Introduction
Since the mid-1990s, pain measurement has been promoted as the “fifth vital sign” to enhance pain management quality, leading to the widespread adoption of pain screening tools across healthcare settings. The numeric rating scale (NRS), a common tool, assesses current pain severity on a 0–10 scale, from “no pain” to “worst pain imaginable.” In outpatient care, NRS scores are frequently recorded in Electronic Health Records (EHRs), enabling tracking of pain intensity over time. This is particularly relevant in regions like Oregon, where healthcare systems are continuously striving to improve patient care, including effective back pain assessment and management.
However, the clinical use of NRS has limitations. Inconsistent administration can skew pain estimation. Concerns exist about the accuracy of clinically collected NRS scores in identifying patients with significant pain that affects function or necessitates medical visits. Research use of clinical NRS data is also challenged by non-normal distribution, irregular measurement intervals, and score variability within individuals over short periods. Despite these drawbacks, longitudinal EHR pain intensity data are used in pragmatic trials and observational studies to evaluate intervention effectiveness. Thus, demonstrating the value of this data for understanding health outcomes remains crucial, especially when considering back pain assessment tool care in Oregon and similar healthcare environments.
Prior research has explored the correlation between clinically administered and research-administered NRS scores. A study of chronic pain outpatient encounters found moderate correlations between clinical and research NRS scores (r = 0.63) and the Brief Pain Inventory severity subscale (r = 0.61). While these findings suggest some utility for administrative NRS data, a unidimensional focus on pain intensity might not fully capture the pain experience. Experts recommend assessing broader aspects of pain, including function, mental health, and quality of life. The relationships between clinically administered NRS scores and these broader pain-related factors, along with the tools used for their assessment, need further investigation, particularly in the context of Back Pain Assessment Tool Care Oregon.
This study expands on existing research by examining the associations of NRS scores with a wider array of pain-related outcomes—pain disability, health-related quality of life (HRQOL), anxiety, and depression. This provides a more comprehensive understanding of the strengths and weaknesses of NRS in clinical pain assessment and the reuse of EHR NRS data for research, especially when considering back pain assessment tool implementation and care pathways in Oregon.
The objective of this study is to determine how well NRS scores, collected during routine back pain assessment tool care oregon, correlate with more robust, validated measures of pain, mental health, and HRQOL. We aimed to understand the association between NRS scores obtained in usual care, both at single points and over time, with (1) research-administered pain intensity measures and (2) measures of pain disability, mental health, and HRQOL. We hypothesized that clinically collected NRS scores would more strongly correlate with pain intensity and less so with pain disability, mental health, and HRQOL. The findings will inform the utility of administrative NRS scores in evaluating the effectiveness of pain-related interventions on crucial clinical outcomes, particularly in systems focusing on optimized back pain assessment tool care oregon.
Methods
Participants
Participants were recruited from the VA Portland Health Care System, a Department of Veterans Affairs hospital and clinic system in Portland, Oregon, between December 2013 and October 2015. This setting is highly relevant to the study of back pain assessment tool care oregon, as it represents a significant healthcare provider in the region. They were part of a larger study on prescription opioid dose escalation for chronic pain. Inclusion criteria for the larger study and this secondary analysis were: a stable opioid prescription for at least 3 months for musculoskeletal pain and English literacy. Exclusion criteria included opioid prescriptions for cancer or palliative care, pending litigation or disability claims related to pain, age under 18, enrollment in an opioid substitution program in the past year, lack of reliable telephone access, current opioid dose exceeding 120 mg morphine equivalents, or opioid prescriptions consisting solely of tramadol or buprenorphine.
Procedures
Research visits were conducted at baseline and at 6, 12, 18, and 24 months post-baseline, either in person or via telephone. NRS scores from outpatient care visits before or after each research visit were extracted from EHRs. For the primary analysis, NRS scores administered within a 4-week interval (2 weeks before and 2 weeks after research assessments) were used. If multiple NRS scores were available within this interval, they were averaged, consistent with prior research methods. When only one score was available, that single score was used.
Secondary analyses examined relationships between research visit data and NRS scores from longer periods around research assessments. These included 8-week (4 weeks before and after), 16-week (8 weeks before and after), and 24-week (12 weeks before and after) intervals. Again, multiple NRS scores were averaged if available.
The Institutional Review Board at the VA Portland Health Care System approved and monitored all study procedures. All participants provided informed consent for study procedures and EHR data extraction, ensuring ethical standards were maintained throughout this study on back pain assessment tool care oregon.
Measures
EHR Data
The NRS, scored from 0–10 (0 = no pain, 10 = worst pain imaginable), is a validated measure of pain intensity. Its widespread use makes it a critical component of back pain assessment tool care oregon and similar systems. NRS scores are recorded as structured vital sign data in the VA EHR and were extracted via the VA Corporate Data Warehouse (CDW), which aggregates EHR data from all VA patients into a relational database.
