Article Text
PDF
Abstract
In the field of acute pain medicine research, we believe there is an unmet need to incorporate patient related outcome measures that move beyond reporting pain scores and opioid consumption. The term “minimal clinically important difference” (MCID) defines the clinical benefit of an intervention as perceived by the patient, as opposed to a mathematically determined statistically significant difference that may not necessarily be clinically significant. The present article reviews the concept of MCID in acute postoperative pain research, addresses potential pitfalls in MCID determination and questions the clinical validity of extrapolating MCID determined from chronic pain and non-surgical pain studies to the acute postoperative pain setting. We further suggest the concepts of minimal clinically important improvement, substantial clinical benefit and patient acceptable symptom state should also represent aspirational outcomes for future research in acute postoperative pain management.
- analgesia
- pain
- postoperative
- pain perception
- regional anesthesia
- pain measurement
Statistics from Altmetric.com
Introduction
Subjective self-reporting remains the primary basis of pain assessment1 that is influenced to varying degrees by biological, psychological and social factors.2 This patient-reported outcome (PRO) is widely accepted as a valid and reliable tool to measure clinically important changes3 4 of patient care in clinical trials and longitudinal epidemiological studies. Pain is typically assessed by one-dimensional pain tools (such as visual analog scale, VAS or numerical rating scale, NRS) along with measures of postoperative opioid requirement consumption usually reported in cumulative oral morphine equivalents.5
Determination of the true value of an analgesic intervention often relies on statistical analysis of data derived from pain assessment in clinical studies. Some argue that a statistically significant improvement conferred by a perioperative analgesic intervention may not necessarily translate into a clinically significant (or meaningful) benefit.6 7 Thus, the term “minimal clinically important difference” (MCID) in pain was introduced8 to define the clinical benefit of an intervention as perceived by the patient and not simply a statistically significant difference in pain response detected in contemporary pain studies. Table 1 shows an overview of ancillary concepts and terminologies discussed in this article.
Definition of common terms
As part of modern day multimodal analgesic pathways, regional analgesia has been shown to reduce pain and opioid consumption with or without a concomitant reduction in opioid-related side-effects.9–11 Several systematic reviews and meta-analyses on new fascial plane blocks or local infiltration analgesia have reported statistically significant analgesic benefits of these recent regional analgesic techniques.12 13 However, quantifying the change in VAS or opioid consumption that represents clinically significant and meaningful patient-perceived improvements in analgesia remains a matter of ongoing discussion. In this article, we aim to review the concept of the MCID and its extrapolations, examine more closely the processes currently used to draw conclusions from clinical studies about regional analgesic effectiveness, address some of the unanswered questions regarding pain assessment and speculate on future directions and novel methodologies in conducting acute postoperative pain research.
Minimal clinically important difference
The term MCID was first coined by Jaeschke in 1989 originally in the context of chronic heart/lung disease improvement,8 and represents “the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side effects and costs, a change in the patient’s management (i.e., he/she would consider repeating the intervention if it were his/her choice to make again).”14 Simply put, MCID can be defined as “the smallest change that is important to patients.”15 This is an important concept as opposed to a statistically significant difference that represents a mathematical difference that does not necessarily translate into a clinically significant one, as it can be of little or no importance to the patient’s health or quality of recovery.14 This applies in particular to clinical studies involving large sample sizes, where statistically significant differences between groups may be found with very small differences that are clinically meaningless.6 16 This is why statistical significance is necessary, but not sufficient, to declare superiority of a measured outcome.17
Determination of the MCID
MCID pain studies have traditionally used anchor-based methods, distribution-based methods or a combination termed the “triangulation method.”18 19 The Delphi methodology (consensus after multiple rounds of opinions of a panel of individuals, often experts in a specific field, along with patients) can also be used to complement anchor or distribution-based methods (explained later).
