Developing Patient-Centered Communication Ecosystems in Integrated Health Care Delivery Organizations

Health care organizations are using an expanding set of automated communication channels to reach out to their patients on a growing array of topics. These communications may be delivered by text message, by standard or secure email, or through interactive voice response technology. They can be initiated by multiple departments within the organization and can be directed to a specific person or to a group of individuals with similar characteristics. The topics can range from clinical concerns to reminders about upcoming appointments, surveys to collect clinical information before those appointments, updates on check-in and billing procedures, population-based health education, marketing, and assessments of satisfaction with care.

Although departments and teams within these organizations often develop and distribute messages independently, patients may justifiably assume that these messages reflect an intentional, systemwide strategy. In other words, patients may already believe, and health care organizations must realize, that they have become partners in an automated communication ecosystem. ¹ The study of ecosystems in nature teaches us that their components interact in complex and often unpredictable ways. Accordingly, a health care organization should consider whether each communication campaign in its ecosystem is achieving its goals and whether these campaigns are mutually reinforcing, duplicative, contradictory, or collectively overwhelming for patients. This question is particularly urgent for organizations that provide clinical care for defined populations, such as the Department of Veterans Affairs, Kaiser Permanente, and other integrated health care systems that make extensive use of communication campaigns to support population health and quality improvement strategies, as well as to facilitate in-person and virtual clinical care. ^1–3 At their best, automated communication interventions can increase patient engagement, improve clinical outcomes, and enhance operational efficiency. ^4,5 Evidence is emerging, however, that these well-intentioned efforts also risk overwhelming patients with a constant barrage of input that induces message fatigue and, ironically, may lead them to opt out of further communications. ³ The Figure, adapted from the experience of patients in one delivery system, illustrates this proliferation of message channels, categories, and types. This commentary will focus on the need for integrated health care organizations to develop a strategic, patient-centered approach to their automated communication ecosystems.

Figure:

Proliferation of automated communications to patients with diabetes in health care organizations. The range of messages an organization may use to communicate with patients living with diabetes is depicted based on chart reviews of actual messages sent week by week over a single year.

Learning from the Past and Understanding the Present

Efforts to establish a patient-centered communication ecosystem within an integrated health care organization must be based on prior communication experiences in clinical care and public health. When organizations began using their electronic health records (EHRs) to send reminders and best practice alerts to clinicians in the 1990s, these interventions, although often effective in isolation, soon proliferated into an undigestible stew of mandates that frustrated clinicians and enhanced burnout. ⁶ Belatedly, health care systems recognized the need to prioritize and coordinate their messages to clinicians. ^7,8 In public health, excessive repetition of health messages to community members can lead to message fatigue and even counterreactions. ^9,10 The potential perils of incessant and often contradictory communications became a concern during the COVID-19 pandemic. ¹¹ Such experiences suggest that a proactive and comprehensive approach to developing a strategy for automated communication with patients may help avoid the cacophony that can result from the proliferation of uncoordinated message campaigns.

In developing its strategy, the organization should begin by describing the range of topics that it addresses or wishes to address. Some message types (such as visit reminders) may coalesce around specific clinical problems or encounters, although others (such as influenza vaccination prompts or marketing messages) may be seasonal, and still others (such as health education on topics in prevention) can be delivered at any time, or in response to new scientific findings or controversies. For every message campaign, designers face choices about the mix of communication channels and frequency of message delivery. The complexity of this process immediately suggests that the organization should establish a robust guidance and governance process, as has been proposed for clinician alerts. ⁸ In the absence of such a process, each clinical and operational department may pursue its immediate goals with little, if any, proactive input or retroactive feedback from patients or other departments within the organization. Even if individual communication campaigns are proven effective through rigorous research or local evaluations, the aggregate burden of messages can easily become overwhelming to patients without ongoing, system-level planning and prioritization.

Communication Goals and Strategy

The authors recently proposed that health care systems wishing to communicate effectively with their patients should “communicate with patients about topics of interest and importance to them, using the communication channels they prefer, at an aggregate frequency they accept, with ongoing evaluation of effectiveness and unintended consequences.” ¹ How health care systems can incorporate each of these components into their communication strategy is discussed next.

Communication content

Defining the content of system-generated communications requires balancing the priorities of patients and the organization. If systems focus too heavily on communication campaigns to improve efficiency or control costs, they may provoke unexpected consequences. In the authors’ research, a widely distributed text message campaign encouraging patients to use in-system urgent care centers over holiday weekends led to a spike in opt-out requests. ³ Because the opt-out system was not unique to each outreach campaign, patients who opted out no longer received even those messages they might have wished to continue, such as reminders before clinic visits. Tailoring message content and channel to specific groups is essential to enhance patients’ perceptions of relevance. ^12,13 For example, messages about care for chronic health conditions may have little appeal to individuals who are healthy and rarely engage the system. Achieving both balance and focus within communication ecosystems requires a communication oversight process that includes representatives from clinical and operational departments, as well as patients themselves, and is empowered to make system-level decisions about the content, priority, number, timing, and continuation of message campaigns.

