Our team's previous research on VA's '5th Vital Sign' found that the measure of current pain intensity moderately corresponded to chronic disabling pain measured with the Brief Pain Inventory, staff delivering the assessment used informal and unstructured pain queries nearly 50% of cases, and there was a link missing between current pain screening approaches and improved care. These findings informed the specific design of this Effective Screening for Pain (ESP) study as well as the proposed research products. The ESP study focused on a qualitative evaluation of provider and Veteran perspectives on pain screening and measures, as well as barriers and facilitators to implementation of higher quality pain care, and a quantitative evaluation of the relative effectiveness of simple alternative measures compared to the current pain screening approach using the Numeric Rating Scale (NRS).
Our Specific Aims included to:
1) Qualitatively evaluate Veteran and multidisciplinary provider perspectives on pain screening and the use of the NRS vs. the PEG items, point of care vs. mobile reporting, clinician-assessed vs. patient-reported pain, and usefulness of medical chart and patient-derived health information in screening, assessment, management, and follow-up of pain.
2a) Quantitatively assess, in a 3-arm randomized controlled primary care clinical team-based trial using tablets, the feasibility and completion rates, validity, and variability of pain information obtained comparing: Tablet-based vs. nurse-documented pain (e.g., 5th vital sign by tablet vs. clinician assessed) and the rate and severity of pain detected in tablet-based 'NRS one week' (Arm 1), tablet-based PEG (Arm 2), and usual 5th vital sign (Arm 3). Also pain, self-reported disability, and Veteran satisfaction with pain screening mode for chronic pain detection.
2b) Informed by Aims 1 and 2a, qualitatively evaluate provider perspectives on different pain reports and features of informatics prototypes to facilitate better pain management.
We did not collect data following the healthcare visit. Because the study was not interventional, we decided it was inappropriate to expect a change in the pain measure. Concurrently, we also decided to collect additional data in the randomized controlled trial on the Defense Veterans Pain Rating Scale (DVPRS) which was supported as an alternative to the NRS and was implemented in some VA sites.
We conducted a two phase mixed method study that builds on prior work to develop and test enhanced pain screening approaches for primary care and Patient Aligned Care Teams (PACTs).
Aim 1: In the first (development) phase of the study, we conducted semi-structured qualitative interviews and focus groups with 60 primary care clinicians and other primary care team members at hospital and community-based PACT sites in diverse settings including Los Angeles (VA GLAHS PACT, downtown CBOC PACTs, Sepulveda PACT) and Oregon (Portland PACT, Vancouver CBOC, Bend CBOC). The groups were presented with questions related to their experiences with the '5th Vital Sign', aspects of teamwork in pain management, queried regarding specific alternative measures (e.g., the PEG vs. the NRS), and asked about how they use data for clinical decision-making including how pain reports can best be integrated into primary care team processes. The focus groups were conducted in person and varied between 6 and 9 participants. Interviews were anonymized and transcribed verbatim, entered into Atlas.ti, codebooks were developed iteratively by two coders to achieve consensus. This informed the development of the enhanced pain screening approaches used in the Aim 2 randomized controlled trial (RCT).
Aim 2a: We collaborated with the University of California; Los Angeles Office of Information Technology to ensure the development of a tablet-based pain screening survey adhered to accessibility standards. Between Phase 1 and 2, the Principal Investigator moved to the VA Palo Alto. The second (testing) phase of the study included a multisite RCT in the PACT clinics at the VA Palo Alto, VA Portland, and VA Minneapolis . Alternatives tested included the use of a tablet-based self reporting, NRS 1 week, PEG, DVPRS, and the Graded Chronic Pain Scale. The tablet-based survey included a number of validated measures that reflected the 'total pain model' (physical, emotional, spiritual, and social wellbeing), and an additional validated functional status measure (e.g., self-reported function compared to one's peers). The total tablet based survey was designed to maximize extreme brevity because patients were approached, consented and enrolled, and interviewed in the entirety before PACT visits. In many cases, the clinic wait durations were extremely short. (e.g., 10 minutes).
We compared nurse to tablet administration of the same tool (e.g., nurse-recorded NRS to tablet-recorded NRS) with nurse-tablet and tablet-tablet administration of different tools (e.g., NRS to GCP). Agreement between the nurse (nurse-recorded NRS) and the tablet documentation of pain (tablet-recorded NRS) was assessed by the interclass correlation (referred to subsequently as "correlation"). Correlations among research ratings (research-recorded NRS and GCP) were used to distinguish tool from measurement differences. Cut points for the NRS are 5 for moderate pain and 7 for severe pain
Aim 2b: We conducted individual semi-structured interviews with patients(n= 36) and providers (n=15) -including VA primary care providers, nurse practitioners, and Clinical pharmacists. The methods used for analysis included qualitative approaches used in Aim 1. Our initial analyses of Aim 1 and Aim 3 content have focused on critical, high value findings.
