A narrative review of data collection and analysis guidelines for comparative effectiveness research in chronic pain using patient-reported outcomes and electronic health records

Utilizing a learning health system to capture real-world patient data: Application of the reliable change index to evaluate and improve the outcome of a pain rehabilitation program

BACKGROUND AND OBJECTIVES

The learning healthcare system (LHS) has been developed to integrate patients' clinical data into clinical decisions and improve treatment outcomes. Having little guidance on this integration process, we aim to explain (a) an applicable analytic tool for clinicians to evaluate the clinical outcomes at a group and an individual level and (b) our quality improvement (QI) project, analyzing the outcomes of a new outpatient pain rehabilitation program ("Back-in-Action": BIA) and applying the analysis results to modify our clinical practice.

METHODS

Through our LHS (CHOIR; https://choir.stanford.edu), we administered the Pain Catastrophizing Scale (PCS), Chronic Pain Acceptance Questionnaire (CPAQ), and Patient-Reported Outcomes Measures (PROMIS)® before and after BIA. After searching for appropriate analytic tools, we decided to use the Reliable Change Index (RCI) to determine if an observed change in the direction of better (improvement) or worse (deterioration) would be beyond or within the measurement error (no change).

RESULTS

Our RCI calculations revealed that at least a 9-point decrease in the PCS scores and 10-point increase in the CPAQ scores would indicate reliable improvement. RCIs for the PROMIS measures ranged from 5 to 8 T-score points (i.e., 0.5-0.8 SD). When evaluating change scores of the PCS, CPAQ, and PROMIS measures, we found that 94% of patients showed improvement in at least one domain after BIA and 6% showed no reliable improvement.

CONCLUSIONS

Our QI project revealed RCI as a useful tool to evaluate treatment outcomes at a group and an individual level, and RCI could be incorporated into the LHS to generate a progress report automatically for clinicians. We further explained how clinicians could use RCI results to modify a clinical practice, to improve the outcomes of a pain program, and to develop individualized care plans. Lastly, we suggested future research areas to improve the LHS application in pain practice.

Hierarchical Clustering Applied to Chronic Pain Drawings Identifies Undiagnosed Fibromyalgia: Implications for Busy Clinical Practice

Currently-used assessments for fibromyalgia require clinicians to suspect a fibromyalgia diagnosis, a process susceptible to unintentional bias. Automated assessments of standard patient-reported outcomes (PROs) could be used to prompt formal assessments, potentially reducing bias. We sought to determine whether hierarchical clustering of patient-reported pain distribution on digital body map drawings predicted fibromyalgia diagnosis. Using an observational cohort from the University of Pittsburgh's Patient Outcomes Repository for Treatment registry, which contains PROs and electronic medical record data from 21,423 patients (March 17, 2016-June 25, 2019) presenting to pain management clinics, we tested the hypothesis that hierarchical clustering subgroup was associated with fibromyalgia diagnosis, as determined by ICD-10 code. Logistic regression revealed a significant relationship between the body map cluster subgroup and fibromyalgia diagnosis. The cluster subgroup with the most body areas selected was the most likely to receive a diagnosis of fibromyalgia when controlling for age, gender, anxiety, and depression. Despite this, more than two-thirds of patients in this cluster lacked a clinical fibromyalgia diagnosis. In an exploratory analysis to better understand this apparent underdiagnosis, we developed and applied proxies of fibromyalgia diagnostic criteria. We found that proxy diagnoses were more common than ICD-10 diagnoses, which may be due to less frequent clinical fibromyalgia diagnosis in men. Overall, we find evidence of fibromyalgia underdiagnosis, likely due to gender bias. Coupling PROs that take seconds to complete, such as a digital pain body map, with machine learning is a promising strategy to reduce bias in fibromyalgia diagnosis and improve patient outcomes. PERSPECTIVE: This investigation applies hierarchical clustering to patient-reported, digital pain body maps, finding an association between body map responses and clinical fibromyalgia diagnosis. Rapid, computer-assisted interpretation of pain body maps would be clinically useful in prompting more detailed assessments for fibromyalgia, potentially reducing gender bias.

Costs of Implementing Electronic Context Factor Assessments and Patient-reported Outcomes in Pain Clinic Settings

BACKGROUND

The Healing Encounters and Attitudes Lists (HEALs) patient-reported measures, consisting of 6 separate context factor questionnaires, predict patients' pain improvements. Our Patient-centered Outcomes Research Initiative-funded implementation project demonstrated success in using HEAL data during clinic consultations to enhance patient engagement, improve patient outcomes, and reduce opioid prescribing.

