How Is the Probability of Reporting Various Levels of Pain 12 Months After Noncatastrophic Injuries Associated with the Level of Peritraumatic Distress?

Abstract

Background

Psychological factors have been shown to be consistent predictors of chronic pain in people with musculoskeletal injuries. However, few prognostic studies have considered multiple risk factors including peritraumatic distress. In addition, previous research has not considered that the associations between peritraumatic distress and pain levels can vary across pain outcomes.

Question/purpose

To determine whether an easily measurable level of baseline distress is associated with pain levels 1 year after noncatastrophic traumatic injuries when the outcome to be assessed is not normally distributed.

Methods

This was a prospective cohort study. The data were captured from two cohorts in London, Ontario, Canada, and Chicago, IL, USA. Participants were adults with acute noncatastrophic (that is, not treated with surgery or hospitalization) musculoskeletal injuries of any etiology with various injury locations (60% [145 of 241] spinal and 40% [96 of 241] peripheral) that presented to local urgent care centers. Other inclusion criteria included English/French speakers and having no other disorder that would affect their pain levels. In total, between the years 2015 and 2018, 241 participants were recruited based on the inclusion criteria. Ninety-three percent (225 of 241) of participants provided baseline data, and of these, 48% (109 of 225) were lost to follow-up in 1 year. Based on a complete case approach, this study included 116 participants who ranged in age from 18 to 66 years and 69% (80) of whom were women. The Traumatic Injuries Distress Scale (TIDS) was used to evaluate distress at baseline (within 4 weeks of injury). The TIDS is a validated, reliable 12-item risk prognosis screening tool that takes less than 3 minutes to complete with questions regarding uncontrolled pain, negative affect, and intrusion/hyperarousal. The minimum and maximum possible scores are 0 and 24, with higher scores indicating higher levels of distress. The Numeric Pain Rating Scale (NPRS) was used to assess the pain level at baseline and again 12 months postinjury. To complete the NPRS, participants rate the severity of their pain on a scale of 0 to 10, with 0 indicating no pain and 10 indicating the worst pain imaginable. NPRS scores of 1 to 3 indicate mild pain, 4 to 6 indicate moderate pain and 7 or higher indicate severe pain. As a preliminary analysis, to assess whether the participants as a group experienced recovery, a paired samples t-test was used to compare NPRS scores at baseline and 12 months. In all, 52% (60 of 116) of participants reported no pain (NPRS = 0), and mean pain intensity scores improved from 4.8 ± 2.1 at baseline to 1.6 ± 1.6 at 12-month follow-up (p < 0.001). Quantile regression was used to describe the association between baseline distress and pain levels 1 year after the injury. This technique was used because it reveals the relationships at different quantiles of the pain outcome distribution.

Results

The results indicate some people (52% [60 of 116]) recovered regardless of their baseline level of distress (30^th quantile of the NPRS: β = 0). However, in those with persisting pain, higher levels of baseline distress are associated with greater levels of pain 12 months after the injury (50^th quantile of the NPRS: β = 0.11; p = 0.01; 70^th quantile of the NPRS: β = 0.27; p < 0.001; 90^th quantile of the NPRS: β = 0.31; p = 0.01). According to this model, with a baseline TIDS score of 5, there is 10% probability that patients will report moderate or greater levels of pain (4 or higher of 10) 12 months later. This probability then increases as the TIDS score increases. For instance, at a score of 13 on the TIDS, the probability of a patient reporting moderate or higher levels of pain increases to 30%, and the probability of reporting severe pain (higher than 6 of 10) is 10%.