Anchor-based methods compare the score change in a given PRO scale (e.g., NRS or VAS) to some other measure of change considered an anchor, for example, a subjective global assessment rating (GAR) scale (patient says that he/she feels worse, the same or better).20 Because a GAR scale is prone to recall biases, it is always necessary to determine the anchor’s validity and reliability by further using other quantifiable anchors for example, the amount of ingested pain medication.21 Nevertheless, in the absence of a gold-standard anchor, it is highly recommended to use multiple independent anchors22 to confirm responsiveness across different samples.22–25 An example of a widely used GAR26 and correlation with psychometric properties is shown in figure 1; for example, +2 and +3 represent minimal clinically important reductions in pain scores, that is, the patient feeling better after treatment.
15-point Global Assessment Rating scale with description of ratings and corresponding psychometric properties. MCID, minimal clinically important difference; MCII, minimal clinically important improvement; SCB, substantial clinical benefit; SCW, substantial clinical worsening.
MCID can also be calculated using the distribution-based methods based on a change in PRO scores using the SEM, the SD, the effect size (ES) or the minimum detectable change (MDC).14 Most commonly, 0.5 SD27 or 1 SEM28 are considered the MCIDs. However, on their own, distribution-based methods are not able to identify important, clinically meaningful outcomes because they lack an anchor that links the numeric scores to an assessment of what is considered important by the studied subjects.6 Thus, the validity of distribution methods remains controversial,6 16 often yielding statistically significant changes that paradoxically are not necessarily perceived as meaningful clinical changes by patients or clinicians.29
In the triangulation method, MCID can be determined by triangulating (averaging) values found using several complementary methodologies.24 For example, 0.5 SD, SEM, and 5%–10% of the instrument range can be calculated and averaged with distribution-based methods, and determination of the minimal change can be obtained with an anchor-based method.18 When combining these, the result is an MCID determination that is not method-bound and is presumably a more accurate reflection of clinically meaningful group-level change.30
Irrespective of the methodology used, the concept of MCID is necessary to provide an objective tool for evaluating (and comparing) the clinical impact of a therapeutic intervention as perceived by the patients and clinicians, and to ultimately guide treatment decisions in clinical practice.23 Similarly, MCID has been applied to assess quality of recovery.19 31 32
Limitations of MCID and its determination
Unfortunately, there are a number of pitfalls with MCID determination. For example, a different MCID value is generated using different methods of determination even for the same patient, thus the reason to use the combined “triangulation method.”18 19 Also, patient self-report of current pain might be judged against expectations rather than initial baseline level symptoms.33 Other limitations such as misunderstanding the context of improvement and retrospective judgements are always subject to recall bias.34
The MCID value is not a constant value even for the same pain condition; it changes with the severity of pain before treatment. That is, the MCID value is higher when a patient reports a higher pain score because a greater magnitude of pain relief to reach a comfortable state after treatment35 will be needed.
Finally, as its name suggests, the MCID is defined as a “minimal” rather than an “optimal” clinical benefit. Therefore, it should not be defined as the standard in clinical success, but the smallest change that has some form of perceived utility. Furthermore, repeatability, validity and test–retest reliability of the MCID currently applied in the postoperative acute pain setting have not been confirmed, pending more robust study data in the future. Because of the inherent limitations of the concept of MCID and its determination, other ancillary concepts are needed to provide a more comprehensive (and absolute) assessment of the effectiveness of a given analgesic intervention.
Extrapolation of MCID values from chronic/non-surgical pain to acute postsurgical pain
Due to the lack of field-specific data, extrapolation of MCID values from different patient populations seems to be a common denominator in acute perioperative pain research. For example, a recent meta-analysis by Hussain et al 36 concluded that pectoralis II block (PECS-II) is a beneficial analgesic technique for breast cancer surgery and is clinically superior to systemic analgesia alone based on an MCID value of −1.1 cm in the VAS for each pain assessment time point for the breast cancer population and opioid sparing of 27 mg of oral morphine equivalents in 24 hours. Unfortunately, the MCID value applied for this determination was extrapolated from chronic breast pain studies37 and not related to acute postoperative pain.
