A Towers Watson analysis found 28 of 29 HCAHPS survey results are directionally correlated with medical malpractice results.
In the new era of healthcare, government and private payers are shifting from volume-based reimbursement to value-based reimbursement. As private payers refine their networks and move away from fee-for-service models, government payers have focused on adjusting payments based on quality, patient satisfaction and value metrics. These metrics eventually become available to the public and researchers through the Medicare website.
The use of these metrics to adjust government reimbursements and their financial and marketing implications has immediately captured the attention of health system senior management. They are driving decisions ranging from human resources to capital investments.
These metrics, particularly when analyzed on an industry-wide basis, provide a window into other aspects of the system's operations previously not fully appreciated. For health systems, direct medical malpractice costs, including actual losses and the costs of insurance, can range from approximately 1 percent to 5 percent of net patient revenues. While much of the difference among systems is attributable to different legal environments around the country, significant differences still exist among entities operating in the same jurisdiction.
We have analyzed Towers Watson's client database of loss experience and the CMS Patient Experience of Care survey results to determine how the survey results relate to medical malpractice loss experience at hundreds of facilities across the country. Advanced analytic methods helped us find which areas measured in the Experience of Care survey are most correlated with medical malpractice experience after adjusting for differences in jurisdiction.
Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS)
The HCAHPS survey is a national, standardized, publicly reported survey of patient perspectives on their hospital experiences. The survey was designed to allow meaningful comparisons among hospitals and to incentivize quality improvements. Furthermore, The Patient Protection and Affordable Care Act of 2010 includes HCAHPS in the metrics used to calculate value-based incentive payments.
The survey contains questions about aspects of patient experiences including: quality of physician and nurse communication, cleanliness and quietness, pain management, discharge instructions and whether patients would recommend the hospital to others.
HCAHPS correlation with medical malpractice results
Towers Watson found that 28 of 29 HCAHPS survey results were directionally correlated with medical malpractice results, as one would expect. For example:
- The percentage of patients who reported that they "always" received help as soon as they wanted is correlated with lower claim costs than expected.
- The percent of patients who reported that their nurses "sometimes" or "never" communicated well is correlated with higher claims costs than expected.
Also, for survey questions that include an element of magnitude, the correlation generally increases as the magnitude of the response increased. For example:
- The percent of patients who reported that their rooms and bathrooms were "sometimes" or "never" clean is more correlated with higher losses than the percent of patients who reported that their rooms and bathrooms were "usually" clean.
It is our view that the experience of care, as measured by the HCAHPS survey, is correlated with medical malpractice results. While correlation does not imply causation, we are hopeful that direct efforts by health systems to improve patients' experiences have the long-term potential to contribute to better medical malpractice results.
Which survey questions are most correlated?
We observed that correlation with favorable results is distributed more broadly across a number of survey areas, while correlation with unfavorable results is more highly concentrated.
Using advanced statistical techniques, the following responses were identified as most correlated with higher malpractice losses:
- Percent of patients who reported that they "sometimes" or "never" received help as soon as they wanted
- Percent of patients who reported that their physicians "sometimes" or "never" communicated well
- Percent of patients who reported "no" when asked if they would not recommend the hospital
The following responses were identified as most correlated with lower malpractice losses:
- Percent of patients who reported that staff "usually" explained about medicines before giving it to them
- Percent of patients who reported "yes" when asked if they would probably recommend the hospital
- Percent of patients who reported that they "usually" received help as soon as they wanted
- Percent of patients who reported that their pain was "always" well controlled
Our final statistical model gives the most weight, by far, to the correlation between poor physician communication and patients not receiving help when needed and higher than expected malpractice losses.
Magnitude of correlation
We have found that facilities with elevated responses (80th percentile) to the survey questions that are most highly correlated with higher medical malpractice experience have losses that are approximately 20 percent higher than those of facilities at the median (50th percentile).
A durable finding
Some publicly available quality of care measures, such as timely and effective care, are intended to show how often hospitals provide care that research shows gets the best results for patients with certain specific conditions. For example, one metric is how often surgical patients receive an antibiotic within one hour before surgery. Hospitals often comply with this guidance 99 percent of the time.
Like an educator "teaching to the test," efforts to improve these discrete metrics may improve the directly measured outcome without changing the underlying culture of quality in a more comprehensive way.
In contrast, measures of the patient's physician and staff communication experience, staff responsiveness and subjective caring (e.g. pain management) are deeper behaviors that are more difficult to change. Therefore, we expect that the correlation found in these areas will be durable over time.
Ed can be reached at ed.lionberger@towerswatson.com