Conclusion

Clinicians and surgeons in orthopaedic settings can screen for the presence of peritraumatic distress using the TIDS, which is an easily administered tool that does not require extensive knowledge of psychology, and by using it they can identify those with higher levels of distress who are more likely to have persistent, long-term pain. A score of 4 or less indicates a low risk of persistent pain, a score between 5 and 12 (endpoints included) indicates moderate risk, and a score of 13 or higher indicates high risk. Future studies should investigate whether certain immediate interventions for peritraumatic distress in the aftermath of trauma can decrease the likelihood that a patient will develop chronic pain after injury. As an analysis technique, quantile regression is useful to assess complex associations in many orthopaedic settings in which a certain outcome is expected to occur in most patients leading to non-normally distributed data.

Level of Evidence

Level II, therapeutic study.

Introduction

After musculoskeletal injuries, most people recover well within the first 3 to 6 months [26]. However, some patients continue to have persistent pain even when the injuries have healed [1]. Various personal and environmental factors can contribute to the inconsistent relationship between injury and outcomes. A mechanistic understanding of the risk factors could potentially reveal new pathways for informed assessments and effective treatments, which could help patients recover more fully and achieve their goals of returning to work or other activities. As a result, a growing body of research has emerged aimed at understanding those variables associated with persistent pain after musculoskeletal injuries. Some of these factors include high levels of baseline pain [22], having more preinjury comorbidities [2, 3], lower education [18, 33], lower income [24, 33], smoking [9, 24], and low levels of physical activity [11, 21].

In addition, psychological factors including depression, anxiety, catastrophic beliefs about pain, low expectations of recovery, and heightened levels of distress have also been shown to be associated with persistent pain [16, 26, 29, 34]. Although these factors are often reported to occur after severe injuries [28], they can also feature in benign injuries such as an ankle inversion injury [30, 36], suggesting injury severity is not the prevailing factor underlining adverse and persistent psychological sequelae. Research on the experience of peri- or posttraumatic distress has been increasing because of its potential role in persistent pain from musculoskeletal injury. It is important to recognize that clinicians in orthopaedic settings need easy-to-use tools to screen for distress that address multiple factors, including symptoms of negative affect, uncontrolled pain, and hyperarousal. Recently, members of our research team found that a new scale, the Traumatic Injuries Distress Scale (TIDS), predicted pain-related outcomes after acute musculoskeletal injuries [35]. However, since recovery rates are high after low-to-moderate severity trauma, traditional statistical methods may not be appropriate due to the skewed outcome distribution. Traditional statistical methods assume a consistent linear relationship over the range of the outcome and rely on the mean of the variables to estimate the association. This can potentially mask the variations in the relationships over the range of outcomes and may provide an incomplete estimate of the true effect. Quantile regression is a relatively new technique in which the average score is not assumed to be a true representation of the entire sample, rather the association is investigated with more detail at various points of the distribution.

Therefore, in this study, we sought to determine whether an easily measurable level of baseline distress is associated with pain levels 1 year after noncatastrophic traumatic injuries when the outcome to be assessed is not normally distributed.

Methods

Study Design and Setting

This is a prospective study drawn from two earlier studies: (1) the Systematic Merging of Biology, Mental Health, and Environment (SYMBIOME, clinicaltrials.gov ID NCT02711085), a longitudinal cohort study collected in London, ON, Canada, and (2) the Neuromuscular Mechanisms Underlying Poor Recovery from Whiplash Injuries (clinicaltrials.gov ID NCT02157038), collected in Chicago, IL, USA.

Participants

In both studies, participants were recruited from local urgent or emergency care centers if they presented with a recent (less than 4 weeks) noncatastrophic musculoskeletal injury such as a fall, motor vehicle collision, or sports injury. We selected the 4-week window owing to theory and feasibility. From a theoretical perspective, common conceptualizations of pain in this field use a period from 0 to 6 weeks to indicate acute, 6 to 24 weeks as subacute, and more than 24 weeks as chronic [32]. The 4-week window was also feasible because although most participants were recruited within the first 2 weeks, in some cases, logistic issues prevented data collection within that window, so occasionally, the longer time frame was necessary to allow complete data collection.

