QuickDASH and PRWE Are Not Optimal Patient-Reported Outcome Measures After Distal Radial Fracture Due to Ceiling Effect

Background: The aim of this study was to determine the floor and ceiling effects for both the QuickDASH (shortened version of the Disabilities of the Arm, Shoulder and Hand [DASH] questionnaire) and the PRWE (Patient-Rated Wrist Evaluation) following a distal radial fracture (DRF). Secondary aims were to determine the degree to which patients with a floor or ceiling effect felt that their wrist was “normal” according to the Normal Wrist Score (NWS) and if there were patient factors associated with achieving a floor or ceiling effect. Methods: A retrospective cohort study of patients in whom a DRF was managed at the study center during a single year was undertaken. Outcome measures included the QuickDASH, PRWE, EuroQol-5 Dimensions-3 Levels (EQ-5D-3L), and NWS. Results: There were 526 patients with a mean age of 65 years (range, 20 to 95 years), and 421 (80%) were female. Most patients were managed nonsurgically (73%, n = 385). The mean follow-up was 4.8 years (range, 4.3 to 5.5 years). A ceiling effect was observed for both the QuickDASH (22.3% of patients with the best possible score) and the PRWE (28.5%). When defined as a score that differed from the best available score by less than the minimum clinically important difference (MCID) for the scoring system, the ceiling effect increased to 62.8% for the QuickDASH and 60% for the PRWE. Patients who had a ceiling score on the QuickDASH and the PWRE had a median NWS of 96 and 98, respectively, and those who had a score within 1 MCID of the ceiling score reported a median NWS of 91 and 92, respectively. On logistic regression analysis, a dominant-hand injury and better health-related quality of life were the factors associated with both QuickDASH and PRWE ceiling scores (all p < 0.05). Conclusions: The QuickDASH and PRWE demonstrate ceiling effects when used to assess the outcome of DRF management. Some patients achieving ceiling scores did not consider their wrist to be “normal.” Future research on patient-reported outcome assessment tools for DRFs should aim to limit the ceiling effect, especially for individuals or groups that are more likely to achieve a ceiling score. Level of Evidence: Prognostic Level III. See Instructions for Authors for a complete description of levels of evidence.

Wrist Evaluation (PRWE), a joint-specific outcome tool, are the most commonly used measures in the literature to assess functional outcome after these injuries 1 . The QuickDASH score was established in 2005 2 and is a shortened version of the DASH questionnaire 3 .
The DASH and PRWE scores are reported to have comparable effect size, standardized response mean, and discriminant ability while the QuickDASH reduces the responder burden 4 .
They have also been extensively tested for validity, reliability, feasibility, and responsiveness in measuring changes after different interventions and injuries of the upper limb 2,4-6 and wrist and hand [7][8][9][10] . Although both have been found to be valid and reliable 11 , concerns have been raised about floor and ceiling effects 12,13 , which limit their ability to assess outcomes. When a score of 0 equates to the best outcome, as is the case with the QuickDASH and PRWE, the minimum score is said to represent a ceiling effect. A ceiling or floor effect may suggest that a PROM is not the optimal tool to detect clinically relevant and statistically significant differences for interventions. However, to date, the studies in this area have involved small cohorts, a specific treatment modality (e.g., surgery), and/or limited follow-up.
The aim of this study was to determine the floor and ceiling effects for both the QuickDASH and the PRWE when assessing the outcome following the management of a DRF. Secondary aims were to determine the degree to which patients with a floor or ceiling effect felt that their wrist was "normal" and if there were patient factors associated with a floor or ceiling effect.

Materials and Methods
T he study cohort was identified from an electronic patient record system. There were 1,380 patients >15 years old who presented to the study center with a DRF during a 12-month period (2016). The center is an academic major trauma center  16,2023 providing care for the population of a defined catchment area. Inclusion criteria were a DRF and residency in the catchment area during the study year. Patients were excluded if they were deceased, were no longer a resident in our catchment, resided in a long-term care facility, had documented cognitive impairment, or were detained in prison. The study was part of a large retrospective study that was approved under the local orthopaedic research database (REC Reference 16/SS/0026). Patients provided informed consent. There were 1,114 patients who met the inclusion criteria and were sent a postal survey, with 526 (47%) responding and making up the study cohort (Fig. 1). Non-responders were significantly younger and were more likely to be male (Table I).

