Development of a Core Outcomes Set for Peripheral Nerve Injury

Abstract

Purpose

The contemporary literature evaluating outcomes after peripheral nerve injury (PNI) does not rigorously and adequately address the domains (motor, sensory, function, and pain) experienced by patients. Our goal was to develop a core outcomes set (COS) to evaluate outcomes after PNI.

Methods

We adhered to recommended guidelines for COS development. Following a systematic review of the literature, we assembled a panel of experts and used a modified Delphi to assess the appropriateness of candidate measures to evaluate recovery after PNI. We convened 20 experts in PNI using two initial electronic surveys, one in-person meeting, and a final electronic survey. We arrived at consensus (≥70% of panelists) for required and recommended measures to evaluate outcomes after PNI.

Results

Our panel arrived at consensus for motor, sensory, function, and pain outcomes in patients after upper and lower extremity nerve injury. We designated the measures to use and the timing and administration of these measures.

Conclusions

We developed a COS that can be used by clinicians and researchers who evaluate patients with PNI. Our goal is to implement the COS in a unified manner, facilitating comparison in the literature as well as collaboration among centers.

Type of study/level of evidence

Diagnostic V.

Peripheral nerve injuries (PNI) can have long-lasting implications for patients, ranging from loss of physical function to more pervasive effects on quality of life. Although many patients improve after both nonsurgical and surgical treatment for PNI, there remains variability in both recovery and treatment strategies after PNI, underscoring the importance of ongoing rigorous research to better understand areas for improvement. However, a systematic review demonstrated that the recent PNI literature does not consistently capture outcomes that reflect the multiple domains affected by PNI (motor, sensory, pain, extremity function, and quality of life).¹ For example, motor recovery is most commonly reported, but domains such as pain and quality of life (which are reported by patients as most important) are least commonly reported.^1,2 This is despite contemporaneous concerted efforts in other areas of clinical investigation to report outcomes in a consistent and comprehensive manner through the use of core outcomes sets (COSs), defined as “an agreed standardized set of outcomes that should be measured and reported, as a minimum, in all clinical trials in specific areas of health or health care”.³ To this end, our purpose was to develop a COS to assess patients after PNI. Having standardized outcomes measurements would facilitate robust comparative analyses of treatment strategies and facilitate collaborative efforts among centers.

To develop a COS for PNI, we followed the standards of development described by Williamson et al⁴ and Kirkham et al.³ Our protocol, which we have previously described, was composed of two phases: (1) a systematic review of the literature to generate a comprehensive list of measures previously for patients with PNI; and (2) an expert consensus panel using Delphi methodology to attain a final COS for PNI.⁵ In this article, we describe the process and outcomes of the Delphi panel to develop a COS to assess patients after PNI.

Materials and Methods

We adhered to recommendations for development of a COS, as delineated by Williamson et al⁴ and Kirkham et al.³ We have previously published our plan for development of this COS.⁵ We performed a systematic review of the last 15 years of literature to fully understand contemporary assessment of outcomes after PNI.¹ This systematic review was registered on an international register of systematic reviews on February 21, 2024 (CRD42024511371). Based on this systematic review, we developed a list of candidate measures to present to a panel of experts.

We assembled a list of experts in PNI to recruit for participation in our Delphi panel. In accordance with best practices delineated by Veugelers et al⁶, we identified experts based on faculty lists at recent international meetings in PNI. We focused on individuals practicing in the United States and Canada to minimize variability based on geographic and cultural variations in practice. Given that there is no consensus on the number of participants to include on a Delphi panel, we based the size of our panel on previous Delphi panels and we aimed for 15–20 participants. In case some experts were not able to complete the assigned work and to ensure an adequate sample size, we invited 25 experts. The panelists were asked to complete a demographic questionnaire including years in practice, subspecialty training, and self-designated level of expertise based on the Giddins and Tang scale.⁷ We also performed a search on PubMed on January 28, 2025, for the panelists’ names and the term “nerve” to gather an estimate of the volume of indexed publications related to nerve injury.