Communication channels

Patients develop their own communication microsystems, based on their access to or comfort with varying types of message content and communication channels. They may prefer to be contacted at different times during the day or on different days of the week. They may designate other individuals as communication proxies. They may prefer to receive different types of messages through different communication channels. Tools for recording such communication preferences are available in many EHRs, but barriers to obtaining, honoring, and updating these preferences can be substantial. ¹⁴ These preferences may change over time, and systems should periodically encourage patients to review and update them. Once identified, these preferences should be shared throughout the health care organization and with external vendors who communicate with patients on behalf of the system. Requests to opt out of specific communication channels should be widely disseminated, because health care systems are required by the Telephone Consumer Protection Act of 1991 to honor such requests. Collection and analysis of data on patient communication preferences can also help an oversight team identify trends that may guide future communication policies.

Communication volume

Health care organizations may not fully appreciate the cumulative burden they impose on patients when they expand their message portfolio. To gain that understanding, systems communication teams must conduct a thorough inventory that enumerates their communication topics, channels and frequency of communication campaigns and protocols. If the prospect of monitoring all communications seems too daunting, organizations can begin with more limited assessments of communications on common topics, such as diabetes care, immunizations, or cancer prevention. The findings of this inventory provide essential information for ecosystem governance. Such an inventory can also identify single messages that are redundant or inaccurate, as well as multiple messages that could be consolidated into a single communication campaign. A communication inventory could also identify message campaigns that can be discontinued because they are obsolete or ineffective. For each campaign and in aggregate, systems communication teams must define the optimal frequency and timing of communication to suit the intended purpose of the message and minimize message fatigue. ^15,16

Communication effectiveness

The effectiveness of communication between the system and its patients should be tracked for each communication campaign and for the ecosystem as a whole. To identify effective campaigns and discontinue ineffective ones, health care systems communication teams should conduct systematic evaluations using tools, such as rapid randomized trials (often referred to as “A/B studies” in nonmedical fields) ^15,17 or time-series analyses to track changes in patient behavior resulting from specific campaigns. At the aggregate level, a communication dashboard can use internal data to assess changes in the rates of opt-outs for text and phone messages and unopened emails. Satisfaction with system-generated communications and perceptions of message burden can also be tracked through periodic patient surveys. Such a dashboard constitutes a crucial information source for teams that develop and implement communications campaigns and for those who provide oversight and governance of the communication strategy.

Patient Engagement and Population Segmentation

Two crosscutting themes should guide the strategy for patient-centered communication ecosystems: 1) engaging patients throughout the communication process, and 2) directing communication campaigns to the subgroup of individuals most likely to benefit.

Patients are ideally positioned to identify topics of interest for system-generated health communications, yet they are too seldom involved in the development of communication campaigns or the design of communication ecosystems. Patient engagement may improve the system’s ability to address common concerns and demonstrate its comprehensiveness across settings and over time. Individuals who rarely engage the delivery system are important to include in this process, because they may have a unique set of information needs and preferences. Patient input into the content of communication can also enhance cultural concordance and accommodate differences in health literacy. Patients may be included in the process through individual qualitative interviews, cognitive interviews, or panels and focus groups, as well as in service on standing organizational committees. Ultimately, patient engagement may even enhance the organizational “bottom line” by improving the quality and safety of care and enhancing patients’ experience of the system.

The identification of patient subgroups that respond differently to health care interventions is a standard practice in clinical and communication research but is too seldom used in operational communication campaigns. ¹³ Although one-size-fits-all communication campaigns are simple and cost efficient to deliver, health care organizations should emulate other industries by directing messages to the subgroups of patients who are most likely to benefit from them. This requires ongoing evaluation of the effectiveness of these campaigns in different population segments, using simple stratification analyses or more complex, multivariable predictive models. ¹⁸ For example, the authors used a predictive model to show that individuals who had missed prior primary care appointments benefited from receiving two reminder text messages or interactive voice response calls, whereas a single reminder was sufficient for individuals with no prior missed appointments. ¹⁹ Patients who are new to the system may also benefit from additional outreach. Predictive analytics should take a larger role in the design of communication campaigns, although communicators must remain aware of potential biases in developing and using predictive models and should use these models judiciously to avoid unintentionally contributing to health disparities.

Conclusion

Although the approach outlined here may seem utopian, a coherent communication strategy is essential to achieving the mission of integrated health care organizations. The goals proposed herein can be translated into tangible, incremental steps that can be evaluated using the principles of the learning health system and quality improvement approaches such as Lean methodology or Plan-Do-Study-Act cycles. ^15,20 Our increasing understanding of ecosystems in the natural world teaches us that such systems are slow to mature and vulnerable to disruption. The iterative process of establishing communication priorities, coordinating communication campaigns, and evaluating the effectiveness of those campaigns can help health care systems develop and sustain their communication ecosystems despite the rapid pace of change in clinical practice and the social environment. By doing so, health care organizations can improve the health and engagement of the patients they serve.