Aim 1: The following summarizes major findings from Aim 1 provider focus groups:
1) The importance of non-pharmacologic management: Current approaches to screening and pain management in PACT lead to bad outcomes in part because pharmacologic approaches are available, but various non-pharmacologic options are difficult to access.
Provider perceptions of barriers to reducing opioids and improving use of nonpharmacologic pain management strategies for patients with chronic pain clustered around availability and access. Barriers to nonpharmacologic therapy access included the following sub-themes: geographical (patient distance from service), financial (out of pocket cost to patient), temporal (treatment time delays), cultural (belief that nonpharmacologic therapies increased provider workload, perception of insufficient training on nonpharmacologic therapies, perceptions of patient resistance to change, confrontation avoidance, and insufficient leadership support), and digital (measure used for pain assessment, older patients hesitant to use technology, providers overwhelmed by information).
2a) Despite understanding that they should respect the patient report, providers struggle with pain and the conflict between how patients describe pain and the conflict between those statements and what they observe.
Providers and staff acknowledge that pain is a highly individualized and subjective experience not adequately captured by the NRS. Providers described a common conflict between the patient's reported experience of pain and the clinician's observations or expectations of pain. Providers attempt to recalibrate the patient's reported experience to fit their own understanding of pain. Providers perceive that patients over report their pain because they do not understand how to use the pain scale to standardize their subjective experience.
2b) Multi-disciplinary team based approaches to screening and managing pain in primary care should incorporate members of the extended PACT team including pharmacy, mental and behavioral health, and social work.
Providers offered solutions to improve integration of ancillary team members for pain. Providers suggested inclusion on case management rounds and joint visits as strategies to better integrate ancillary team members and specialty providers in the team process. In reference to working more closely with different types of providers involved in pain care, PCPs suggested that "maybe you can, as a team, come up with a new approach."
Case management rounds: "I think if you have a team set up that's very highly functional where they actually have, you know, case management rounds with the nurses and maybe social workers and the providers and the teams have a good understanding of the patient and are talking about the patient's limitations and how they can help." [PCP]
Joint visits with core team members and specialists: "Joint visits between provider, nurse care manager, patient, and also mental health provider and patient, is something that we don't do very often at all but is very, very useful. it's something that we sometimes use and recommend for the tough cases that we have. And that is, not readily available enough, that kind of a [joint] visit.. If we had more of the team-the doctor, the RN, mental health, social services or whatever-in with the group meeting. that way the patient is more involved with the care plan." [PCP]
Clinical Pharmacists were identified as participating with refill management, state prescription drug monitoring, medication education, naloxone use, and managing laboratory screening for opioids, and narcotic tapering. Specific barriers to clinical primary care pharmacists taking on expanded roles around pain care include limitations of local scopes of practice, support from the institution (insufficient training, insufficient time to get training, low staffing), physicians lacking awareness of pharmacists' capabilities, and the demands of competing medical conditions. Additionally, decentralization of care for some conditions (e.g. all pharmacists covering some anticoagulation patients) versus use of pharmacist dedicated clinic time for other conditions (e.g. pain) causes tension around primary care pharmacists workload distribution.
Aim 2: The following summarizes quantitative results of the Aim 2 randomized controlled trial of NRS with a one week look back, PEG, and DVPRS.
We enrolled a random outpatient visit-based sample of 581 patients. The sample came form 3 VA health systems (Palo Alto, Portland, Minneapolis) and included 5 sites ( VA hospital and community based outpatient clinics). All surveys were completed February - June 2016. Of the 2258 Veterans approached, 632 / 1323 eligible Veterans participated. Of the 632 recruited 581 surveys were completed, 4 withdrew and 47 were incomplete. The main reason for an incomplete survey was patients getting called back to their visit during the middle of the survey and not coming back to finish in after their visit. Veterans were broadly representative of the three sites from which they were recruited and approximately balanced by study Arm. Among the three study Arms (NRS one week, PEG, and DVPRS), age varied between 60-62 years, white race characterized 73-79% and male gender 93%-81% of respondents. Items (e.g., wellbeing, self-reported physical activity) were similar across groups except back pain was lower in the DVPRS Arm (7% vs. 13-14% in Arm 1,2).