Following approval from our institutional review board and participant informed consent, panelists were invited to participate in an electronic survey. The panelists graded each candidate measure on a scale of 1 (not critical to include in the core outcome set) to 9 (critical to include in a core outcome set). Panelists were instructed that ratings would be interpreted in the following manner: 7–9 is critical, 4–6 is important but not critical, 1–3 as not critical. Our criteria for establishing that a measure has reached consensus as being critical was ≥ 70% of the 20 panelists rate the measure 7–9. The first two rounds were administered by electronic survey. The third round was performed as an in-person discussion, during which the panel agreed to designate items as either “critical” to include in a COS or “recommended” to collect during outcomes assessment. The fourth round was administered by electronic survey.

Results

Of the 25 invited experts, 20 participated in the first round (electronic survey). The same 20 individuals participated in the second round (electronic survey). Sixteen of the 20 participated in a third round, which was in-person. Detailed meeting notes were distributed to the 20 original participants, all of whom participated in the fourth and final round (electronic survey). Participant demographics are included in Table 1. There were eight females among the 20 panelists. In total, 9 completed core training in plastic surgery, 1 in physical therapy, and 10 in orthopedic surgery. The mean years in practice for participants was 14 (range 3–41 years; median 10 years) Of the 20 participants, 5 classified themselves as Level 5 on the Giddins Tang scale for expertise; 8 classified themselves as Level 4; 5 classified themselves as Level 3; and 1 classified themselves as Level 2. The mean number of nerve-related PubMed publications was 45 (range: 8–178; median 38).

Table 1.

Consensus for Motor

Nerve	Decision for Core Outcomes Set
Axillary	Required: MRC + AROM shoulder abduction
Musculocutaneous	Required: MRC + AROM elbow flexion
Radial	Required: MRC + AROM elbow extension + wrist extension Recommend: MRC and AROM index metacarpophalangeal extension Dynamometry: grip (position 2)
Median	Required: MRC + AROM thumb interphalangeal flexion, palmar abduction Recommend: MRC and AROM middle finger flexor digitorum superficialis Required: Dynamometry: grip (position 2), lateral pinch
Ulnar	Required: MRC + AROM flexor digitorum profundus for small finger and first dorsal interosseous Required: Dynamometry: grip (position 2), lateral pinch
Sciatic	Required: MRC + AROM knee flexion Required: MRC ankle dorsiflexion extension, eversion Required: MRC ankle plantarflexion, inversion
Femoral	Required: MRC + AROM knee extension, hip flexion
Peroneal	Required: MRC ankle dorsiflexion extension, eversion
Tibial	Required: MRC ankle plantarflexion, inversion

Motor

In the first round, there was consensus that the Medical Research Council (MRC) grading of manual muscle testing and active range of motion (AROM) were critical to include in the core outcomes set. For upper extremity (UE), although not reaching consensus, there were enough respondents selecting grip strength and pinch strength as “critical” or “important but not critical” to consider it in the next round of voting. There was insufficient interest in EMG or compound motor action potential amplitude from nerve conduction studies as being critical to include, so these were not advanced to Round 2 of the voting. In the open-ended comments for Round 1, panelists suggested consideration of dynamometry for individual tested joint motions. This was evaluated by panelists in Round 2, and there was consensus that dynamometry-measured strength was not critical for all nerves with the exception of median and ulnar nerves. In free text comments, panelists frequently referenced the logistical challenges with performing dynamometry in predominantly clinical settings aside from grip and pinch. There was consensus that grip and lateral pinch strength measured with dynamometry was critical for assessment after median and ulnar nerve injury. In the third (in-person) round, the panel discussed the merits of including grip strength as a measure of radial nerve function. It was agreed upon that grip strength would be considered critical to outcomes assessment after radial nerve injury to facilitate simplicity among the three most distal UE peripheral nerves and the impact of lack of grip strength on quality of life. For lower extremity (LE) motor function, the panel reached consensus in Round 1 that MRC and AROM were critical to assess. In Round 2, the panel reached consensus that MRC and AROM for hip flexion and knee extension were critical to include for femoral nerve palsies; that MRC and AROM for knee flexion were critical to include for sciatic nerve palsies. In Round 3, the panel discussed that AROM for ankle motion was not practical or reliable to measure in predominantly clinical settings. Therefore, the following measures were deemed critical: sciatic nerve–MRC ankle dorsiflexion extension and eversion, MRC ankle plantarflexion and inversion; peroneal nerve–MRC ankle dorsiflexion extension and eversion; and tibial nerve–MRC ankle plantarflexion and inversion. All recommendations for motor assessment (for both UE and LE) provided by the panel in Round 3 reached consensus approval by the panel in Round 4 (Table 1). Appendix S1 (available online on the Journal’s website at https://www.jhsgo.org) includes recommended techniques (with photographs) for assessing motor strength.