Survey Data
The Chronic Pain Grade (CPG), a 7-item self-report measure, assesses pain intensity and pain disability. It evaluates current pain and pain over 3 months and is widely used and validated.
Depressive symptoms over the past 2 weeks were assessed with the Patient Health Questionnaire-8 (PHQ-8), an 8-item self-report measure. Higher scores indicate more severe depressive symptoms, with scores ≥10 indicating moderate depression.
Anxiety over the past two weeks was measured using the Generalized Anxiety Disorder 7-item Scale (GAD-7), a brief self-report measure of anxiety symptom severity. Scores ≥10 indicate moderate anxiety.
Health-related quality of life (HRQOL) over the past 4 weeks was measured using the Short-Form Health Survey, Version 2 (SF-12v2), a 12-item self-report measure of physical and mental health functioning. Higher scores indicate better functioning.
Demographic data, including age, gender, race, marital status, employment status, socioeconomic status, and disability status, were collected via self-administered surveys to provide context for understanding back pain assessment tool care oregon patient populations.
Statistical Analyses
Data completeness and accuracy were verified. Patient data spanned up to 5 assessment periods. The number of unique patients completing assessments at each time point were: Baseline (n = 184), 6 months (n = 171), 12 months (n = 167), 18 months (n = 162), 24 months (n = 154). All available data were used in analyses. Generalized estimating equations (GEEs) with model-based variance estimation and unstructured correlation matrices were used to examine bivariate associations between EHR-derived NRS data and patient-reported outcomes: pain intensity and disability (CPG), depression (PHQ-8), anxiety (GAD-7), and mental and physical functioning (SF-12v2). GEEs controlled for longitudinal associations of patient-reported outcomes from individuals with ≥2 assessments. The primary model examined associations between patient-reported outcomes and NRS scores from the 4-week interval. Subsequent models assessed associations with NRS scores averaged over 8-, 16-, and 24-week intervals. All inferential analyses were two-tailed with alpha = 0.05. Associations are reported as standardized beta coefficients (range: −1 to 1). These statistical methods are crucial for rigorously evaluating back pain assessment tool care oregon data.
Results
Figure 1 displays participant retention and reasons for missing data across all five time points. The sample comprised 186 patients, with a mean age of 59.9 years (SD = 11.9 years). Most were male (89.8%), white, non-Hispanic (76.9%), and had some college education (60.2%). Common musculoskeletal pain conditions included back pain (63%), neck or joint pain (53%), and arthritis (50%). The baseline average daily opioid dose was 33.8 mg morphine equivalents (SD = 24.7). Sociodemographic and clinical characteristics are detailed in Table 1, which also presents average scores for self-reported outcomes across all time points. Table 2 shows summary scores of clinical measures at each time point, providing a comprehensive overview of the study sample relevant to back pain assessment tool care oregon.
Figure 1.
Participant flow diagram.
Alt text: Participant flow diagram illustrating patient retention in a study assessing back pain and pain management in Oregon.
Table 1.
Sample sociodemographic and clinical characteristics
| Demographic Variables | M (SD) | % (n) |
|---|---|---|
| Age | 59.9 (11.9) | |
| Male | 89.8 (167) | |
| Race/ethnicity | ||
| White, Non-Hispanic | 76.9 (143) | |
| Native American, Non-Hispanic | 11.8 (22) | |
| Hispanic, any race | 6.9 (13) | |
| African American, Non-Hispanic | 4.3 (8) | |
| Education | ||
| High school or less | 16.1 (30) | |
| Some college | 60.2 (112) | |
| College or graduate school | 23.7 (44) | |
| Marital status | ||
| Married/living with partner | 52.7 (98) | |
| Separated/divorced | 34.9 (65) | |
| Widowed | 7.5 (14) | |
| Never married | 4.8 (9) | |
| Employment | ||
| Disabled | 37.6 (70) | |
| Retired | 31.2 (58) | |
| Working/homemaker | 30.1 (56) | |
| Other | 1.1 (2) | |
| Clinical variables | ||
| Opioid dose in MME | 33.8 (24.7) | |
| Arthritis | 50.0 (93) | |
| Back pain | 61.8 (115) | |
| Neck or joint pain | 53.2 (99) | |
| Nicotine use | 37.6 (70) | |
| Self-reported outcome measures | ||
| Pain intensity (CPG) | 64.53 (15.37) | |
| Pain disability (CPG) | 50.06 (27.65) | |
| Depression severity (PHQ-8) | 9.13 (6.06) | |
| Proportion with moderate depression | 48.5 (90) | |
| Anxiety severity (GAD-7) | 6.82 (6.17) | |
| Proportion with moderate anxiety | 33.9 (63) | |
| Physical functioning (SF-12) | 37.71 (33.21) | |
| Mental functioning (SF-12) | 58.82 (25.43) | |
| NRS scores | ||
| 4-week interval | 4.91 (2.75) | |
| 8-week interval | 4.86 (2.61) | |
| 16-week interval | 4.85 (2.59) | |
| 24-week interval | 4.64 (2.47) |
Alt text: Table showing sociodemographic and clinical characteristics of patients with musculoskeletal pain undergoing back pain assessment and care in Oregon.
