Does Sharing Patient-Reported Outcomes with Doctors Improve Patient Symptoms? [Internet]

Excerpt

Background: Symptoms account for more than 400 million clinic visits annually in the United States. Among the most prevalent, chronic, disabling, and undertreated symptoms in clinical practice are SPADE symptoms—sleep, pain, anxiety, depression, and low energy/fatigue. Recently, there has been a push to integrate patient-reported outcome measures such as PROMIS into clinical practice, with the belief that providing this information to clinicians will result in improved patient outcomes. The purpose of this study was to determine the usefulness of providing PROMIS symptom scores to primary care clinicians on patients' symptom outcomes.

Objectives: The primary objective was to determine the effectiveness of providing PROMIS symptom scores to clinicians on symptom improvement. A secondary aim was to determine whether providing PROMIS scores to clinicians increased symptom documentation, symptom-specific clinician actions, and patient satisfaction. Last, this study explored the correspondence between PROMIS subscales and brief non-PROMIS legacy measures.

Methods: We recruited participants from primary care clinics located in an urban setting. In clinic, prior to the encounter, patients who screened positive for at least 1 SPADE symptom and provided informed consent were asked to complete the 20-item PROMIS questionnaire (4-item scales for each of the SPADE symptoms). We then randomized the 300 study participants (72% female; 49% African American) to a feedback group in which their clinician received a graph of their symptom scores, or a control group. We reviewed electronic medical records of the baseline visit for documentation of symptoms. We reassessed symptom outcomes 3 months later using PROMIS, in addition to brief measures of sleep, (2-item Pittsburgh Insomnia Rating Scale), pain (3-item PEG [Pain intensity, Enjoyment of life, and General activity]), anxiety (GAD-2), depression (PHQ-2), fatigue (4-item Short-Form-36 Health Survey [SF-36] Vitality Scale), and treatment satisfaction.

Results: Most patients (84%) had multiple clinically significant SPADE symptoms at baseline (T-score ≥55; that is, ≥0.5 SD above the population norm of 50). Symptoms improved moderately in both groups. The group difference was not statistically significant, but favored the feedback compared with control group (effect sizes, 0.46 vs 0.37, P = .17) (effect sizes, 0.46 vs 0.37, P = .17). Symptoms present at baseline resolved at 3-month follow-up only one-third of the time. Clinically significant symptoms were documented in the visit note in 20% to 87% of cases, with pain most often documented, and fatigue least often. Symptom documentation, symptom-specific clinician actions, and patient satisfaction did not differ between treatment arms. Subgroup analyses indicated that male sex, White race, and greater medical comorbidity predicted worse symptom outcomes. Each PROMIS subscale strongly correlated with its corresponding brief non-PROMIS measure (r = 0.70-0.86). A moderate change (0.5 SD) on each non-PROMIS measure translated into approximately a 4-point T-score change on its corresponding PROMIS measure.

Conclusions: Simply providing feedback of symptom scores is insufficient to improve symptom outcomes in a primary care population. Factors that might enhance the use and utility of PROMIS include (1) improved clinician training in symptom management, (2) extra visit time or resources (eg, telecare support) for managing symptoms, and (3) incentives (eg, reimbursement, quality indicators) for optimized symptom management.

Limitations: We drew a convenience sample from urban clinics providing care to an inner-city population. Replication of study findings in other primary care settings is warranted.