Fracture Classification and Management
Fractures were classified with the AO/OTA classification system 14 by a single author (K.R.B.) using preoperative posteroanterior (PA) and lateral radiographs. Management was determined by the treating orthopaedic trauma surgeon. Nonsurgical management routinely involved 6 weeks in a below-the-elbow cast. Surgical management was defined as an operation within 4 weeks after injury.

Outcome Measures
Outcome measures included the QuickDASH 2 , PRWE 7 , and EuroQol-5 Dimensions-3 Levels (EQ-5D-3L) 15 . A Normal Wrist Score (NWS) was also calculated by asking the patients the question "Taking into consideration the look, feel, and function of your wrist, how normal do you think your wrist is?", to which they responded by marking an "x" on a visual analogue scale numbered 0 to 100, with 100 meaning the most normal and 0 meaning the least normal 16,17 . The QuickDASH includes an 11-item disability and symptom section (with each item scored from 1 to 5) and 2 optional sports/performing arts and work modules 2 . The main disability and symptom section is scored 0 to 100, with higher scores reflecting increased disability 2 . The PRWE was developed in 1996 and was later modified to include the hand/wrist 7,18 . It is a 15-item questionnaire that asks patients to rate pain or disability on a scale from 1 to 10. It is scored from 0 to 100, with lower scores equating to better outcomes 18 . The EQ-5D-3L includes 5 dimensions (mobility, self-care, usual activities, pain/discomfort, anxiety/ depression) graded on 3 levels (no problems, some problems, extreme problems) and is a recognized and validated measure of quality of life 15 . A 1-digit number is given for each dimension, and the 5 answers can be combined into a 5-digit number expressing the patient's general health state 15 . This number can then be converted into a single value, which in the United Kingdom ranges from 20.594 to 1.000 19 . Full health is represented by a score of 1; death, by a score of 0; and a state worse than death, by a negative score 19 .

Floor and Ceiling Effects
A floor effect is present when >15% of respondents score the worst possible value 20 . Similarly, a ceiling effect is present when >15% of respondents reach the best possible score 20 . For both the QuickDASH and the PRWE, a lower score equates to a better outcome, so a ceiling score was present if a patient scored 1272 16,2023 0 (best outcome) and a floor score, if they scored 100 (worse outcome).
A floor or ceiling score has recently been defined as a score that differs from the minimum or maximum score by less than the minimally clinical important difference (MCID) for the scoring system 21 . The MCID is defined as the smallest difference in score that results in what patients consider to be an important change 22 . For the QuickDASH, this is 15.91 6 . The MCID for the PRWE varies depending on the condition being studied, but for DRFs it is 11.5 9,23 . Therefore, for the QuickDASH, a ceiling score was defined as being within the range of 0 to 15.90 and a floor score was within the range of 84.10 to 100. For the PRWE, a ceiling score was defined as within the 0 to 11.4 range and a floor score was 88.6 to 100.

Statistical Analysis
Statistical analysis was performed using IBM SPSS Statistics (28.0). RStudio IDE (1.4.1717; RStudio) was used to produce kernel density plots. Descriptive statistics were used to summarize the distribution of the data. The Cronbach alpha was calculated as a measure of internal consistency 24 , with 0 as the best and 100 as the worst score for both PROMs. The Shapiro-Wilk test was used to assess the data distribution. As the data were not normally distributed, a Mann-Whitney U test was used. The chi-square test was used for categorical data. Binary logistic regression was used to measure the relationship between various patient characteristics and the achievement of a ceiling score for both the QuickDASH and the PRWE. A p value of £0.05 was defined as significant.

Source of Funding
While no external funding was provided directly for the study itself, the Scottish Orthopaedic Research Trust Into Trauma (SORT-IT), our research charity, supported the salary of our trauma research fellow (the first author).