Noncatastrophic was defined as injuries that did not require inpatient hospital admission or surgery. A research assistant provided all the information to potential participants, answered any questions, screened potential participants against the inclusion/exclusion criteria, and recruited participants after obtaining informed written consent. Consistent inclusion criteria in both studies were recent musculoskeletal trauma, the ability to understand and speak conversational English or French, and being of working age (18 to 70 years old). Exclusion criteria were any cognitive, systemic, or neuromuscular disorders that would interfere with recovery, following instructions, or participating in the study. The primary differences between the two cohorts were that the Chicago, IL, USA, cohort recruited exclusively those with whiplash-associated disorder after motor vehicle collisions, and the London, ON, Canada, cohort recruited all-cause musculoskeletal injuries (including both axial and extremity injuries). In addition, there were medicolegal contextual differences between the American and Canadian healthcare systems. Participants were given a package of questionnaires and were asked to complete them within 24 hours. Data were collected at baseline and again at regular intervals up to 12 months postinjury. For the purposes of the current analysis, only the baseline and the 12-month data are used. The study protocols were approved by the respective research and hospital ethics board before participant recruitment.

Outcome Measures

We collected demographic information such as self-reported gender and year of birth using study-specific forms at baseline (defined as the point of entry to the study). The pain level experienced by participants was captured using a Numeric Pain Rating Scale (NPRS) with 0 indicating no pain and 10 indicating extreme pain. Generally, scores of 1 to 3 indicate a mild pain level, 4 to 6 indicate a moderate pain level, and 7 or more indicate severe pain in musculoskeletal conditions [4]. There is general agreement that the NPRS has a minimum clinically important difference (MCID) of 2 of 10 points [23, 31]. Distress was captured in both cohorts using the TIDS, a screening tool specifically designed to capture distress experienced after acute musculoskeletal injuries [36]. The TIDS has 12 items with questions regarding uncontrolled pain, negative affect, and intrusion/hyperarousal [36]. Each item is scored on a scale of 0 (never or not at all) to 2 (often or all the time), and the maximum possible score is 24 (higher scores indicate higher levels of distress). It is an easy-to-use and understand tool that takes less than 3 minutes to complete [36]. Currently, there is no set MCID for the TIDS. The psychometric properties of the TIDS, including predictive validity and discriminative accuracy in this patient population, have been reported but only through the use of statistical techniques that assume linear associations and normal distribution of outcomes [35, 36], which might not provide the most complete estimates for populations that are not normally distributed. The TIDS is a freely accessible tool, and it has been translated to French, Spanish, and Persian [25].

Sample Description

Participants’ characteristics and baseline patient-reported scores were explored descriptively (frequencies or medians and interquartile ranges). Across both cohorts, between the years 2015 and 2018, 241 patients consented and were included in the study based on the above-mentioned inclusion/exclusion criteria. The cause of injury was a mixture of motor vehicle collisions and sports injuries, with 60% (145 of 241) having spinal and 40% (96 of 241) having peripheral injuries. Of those, 93% (225) of patients provided baseline data. Of those, 52% (116 of 225) provided complete data through 12 months, meaning 48% (109 of 225) were lost to follow-up at 12 months. Patients received various pharmacological treatments such as prescription and over-the-counter pain medications and nonpharmacological treatments such as physical therapy. The patients’ median age was 44 with an interquartile range of 26. Most were women (69% [80 of 116]), and the median baseline TIDS score was 6.0 with an interquartile range of 6.0 (Table 1).

Table 1.

Demographic information and descriptive statistics of NPRS and TIDS scores

Parameter	Participants (n = 116)
Age in years	44 (26)
Female, % (n)	69% (80)
Pain (measured using Numeric Pain Rating Scale) at baseline	5 (3)
Pain (measured using Numeric Pain Rating Scale) at 12 months	4 (3)
Score on the Traumatic Injuries Distress Scale at baseline	6 (6)

Data are presented as median (interquartile range) unless otherwise noted; two cohorts were collected one from London, ON, Canada (n = 54) and the other from Chicago, IL, USA (n = 62).

