Many complementary and integrative health approaches have been shown to be effective for chronic pain and included in guidelines. This evidence of effectiveness is built on hundreds of studies representing millions of research dollars, and the ability to analyze and better compare results across these studies is essential to obtain the full value of this investment. However, useful across-study comparisons which would allow better understanding and targeting of these interventions are hampered by at least two challenges: the lack of common outcome measures and the inability to meaningfully stratify or classify patients.

The first two Aims of this study each address one of these two challenges. The last Aim takes advantage of this study to further test the viability of MTurk as a data collection platform.

Specific Aim 1 addresses the lack of common outcome measures. This Aim is to develop and evaluate links or crosswalks between the PROMIS-29 and other common (legacy) measures used for chronic pain so that the results of studies using different measures can be compared. Our hypothesis here is that we will be able to create some sort of linkage between PROMIS and legacy measures that will allow for each pair estimates of one to be made from the other.

Specific Aim 2 addresses the inability to meaningfully stratify or subclassify chronic pain patients. This Aim is to evaluate and refine the chronic pain impact stratification scheme (ISS) proposes by the NIH Research Task Force on chronic low back pain, which is based on 9 PROMIS-29 items. Our hypothesis here is that the ISS can be refined so that it can identify subgroups of chronic pain patients with different levels of morbidity and different needs for treatment.

Specific Aim 3 addresses the potential for more efficient data collection using the MTurk platform. This Aim is to evaluate MTurk as a cost- and time-efficient method to collect quality data on individuals with chronic pain. Our hypothesis here is that we find that MTurk produces data of sufficient quality to be of use to future research projects.

While there is an extensive literature on the effectiveness of pharmaceutical and nonpharmacologic interventions for chronic low back pain (CLBP), it is challenging to synthesize the findings because of differences in the CLBP samples and in the outcome measures used. The NIH Research Task Force (RTF) on CLBP noted that these differences make it "difficult to compare epidemiologic data and studies of similar or competing interventions, replicate findings, pool data from multiple studies, resolve conflicting conclusions, develop multidisciplinary consensus, or even achieve consensus within a discipline regarding interpretation of findings." To this list we would add that these differences also prevent the use of the results to answer questions such as 'Which therapies work best? And for whom?' This project tackles two things that are needed to address these differences and allow for better cross-study comparisons: we will develop empirical links between different outcome measures to allow their expression in similar units (Aim 1), and we will refine the RTF's proposed stratification (subgrouping) of patients by the impact of their CLBP (Aim 2).

To address the first challenge, in Aim 1 this study will develop and evaluate crosswalks or links between components of the 29-item Patient-Reported Outcomes Measurement Information System (PROMIS®) short form (PROMIS-29) and common legacy measures used for chronic pain. The purpose of these crosswalks or links is to allow researchers who measured outcomes using one instrument to estimate what the outcome would be if it had been measured using the other instrument. In particular, we will create crosswalks/links for the two most commonly used instruments used to measure outcomes for CLBP: the Roland-Morris Disability Questionnaire (RMDQ) and the Oswestry Disability Index (ODI). In addition, depending on data availability and input from our Advisory Council we will create at least two other crosswalks/links between the PROMIS-29 and other legacy measures for CLBP (e.g., the Back Pain Functional Scale) or legacy measures for other types of chronic pain (e.g., the Neck Disability Index for chronic neck pain).

To address the second challenge, in Aim 2 we will evaluate and refine the chronic pain impact stratification scheme proposed by the NIH Research Task Force on chronic low back pain. The proposed scheme uses the Impact Stratification Score (ISS) which is calculated using 9 items from the PROMIS-29. This ISS was intended to identify and categorize patients with chronic pain into groups based on the severity of their condition so that treatment can be better targeted. We will first evaluate the ISS and its properties to determine whether they are stable across different samples and determine whether they can be improved. After we have finalized the components and calculation of the ISS, we will examine its effect on the impacts of chronic pain (e.g., health-related quality of life, healthcare utilization, worker productivity) to identify meaningful cut-points to use to stratify chronic pain patients into subgroups who exhibit different levels of chronic pain impact.

Three types of data will be used in the analyses to address Aims 1 and 2: data from existing datasets built during other studies, data collected from an anonymous national convenience sample using Amazon's Mechanical Turk (MTurk) crowdsourcing platform, and data from members of the probability-based nationally representative KnowledgePanel. Aim 3 will evaluate whether MTurk is a reliable, efficient method to collect data on individuals with chronic pain. One part of this evaluation will involve comparison of the results from MTurk to what was found using KnowledgePanel.

Inclusion Criteria:

• adult
• self-described chronic pain
• member of one of the panels (MTurk or KnowledgePanel) we are using for data collection

Exclusion Criteria:

• unwillingness to participate

We have two cohorts. Both will have similar inclusion and exclusion criteria. The first from MTurk is a non-probability sample, and the second is from KnowledgePanel and is a probability-based sample, but one that has been established by the group who owns this panel.

All data collected is shared.

Data were made available 3/8/24

All researchers with access to ICPSR - openicpsr-198049