Results
T he mean age of the cohort (n = 526) was 65 years (range, 20 to 95 years; standard deviation [SD], 15.9), and 421 (80%) of the patients were female (Table I). The most common mechanism of injury was a fall from a standing height (69%, n = 361), followed by a fall from greater than standing height (12%, n = 65), sports (14%, n = 72), and a road traffic accident (3%, n = 16). The majority of patients were managed nonsurgically (73%, n = 385. Most operatively managed patients underwent open reduction and internal fixation (ORIF) (n = 133 of 141). Six patients underwent closed manipulation and percutaneous Kirschner wire fixation, and 2 underwent bridging external fixation.

Floor and Ceiling Effects
All outcomes assessed showed a high internal consistency with acceptable Cronbach alpha values of ‡0.95 (Table II). A ceiling effect was observed for both the QuickDASH (22.3%) and the PWRE (28.5%) (Table III). A ceiling effect was more likely in the PWRE than the QuickDASH (odds ratio [OR] = 1.39, 95% confidence interval [CI] = 1.04 to 1.85, p = 0.024). The score distributions for the main QuickDASH, the optional sport/performing arts and work QuickDASH modules, and the PRWE were skewed to the right (Figs. 2-A, 2-B, 2-C, and 3). The ceiling effects were greater for the optional QuickDASH sport/performing arts and work modules, with 55.3% and 60.3% of patients having the best score, respectively.
When ceiling and floor effects were defined as differing from the best or worst possible score by less than the MCID, the ceiling effect increased to 60.0% for the PRWE score and 62.8% for the QuickDASH score (Table IV).

NWS Associated with Floor and Ceiling Effects
The NWS was available for 516 (98%) of the patients. The median NWS was 90 (IQR, 70 to 96) and the distribution was skewed (Fig. 4). In the groups of patients achieving ceiling scores in the QuickDASH and the PWRE, the median NWS was 96 (IQR, 90 to 99) and 98 (IQR, 91 to 99), respectively. In the group of patients achieving a score within one MCID of the ceiling score, the median NWS was 91 (IQR, 84 to 98) for the QuickDASH and 92 (IQR, 88 to 98) for the PRWE.
Factors Associated with a Ceiling Score (Defined as Best Possible Score of 0) Male sex (p < 0.001), younger age (p = 0.007), a dominanthand injury (p = 0.002), and better health-related quality of life according to the EQ-5D-3L (p < 0.001) were associated with achieving a ceiling score for the QuickDASH (Table V). Older age (p = 0.023), a dominant-hand injury (p = 0.017), better health-related quality of life according to the EQ-5D-3L (p < 0.001), and an AO/OTA type-B fracture (p = 0.016) were associated with achieving a ceiling score for the PRWE (Table V). When adjusting for confounding factors using logistic regression analysis, male sex (p < 0.001), a dominant-hand injury (p = 0.006), and better health-related quality of life (p < 0.001) were associated with achieving a ceiling score for the QuickDASH (Table VI). On logistic regression analysis for the PRWE, increasing age (p < 0.001), a dominant-hand injury (p = 0.038), an AO/OTA type-B fracture (p = 0.015), and better health-related quality of life (p < 0.001) were associated with achieving a ceiling score (Table VI).