N = 186. MME = morphine milligram equivalent; CPG = Chronic Pain Grade; PHQ-8 = Patient Health Questionnaire- 8 item; GAD-7 = Generalized Anxiety Disorder- 7 item; SF-12 = Short Form-12 item.
Table 2.
Scores of self-reported pain, mental health and quality of life outcomes
| Self-Report Variable | Baseline M (SD) N = 184 | 6 Months M (SD) N = 171 | 12 Months M (SD) N = 167 | 18 Months M (SD) N = 162 | 24 Months M (SD) N = 154 |
|---|---|---|---|---|---|
| Pain intensity (CPG) | 65.63 (13.38) | 64.58 (15.41) | 64.19 (16.47) | 64.30 (16.24) | 63.77 (15.55) |
| Pain disability (CPG) | 53.19 (24.54) | 48.42 (28.57) | 48.98 (28.46) | 49.69 (28.90) | 48.62 (27.96) |
| Depression severity (PHQ-8) | 9.84 (9.03) | 9.23 (5.90) | 9.08 (6.00) | 8.77 (6.16) | 8.57 (6.23) |
| Anxiety severity (GAD-7) | 7.23 (6.28) | 7.29 (6.33) | 6.71 (6.26) | 6.55 (6.11) | 6.23 (5.79) |
| Physical functioning (SF-12) | 35.19 (25.00) | 36.37 (32.80) | 39.07 (31.23) | 39.20 (35.01) | 38.80 (33.95) |
| Mental functioning (SF-12) | 57.13 (25.79) | 58.70 (25.69) | 58.83 (25.68) | 59.95 (25.24) | 59.74 (24.89) |
Alt text: Table displaying self-reported pain, mental health, and quality of life scores in patients receiving back pain care in Oregon, assessed using a pain assessment tool.
As shown in Table 3, the 4-week interval NRS pain intensity score from clinical practice was moderately associated with CPG-measured pain intensity (B = 0.59). Associations remained similar across 4-, 8-, 16-, and 24-week NRS averaging intervals. Similar temporal stability was observed for associations between EHR-derived NRS pain scores and other patient-reported outcomes, though magnitudes were smaller than for pain intensity. Specifically, 4-week interval NRS associations were: pain disability (B = 0.33), depression (B = 0.26), anxiety (B = 0.26), physical functioning (SF-12v2) (B = −0.15), and mental functioning (SF-12v2) (B = −0.13). The weakest associations were between EHR-derived NRS pain scores and physical and mental health functioning; the 16-week NRS scores did not significantly associate with SF-12 mental functioning. These results shed light on the effectiveness of back pain assessment tool care oregon practices.
Table 3.
Associations between NRS scores and clinical variables
| Clinical Variable | 4-week Avg NRS | 8-week Avg NRS | 16-week Avg NRS | 24-week Avg NRS |
|---|---|---|---|---|
| Pain intensity | 0.59*** | 0.54*** | 0.56*** | 0.53*** |
| (n = 232) | (n = 375) | (n = 531) | (n = 628) | |
| Pain disability | 0.33*** | 0.35*** | 0.34*** | 0.36*** |
| (n = 231) | (n = 373) | (n = 528) | (n = 622) | |
| Depression severity | 0.26*** | 0.20*** | 0.21*** | 0.26*** |
| (n = 230) | (n = 372) | (n = 528) | (n = 625) | |
| Anxiety severity | 0.26*** | 0.19*** | 0.19*** | 0.23*** |
| (n = 232) | (n = 375) | (n = 529) | (n = 625) | |
| Physical functioning | −0.15* | −0.16* | −0.16*** | −0.16*** |
| (n = 232) | (n = 374) | (n = 530) | (n = 626) | |
| Mental functioning | −0.13* | −0.14* | −0.09+ | −0.12* |
| (n = 232) | (n = 375) | (n = 531) | (n = 626) |
Alt text: Table showing the associations between Numeric Rating Scale (NRS) scores and clinical variables in a back pain assessment study conducted in Oregon.
n in parentheses denotes number of observations for the specified correlation. Each participant could contribute up to five observations spread over a 24-month period to the generalized estimating equation models. Unique observations at each time point for the self-reported outcome measures were as follows: Baseline: n = 184; 6 months: n = 171; 12 months: n = 167; 18 months: n = 162; 24 months: n = 154.