Most patients (52% [60 of 116]) had no pain (NPRS score 0 of 10) 12 months postinjury. The data normality was assessed using the Shapiro-Wilk test, and the results indicated that our continuous data including NPRS scores at 12 months (p < 0.001) were not normally distributed.

Ethical Approval

Ethical approval for this study was obtained from Northwestern University Institutional Review Board (STU00090769-MODCR0002) and Western University Health Sciences Research Ethics Board (HSREB 106140).

Analytic Approach

Preliminary Analysis

To assess the recovery status of the sample, we compared the mean NPRS scores at baseline and 12 months using the paired-samples t-test. This analysis informs us whether recovery has occurred as a whole before investigating the association with the variables of interest. The result of this analysis indicated that the mean NPRS score improved from 4.8 ± 2.1 at baseline to 1.6 ± 1.6 at 12 months (t = 12.8; p < 0.001). This means that the patients, as a sample, experienced recovery.

We used the Spearman rho to assess the correlation between the continuous variables. There was no correlation between age and NPRS at 12 months (ρ = -0.10; p = 0.30). There was a slight correlation between baseline NPRS scores and NPRS at 12 months (ρ = 0.18; p = 0.05). The paired-samples t-test was used to compare mean NPRS scores at 12 months between men and women, and we observed no difference (t = 1.80; p = 0.07). Based on these results, gender and age were not included in the subsequent quantile regression analysis, but baseline NPRS scores were initially included.

Quantile Regression

One methodological limitation of studies investigating the association between psychological pathologies and pain is that they rely on the mean of the variables, which can lead to masking variations in the distribution of the outcome variable. This is especially true if the data are skewed as happens in populations such as musculoskeletal injuries where most people experience good outcomes. This in turn can lead to over- or underinflation of the prognostic associations. Although such data could be dichotomized and explored through logistic regression, that approach leads to substantial loss of potentially important information. Another method of analyzing the data is through quantile regression, which offers flexibility to identify associations at different quantiles of the distribution of the outcome variable. The unique aspect of quantile regression is that it models an entire conditional distribution and not just the mean, as is the goal of ordinary least square (OLS) regression [6]. It should be noted that for each quantile the entire dataset is used, and there is no subsampling as the name may suggest [19].

Therefore, in the next step, we employed the quantile regression approach to evaluate the association between peritraumatic distress (using TIDS scores) and pain levels 12 months later, using the 12-month NPRS score as the primary outcome.

In this study, the coefficients for the variables of interest (TIDS score at baseline and baseline NPRS scores) were estimated from the 10^th to 90^th quantile of the 12-month NPRS scores, with an increment of 20 percentile points per step. The coefficient values for baseline NPRS scores did not reach statistical significance in all quantiles (that is, p values were greater than 0.05) (30^th quantile: β = 0, no p value; 50^th quantile: β = -0.67; p = 0.61; 70^th quantile: β = 0; p > 0.99; 90^th quantile: β = 0.41; p = 0.11). Therefore, this variable was subsequently removed from the quantile regression. For the 70^th and 90^th quantiles, we used the equation of the line to find corresponding TIDS scores for NPRS values. The regression lines were plotted for each quantile to visualize the slopes (Fig. 1). A second plot was also constructed to visualize the slope coefficients and confidence intervals with respect to the quantiles (Fig. 2).

Fig. 1.

Prediction lines for all the estimated quantiles. A color image accompanies the online version of this article.

Fig. 2.

Parameter estimates along with 95% confidence intervals are shown as a function of the quantile level. A color image accompanies the online version of this article.

