Developing and Testing the Health Care Safety Hotline: A Prototype Consumer Reporting System for Patient Safety Events

It is important to emphasize not only how much was known, but also how much was not known at the outset of this project. What was known after two decades of prior research is that adverse events, errors, and even near-miss events can be identified, a nomenclature can be created to structure and catalog reports of such events, and systems can be built to enable health care professionals to reliably report these events in real time. What was less well known was whether patients could contribute unique information that health care organizations did not already have—that is, whether a parallel system of reporting could allow patients to help identify the health safety problems they experienced in health care institutions. Also unanswered was the question, “Is it worth getting patient reports about safety events?” That is, is obtaining patient reports an important thing to do?

Even stakeholders who (provisionally) supported an attempt to gather patient-reported data on safety problems had concern that patients might overwhelm health care organizations with reports, or that patients who complained about quality and safety issues might face retribution from health care providers. Neither of these fears was realized in the deployment of the Health Care Safety Hotline prototype; likewise, risk management operations were not overrun or compromised in any way.

The experience of developing and deploying the hotline answered some of the outstanding questions. There clearly is an appetite among health care organizations for collecting information from consumers about things that go wrong. The prototype successfully elicited information that was considered useful by our partner health care organizations and that had little overlap with the information received from other sources (such as internal adverse event reporting systems). The partner organizations were not overwhelmed by patient reports and valued the detail and complexity captured by both the narrative and structured portions of the hotline instruments. However—and this is a big however—the hotline prototype elicited far fewer reports than the project team, the health care organizations, the TEP members, and AHRQ had hoped for.

How, then, do we begin to address the discrepancy between the potential high value of the hotline tool we developed and the low response rate among patients and caregivers in this test of the prototype? The project team, partner organizations, and TEP members identified a range of possibilities that could account for the low number of reports. We think there were three principal problems.

First, although we employed a variety of means of marketing and outreach (brochures, hospital Web sites, pamphlets included in packets of discharge materials), we assumed that when given the opportunity and invitation, patients would reach out to institutions to report problems. We employed a passive outreach strategy (in which patients and caregivers must choose to opt in and make a report). We did so believing that we had avoided the problems documented in other such efforts. We learned, however, that affording patients an opportunity to report their patient safety concerns in real time did not generate a large volume of reports.

Second, based on Web traffic data, we know that more individuals accessed the Web site than actually made reports; many potential users dropped off after reading the introductory materials or completing only the narrative portion of the report. This suggests that despite the time and care that were taken in developing the Web site, there most likely were problems with the platform itself.

Third, the presence of alternative mechanisms for reporting at the hospital level (discharge surveys) or community level (State regulatory agencies) could have depressed the number of reports. A next natural step in the evolution of the prototype would be to determine whether a modified patient event reporting form, together with a prompted deployment effort (such as contacting patients at discharge) would substantially increase the volume of useful reports.

While we are not at this time recommending a full scale-up of the prototype, we urge AHRQ to consider a follow-on effort to this project designed to answer the specific questions raised here. For example, a full usability study of the Web tool would address the content and platform issues we have identified. After modifications are made to the hotline tools, additional partners could easily be identified and recruited to help test a prompted deployment strategy. Another alternative AHRQ might explore would be to expand the patient experience survey to incorporate patient safety screeners or questions.

A second, related issue we struggled to resolve was whether the prototype was, perhaps, trying to serve too many purposes. The Research Triangle Institute (RTI) report commissioned by AHRQ (which gave rise to this project) recommended that patient reports be collected locally but also communicated to a centralized (national) level, where they would be aggregated, analyzed, and triaged or distributed to State and local levels for action. On the basis of our experience with the hotline prototype, we now question whether all of these objectives (e.g., improving patient safety at the organizational level, aggregating data for public monitoring and accountability, conducting research on patient-reported errors) are compatible and whether it is realistic to think that they can all be addressed with a single reporting system. Our attempt to address all of these goals in one prototype resulted in an instrument and a process that were probably much more cumbersome than would be necessary to serve a single objective (in this case, patient safety reporting within a single health care system). The need to serve too many “masters” also may have contributed to the lower-than-hoped-for response rates.

We believe that one of the most important lessons learned from this experience is that it is possible to build a single technology platform that can meet diverse objectives, but the technology has to be rolled out differently to meet different objectives. The platform we developed has a number of components that can be used separately, depending on the specific goals of the end user. Which components are employed and how they are adapted are highly dependent on a project’s purpose. For example, if the hotline is used as an internal patient-reporting mechanism, it would not be necessary to work through an external partner (a research organization such as RAND), nor would events (necessarily) need to be classified by type, severity, and contributing factors for the reports to be useful for patient safety improvements.