We compared the three approaches to pain screening with the Graded Chronic Pain Scale. Because of errors in programming of the tablet and an error in data cleaning after survey piloting of the initial sample, a non-random error in the tablet algorithm for GCPS function went undetected. Corruption of the GCPS functional scale was identified after recruitment closure. We therefore used the GCPS pain intensity as the gold standard comparator to the screening measures. That comparison reveals:
Agreement of the GCPS pain intensity to simple pain screening measures
N (% agreement with the GCPS)None Mild Mod Severe
Tablet NRS one week 8 (4.42) 20 (11.05)14 (7.73)83 (45.86)
PEG 9 (5.33)14 (8.28)11 (6.51)76 (44.97)
DVPRS 4 (2.38) 23 (13.69) 11 (6.55) 66 (39.29)
Because we included an alternative validated functional status measure, we will use it to compare functional differences associated with pain screening. Our preliminary results include that all 3 measure detected pain intensity at roughly equivalent rates. The Tablet NRS and PEG detected clinically important moderate to severe pain with higher GCPS agreement than the DVPRS.
Aim 3: The following summarizes key messages from qualitative analyses of Veteran perspectives on pain screening and provider perspectives on
1) Barriers to improving uptake of non-pharmacologic pain management therapies (NPTs) in a primary care context.
Emergent Veteran identified themes include 1) Limited provider offerings: "There wasn't a replacement regimen and that was frustrating" and "I've wanted to explore alternatives to [medication] and they have not done that for me." 2) Competing demands on provider time: "Often times doctors go, 'Here, have a narcotic. Have a nice day.' they have too much caseload to pay attention." 3) Multi-dimensional access barriers: geographical (distance from service), financial ("its cost-prohibitive.it's not that PT's a problem, it's the system"), temporal (long wait times: "So I've kind of given up on the idea of physical therapy with the VA.I've had to wait hours to see the PT") and digital (provider turnaround time: "E-mailed my doctor and said, 'Hey, I need an alternative to [Medication] for my back and shoulder pain.' After three weeks I got no response.") 4) Patient expectation to be pain free: "We believe as a society we should be inoculated from pain." 5) Doctor shopping: "we have shifted the focus to "if this doctor doesn't relieve me of my pain I will find someone who does." 6) Patient activation: currently there is not enough "educating people that they're a part of their pain management."
2) Providers identify unique challenges associated with each pain measure and our thematic content analysis revealed 5 themes around how providers evaluate pain screening tools.
Providers shared five unique themes regarding how they view the PEG and DVPRS as brief pain screening alternatives to the '5th Vital Sign'. 1) Item abstractness or perception of how bounded and concrete a patients interpretation of an individual item will be "I suspect how does pain interfere with sleep is easier for patients to understand than how does pain interfere with enjoyment in life". 2) Item distinctness or the patients ability to differentiate between the meaning of various items in a pain measure "Those last two [PEG] questions - how much does pain interfere with enjoyment of life and how much does pain interfere with general activity; they struggle to differentiate between what those two things are." 3) item anchoring: of the need for descriptions to accompany numbers making the meaning of pain screening items explicit. 4) Item look-back period or the retrospective time over which patients are asked to remember and comment on their pain. 5)Parsimony: Identifying the shortest and simplest approach possible to acquire desired information "shorter is better, so PEG is shorter- I like it better".
In addition to health system purposes (e.g., payment and accountability, population health) clinicians and patients remind us that pain measures have to serve the purposes of clinical assessment and as informative tools for management.
1.Patients and providers affirmed an informative role for pain screening, even as providers suggested targeted screening to balance population health goals with the demands screening places on clinical practice.
2.Screening is a limited part of the pain care ecosystem, and it has downstream effects that providers perceive as contributing to medication overuse because of management gaps including resources to address pain.
3.Each measure had strengths and limitations, but neither one of the three simple screening measures evaluated addresses the universe of issues critical to different goals of measurement.
a.Regarding population health implications, the NRS has the lowest rate of missingness at 2%; whereas, both the PEG and DVPRS had rates of missingness around 6%.
b.Regarding clinical care and intervention, brevity favors the existing NRS or the PEG, but providers find some items on both PEG and DVPRS functional scales challenging to interpret. Nevertheless, function is viewed as an important indicator of patient response to treatment.
4.Pain management may benefit from improvements to the pain measure and relevant clinical data, team-based processes, non-pharmacologic resources, and followup care. Providers endorse inter-relationships among aspects of pain management which has implications for planning and evaluating research interventions, and implementing improvements in pain care including various opportunities and limitations of strategies to mitigate opioid overuse.
External Links for this Project
Grant Number: I01HX000938-01
None at this time.