Missing Data

Rather than attempting to impute missing data, we elected to use the complete case method and include only those participants who provided complete data sets for NPRS and TIDS scores for this novel analysis. Since the missing data were at random (χ² = 0.52 [df = 2]; p = 0.77) , this method ensures more statistical precision in estimates [15]. When only participants with complete data for NPRS and TIDS scores were included in the dataset, there were no missing values for gender. The missing data for age was less than 10% of the total dataset.

Statistical Analysis Software

All analyses were conducted using the Statistical Package for the Social Sciences (version 26.0) program (SPSS Inc), accepting a p value of 0.05 or less to indicate statistical significance.

Results

In general, the higher the level of baseline distress in patients with noncatastrophic traumatic injuries, the greater the pain levels appeared 12 months after injury. This held true for patients in whom NPRS scores were at or beyond the 50^th quantile (50^th quantile: β = 0.11; p = 0.01; 70^th quantile: β = 0.27; p < 0.001; 90^th quantile: β = 0.31; p = 0.01) (Table 2). However, for patients in the 30^th quantile of NPRS, there was no relationship with the numbers available between baseline distress and pain 12 months later (β = 0; no p value) (Table 2). Stated another way, some people recovered from their injury regardless of their level of distress at baseline, but as the level of baseline distress increased, this became less likely, suggesting a possible dose-response relationship between moderate to high levels of baseline distress and persistent pain 12 months after injury. Since the quantiles below the 50^th quantile (the median) have β = 0, we can deduce that approximately half of the participants in the dataset fully recovered (reported no pain) regardless of their level of baseline peritraumatic distress. This is in line with the descriptive statistics that show 52% reported NPRS value of 0 at 12 months. But for the other half (quantiles greater than the median), there was a positive association between NPRS scores at 12 months and baseline distress. We observed that higher quantiles were more severely affected by distress as indicated by the increasing slope; that is, the β values increase with subsequent quantiles (Fig. 1). When visualizing the results, the regression line for the 30^th quantile overlaps with the x-axis line as there was no association (Fig. 1). With increasing quantiles, the slope of the association increases as well, indicating a possible dose-response relationship for those that continue to report pain at 12 months after injury (Fig. 2).

Table 2.

Parameter estimates of all quantiles of the Numeric Pain Rating Scale distribution with scores on the Traumatic Injuries Distress Scale as the independent variable

Quantile	Coefficient β (95% CI)	p value
30th	0
50th	0.11 (0.02-0.2)	0.01
70th	0.27 (0.14-0.40)	< 0.001
90th	0.31 (0.07-0.54)	0.01

According to this model, we found that with a baseline TIDS score (distress level) of 5, the probability that a patient will report moderate or higher levels of pain (4 or higher of 10) 12 months later is 10%. This probability then increases as the TIDS score increases. At a score of 13 on the TIDS, the probability of a patient reporting moderate or higher levels of pain increases to 30%, and the probability that a patient will report severe pain (at least 6 of 10) is 10%. In other words, a score of 4 or less on the TIDS would indicate low risk, a score between 5 and 12 (endpoints included) would indicate moderate risk, and a score of 13 or higher would indicate a high risk of having persistent pain 1 year after noncatastrophic musculoskeletal injuries.

Discussion

Understanding factors that are associated with persistent pain after musculoskeletal injuries is essential to minimize the number of patients who will develop chronic pain. Psychological factors are widely acknowledged to play an important role in this. However, less is known regarding how acute distress after an injury can be associated with various pain levels 1 year later. We were able to carry out this investigation using a statistical technique (quantile regression) that assessed the association across the entire distribution one quantile at a time. We found that most participants did not report any injury-related pain 1 year later. But among those who continued to have pain, those with higher baseline distress generally had greater pain levels 1 year later. In this study, we assessed baseline distress levels using an easy-to-use tool called TIDS. Orthopaedic surgeons—even those with no special expertise in psychology—can easily use the TIDS to screen for distress levels, and our findings suggest that they ought to pay more attention to those with higher levels of distress, as they appear to be at special risk. A score of 4 or less indicates a low risk, a score between 5 and 12 (endpoints included) indicates moderate risk, and a score of 13 or higher indicates a high risk of having persistent pain 1 year after noncatastrophic musculoskeletal injuries. Another advantage is that it does not label patients or take a diagnostic approach, but it does include items that are potentially modifiable. Future studies are recommended to investigate possible intervention strategies for those with higher levels of distress at baseline to assess the mitigating effects on pain levels 1 year later.

