline characteristics on MCID

The MCID may vary according to the demographic factors of the participants and different disease processes. According to Bartley et al. women showed greater pain sensitivity and less pain inhibition, although the magnitude of gender differences varies across studies [24]. Kelly found no statistically significant difference between gender, age, or pain etiology in this context [25]. This conclusion was consistent with our findings which failed to show any significant difference between various demographic categories. However, higher sample size is needed to make subgroup analyses interpretable. A systematic review by Olsen et al did not report a significant difference between the MCID and baseline pain intensity but found a strong correlation between high baseline pain and larger pain changes [23]. Other researchers also concluded that the severity of pain did not influence the minimally important acute pain or chronic health condition change perceived by patients [11,25,26]. Seemingly, no statistically significant difference was found in our study between the severe and non-severe pain groups (data is not reported due to low power in some subgroup analyses).

Findings by polynomial regression

The scores of VAS and NRS were innovatively regressed by the corresponding clinical pain changes to verify the relationship between clinical pain changes and pain scores. There was a moderate goodness of fit for clinical pain changes of NRS and VAS. Eliminating the illogical pain changes, we reached better goodness of fit in VAS and NRS which depicts better fit of the regression model. Therefore, VAS and NRS are possibly appropriate tools for the determination of “a little change” in pain intensity. Several investigators have applied polynomial regression with this purpose and have analyzed various outcomes. Lauridsen et al. applied regression of a pain scale within each subpopulation and concluded that there was a correlation between the function loss due to low back pain and the Treatment outcome [16]. In a study on chronic pain, MCID was regressed on the patient acceptable symptom state and the authors found it useful to determine the patient’s function for daily practice [27]. Bijur et al. also regressed VAS on NRS scores and pointed out a strong correlation between the two scales [10].

In conclusion, this study shows that our patients needed a change in pain score of 1.65 +- 1.58 by NRS and 16.55 +- 17.53 by VAS to be able to perceive a change in pain severity. This extent of change in pain score that matters to patients can be considered as MCID for the evaluation of treatment effect in pain reduction. There was no difference in MCID between “a little better” and “a little worse” groups.

Limitations and recommendations

In the case of analgesic administration, each patient may receive several treatments over a relatively Short period of time due to the intermittent and recurrent nature of breakthrough pain. Similar to most related studies, pain scores were asked at the patient entry and 30 and 60 min after, and pain changes were asked based on a 5-item Likert scale: much better or worse, little better or worse or the same although using other time frames and different scales may affect these findings. Of the participants, 79% were male. Subgroup comparisons require higher sample sizes to have sufficient power. This can be addressed in future studies. In addition, we did not record the exact time and type of the administration of pain control treatments or their temporal relation to the time of documentation of pain episodes. The presence of recall bias can alter the meaning and analysis of MCID.

Contributions

MB and MJ conceived the study, designed it and supervised the conduct of the study and collect data. MB, MJ and AS undertook recruitment of participants and managed the data, including quality control. MT, MJ and AS provided statistical advice on study design and analyzed the data; MB, MJ, HM drafted the manuscript, and all authors contributed substantially to its revision. MJ takes responsibility for the paper as a whole.

Disclosures

MB, AS, HM and MJ report no conflict of interest.

Funding/grant

None.

Acknowledgment

The authors would like to thank Dr. Akbar Fotouhi for critically reviewing the manuscript and his insightful comments.

References

  1. Torrance N, Elliott AM, Lee AJ, Smith BH. Severe chronic pain is associated with increased 10 year mortality. A cohort record linkage study. Eur J Pain. 2010;14: 380-6.
  2. Ferraz MB, Quaresma MR, Aquino LR, et al. Reliability of pain scales in the assessment of literate and illiterate patients with rheumatoid arthritis. J Rheumatol. 1990;17:1022-4.
  3. Peters ML, Patijn J, Lame I. pain assessment in younger and older pain patients: psychometric properties and patient preference of five commonly used measures of pain intensity. Pain Med. 2007;8:601-10.
  4. Ferreira-Valente MA, Pais-Ribeiro JL, Jensen MP. Validity of four pain intensity rating

    scales. Pain. 2011;152:2399-404.

    Bahreini M, Jalili M, Moradi-Lakeh M. A comparison of three self-report pain scales in adults with acute pain. J Emerg Med. 2015;48:10-8.