* P
** P
*** P
+Not significant at P
Discussion
This study found that NRS scores from routine clinical care, extracted from EHRs within the VA Portland Health Care System, a key provider of back pain assessment tool care oregon, show a moderate association with a standardized measure of pain intensity. EHR NRS scores also exhibited modest associations with pain disability scores and weak associations with validated measures of depression, anxiety, and HRQOL. This was observed across 4-, 8-, 16-, and 24-week intervals around research assessments over two years. The consistent association of pain intensity with NRS over these intervals suggests stability in patients’ pain intensity experiences over time, aligning with previous research.
The weak associations between NRS scores and depression, anxiety, and quality of life highlight a critical point for healthcare systems, including those focused on back pain assessment tool care oregon. To better capture the comprehensive pain experience, standardized, brief, multidimensional assessments should be integrated into routine clinical care and their results documented in EHRs. These brief measures can capture a wider range of pain-related factors without significantly increasing burden during clinical encounters. Examples include the Patient Health Questionnaire-2 item scale (PHQ-2) for depression and the Pain, Enjoyment, General Activity (PEG) scale, assessing pain intensity, enjoyment interference, and activity interference. Quality of life measures are also essential. These tools can be administered electronically before appointments and made accessible to clinicians via EHRs, enhancing back pain assessment tool care oregon and beyond. A two-step NRS screening process could also be beneficial: NRS scores ≥5 would trigger a more detailed assessment of function, mood, and quality of life, while lower scores would not, streamlining back pain assessment tool care oregon protocols.
Our finding of a moderate association between clinical and research-administered pain intensity scores aligns with prior research (r = 0.63 in a similar study). This reinforces the utility of EHR-derived NRS scores from administrative data in comparative effectiveness research. Well-controlled observational studies using administrative data can offer preliminary evidence of intervention effectiveness on pain intensity, especially when large RCTs are impractical. However, our findings support augmenting NRS scores with measures of disability, mental health, and quality of life, which are recommended outcomes in pain trials, particularly when aiming for holistic back pain assessment tool care oregon.
Limitations in using clinically collected NRS scores for comparative effectiveness studies include variations in clinical administration, potentially reducing NRS accuracy, and incomplete or non-randomly missing data for individuals not receiving care. Statistical methods and careful case selection from EHR data can mitigate some of these issues. However, the inherent incompleteness and inconsistencies of administrative data must be considered in research projects using EHR data. Researchers using NRS alone should recognize its limitations in reflecting broader pain-related outcomes. Learning health care systems infrastructure, such as the Collaborative Health Outcomes Information Registry (CHOIR), could bridge EHR NRS scores with more robust pain outcomes regularly documented in EHRs and usable for research, potentially improving back pain assessment tool care oregon research capabilities. Embedding patient-reported outcomes like pain and quality of life, as demonstrated in mental health clinics, allows for comparisons of retrospective outcomes across healthcare systems.
Study limitations include a sample from a single VA medical center, primarily male and White, Non-Hispanic, receiving long-term opioid therapy, which may limit generalizability to other pain populations and diverse demographics. The asynchronous collection of EHR and research tools, and differing recall periods of NRS and research measures, are also limitations. Variable pain, mood, and functioning over recall windows could attenuate associations. Finally, while validated measures were used, findings might vary with other measures of pain-related outcomes. Future research in back pain assessment tool care oregon should address these limitations to enhance the applicability of findings.
Conclusion
EHR-derived NRS scores show a moderate association with pain intensity, a modest association with pain disability, and weak associations with mood and quality of life. These findings support using EHR data for retrospective studies of pain intensity changes but suggest caution when examining other pain-related outcomes. This study enhances understanding of the associations between clinically collected NRS scores and broader pain-related variables, including pain disability, mental health, and physical and mental functioning, within systems like back pain assessment tool care oregon. Clinicians and researchers using EHR data for comparative effectiveness would benefit from incorporating more robust measures of pain-related function, mood, and quality of life into routine patient care, thereby improving the comprehensiveness of back pain assessment tool care oregon and similar healthcare models.
Disclosures and Funding Sources
This work was supported by the National Institute on Drug Abuse (grant number 034083, PI: Morasco), VA Health Services Research and Development-funded Center to Improve Veteran Involvement in Care (CIN 13-404, PI: Dobscha), and an American Cancer Society Mentored Research Scholar Grant (132817 MSRG-18-216-01-CPHPS, PI: Nugent). The content is the authors’ responsibility and does not represent the views of the Department of Veterans Affairs or NIDA.
Conflicts of interest
The authors declare no conflicts of interest.