Limitations

The first limitation is that we used two different cohorts of participants. Although this may have introduced some variation in samples and test procedures for which we did not account, this would only have undermined our ability to detect associations, and so we are confident in the relationships that we identified. It is important to note that the inclusion/exclusion criteria were similar in the two cohorts and the same outcome measures were used. A more diverse sample would improve the generalizability of our findings. In addition, when we ran the analysis separately for each cohort we found the same trends. The second limitation is that we did not account for the severity and type of injury in our analyses. This may have limited the accuracy of the estimated association. However, we assumed that all injuries were similar in severity because one of the exclusion criteria was that the injury should not be catastrophic. We defined noncatastrophic injuries as those not requiring hospitalization or surgery. An advantage of using the tools we used in this patient population is that these patients often present to high-volume clinics and using various tools for different injuries would be impractical. However, we recognize that future studies should investigate whether differences in tools or prediction rules for different clinical populations are more predictive. Another limitation is that one of the inclusion criteria for this study was that participants were to be recruited within 4 weeks from injury. Although this decision was based on both theory and feasibility, we acknowledge that some prior trajectory studies indicate that this is a labile time in the postinjury period and that some, most notably those who are recovering quickly, will experience change to their pain and distress over this window. However, the choice of a 4-week window aligns well with other similar types of prognostic association studies [14] and is consistent with the window during which injured patients are most likely to present to their primary care physicians. We do not have enough data to permit for stratified analyses, but readers are encouraged to consider that distress rated within 3 days from injury may be different from that rated 3 weeks from injury.

Increasing Distress at Baseline Is Associated with More Severe Pain 12 Months After Injury

Most patients recovered with no pain 12 months after their injury. This is in line with previous studies which showed that only a subset of patients develop chronic pain after musculoskeletal injuries [20, 27]. These patients recovered from their injury regardless of their level of distress at baseline. However, as we moved up in the quantiles, the effect of distress became more pronounced, which indicates a possible dose-response relationship; that is, higher distress levels were associated with higher amounts of pain 12 months postinjury. Specifically, patients with a TIDS score of 4 or lower can be categorized as low risk, between 5 and 12 (endpoints inclusive) can be categorized as moderate risk, and 13 or higher can be categorized high risk. These thresholds are similar to those reported by Walton et al. [35], who used linear discrimination analysis to identify cut-off scores of 3 and 11 for low and high risk, respectively. This difference reflects the analytic techniques used. The current study also has the added benefit of providing the probability of patients falling in each of the categories of mild, moderate, and severe pain according to the distress levels. In addition, this study adds to the prior analysis of the prognostic value of the TIDS, which relied on linear associations and OLS regression to identify potential predictors [35], by showing that peritraumatic distress after musculoskeletal injuries is associated with pain levels in those who have higher pain scores at 12 months but not in those with low pain scores. Based on this, we believe that targeting higher levels of peritraumatic distress may benefit patients over the long-term clinical course. With an aim to intervene accordingly, clinicians in orthopaedic settings can carefully monitor and screen for distress after acute musculoskeletal injuries using the TIDS, which is an easy-to-use tool that takes less than 3 minutes to complete.