  5. Mehrpour O, Karrari P, Abdollahi M. Chronic lead poisoning in Iran; a silent disease. Daru. 2012;20(1):8 In.
  6. Copay AG, Subach BR, Glassman SD, et al. Understanding the minimum clinically important difference: a review of concepts and methods. Spine J. 2007;7:541-6.
  7. Lee JS. Pain measurement: understanding existing tools and their application in the emergency department. Emerg Med (Fremantle). 2001;13:279-87.
  8. Turk DC, Okifuji A. Psychological factors in chronic pain: evolution and revolution. J Consult Clin Psychol. 2002;70:678-90.
  9. Bijur PE, Latimer CT, Gallagher EJ. Validation of a verbally administered numerical rating scale of acute pain for use in the emergency department. Acad Emerg Med. 2003;10:390-2.
  10. Gallagher EJ, Liebman M, Bijur PE. Prospective validation of clinically important changes in pain severity measured on a visual analog scale. Ann Emerg Med. 2001;38:633-8.
  11. Kelly A-M. The minimum clinically significant difference in visual analogue scale pain score does not differ with severity of pain. Emerg Med J. 2001;18:205-7.
  12. Farrar JT, Pritchett YL, Robinson M, et al. The clinical importance of changes in the 0 to 10 numeric rating scale for worst, least, and average pain intensity: analyses of data from clinical trials of duloxetine in pain disorders. J Pain. 2010;11:109-18.
  13. Kendrick DB, Strout TD. The minimum clinically significant difference in patient- assigned numeric scores for pain. Am J Emerg Med. 2005;23:828-32.
  14. McGlothlin AE, Lewis RJ. Minimal clinically important difference: defining what really matters to patients. JAMA. 2014;312:1342-3.
  15. Lauridsen HH, Hartvigsen J, Manniche C, et al. Responsiveness and minimal clinically important difference for pain and disability instruments in low back pain patients. BMC Musculoskelet Disord. 2006;7:82.
  16. Angst F, Aeschlimann A, Angst J. The minimal clinically important difference raised the significance of outcome effects above the statistical level, with methodological implications for future studies. J Clin Epidemiol. 2017;82:128-36.
  17. Kovacs FM, Abraira V, Royuela A, et al. Minimum detectable and minimal clinically important changes for pain in patients with nonspecific neck pain. BMC Musculoskeletal disorders. 2008;9:43.
  18. Krebs EE, Carey TS, Weinberger M. Accuracy of the pain numeric rating scale as a screening test in primary care. J Gen Intern Med. 2007;22:1453-8.
  19. Todd KH. Clinical versus statistical significance in the assessment of pain relief. Ann Emerg Med. 1996;27:439-41.
  20. Myles PS, Myles DB, Galagher W, Boyd D, Chew C, MacDonald N, Dennis A. Measuring acute postoperative pain using the visual analog scale: the minimal clinically important difference and patient acceptable symptom state. Br J Anaesth. 2017;118(3):424-9. https://doi.org/10.1093/bja/aew466.
  21. Lee JS, Hobden E, Stiell IG, Wells GA. Clinically important change in the visual analog scale after adequate pain control. Acad Emerg Med. 2003;10:1128-30.
  22. Olsen MF, Bjerre E, Hansen MD, et al. Pain relief that matters to patients: systematic review of empirical studies assessing the minimum clinically important difference in acute pain. BMC Med. 2017;15:35.
  23. Bartley EJ, Fillingim RB. sex differences in pain: a brief review of clinical and experimental findings. Br J Anaesth. 2013;111:52-8.
  24. Kelly AM. Does the clinically significant difference in visual analog scale pain scores vary with gender, age, or cause of pain? Acad Emerg Med. 1998;5:1086-90.
  25. Jones PW. Interpreting thresholds for a Clinically significant change in health status in asthma and COPD. Eur Respir J. 2002;19:398-404.
  26. Tubach F, Ravaud P, Martin-Mola E, et al. Minimum clinically important improvement and patient acceptable symptom state in pain and function in rheumatoid arthritis, Ankylosing spondylitis, Chronic back pain, hand osteoarthritis, and hip and knee osteoarthritis: Results from a prospective multinational study. Arthritis Care Res (Hoboken). 2012;64:1699-707.

Leave a Reply

Your email address will not be published. Required fields are marked *