University of Pittsburgh


Grant Number: 2U01AR052155
Project Title: Validation and Domain Development in Mental Health and Sleep–Wake Function
PI Name, Title and Contact Information: Paul A. Pilkonis, PhD, Principal Investigator, Professor of Psychiatry and Psychology, University of Pittsburgh School of Medicine
Other Key Personnel: Paul A. Pilkonis, PhD, Principal Investigator
Daniel J. Buysse, MD, Co-Investigator
Andrew Krystal, MD, Co-Investigator
Matthew Schonlau, PhD, Co-Investigator
Web Site: http://www.wpic.pitt.edu

Abstract:

The goal of the Patient-Reported Outcomes Measurement Information System (PROMIS) is to develop an adaptive system for measuring outcomes in chronic diseases. The University of Pittsburgh has served as a PROMIS research site since 2004, focusing on the domains of emotional distress and sleep-wake function. We developed and calibrated five PROMIS item banks (depression, anxiety, anger, sleep disturbance, and sleep-related impairment) of the 11 currently available from the PROMIS Network. This renewal application will continue and extend our earlier work by validating further the item banks we have developed using new samples, underrepresented groups (older adults), and longitudinal analyses. Our specific aims are

  1. to validate further the PROMIS item banks for emotional distress by demonstrating the feasibility and value of CAT administration and by establishing the greater sensitivity to change of the PROMIS measures compared to conventional legacy instruments in a high-risk population–geriatric patients with depression; and
  2. to validate further the PROMIS sleep disturbance and sleep–related impairment item banks with work on construct validity (comparing known groups with clinically diagnosed sleep disorders versus no sleep disorder) and work on responsiveness to treatment and minimally important differences by following patients with sleep disorders from initial evaluation through six to eight weeks of treatment. In all studies, we propose to investigate the value of the comprehensive PROMIS health status profile (i.e., assessments across all 11 current PROMIS item banks) for identifying novel associations between domains and novel profiles of patients with chronic diseases. Our work will involve collaboration with various research groups within the University of Pittsburgh and across the country (RAND, Duke University).

Specific Aims

Patient-reported outcomes (PROs) are important measures of health status that capture the impact of chronic diseases. PROs assess subjective experience in ways distinct from physiological outcomes, and they include such indicators as symptom frequency and severity, emotional and social well–being, and perceived level of health and functional ability. The goal of the Patient–Reported Outcomes Measurement Information System (PROMIS) is to develop and test a psychometrically validated, adaptive system to measure PROs in respondents with a wide range of chronic diseases and demographic characteristics. The University of Pittsburgh has served as a principal research site for PROMIS since 2004, focusing on the areas of emotional distress and sleep-wake function with demonstrated productivity during the initial funding period. We developed five PROMIS item banks (depression, anxiety, anger, sleep disturbance, and sleep-related impairment) of the 11 currently available through a comprehensive process of systematic literature searches, review by content experts, qualitative research with patients, and rigorous psychometric analyses using models from item response theory (IRT).1, 2 We have also demonstrated the value of computerized adaptive testing (CAT) with these item banks.3

Nonetheless, important steps remain if the PROMIS item banks are to be disseminated and applied effectively in research and clinical settings. This renewal application will continue and extend our earlier work by validating further the item banks we have developed for emotional distress and sleep–wake function using new samples, underrepresented groups (older adults), and longitudinal analyses. The new studies will rely on methods we have refined during the previous funding period, but they will also be innovative in several ways. Specifically, we will include new samples of older adults; examine the clinical utility of PROMIS measures in applied treatment contexts; utilize CAT to maximize efficiency in real-world settings; and demonstrate the value of the overall health status profile that can be derived from concurrent administration of all available PROMIS item banks. We will pursue these general goals with an aggregate sample of over 1100 new participants. These participants will come from multiple sources: the RAND American Life Panel (ALP), a probability-based internet survey panel (n = 800) and two university sites–the University of Pittsburgh and Duke University (n = 280). We regard the size and diversity of our research team, the multiplicity of our aims, and the variety of our sampling frames as strengths that increase the value of the work we will undertake during the second cycle of PROMIS funding. We will address the following specific aims.

