Non-response bias on the HCAHPS surveys can hinder a hospital’s ability to track improvements in the patient experience.
HCAHPS survey response rates are on the decline, introducing an element of non-response bias that may impact organization efforts to improve the patient experience, according to a new study published in the journal Patient Experience.
HCAHPS scores and other patient satisfaction assessments have become increasingly important in healthcare. These surveys are key for influencing some value-based reimbursements, inform the Hospital Compare Star Ratings, and can shape organizational business decisions.
“Patient experience has been elevated in the priorities of healthcare leaders in recent years as organizations have come to understand the rationale for and benefits of improving patient experience,” the researchers, who hail from healthcare consulting group PRC, said.
“Not only has a patient experience focus been shown to yield better financial outcomes in terms of both CMS reimbursements and increased patient loyalty and market share, it is also increasingly regarded as a crucial quality measure.”
But survey response rates are decreasing, which can in most cases influence the data with which payers, providers, and patients have to make decisions.
In an analysis of HCAHPS response rates and scores from 2008 to 2017, the researchers found that patient responses tapered off considerably. In 2008, the first year all hospitals had to administer HCAHPS, patient response rates averaged at 33.3 percent. By 2017, that rate dropped to 26.7 percent.
These lower response rates were tied to lower HCAHPS scores across each dimension included in the survey. When there were fewer patient responses, the hospital did not have as good of a report card, the researchers said. When a hospital did receive more patient responses, it had higher HCAHPS scores.
This finding goes against conventional wisdom, the researchers pointed out, revealing implications of non-response bias.
“Typically, if a surveyed sample is representative of the population, a correlation does not exist between response rate and survey results,” they said.
Instead, their findings suggested complications with the types of patients responding to the surveys.
“The moderate correlations observed suggest that, at a national level, the HCAHPS data being collected are not capturing a representative sample of the patient population for these hospitals,” the team explained.
Obtaining a representative patient sample is complex, the researchers continued. Healthcare organizations across the country are struggling to motivate patients to adopt healthier behaviors, never mind respond to a patient satisfaction survey.
One potential solution is to simply increase the number of responses the hospital receives, the researchers said. If a hospital were to administer 200 HCAHPS surveys and see the national average response rate for 2017 (26.7 percent), it would receive 53 completed surveys from patients.
That hospital could increase the number of completed surveys received by increasing the number of surveys they administer to 300. This would yield about 80 completed surveys received.
“But simply calling or mailing more patients will not improve the response rate,” the researchers cautioned. “With increased response rates, hospitals receive valuable feedback from a higher proportion of patients, creating a more representative sample of the total hospital patient population.”
In other words, increasing the volume of surveys administered will not yield a representative sample. Instead, it will continue to create responses from the most satisfied or dissatisfied patients.
Organizations must instead change their surveying methods to compel a higher percentage of patients to respond to the HCAHPS surveys. There is less information about which surveying techniques are most effective.
Through qualitative investigation, the researchers did learn that hospitals that only mail out HCAHPS surveys were less successful at improving their response rates than those that mailed and telephoned about the surveys. This suggests a multimodal approach to surveying.
But there may be other factors at play.
“While this research analyzed methodology as a primary influencer in the relationship between response rates and dimension scores, it should be acknowledged that there are additional variables external to the care experience that can impact scores,” the researchers said.
Patient demographics, geographic differences, and other survey modes (such as email or online surveying) could influence response rates but were not included in this investigation.
The Centers for Medicare & Medicaid Services and the Agency for Healthcare Research and Quality have likewise offered few suggestions in this area.
Neither agency has been able to pinpoint the best surveying mode to yield a more representative response rate, which the researchers argue means they should account for lower response rates when calculating care quality scores.
Ultimately, organizations should understand the needs and habits of their own patient populations and shape their strategies to best fit their patients. Driving patient survey responses should not be a one-size-fits-all approach, and therefore should leverage multiple strategies that best fit a unique population.
Date: May 07, 2019