Discussion
T his study found that the QuickDASH and PRWE demonstrate ceiling effects for patients with a DRF. Despite achieving the best possible functional outcome scores (ceiling effect) according to the QuickDASH and PRWE, some patients felt that their wrist was not normal compared with their pre-injury function. These findings suggest that these outcome scores are failing to detect dysfunction in some patients following DRF. The QuickDASH and the PWRE may not be the optimal PROMs for assessing upper-limb or wrist-specific outcomes following a DRF. These findings have implications for future studies of DRF outcomes. Future development of PROMs to assess DRF outcome should aim to limit ceiling effects, so that scores will be more sensitive to differences in outcomes.
The presence or absence of a ceiling effect depends on the definition used. The current literature examining the ceiling effects of the QuickDASH and PRWE in the context of upper-limb function typically defines them as scores equating to the best functional outcome (0 for both scores) 12,25,26 . More recently, this definition has been expanded to include scores that differed from the best available score by less than the MCID for the scoring system 27 . Given the reliance on the use of the MCID to both plan and interpret study findings, this may be the logical next step when considering potential ceiling effects. Unsurprisingly, our study found that expanding the definition increased the number of patients with a ceiling score. Furthermore, it is clear from the use of the NWS that not all patients achieving ceiling scores consider themselves to have a normal wrist in terms of look, feel, and function. This illustrates that even when patients have   16,2023 scores indicating optimal function according to the PRWE and QuickDASH, this may not be the case and there is clearly room for improvement in the PROMs used to assess this injury.
Ceiling effects are already known to be a particular issue in patients with higher functional demand, such as intercollegiate athletes 25,26 . Hsu et al. 25 found that 65% of 321 healthy intercollegiate athletes achieved a ceiling score for the DASH, more than double the proportion of patients in the current study, which may be expected in a young, healthy athletic population. In contrast, Kim et al. 12 reported that 23% of their patients achieved a ceiling effect for both the DASH and the PRWE at 1 year following surgery for a DRF, although their study was smaller, only included female patients aged ‡50 years, and all patients had undergone surgery. Our study demonstrated unacceptable ceiling effects in a much larger representative DRF population that was managed both operatively and nonoperatively and had outcomes measured at a mean of almost 5 years following injury 28 . Furthermore, current literature has shown that the proportion of patients achieving a ceiling score for both the PRWE and the DASH increases over time 13,29 .
We found better health-related quality of life and a DRF of the dominant limb to be associated with ceiling scores for the QuickDASH and PRWE. Prior studies examining the ceiling score in a DRF population did not investigate predisposing factors 12,13,25,26,29 . However, a study of normative population values showed males to be more likely to achieve ceiling scores 26 , a finding that was supported by our study. These data could be useful in future PROM development. Computer-adaptive designs such as the Patient-Reported Outcomes Measurement Information System (PROMIS) also have the potential to reduce these effects.
There is an increasing focus on patient-centered care, and large funding bodies recommend that researchers consider including PROMs as part of their outcome measures 30 . The PRWE is the primary outcome measure listed for three National Institute for Health and Care Research (NIHR)funded trials since 2013, two of which studied DRFs 31 , and the DASH and PRWE account for the majority of the primary outcome measures used in other randomized controlled trials examining interventions for DRFs. The observed ceiling effect reported in the current study of a representative DRF patient group may bring into question whether these are in fact the most appropriate outcome measures to use. Kim et al. 12 observed no ceiling effect when using the Modified Mayo Wrist Score (MMWS) for assessing DRF outcomes, although this score is physician-assessed, not patient-reported. Fang et al. 29 investigated the use of the Green and O'Brien score (Cooney modification) (CGNO) and the Gartland and Werley score (Sarmiento modification) (SGNW), and neither was thought to be adequate beyond 6 months post-injury, mainly due to a lack of responsiveness.  16,2023 Limitations of the present study include differences in demographic characteristics between the survey responders and non-responders. Responders were significantly more likely to be older and female. This is a recognized phenomenon 32 and the ceiling effect that we identified may actually be an underestimate because younger males, a group usually considered to have higher functional demands, were underrepresented. Additionally, responders were significantly more likely to have undergone surgical management. Subjects who underwent nonsurgical management may have had less severe injuries or lower functional demands and therefore did not require surgery, which may have caused selection bias. The retrospective study design also limited the ability to assess how scores changed over time, although the validity of pre-injury PROMs is debated 33 . In relation to the design, despite the number of patients in the study, there is also the possibility of restricted clinical variation and truncation in the cohorts at the extremes of the PROM scores and this could have impaired estimates of the floor and ceiling effects. The analysis of the floor and ceiling effects as defined by the MCID was limited by the lack of an established consensus in the literature regarding the MCID for these outcome measures in this patient group and more broadly by issues regarding how the MCID is determined in the trauma setting. Finally, it could be argued that a ceiling effect would be expected at a mean of almost 5 years post-injury. However, literature suggests that there is only a marginal improvement in PROMs past a year following a DRF 34 .

1276
In conclusion, this exploratory study found that the QuickDASH and PRWE demonstrate ceiling effects when used to assess the outcome of DRFs. A proportion of patients achieving ceiling scores do not consider their wrist to be "normal." Future work should aim to develop PROMs for  16,2023 these injuries that limit the ceiling effect, especially for patients with predisposing factors for achieving a ceiling score. n