AIM 1. To validate further the PROMIS item banks for emotional distress (with a primary emphasis on depression).

Validation work will focus on two issues: demonstrating the value of CAT administration by verifying empirically our work on CAT simulation in the initial PROMIS calibration sample, and establishing the greater sensitivity to change of the PROMIS measures compared to conventional legacy instruments in a high-risk population—geriatric patients with depression. The first issue will be addressed using a subsample (n = 800) from the American Life Panel, an internet survey panel with over 2000 members across all 50 states. The second issue (n = 75) will involve collaboration with the Advanced Center for Intervention and Services Research (ACISR) on late-life mood disorders in the Department of Psychiatry at the University of Pittsburgh (PI: C. F. Reynolds, III, MD, P30MH071944).

  • Hypothesis 1a: Actual CAT scores will correlate as highly with full-bank scoring as did simulated CAT scores (r = .98 for depression) from the initial PROMIS calibration sample.
  • Hypothesis 1b: PROMIS measures will show greater sensitivity to change (i.e., larger effect sizes) than conventional legacy measures, and they will correlate more highly with both patient and clinician reports of global improvement related to treatment.

AIM 2: To validate further the PROMIS sleep disturbance and sleep-related impairment item banks.

Validation of the sleep disturbance (SD) and sleep-related impairment (SRI) item banks will include work on construct validity (comparing known groups with clinically diagnosed sleep disorders versus no sleep disorder) and determination of treatment responsiveness and minimally important differences by following patients with sleep disorders from initial evaluation through 6–8 weeks of treatment. Participants will be 280 individuals with obstructive sleep apnea (OSA), chronic insomnia (CI), or good sleep (GS), recruited from two sleep medicine centers (University of Pittsburgh and Duke University).

  • Hypothesis 2a: Scores on the PROMIS SD and SRI measures will differ significantly among the three groups, with patients with CI scoring highest on SD and patients with OSA scoring highest on SRI.
  • Hypothesis 2b: Scores on the PROMIS SD and SRI measures will display moderate to large correlations (r’s > .50) with similar retrospective (qualitative) self-report measures, supporting convergent validity.
  • Hypothesis 2c: Both SD and SRI measures will show greater sensitivity to change (i.e., larger effect sizes) than other common retrospective self-report measures, and they will correlate more highly with both patient and clinician reports of global improvement related to treatment.

Public Health Relevance

Our proposed new studies will enhance the value and public health impact of PROMIS by including a new sample of older adults; examining the utility of PROMIS measures in applied treatment contexts; utilizing CAT to maximize efficiency in real–world settings; and demonstrating the value of the overall health status profile that can be derived from all available PROMIS item banks

Populations

General population sample (n = 800) from the RAND American Life Panel

Geriatric outpatients suffering from depression (n = 75)

Patients with chronic insomnia (n = 110) or sleep apnea (n = 110)

Individuals from the general population with no sleep disorder (n = 60)

Disease Groups

Depression

Sleep disorders (insomnia, sleep apnea)

References

  1. Pilkonis, P. A., Choi, S. W., Reise, S. P., Stover, A. M., Riley, W. T., & Cella, D. (2010). The development of item banks for emotional distress from the Patient-Reported Outcomes Measurement Information System (PROMIS): Depression, anxiety, and anger. Manuscript under review.
  2. Buysse, D. J., Yu, L., Moul, D. E., Germain, A., Stover, A., Dodds, N. E., Johnston, K. L., Shablesky-Cade, M. A., & Pilkonis, P. A. (2010). Development and validation of patient-reported outcome measures for sleep disturbance and sleep-related impairments. Sleep, 33, 781-792.
  3. Choi, S. W., Reise, S. P., Pilkonis, P. A., Hays, R. D., & Cella, D. (2010). Efficiency of static and computer adaptive short forms compared to full length measures of depressive symptoms. Quality of Life Research, 19, 125-136.

For questions concerning the role of the University of Pittsburgh in the PROMIS network, please contact Paul A. Pilkonis, PhD at pilkonispa@upmc.edu. For more information about the PROMIS network, go to http://www.nihpromis.org.