Another meta-analysis by Abdallah et al 38 examined the impact of single shot interscalene block on analgesic outcomes after shoulder surgery. They concluded that compared with the control group, interscalene block offers no demonstrable benefits for VAS pain score reduction at rest and opioid consumption after the 8th and 12th postoperative hours, respectively.38 The value of MCID in VAS pain score reduction used for this study (0.9–1.1 cm on a 10 cm VAS)38 was derived from studies of acute pain in the emergency department due to trauma17 and other non-surgical causes.39 The MCID value for acute pain after breast cancer or shoulder surgery is currently unknown, and so is the validity of MCID extrapolation from chronic and non-surgical pain conditions to acute perioperative pain.
Past MCID studies undertaken in chronic pain conditions of different types and different baseline pain intensities have found highly variable MCID values (figure 2).35 For example, MCID ranges from −8 mm for scleroderma40 to −82 mm out of 100 mm for trigeminal neuralgia.41 There are very few studies of MCID in the acute postoperative pain setting,18 19 and most available ones do not comply with existing standards17 42 43 as they have included mixed patient groups/surgical procedures,44 different analgesic interventions, pre-existing treatments and baseline pain.45 Additionally, past MCID studies have focused mostly on moderate or severe acute pain conditions with little information for mild pain conditions (VAS<3). At the present time, we do not know if MCID for acute pain is pain intensity-specific, procedure-specific or both. Given that MCID and responsiveness to treatment depend on population characteristics and study context, the MCID value may not be the same for all types of surgery and patients,23 similar to findings of chronic pain conditions. It is also possible that MCIDs for postoperative pain relief in patients with chronic pain (e.g., hip and knee osteoarthritis) are different from those without preoperative pain.
Variability of the minimal clinically important difference (MCID) (absolute, mean change for visual analog scale pain intensity scores) across different studies for chronic pain conditions.35 MCID value is shown in the upper portion of each bar.
MCID in opioid consumption
Currently there is no clear definition of the opioid sparing impact of an analgesic intervention to be considered clinically important or significant. Intuitively, a treatment minimizing opioid consumption, opioid related side effects in the short and/or long-term and favoring improvement in functional outcomes is considered clinically important or meaningful. While many regional analgesic studies report statistically significant reductions in opioid consumption, few have shown a concomitant reduction in adverse opioid-related outcomes. This raises the question: how many milligrams of change in opioid consumption constitutes a clinically significant and meaningful improvement? Furthermore, patient’s request for additional analgesic is often influenced by other factors besides uncontrolled pain. For example, it is only when a patient feels that the analgesic benefit of a pain treatment outweighs the potential side effects for example, nausea and vomiting, that a request for supplemental opioid is actually made.5 It has been clearly shown that patients are willing to trade-off pain relief for less upsetting and/or less severe side effects.46 Thus, the observed opioid-sparing effect might not be entirely attributed to effectiveness of an analgesic treatment.
Minimal clinically important improvement, substantial clinical benefit and patient acceptable symptom state in pain
Ideally, the goal of pain treatment is to make the patient feel good and not just to feel slightly better (MCID).47 As defined by Tubach et al, the minimal clinically important improvement (MCII) is another key concept that represents the smallest change in measurement that signifies an important improvement in a patient’s symptoms, sufficient enough to achieve a patient acceptable symptom state (PASS).47 That is, the patient now feels good and not just better after pain treatment for example, an improvement of >+4 (MCII)47 in contrast to +2 or +3 (MCID).19 Substantial clinical benefit (SCB) is an alternative psychometric value reflecting the amount of change associated with substantial or large patient-perceived meaningful improvement48 49 and may be considered equivalent to MCII. SCB thresholds can be calculated with the same anchor-based method used for MCID calculation. For example, a retrospective study in chronic pain by Martin et al found pain VAS MCID and change SCB scores 1 year after hip arthroscopy to be −15.0 and −22.7 mm, respectively, using anchor and distribution-based methods.50 MCID, MCII and SCB all represent a change in pain intensity after treatment but the magnitude of change varies (figure 3).