Sensory

The panel reached consensus that static two-point discrimination was critical to include in a COS. There were numerous respondents who believed that monofilament testing and the Ten test warranted inclusion in a COS, so these were assessed in the subsequent round.⁸ There was insufficient interest in including moving two-point discrimination (2PD), measures of locognosia, sensory nerve conduction testing, or Tinel sign to warrant consideration of these measures in the subsequent round. Pinprick sensation was recommended for consideration in open-text comments in Round 1, so this was brought to the panel for Round 2. In Round 2, the panel agreed that pinprick sensation was not critical to include. In Round 2, there was sufficient interest to include monofilament testing and Ten test, although these did not reach consensus. In Round 3, panelists discussed that obtaining static 2PD for nerves other than median and ulnar nerves was not reliable with established normative values and discussed the use of the Ten test in the other UE nerves. The panel reached consensus that static 2PD was critical for assessment of median and ulnar nerves and that the Ten test was recommended for all UE nerves. The panel reached consensus that monofilament testing was not critical for UE nerve assessment but agreed to consider it for LE nerve assessment. Given the paucity of literature for LE nerve assessment and the presence of recommendations from a multidisciplinary panel convened by the American Diabetes Association, the panel agreed to follow recommendations that monofilament testing was critical for LE nerve assessment and that one other assessment (options: 128 Hz tuning fork, pinprick sensation, ankle reflexes, vibration perception threshold testing) would be recommended.⁸ These recommendations for UE and LE nerve sensory assessment reached consensus approval in Round 4 (Table 2). Appendix S1 includes instructions for UE and LE sensory assessment.

Table 2.

Consensus for Sensory

Nerve	Decision for Core Outcomes Set
Median and ulnar nerves	Required: Static two-point discrimination
Axillary, MC/LABC, Radial/SRN, median, and ulnar	Recommended: Ten test
All LE nerves	Required: Monofilament testing Recommended: one additional test (options listed: 128Hz tuning fork, pinprick sensation, ankle reflexes, vibration perception threshold testing)

LABC, lateral antebrachial cutaneous nerve; MC, musculocutaneous nerve; SRN, superficial radial nerve.

Function

In Round 1, the Patient-Reported Outcomes Measurement Information System (PROMIS) UE and LE function assessments both reached consensus as critical to include in a COS. The Disabilities of the Arm, Shoulder, and Hand (DASH) questionnaire did not reach consensus but was more favorably rated than the QuickDASH. The DASH was advanced to Round 2 and the QuickDASH was dropped from further evaluation. The Michigan Hand Questionnaire and Patient Specific Function Scale did not reach consensus, but there was sufficient interest in Round 2. In the open-ended comments, many participants expressed unfamiliarity with the Sollerman Hand Function Test and Moberg Pick Up Test, so links to descriptions of these evaluations were included in Round 2. For LE function, there was a similar lack of familiarity with the Walking Impairment Questionnaire and the Lower Extremity Functional Scale, so links to these questionnaires were included in Round 2. In Round 2, for UE, the DASH reached consensus as critical for inclusion, and there was consensus that the other measures were not critical for inclusion. For LE, there was consensus that the Walking Impairment Questionnaire and the Lower Extremity Functional Scale were not critical to include. In Round 3, the panel considered the prior designations as PROMIS UE and DASH as critical to assess after UE nerve injury and PROMIS LE as critical to assess after LE nerve injury and agreed with prior recommendations. It should be noted that there was discussion about survey burden to participants with inclusion of both the DASH and PROMIS UE, but this concern did not rise to the level of overriding the panel’s prior consensus. These recommendations were agreed upon in Round 4 (Table 3).

Table 3.