OBJECTIVE

We aimed to determine the resources needed for additional sites to implement HEAL to improve pain care treatment.

RESEARCH DESIGN

An observational study from March 1 to November 30, 2021, assessing implementation cost data from invoices, time and salary requirements for clinic personnel training, estimates of non-site-based costs, and one-time resource development costs.

SUBJECTS

Unique patients eligible to complete a HEAL survey (N=24,018) and 74 clinic personnel.

MEASURES

The Stages of Implementation Completion guided documentation of preimplementation, implementation, and sustainment activities of HEAL pain clinic operations. These informed the calculations of the costs of implementation.

RESULTS

The total time for HEAL implementation is 7 months: preimplementation and implementation phases (4 mo) and sustainment (3 mo). One hour of HEAL implementation involving a future clinical site consisting of 2 attending physicians, 1 midlevel provider, 1 nurse manager, 1 nurse, 1 radiology technician, 2 medical assistants, and 1 front desk staff will cost $572. A 10-minute time increment for all clinic staff is $95. Total implementation costs based on hourly rates over 7 months, including non-site-based costs of consultations, materials, and technology development costs, is $28,287.

CONCLUSIONS

Documenting our implementation costs clarifies the resources needed for additional new sites to implement HEAL to measure pain care quality and to engage patients and clinic personnel.

Call to focus on digital health technologies in hospitalized children's pain care: clinician experts' qualitative insights on optimizing electronic medical records to improve care

ABSTRACT

Most hospitalized children experience pain that is often inadequately assessed and undertreated. Exposure to undertreated childhood pain is associated with negative short-term and long-term outcomes and can detrimentally affect families, health services, and communities. Adopting electronic medical records (EMRs) in pediatric hospitals is a promising mechanism to transform care. As part of a larger program of research, this study examined the perspectives of pediatric clinical pain experts about how to capitalize on EMR designs to drive optimal family-centered pain care. A qualitative descriptive study design was used and 14 nursing and medical experts from 5 countries (United States, Canada, United Kingdom, Australia, and Qatar) were interviewed online using Zoom for Healthcare. We applied a reflexive content analysis to the data and constructed 4 broad categories: "capturing the pain story," "working with user-friendly systems," "patient and family engagement and shared decision making," and "augmenting pain knowledge and awareness." These findings outline expert recommendations for EMR designs that facilitate broad biopsychosocial pain assessments and multimodal treatments, and customized functionality that safeguards high-risk practices without overwhelming clinicians. Future research should study the use of patient-controlled and family-controlled interactive bedside technology to and their potential to promote shared decision making and optimize pain care outcomes.

Using Natural Language Processing and Machine Learning to Identify Opioids in Electronic Health Record Data

Purpose

This study evaluates the utility of machine learning (ML) and natural language processing (NLP) in the processing and initial analysis of data within the electronic health record (EHR). We present and evaluate a method to classify medication names as either opioids or non-opioids using ML and NLP.

Patients and Methods

A total of 4216 distinct medication entries were obtained from the EHR and were initially labeled by human reviewers as opioid or non-opioid medications. An approach incorporating bag-of-words NLP and supervised ML classification was implemented in MATLAB and used to automatically classify medications. The automated method was trained on 60% of the input data, evaluated on the remaining 40%, and compared to manual classification results.

Results

A total of 3991 medication strings were classified as non-opioid medications (94.7%), and 225 were classified as opioid medications by the human reviewers (5.3%). The algorithm achieved a 99.6% accuracy, 97.8% sensitivity, 94.6% positive predictive value, F1 value of 0.96, and a receiver operating characteristic (ROC) curve with 0.998 area under the curve (AUC). A secondary analysis indicated that approximately 15-20 opioids (and 80-100 non-opioids) were needed to achieve accuracy, sensitivity, and AUC values of above 90-95%.

Conclusion

The automated approach achieved excellent performance in classifying opioids or non-opioids, even with a practical number of human reviewed training examples. This will allow a significant reduction in manual chart review and improve data structuring for retrospective analyses in pain studies. The approach may also be adapted to further analysis and predictive analytics of EHR and other "big data" studies.