Visual representation of the concepts of MCID, MCII/SCB and PASS for a VAS pain scale, assuming: (A) 80/100 mm mean (i.e., high baseline pain) VAS pain intensity, an MCID of 10 mm and pass of 33 mm; and (B) 55/100 mm mean (i.e., moderate baseline pain) VAS pain intensity, an MCID of 8 mm and pass of 22 mm. The MCID (dark yellow) can be considered a subset of the MCII/SCB (dark + light yellow). The MCII/ SCB can be defined as the 75th percentile of the change in score between the baseline and final pain scores among patients describing improvements above “moderately better” (>+4) in a 15-point Global Assessment Rating after analgesic intervention. Final pain scores are represented in blue (31 mm VAS for (A) (high baseline pain), and 21 mm VAS for (B) (moderate baseline pain)). Note that MCID, MCII/SCB and PASS often change with different baseline pain intensities (A and B). In contrast to MCII/SCB and MCID, the PASS is an absolute value in the acute pain setting, often noted in the range of 30–40 out of 100 mm of a VAS pain scale. Initial pain intensity minus MCII/SCB should (roughly) equal PASS, as patients with final pain scores lower than the PASS have now achieved what they consider to be a “satisfactory pain score” (concept adapted from references44 47). MCID, minimal clinically important difference; MCII, minimal clinically important improvement; SCB, substantial clinical benefit; PASS, patient acceptable symptom state; VAS, visual analog scale.
The PASS represents a “satisfactory symptom score” which can be determined by the 75th percentile of the final score for patients who considered their pain relief satisfactory and well in themselves after an analgesic intervention (i.e., at least 75% of patients have achieved what they consider to be a “satisfactory pain score”).47 In contrast to MCII/SCB and MCID, the PASS is an absolute value in the acute pain setting, often noted in the range of 30–40 out of 100 mm of a VAS pain scale (i.e., mild pain range).18 Because PASS is an absolute value (“yes” or “no” answer), not a change from baseline, and is well validated in chronic pain studies, it can be a clinically relevant treatment target51 to be used in the acute postoperative pain setting. Thus, MCID can be considered a minimum target for outcome improvement, MCII and SCB optimal targets for outcome improvement, and PASS a comparatively higher outcome status rooted in the satisfaction domain52 (figure 3).
Future directions of research
Determining which statistically significant improvements translate into clinically significant (or meaningful) benefits should be prioritized in perioperative pain research, and our reliance on statistically significant outcomes must be averted. While numeric pain scores and opioid consumption are useful outcomes to measure, one may argue that a clinically meaningful change after an analgesic intervention represents a significant improvement in pain relief that is considered satisfactory or acceptable by the patient and not just a statistically significant difference. Preferably, interpretation of pain scores and opioid consumption should also take into consideration other patient comfort measures such as adverse side effects,53 quality-of-recovery,31 32 gastrointestinal recovery, mobilization and sleep quality.54 In the future, it may be wise to replace isolated pain scores with GAR scores, allowing for direct incorporation of the key clinical measurable metrics. The proportion of patients reaching MCII, SCB and PASS may be considered better goals for postoperative pain management strategies. Bearing in mind that the concepts of MCID, MCII, SCB and PASS have been mostly studied in chronic pain conditions and disability, it is now time to define these thresholds in acute perioperative pain and postoperative recovery settings.
Future pain research studies need to individually determine MCIDs for different surgeries and patient populations. Until MCID and ancillary values are determined for acute perioperative pain, studies using extrapolated MCIDs for chronic pain must clearly disclose the method of MCID determination and study context, which includes the type of chronic pain condition, patient characteristics and baseline pain level. Future determination of these values will likely allow a better tailored calculation of effect and sample sizes for future randomized controlled trials and a more contextualized interpretation of findings for future meta-analyzes in acute perioperative care. Furthermore, this will also help avoid research pitfalls deriving from the issue of the non-generalizability of the MCID, such as falsely accepting or rejecting research hypotheses. We believe adopting these newer approaches will eventually shed the necessary light for a more accurate assessment of what constitutes a meaningful patient-rated relief in perioperative pain.
References
Footnotes
Twitter @elboghdadly
Contributors FM-L drafted versions of the manuscript, provided figures and edited the manuscript. VC designed the concept of the manuscript, provided figures, drafted versions of the manuscript and edited the manuscript. KE-B provided figures and edited the manuscript. All authors contributed significantly to the elaboration of the final version of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.