The association between psychological distress and the development of chronic pain has been reported [10]. However, distress has been mostly shown to be strongly associated with persistent pain after major injuries, such as whiplash-associated disorder as a result of road traffic accidents or after severe injuries that were treated surgically [5]. For example, Edgley et al. [8] found that acute emotional distress is associated with persistent pain in a sample of 303 patients after severe orthopaedic injuries who underwent surgical interventions. Recently, Gopinath et al. [13] reported that distress is associated with chronic pain 1 year after noncatastrophic road traffic collisions. However, persistent pain can also follow seemingly innocuous injuries such as ankle sprains [30], and distress can also happen after such mild injuries [14, 36]. Haahr and Andersen [14] reported that most of their participants (83%) with elbow pathologies improved after 1 year, but distress was associated with poor prognosis. This is important to highlight, since persistent pain and the attendant ongoing psychological consequences for the patient, as well as economic consequences for the patient and society, therefore apply not only to severe injuries but also to patients with milder injuries. Compared with patients with severe injuries, less research has been conducted in this specific population of patients with noncatastrophic injuries. In the research that has focused on this population, findings generally were in line with ours [14].

Our study extends what is known by including patients with injuries of mixed etiologies; we observed that acute distress can occur in many kinds of acute musculoskeletal injury. Identifying factors that are associated with the persistence of pain at any point in the recovery process can allow clinicians to better direct their clinical decisions and choose the most appropriate treatment strategies. Since the TIDS has questions regarding uncontrolled pain, negative affect, and intrusion/hyperarousal, identification of the most salient problem based on scores on these questions and appropriately targeting a potential intervention is a potential advantage of the TIDS. Possible intervention options to address the issues identified by the TIDS include re-examining the pain management plan, cognitive reshaping and behavioral techniques, relaxation techniques, graded (gradual) return to activity, and others. Individuals at higher risk for persistent pain based on screening with the TIDS may benefit from adding cognitive-behavioral therapy (CBT) to the physical therapy rather than traditional therapy, and referral to mental health professionals may be appropriate. Previous research has endorsed the use of techniques such as CBT in people with chronic pain to reduce catastrophic thinking, pain-related fear, and negative appraisals of pain to improve both psychological and physical symptoms [12, 17]. Psychological interventions studied to date have mainly focused on people who have suffered catastrophic injuries [7]. Future research is needed to discover specific biopsychosocial interventions that can be implemented for people after acute noncatastrophic musculoskeletal injuries. Our research has identified this potential target population of patients who might be especially appropriate to include in such research: Patients with noncatastrophic traumatic injuries (defined as injuries that do not require inpatient hospital admission or surgery) whose TIDS scores within 1 month after injury were greater than or equal to 5.

The novel findings of this study were possible through the use of quantile regression, as an advanced approach to prognosis that offers advantages over traditional linear regression modeling. This approach is useful when the distribution of the outcome variable is skewed, and the association can be assessed at various points (quantiles) in the distribution of the outcome variable. This technique may be useful across orthopaedic populations where most people are expected to recover within a specific timeframe but where a subset of people continue to have pain and disability. As computing power has increased, quantile regression functions are now available in many statistical packages, and statisticians believe that the use of this method will be increasing in future studies [19].

Conclusion

We found that about half of our study group reported no pain 1 year after noncatastrophic musculoskeletal injuries, regardless of their level of baseline distress. However, some people continued to have persistent pain, and in those people, individuals with higher baseline distress levels generally had greater pain levels 12 months later. The TIDS, which is a unique patient-reported outcome measure specifically designed for assessing acute distress in musculoskeletal injuries of any etiology, was used to assess distress levels. In this study, we showed that patients with a TIDS score of 4 or lower can be categorized as low risk of persistent pain, between 5 and 12 (endpoints inclusive) can be categorized as moderate risk, and 13 or higher can be categorized high risk of having persistent pain 1 year after their injury. Orthopaedic surgeons and therapists can use the TIDS, as it is quick, easy to administer even by individuals with no specific training in psychology, and easy to interpret (higher scores indicate higher distress). Potential therapeutic options can be identified based on the scores from the three subscales of the TIDS, though the degree to which early, targeted intervention changes recovery trajectory is an area for further study.