1 Introduction

Millions of orthopaedic surgeries are performed annually in the United States.^1–3 Despite the increasing number of orthopaedic surgeries, there is a significant lack of data reporting on patient experience and recovery following surgery. Pain is an especially important patient reported outcome because it is often the predominant complaint of patients seeking orthopaedic surgery consultation. Thus, pain must be assessed and managed postoperatively to optimize care. Persistent pain is an immense personal burden—affecting function, emotional wellbeing, and longevity—as well as a public health burden.^4–6

This study used the Patient-Reported Outcomes Measurement Information System (PROMIS) developed by the National Institutes of Health (NIH). PROMIS is comprised of question banks related to domains, including Pain Interference. The system utilizes computerized adaptive testing (CAT), in which subsequent questions are selected from a question bank based on responses to previous questions to create efficient, individualized surveys with fewer questions.⁷ The validity of the complete question banks and the shorter CAT versions for individual PROMIS domains was evaluated among a sample of 21,133 people from the general population, and PROMIS scores had strong correlations with results from previously accepted measures of each domain.⁸ PROMIS therefore provides an assessment of patient-reported outcomes that is cost-effective, time-effective, and generalizable over a wide variety of clinical patients.

The PROMIS Pain Interference (PI) domain has already been applied to procedures involving the knee; however, results focus on later postoperative values.^9,10 Pain Interference is a measure of the degree to which pain impacts relevant aspects of patients' lives, including the impact on social, cognitive, emotional, and recreational activities.¹¹ Other surgical fields have examined early postoperative pain and have found an association with negative long-term outcomes.^10,12–14 However, more work is needed in quantifying pain interference two weeks postoperatively in orthopaedics specifically. Aside from the ethical and economic implications of reducing pain, early pain can predict the development of chronic postsurgical pain.^13,15–19 Chronic postsurgical pain is defined as pain lasting more than two months following a surgical procedure that cannot be attributed to another cause such as infection or malignancy.²⁰ Studies of elective orthopaedic procedures have already documented the emerging incidence of chronic postsurgical pain.^21–23 Proper management of early pain may deter the development of chronic pain, and analysis of factors that predict early pain could assist physicians in counseling patients and understanding their respective risks. Furthermore, a better understanding of patients’ postsurgical pain could be instrumental in tailoring analgesic regimens.¹⁹

The objective of this study was to identify factors associated with greater Pain Interference two weeks following orthopaedic surgery. We hypothesized that greater postoperative Pain Interference two weeks following orthopaedic surgery will be associated with worse preoperative pain, function, and general health status.

2 Methods

All patients undergoing elective outpatient orthopaedic surgery at a single urban academic institution were evaluated preoperatively from August 2016 to March 2018. Inclusion criteria were as follows: (1) patient undergoing surgery, (2) age 17 years and older, and (3) English speaking. Patients who were unable to read or write English, and patients who were incarcerated or wards of the state were excluded. Eligible patients were enrolled into an Institutional Review Board (IRB) approved, prospective Web-based registry.²⁴ Study data was collected and managed using REDCap™ (Research Electronic Data Capture) electronic data capture tools hosted at the University of Maryland School of Medicine.²⁵ Eligible patients were approached by research staff preoperatively, either when the decision to proceed with surgery was made or on the day of surgery in the preoperative registration area. Patients who provide written, informed consent were enrolled in the Maryland Orthopaedic Registry.

Within one week of surgery, participants completed an electronic baseline assessment. Specifically, questionnaires include the following: PROMIS Computer Adaptive Testing in 6 domains (Physical Function, Pain Interference, Social Satisfaction, Fatigue, Anxiety, Depression), Numeric Pain Scales (NPS) for pain in the whole body and the location of surgery, International Physical Activity Questionnaire (IPAQ), and Tegner Activity Rating Scale. Patients then complete the same questionnaires two weeks after surgery, along with assessments of Improvement and Satisfaction. Demographic data was self-reported by each respondent, and each their medical records were reviewed for relevant medical history, including American Society of Anesthesiologists (ASA) score, smoking status, alcohol use, and current medications.

Data from pre and postoperative questionnaires, as well as clinical data from participants’ medical records, were collected and managed through REDCap™. De-identified patient data was extracted and analyzed. Statistical analysis included parametric and/or non-parametric bivariate tests depending on normality of the data as well as multivariable analysis.

3 Results

A total of 435 patients were analyzed. There were of 232 (53%) males and 203 (47%) females (Table 1). The mean age of patients sampled was 41.1 ± 15.7 (Table 2). The mean preoperative PROMIS PI was 60.1 ± 7.0, and mean PROMIS PI at two weeks increased to 61.7 ± 7.6 (Table 2). Worse postoperative PI was significantly associated with no college degree, lower income, preoperative opioid use, higher ASA score, history of depression or anxiety symptoms, previous surgery on the operative joint, and surgery on the lower extremity (Table 1).

Better 2 week PI was correlated with better baseline scores on all outcome measures (Table 2), and was most strongly correlated for NPS in the location of surgery (rs = 0.27, p < 0.001). Similarly, better postoperative scores on all outcome measures were associated with better 2 week PI scores and the correlations were stronger than baseline measures with strongest correlation with 2 week PROMIS Fatigue ((rs = 0.62, p < 0.001) (Table 2).