Consensus for Function

Nerve	Decision for Core Outcomes Set
UE nerves	Required: DASH and PROMIS upper extremity function
LE nerves	Required: PROMIS lower extremity function

Pain

In Round 1, all panelists agreed that the visual analog scale (VAS) for assessment of pain was critical for inclusion in the COS. There was sufficient interest in the PROMIS pain interference, PROMIS pain intensity, neuropathic pain scale, and cold intolerance measures to include in Round 2. Open-ended responses in Round 1 identified the douleur neuropathique and body diagram, so these were included in Round 2 (with links to their descriptions). In Round 2, PROMIS pain interference and PROMIS pain intensity reached consensus for inclusion. There was consensus that the other measures were not critical to include in the COS. In Round 3, there was discussion about the merits of including a body diagram, but the panel agreed that pain-focused researchers would likely select additional measures beyond the core measures selected by our panel. The panel’s recommendations from Round 3 were accepted in Round 4 (Table 4).

Table 4.

Consensus for Pain

Nerve	Decision for Core Outcomes Set
All Nerves	Required: VAS pain, PROMIS pain intensity, PROMIS pain interference

Other Measures

The open-ended comments in Round 1 demonstrated numerous panelists suggesting that quality of life measures should be considered, so these were included in Round 2. In Round 2, panelists evaluated the Euro-Qol 5D and Short Form-36 but did not come to consensus that these measures should be deemed as critical to include in a COS. In Round 3, panelists discussed the merits and drawbacks of including one or both of these measures in the COS but did not come to consensus that one should be included. This was agreed upon by panelists in Round 4.

Discussion

Outcomes after PNI are not consistently reported and may not fully reflect the patient experience. Our Delphi panel provides guidelines for outcome reporting for PNI that include the multiple domains affected by PNI, including motor and sensory function, pain, and limb/domain-specific patient-reported outcome measures.¹ Standardization of outcome measures to assess recovery after nerve injury or following peripheral nerve surgery is essential to ensure that comparisons can be made among studies and different patient populations.

For assessment of motor function, our Delphi panel agreed on the MRC grading scale and AROM at different joints as the outcome measures.⁹ MRC grading of muscle strength remains the most common measure used to assess motor function in patients with PNI or compression. Although this grading scale has been criticized for the subjective qualities of the examiner and the lack of standardized definitions, its widespread use and ubiquity in the literature are noted.⁹ To reduce inter-rater variability, standardization of grading and scoring MRC is provided through appended instructions (See Appendix S1). The inclusion of dynamometry for measurement of grip and lateral pinch strength as recommended outcome measures for assessment of median and ulnar nerve motor function follows previous reports in hand surgery and have been adopted for quantitative serial assessment of motor function.

For assessment of sensory function, our Delphi panel believes that static two-point 2PD is critical to include after median and ulnar nerve injuries. Static 2PD has been shown to be reliable and valid.¹⁰ However, like MRC grading, it is an imperfect measure, with an age-related decline in the ability to discriminate between two points.¹¹ The panel did not come to consensus that 2PD was appropriate for other PNI. For other UE PNI, the Delphi panel came to consensus that using 2PD was not critical to include for nonmedian/nonulnar upper extremity PNI and agreed that use of the Ten test was recommended (but not critical) for reporting sensation after axillary, lateral antebrachial cutaneous, and superficial radial nerve injuries. The Ten test was initially described by Strauch et al¹², has been noted to have excellent reliability parameters, and has been used widely after hand injuries.^13,14 For lower extremity PNI, review of the literature demonstrated that a multidisciplinary task force convened by the American Diabetes Association believed that monofilament testing was critical to include after LE nerve injury.⁸ The same task force also recommended use of one additional test (options included 128 Hz tuning fork, pinprick sensation, ankle reflexes, vibration perception threshold testing) in addition to the required monofilament testing. Our panel did not believe it was productive to deviate from these recommendations given the acknowledged heterogeneity in how LE PNI are treated, which clinicians treat them, and how outcomes and reported.

