Computer-Administered Patient Reported Outcomes (PRO) and Psychiatric Screening in Outpatient Pain Patients: Effect of a Point-of-Care Biopsychosocial Patient Health Report on Treatment Outcomes




Swanholm, Eric Neal

Journal Title

Journal ISSN

Volume Title



RATIONALE: Chronic pain is a widespread health problem that carries steep costs for both individuals and society. Pain-related complaints represent one of the most common presenting symptoms across ambulatory care settings. Individuals with chronic pain often have comorbid psychiatric symptoms and/or psychosocial dysfunction. Given the related impact on treatment and health-care costs, tracking psychiatric and psychosocial outcomes is beneficial for chronic pain patients, their health care providers, and service providers. Outcome-tracking interventions that could positively affect treatment outcomes hold potential benefits for patient care.
PURPOSE: The purpose of the present study was to examine the effects of the regular collection and results feedback prior to the point-of-care on multiple patient-reported outcome (PRO) domains in outpatient pain patients. Maximizing ecological validity and non-disruption of clinic flow were given significant focus in the study design and process.
PATIENTS AND METHODS: This study used repeated measures and was conducted in an outpatient pain management and interdisciplinary treatment clinic (2 anesthesiologists, 1 psychiatrist, 1 psychologist, 1 counselor, 1 physical therapist). A total of 69 pain patients were randomly assigned to one of two protocol-based PRO feedback intervention groups (separate feedback to both patients and providers [Dual Feedback]; Provider-Only Feedback) or a non-intervention group (Chart-Review Only). Assessments were completed prior to the point-of-care; feedback for intervention groups was based upon a real-time, automated report generated from their PRO data. Data were gathered on touch-screen tablet-pc’s using multiple computer-adaptive-tests from the NIH-sponsored Patient Reported Outcomes Measurement Information Systems (PROMIS) Assessment Center platform; outcome domains included pain-related functioning (Pain Disability Questionnaire [PDQ]; VAS pain rating; PROMIS Physical Functioning, Pain Behavior, and Pain Interference), psychological symptoms and psychosocial variables (PROMIS Depression, Anxiety, Sleep-Related Impairment, Fatigue, Social Functioning; hypomania history screen; Pain Medication Questionnaire [PMQ; opioid misuse risk]), global HRQoL (PROMIS Mental and Physical Health domains), treatment alliance (Working Alliance Inventory [WAI]; ratings from both patients and providers]), and illness perception and optimism (Brief Illness Perception Questionnaire [BIPQ]; Life Orientation Test- Revised). Performance-based data (walking time, grip strength, range-of-motion/flexibility) were collected by the physical therapist for study patients whose treatment included a PT component (e.g. interdisciplinary pain program, individual services). Significant covariates were identified and incorporated into the primary analyses. Primary outcomes were the individual measures within each outcome domain. Analyses utilized mixed-effects modeling with random coefficients and multiple regression in comparisons of all three study groups. Secondary analysis included tabulation of completion time and comparisons between a Combined Feedback group (both intervention groups) and Chart-Review Only.
RESULTS: Significant covariates included treatment type, history of psychiatric diagnosis, and a biological family history of psychiatric diagnosis. Comparing Dual Feedback vs. Chart Review Only, patients in the Dual Feedback intervention had significantly better outcomes over time for a number of domain outcomes; specifically, in pain-related functioning/symptoms (PDQ [P = .003]; PROMIS Pain Interference [P = .023]; VAS pain [P = .03]), psychological and psychosocial variables (PROMIS Anger [P = .001]; PROMIS Anxiety [P = .012]; PROMIS Depression [P = .029]; PROMIS Sleep-Related Impairment [P = .001]; PROMIS Social Functioning – Satisfaction with Discretionary Social Activities [P = .047]), PROMIS Global HrQOL (Mental Health [P = .021]; Physical Health [P = .032]), treatment alliance (WAI – Bond [patient-rated][P = .046]), illness perceptions (BIPQ – Consequence [P = .017]; BIPQ – Timeline [P = .011]; BIPQ – Treatment Control [P = .029]), and one performance-based measure (Walk Time [P = .007]). Similarly, patients in the Provider-Only group had better outcomes over time for multiple outcome domains; including, pain-related functioning/symptoms (PDQ [P = .033]; PROMIS Pain Interference [P = .031]; PROMIS Fatigue [P = .036]; PROMIS Physical Functioning [P = .049]), psychological and psychosocial variables (PMQ [opioid misuse risk][P = .041]), treatment alliance (WAI – Bond [patient-rated][P = .076]; WAI – Bond [provider-rated][P = .008]), illness perceptions (BIPQ – Timeline [P = .048]; BIPQ – Personal Control [P = .027]), and one performance-based measure (Walk Time [P = .035]). Comparisons between patients in the Dual Feedback and Provider-Only Feedback groups were significant for a few domain outcomes. Compared to Provider-Only Feedback, Dual Feedback had better outcomes over time for multiple domain measures; including, the PDQ (P = .085), PROMIS Anger (P = .000), PROMIS Anxiety (P = .018), and BIPQ – Treatment Control (P = .015). Conversely, the Provider-Only group had better outcome scores over time for PROMIS Global HrQOL (Mental Health (P = .032); Physical Health (P = .074). Analyses of process variables showed a mean completion time of 15.8 minutes for the entire assessment; completion-time statistics were also calculated for the 11 PROMIS computer-adaptive-tests (M = 7.57 minutes [all PROMIS CAT’s]; M = 41.3 second per measure, SD = 9.3 seconds) and other primary outcomes (PDQ, PMQ, BIPQ) (M = 8.23 minutes total; M = 2.74 minutes per measure, SD = .99 minutes). CONCLUSION: The provision of dual feedback (patient and providers) from PRO data collected prior to the point-of-care had an impact on several outcomes from multiple domains (pain-related functioning, psychological symptoms, psychosocial variables, illness perception, walking performance) over time, compared to patients who received no point-of-care feedback. To a lesser extent, group by time effects were also observed in comparisons between patients receiving provider-only feedback and those with no feedback. Brought together, high ecological validity was maintained with minimal disruption of clinic flow; likely contributing factors include the use of a set framework for outcome-tracking, protocol-based delivery of feedback, and efficiency of administration. This is the first study to show the potential benefits of providing PRO data feedback to both patients and providers prior to the point-of-care.

General Notes

Table of Contents


Related URI