Table 3 displays the ten most common surgeries performed on upper and lower extremities that were included in this study. There were significant differences between surgeries performed on the lower extremity for both baseline PI and postoperative PI. The lowest levels of baseline and two-week PI were associated with arthroscopic removal of foreign bodies in the knee. Conversely, total hip arthroplasty was associated with the highest levels of pre- and postoperative PI, though pain was on average decreased following surgery. There were no significant differences between surgeries performed on the upper extremities for either baseline or postoperative PI. Analogous to lower extremity patients, total shoulder arthroplasty patients experienced the worst PI preoperatively (Table 3).

Multivariable regression confirmed that better preoperative pain assessment scores were independent predictors of better 2-week PI scores, while lower extremity surgery predicted worse pain (Table 4).

4 Discussion

The findings of our study support our hypothesis as greater postoperative Pain Interference two weeks following orthopaedic surgery was associated with worse preoperative pain, function, and general health status. Furthermore, we found that postoperative pain interference was associated with multiple demographic and surgical factors, and undergoing lower extremity surgery was an independent predictor of greater pain interference. These findings are congruent with assessment of physical function, as patients undergoing lower extremity surgery also experienced the greatest impedance on physical function.²⁶ Our results provide further evidence that Pain Interference is an important metric that takes into account the impairment due to pain.¹¹ An understanding of the intertwined nature of the pain and function is important for considering the effects that surgery may have on mental wellbeing and general health, both short-term and long-term.

Mental health was assessed in several ways in this study. A history of depression or anxiety as assessed by the preoperative survey showed significant association with worse postoperative PI. Additionally, PROMIS Depression, Anxiety, and Fatigue were significantly positively correlated with postoperative PI scores both preoperatively and postoperatively. Together, these findings demonstrate the relationship between mental health and pain. While the coincidence of pain and depression is well documented in patients experiencing chronic pain, this study provides evidence that the troublesome relationship is present even in the acute phase of pain.^27–29 Overall, worse pain was correlated with worse scores on all outcome measures, illustrating the broad effects of pain on well-being and health. These results are consistent with results from analysis of preoperative pain in orthopaedic patients that found significant correlations amongst PROMIS domains and legacy outcome measures.^30,31 Furthermore, surgical satisfaction and improvement were both negatively correlated with postoperative PI, indicating more satisfaction and improvement are seen when there is less postoperative PI.

Traditional methods of assessing patient reported pain, such as the numerical pain scale, have failed to encompass the broad effects of pain. These measures are susceptible to their subjective nature and the risk of ceiling effects, inflating pain intensity and therefore escalating pain management.³² Proper management of acute pain is imperative for function, recovery, wellbeing, and may stave off development of chronic pain.¹⁹ However, the United States is currently in the midst of an opioid epidemic,³³ a crisis which has not spared orthopaedists: over 25% of conditions associated with prescription of opioids leading to sustained use were due to orthopaedic or spine conditions.³⁴ The use of PROMIS PI in orthopaedics setting may help curb this trend in opioid prescription by providing a more accurate representation of patients’ pain. However, it is important to recognize that PROMIS PI may perform differently in different patient populations. We found that greater pain interference after surgery was associated with prior surgery on the joint, preoperative opioid use, lower income, lower education, and higher ASA score. Consequently, these factors should be taken into account when considering outcomes. Nevertheless, the significant correlations with all other outcome measures suggest that PROMIS PI may be useful as a single measure of patient status two weeks after orthopaedic surgery.

This study is possibly limited by its observational nature and all the conventional limitations associated with an observational study design. Thus, this study cannot determine causality with the degree of certainty that can be achieved with experimental studies, nor is it possible to determine if all potential confounding factors were known, collected, and controlled. Furthermore, our study is limited to a degree by response bias. Within the Maryland Orthopaedic Registry, there were significant differences between patients who filled out early postoperative questionnaires and those who did not. Specifically, patients with more fatigue and pain preoperatively were less likely to fill out their postoperative questionnaires.³⁵ Nonetheless, our study is strengthened by its sample size and the diversity of the patient population. To the extent of our knowledge, this is the largest sample on which PROMIS PI CAT has been applied in a prospective manner.

5 Conclusion

Postoperative PROMIS Pain Interference showed significant associations with measures of pain, function, general health, and psychosocial status. PROMIS Pain Interference may be useful as a single measure of patient status two weeks after orthopaedic surgery.

Funding

This work was supported by a grant from The James Lawrence Kernan Hospital Endowment Fund, Incorporated.

CRediT authorship contribution statement

Jamie Kator: Writing - original draft, Data curation, Visualization, Investigation. Ali Aneizi: Formal analysis, Writing - review & editing, Data curation, Visualization. Vidushan Nadarajah: Methodology, Writing - review & editing. Patrick M.J. Sajak: Visualization, Writing - review & editing. Min Zhan: Formal analysis, Data curation. Mohit N. Gilotra: Investigation, Writing - review & editing. Ngozi M. Akabudike: Investigation, Writing - review & editing. Jonathan D. Packer: Investigation, Writing - review & editing, Supervision. R. Frank Henn: Supervision, Funding acquisition, Project administration, Writing - review & editing.