The current version of the event reporting form may request information that is not necessary for public monitoring, so the number of elements could be reduced, making the reporting burden on consumers lighter. In addition, research or public monitoring uses would not require patients to identify themselves, removing the need for that portion of the Web page and perhaps encouraging patients who may fear retribution for themselves or their providers to make a report for research or public accountability purposes.

Table 6.1 shows the components of the hotline prototype and discusses how they might be employed in future research, patient safety improvement, and public monitoring efforts.

We urge AHRQ to put all of these materials, including Appendix C, the Operations Manual (which includes all the forms, surveys, and Web site specifications), into the public domain and to encourage health care organizations, regulatory agencies, patient safety organizations, and others to use these tools (or pieces of them) for a variety of research, safety improvement, and public reporting purposes.

AHRQ is tasked by Congress to produce evidence to make patient care safer and to ensure that such evidence is understood and used. Over the past two decades, AHRQ has sponsored a plethora of activities to improve the quality and safety of health care systems through research and implementation of evidence. Those projects have included sponsoring user-driven research to improve measurement and event reporting and to reduce and remediate patient safety risks. Much of this research has resulted in the development of practical and widely used tools for improving patient safety in hospitals and other health care settings, including surveys to assess patient safety culture, training to improve communication and reduce errors by health care teams (TeamSTEPPS), and CAHPS to assess patient experience with health care. AHRQ is also responsible for the development of the Common Formats (a voluntary effort to help standardize reporting of adverse events) and PSOs to confer privilege and confidentiality protections on providers in order to promote shared learning to enhance patient safety nationally. In the present project, many of these threads related to reporting and patient safety improvement came together.

As AHRQ considers scaling up the prototype and generalizing from the pilot project findings, it will be important to assess both the relative value of this particular method for gathering consumer input and how continued efforts to promote the patient voice in improving patient safety could be integrated with other ongoing AHRQ efforts—in particular, the support of CAHPS and the PSOs.

Finally, we would note that AHRQ’s support for a culture change around patient safety in health care institutions (i.e., moving away from blaming individuals toward identifying and addressing systemic risks) was critical to the agency’s efforts to promote adverse event reporting by health care professionals. When adverse event reporting systems were first built, they did not receive many reports. It was not until the culture in hospitals changed around adverse event reporting that reports started flowing in from health care professionals. It may be that a complementary effort to change the culture around consumer reporting is necessary, in addition to building a prototype model to effectuate reporting.

A decade of research on consumer engagement in patient safety has demonstrated the capacity of patients and their caregivers to identify errors and injuries experienced in the course of medical care.^{43, 44} While patients and their proxies may sometimes be reluctant to report, especially if they believe that reporting may be futile or could alienate providers, patients have knowledge that professionals do not possess, and this information can reveal vulnerabilities in health care delivery organizations, inform and motivate improvements, and convey a degree of respect for patients and professional humility that are sometimes lacking in patient-provider interactions.^{2, 13, 15, 18, 45-49}

The design and development of the Health Care Safety Hotline represents a significant effort on the part of the research team and AHRQ to move promising research findings “from bench to bedside.” The effort yielded important information about the feasibility of and constraints upon consumer reporting of patient safety events. Overall, the number of reports received was disappointing, given the potential reservoir of adverse events and the yield of research-oriented methods. Nevertheless, the data provided were rich compared to the information available from standard data collection methods. The events reported had previously been largely unknown to the well-informed participating health systems, and the reports prompted them to reach out to those patients and directly address their concerns through patient safety improvement mechanisms.

The hotline demonstrated that phone and online reporting are feasible methods for collecting patient safety information, that patients and caregivers welcome the opportunity to clarify the information in their reports, and that they are largely willing to share this information with the organizations that delivered the care they found problematic. While the logistics of managing patient-reported safety information pose certain challenges related to confidentiality, timeliness, and peer review, these potential barriers can be addressed in a way that advances consumer engagement in patient safety. We are confident that the hotline platform developed in this project, with some modifications, is scalable to diverse settings and will be a valuable piece of the patient safety improvement puzzle in years to come.

Table 6-1 Application of Prototype Components to Future Research, Patient Safety Improvements, and Public Monitoring

Note: CAHPS = Consumer Assessment of Healthcare Providers and Systems; PSES = patient safety evaluation system; PSO = patient safety organization; RCA = root cause analysis.