For assessment of UE function, the Delphi panel agreed on use of the PROMIS UE function and the DASH as recommended outcome measures.^15,16 In the hand surgery literature, these are the most commonly used patient-reported outcome measures. Although the DASH has some known disadvantages such as lack of specificity for assessment of peripheral nerve conditions, ambiguity with laterality and some antiquated functions in the questions comprising the questionnaire, it remains the most commonly used patient-reported outcome measure in the UE and is focused on evaluating the patient’s view of disability.¹⁷ PROMIS, developed by the National Institutes of Health, has greater precision, a larger range of measurement and fewer items than most conventional outcome measures.¹⁸ As such, adoption of PROMIS has increased dramatically in many disciplines in recent years, including after PNI.19, 20, 21, 22 During our Delphi panel process, there was discussion about the potential survey burden resulting from use of both DASH and PROMIS. Although the combined use of these instruments has not been examined in detail, it is the panel’s expectation that the two instruments will complement each other. Inclusion of the DASH will facilitate comparison to the historical literature, whereas PROMIS UE can protect against both ceiling and floor effects given its computer-adaptive nature. Furthermore, it is the lead author’s experience that both the DASH and PROMIS UE can be administered to PNI research participants in an acceptable amount of time without excessive respondent burden or fatigue.²³ For LE function, there was consensus that the PROMIS Lower Extremity scale was critical to include. This recommendation was largely informed by (1) the lack of a LE functional score that had previously been broadly adopted for use in patients after PNI and the (2) general sentiment that it was important to include a patient-reported measure of function. Other measures, such as Walking Impairment Questionnaire and the LE Functional Scale, were considered but not selected because they were developed and exclusively used after pathophysiological processes (such as arthritis and peripheral vascular disease) that materially differed from traumatic PNI.

To assess pain after PNI (both UE and LE), the panel thought that use of the VAS, PROMIS pain intensity, and PROMIS pain interference were critical. Prior systematic reviews have shown that pain is the domain that is most inconsistently reported after PNI, yet it is the domain that it is of paramount concern to many patients.^1,2,24,25 The VAS is the most commonly used scale after PNI and has been broadly used in other conditions, making its inclusion in this COS useful for comparison to the existing literature.¹ Pain intensity refers to how severe the pain is perceived, while pain interference reflects how the pain has affected the patient’s daily activities.²⁶ These concepts are thought to be complementary in understanding the patient experience of pain and are becoming increasingly adopted in the PNI literature.^23,27,28

Another important consideration for assessing outcomes after PNI is standardizing follow-up intervals to improve consistency in the reported literature. Although our panel recognized the importance of recommending specific follow-up intervals, we refrained from deeming these intervals as “critical”/required given the heterogeneity in surgeon/institution preferences and inconsistency in patient follow-up after trauma. With these considerations in mind, our panel recommended standardized follow-up at baseline (either injury or date of surgery), 3, 6, 12, 18 and 24 months. With peripheral nerve trauma or surgery, even in the best of circumstances, the rate of recovery is slow. Additionally, in trauma populations, loss to follow-up is a challenge in many centers. Although 3-month and 6-months visits may be too early to see meaningful improvement in function for some PNI, the panel agreed that these intervals are aligned with their clinical practices and would provide opportunities for establishing trajectory of recovery. The 12- and 24-month follow-up intervals were intended to maintain consistency with how outcomes have been reported in the published literature and allow assessment of mid to long-term outcomes after PNI.

The limitations of this study include the Delphi process which may introduce bias based on the selection and inclusion of the participants and the methodology.²⁹ To minimize the bias, our study included 20 peripheral nerve experts in the study from diverse specialty backgrounds, geographical region of practice and years in practice to provide a broad range of experience and perspective. Iterative, anonymous survey rounds were performed and controlled feedback was provided to panel members between rounds. We followed established guidelines for the development of a COS and published our study protocol a priori.⁵ In accordance with this protocol, a systematic review of the literature was performed and used to inform the Delphi consensus panel.¹ Three rounds were used to facilitate responses and a methodology expert with experience in designing and conducting Delphi panels was consulted throughout the process.⁵

Our goals in developing a COS for PNI are to improve the validity and reliability of outcome reporting. We designed this COS for broad adoption by PNI surgeons and will strongly encourage use at our institutions, with the goals of raising the scientific rigor of studies performed in our field and facilitating collaboration across centers.

Conflicts of Interest

Dr. Stepan reports receiving speaker fees from Checkpoint and research funding from Neuraptive. Dr. Dy reports receiving research support and speaker fees from Checkpoint Surgical and speaker fees from Johnson and Johnson/DePuy Synthes (speaker fees); he is on the medical advisory board of Orthocell and was an ad hoc consultant for Axogen (not ongoing). Dr. Kozin is a consultant for Biocircuit and Sientra. No benefits in any form have been received or will be received by the other authors related directly to this article.