Supplemental Items Reduce HCAHPS Response Rates, but Response Rates Do Not Affect HCAHPS Scores: A Randomized Experiment.

Background: Longer surveys can reduce response rates (RRs). Observational data suggest a positive hospital-level association between RRs and HCAHPS scores, but a negative patient-level relationship. There is no experimental evidence of whether a hospital's RR affects its HCAHPS scores.

Objectives: Estimate the effect of the 59-item versus 32-item survey on RR by survey mode; assess whether any reductions in RR change HCAHPS scores.

Research design: Patients randomized within hospitals to 4 survey protocols [mail-only, mixed mode (MM; mail with telephone follow-up of nonrespondents)] and survey length [32-item or 59-item (32 core + 27 supplemental items)]. Regression models predicted (1) whether a patient responded to a survey and (2) responses to patient experience measures within hospital and survey mode from survey length, controlling for patient characteristics.

Subjects: A total of 10,099 adult patients from 51 nationally representative US hospitals participating in a randomized HCAHPS mode experiment.

Principal findings: Adjusted MM RRs were 21.9 percentage points higher than mail-only RR for the 32-item survey. The adjusted RR for the 59-item survey was 5.6 percentage points lower than the 32-item survey in MM and 2.9 percentage points lower in mail-only mode. The lower RR caused by greater length had no effect on standard adjusted HCAHPS scores in either mode.

Conclusions: A longer survey reduced RRs, especially by telephone follow-up after a mail survey. HCAHPS hospitals and vendors may want to consider trade-offs when adding many supplemental items and the RR advantage of MM. Reducing RRs within a hospital does not change standard adjusted HCAHPS scores.