CHOIRBM: An R package for exploratory data analysis and interactive visualization of pain patient body map data

Body maps are commonly used to capture the location of a patient's pain and thus reflect the extent of pain throughout the body. With increasing electronic capture body map information, there is an emerging need for clinic- and research-ready tools capable of visualizing this data on individual and mass scales. Here we propose CHOIRBM, an extensible and modular R package and companion web application built on the grammar of graphics system. CHOIRBM provides functions that simplify the process of analyzing and plotting patient body map data integrated from the CHOIR Body Map (CBM) at both individual patient and large-dataset levels. CHOIRBM is built on the popular R graphics package, ggplot2, which facilitates further development and addition of functionality by the open-source development community as future requirements arise. The CHOIRBM package is distributed under the terms of the MIT license and is available on CRAN. The development version of the package with the latest functions may be installed from GitHub. Example analysis using CHOIRBM demonstrates the functionality of the modular R package and highlights both the clinical and research utility of efficiently producing CBM visualizations.

Whole Body Pain Distribution and Risk Factors for Widespread Pain Among Patients Presenting with Abdominal Pain: A Retrospective Cohort Study

Introduction

Abdominal pain frequently co-occurs with pain in other body sites. Chronic overlapping pain conditions (COPCs) represent a group of widespread pain diagnoses. Our study characterized how patterns of somatic pain distribution are associated with COPCs and aimed to characterize predictors of widespread pain among patients with chronic abdominal pain.

Methods

This retrospective cohort study included adults presenting to a tertiary pain clinic, reporting abdominal pain at their initial visit, and with a follow-up visit at 12 months. Body maps divided patients into localized, intermediate, and widespread pain distribution patterns. Diagnostic and psychosocial measures were assessed across groups at the initial and follow-up visits. We analyzed the association of baseline diagnoses and demographics and time-varying changes in psychosocial measures from initial to follow-up visit with changes in pain distribution over time with alternating logistic regression (ALR).

Results

Among 258 patients, most were female (91.5%) and reported widespread pain (61.5%). Those with widespread pain at baseline reported elevated anger and 60.0% of patients remained in the same pain category at follow-up. Multivariable ALR demonstrated higher pain interference (AOR 1.06, 95% CI 1.02–1.10, P = 0.002), higher anxiety (AOR 1.05, 95% CI 1.01–1.09, P = 0.01), more than one COPC at initial visit (AOR 2.85, 95% CI 1.59–5.11, P = 0.0005), and initial visit widespread pain categorization (AOR 4.18, 95% CI 2.20–8.00, P < 0.0001) were associated with an increased risk of widespread pain at the follow-up visit.

Conclusion

Most patients with abdominal pain report additional pain locations at multiple other body sites, and non-localized pain persists 12 months after pain treatment. Screening for widespread pain and COPC at the initial visit may identify patients at higher risk for persistent or new-onset widespread pain, and interventions to reduce pain interference and anxiety may promote reversal of widespread pain.

Determining chronic pain data elements as a first step towards improving quality of care and research in chronic pain

Background:

Chronic pain is a significant clinical problem in the world. There is still no quite effective treatment for this pain due to its complex nature. Timely retrieval of accurate and comprehensive information through organized clinical and epidemiological studies is an essential prerequisite for providing high quality clinical care and more accurate health planning. We aimed to determine minimum set of data needed as a first step in design and development of a chronic pain registry system.

Materials and Methods:

This descriptive-applied study was carried out in three phases; identifying necessary minimum data, preparing a primary minimum dataset, and surveying expertsby questionnaire.

Result:

The literature review revealed that, theprimary minimum dataset consisted of 51 elements, which were reduced to 41 after applying the experts’ opinion. This dataset covered six areas:demographic information(8 elements), initial pain assessment(12 elements), medical history (8 elements), mental health and well-being(6 elements), diagnostic measures(3elements), and diagnosis and treatment plan (4 elements).

Conclusion:

Determining minimum set of chronic pain data will be an effective step towards integrating and improving information management of patients with chronic pain. It will also allow for proper storage and retrieval of information related to these patients.

PROMIS Neuropathic and Nociceptive Pain Quality in Musculoskeletal Pain Presentations

OBJECTIVE

Musculoskeletal pain is a significant contributor to the global disease burden. Management of musculoskeletal pain where a neuropathic component is present can be challenging. This study evaluated the internal structure of the Patient-Reported Outcome Measures Information System (PROMIS) pain quality scales, explored the prevalence of neuropathic and nociceptive pain, and identified health demographics and behaviors related to musculoskeletal pain presentations.

METHODS

Patients presenting to the Victoria University Osteopathy Clinic (Melbourne, Vic., Australia) were invited to complete a health demographics and behaviors questionnaire, and the PROMIS Neuropathic (NeuroPQ) and Nociceptive (NociPQ) pain quality scales, before their initial consultation. Descriptive, inferential, and correlation statistics were used to evaluate the PROMIS scales, health demographics, and behaviors. Mokken scale analysis was used to evaluate the internal structure and dimensionality of the NeuroPQ and NociPQ scales.

RESULTS

Three hundred eighty-three (N=383) patients completed the measures. Mokken scaling suggested the PROMIS scales demonstrated acceptable internal structure and were unidimensional. Over 22% of patients demonstrated cutoff scores above 50, suggesting a substantive neuropathic pain component to their musculoskeletal presentation. Patients who reported cigarette smoking, not being born in Australia or not speaking English at home, demonstrated higher NeuroPQ scores. Females demonstrated significantly higher NociPQ scores than males. Pain intensity demonstrated small to medium correlations with NeuroPQ and NociPQ scores.

DISCUSSION

This study provides support for the use of the NeuroPQ and NociPQ scales in musculoskeletal pain patients. Associations with health demographics and behaviors were identified, and patients typically experienced a combination of neuropathic and nociceptive pain.

Hierarchical clustering by patient-reported pain distribution alone identifies distinct chronic pain subgroups differing by pain intensity, quality, and clinical outcomes

Background

In clinical practice, the bodily distribution of chronic pain is often used in conjunction with other signs and symptoms to support a diagnosis or treatment plan. For example, the diagnosis of fibromyalgia involves tallying the areas of pain that a patient reports using a drawn body map. It remains unclear whether patterns of pain distribution independently inform aspects of the pain experience and influence patient outcomes. The objective of the current study was to evaluate the clinical relevance of patterns of pain distribution using an algorithmic approach agnostic to diagnosis or patient-reported facets of the pain experience.

Methods and findings

A large cohort of patients (N = 21,658) completed pain body maps and a multi-dimensional pain assessment. Using hierarchical clustering of patients by body map selection alone, nine distinct subgroups emerged with different patterns of body region selection. Clinician review of cluster body maps recapitulated some clinically-relevant patterns of pain distribution, such as low back pain with radiation below the knee and widespread pain, as well as some unique patterns. Demographic and medical characteristics, pain intensity, pain impact, and neuropathic pain quality all varied significantly across cluster subgroups. Multivariate modeling demonstrated that cluster membership independently predicted pain intensity and neuropathic pain quality. In a subset of patients who completed 3-month follow-up questionnaires (N = 7,138), cluster membership independently predicted the likelihood of improvement in pain, physical function, and a positive overall impression of change related to multidisciplinary pain care.

Conclusions

This study reports a novel method of grouping patients by pain distribution using an algorithmic approach. Pain distribution subgroup was significantly associated with differences in pain intensity, impact, and clinically relevant outcomes. In the future, algorithmic clustering by pain distribution may be an important facet in chronic pain biosignatures developed for the personalization of pain management.

A Narrative Review on Perioperative Pain Management Strategies in Enhanced Recovery Pathways-The Past, Present and Future

Effective pain management is a key component in the continuum of perioperative care to ensure optimal outcomes for surgical patients. The overutilization of opioids in the past few decades for postoperative pain control has been a major contributor to the current opioid epidemic. Multimodal analgesia (MMA) and enhanced recovery after surgery (ERAS) pathways have been repeatedly shown to significantly improve postoperative outcomes such as pain, function and satisfaction. The current review aims to examine the history of perioperative MMA strategies in ERAS and provide an update with recent evidence. Furthermore, this review details recent advancements in personalized pain medicine. We speculate that the next important step for improving perioperative pain management could be through incorporating these personalized metrics, such as clinical pharmacogenomic testing and patient-reported outcome measurements, into ERAS program.

Comparative Effectiveness of Embedded Mental Health Services in Pain Management Clinics vs Standard Care

OBJECTIVE

Embedded behavioral medicine services are a common component of multidisciplinary chronic pain treatment programs. However, few studies have studied whether these services are associated with improved treatment outcomes.

METHODS

Using a retrospective, matched, two-cohort study design, we examined patient-reported outcomes (PROs), including Patient-Reported Outcomes Measurement Information System pain, mental health, and physical function measures, collected at every clinic visit in every patient. Changes from baseline through 12 months were compared in those receiving embedded Behavioral Medicine in addition to usual care to a Standard Care group seen in the same pain practice and weighted via propensity scoring.

RESULTS

At baseline, Behavioral Medicine patients had worse scores on most pain, mental health, and physical health measures and were more likely to be female, a member of a racial minority, and have lower socioeconomic status. Regardless of having a worse clinical pain syndrome at baseline, at follow-up both Behavioral Medicine (N = 451) and Standard Care patients (N = 8,383) showed significant and comparable improvements in pain intensity, physical function, depression, and sleep disturbance. Behavioral Medicine patients showed significantly greater improvements in their global impressions of change than the Standard Care patients.

CONCLUSIONS

Despite worse pain and physical and psychological functioning at baseline, Behavioral Medicine patients showed improvements comparable to patients not receiving these services. Further, Behavioral Medicine patients report higher global impressions of change, indicating that embedded mental health services appear to have the additive value of amplifying the benefits of multimodal pain care.

Development and validation of the Collaborative Health Outcomes Information Registry body map

Supplemental Digital Content is Available in the Text.

Introduction:

Critical for the diagnosis and treatment of chronic pain is the anatomical distribution of pain. Several body maps allow patients to indicate pain areas on paper; however, each has its limitations.

Objectives:

To provide a comprehensive body map that can be universally applied across pain conditions, we developed the electronic Collaborative Health Outcomes Information Registry (CHOIR) self-report body map by performing an environmental scan and assessing existing body maps.

Methods:

After initial validation using a Delphi technique, we compared (1) pain location questionnaire responses of 530 participants with chronic pain with (2) their pain endorsements on the CHOIR body map (CBM) graphic. A subset of participants (n = 278) repeated the survey 1 week later to assess test–retest reliability. Finally, we interviewed a patient cohort from a tertiary pain management clinic (n = 28) to identify reasons for endorsement discordances.

Results:

The intraclass correlation coefficient between the total number of body areas endorsed on the survey and those from the body map was 0.86 and improved to 0.93 at follow-up. The intraclass correlation coefficient of the 2 body map graphics separated by 1 week was 0.93. Further examination demonstrated high consistency between the questionnaire and CBM graphic (<10% discordance) in most body areas except for the back and shoulders (≈15–19% discordance). Participants attributed inconsistencies to misinterpretation of body regions and laterality, the latter of which was addressed by modifying the instructions.

Conclusions:

Our data suggest that the CBM is a valid and reliable instrument for assessing the distribution of pain.

Identifying Predictors of Recommendations for and Participation in Multimodal Nonpharmacological Treatments for Chronic Pain Using Patient-Reported Outcomes and Electronic Medical Records

OBJECTIVE

High-quality chronic pain care emphasizes multimodal treatments that include medication and nonpharmacological treatments. But it is not clear which patients will participate in nonpharmacological treatments, such as physical therapy or mental health care, and previous research has shown conflicting evidence.

METHODS

We used the Patient Outcomes Repository for Treatment (PORT) registry, which combines patient-reported outcomes data with electronic medical records. In this retrospective observational study, we performed two separate multinomial regression analyses with feature selection to identify PORT variables that were predictive of 1) recommendation of a nonpharmacological treatment by the provider and 2) patient participation in nonpharmacological treatments. Two hundred thirty-six patients were recommended (REC) or not recommended (NO REC) a nonpharmacological treatment, and all REC patients were classified as participating (YES) or not participating (NO) in the recommendations.

RESULTS

Female gender and a diagnosis of Z79 "Opioid drug therapy" were significant positive and negative predictors of nonpharmacological treatment recommendations, respectively. Schedule II opioid use at initial presentation and recommendations for rehabilitation therapy were significant predictors of nonparticipation.

CONCLUSIONS

Patients using opioids are less likely to be recommended nonpharmacological treatments as part of multimodal chronic pain care and are less likely to participate in nonpharmacological treatments once recommended. Males are also less likely to be recommended nonpharmacological treatments. Patients referred for rehabilitation therapies are less likely to comply with those recommendations. We have identified patients in vulnerable subgroups who may require additional resources and/or encouragement to comply with multimodal chronic pain treatment recommendations.

Information and communication technology for increasing healthy ageing in people with non-communicable diseases: identifying challenges and further areas for development

Information and communication technology (ICT) within healthcare covers a range of technologies that aim to improve disease management or help modify health behaviors. We discuss clinical practice and system-related ICT challenges in Europe in relation to healthy ageing in people with non-communicable diseases (NCD). Although ICT use within healthcare is increasing, several challenges remain, including: (i) variations in ICT use within Europe; (ii) under-use of electronic health records; (iii) frequent use of single domain outcomes; (iv) shortage of clinical trials on current technologies; (v) lack of involvement of patients in ICT development; (vii) need to develop and adapt ICTs for people with cognitive or sensory impairment; and (viii) need to use longitudinal big data better. Close collaboration between key stakeholders (academia, biopharmaceutical and technology industries, healthcare, policy makers, patients, and caregivers) should foster both technological innovation and innovative models to facilitate more cost-effective approaches, ultimately leading to increased healthy ageing.