Patient Experiences, Trust, and Preferences for Health Data Sharing

PURPOSE:

Scholars have examined patients' attitudes toward secondary use of routinely collected clinical data for research and quality improvement. Evidence suggests that trust in health care organizations and physicians is critical. Less is known about experiences that shape trust and how they influence data sharing preferences.

MATERIALS AND METHODS:

To explore learning health care system (LHS) ethics, democratic deliberations were hosted from June 2017 to May 2018. A total of 217 patients with cancer participated in facilitated group discussion. Transcripts were coded independently. Finalized codes were organized into themes using interpretive description and thematic analysis. Two previous analyses reported on patient preferences for consent and data use; this final analysis focuses on the influence of personal lived experiences of the health care system, including interactions with providers and insurers, on trust and preferences for data sharing.

RESULTS:

Qualitative analysis identified four domains of patients' lived experiences raised in the context of the policy discussions: (1) the quality of care received, (2) the impact of health care costs, (3) the transparency and communication displayed by a provider or an insurer to the patient, and (4) the extent to which care coordination was hindered or facilitated by the interchange between a provider and an insurer. Patients discussed their trust in health care decision makers and their opinions about LHS data sharing.

CONCLUSION:

Additional resources, infrastructure, regulations, and practice innovations are needed to improve patients' experiences with and trust in the health care system. Those who seek to build LHSs may also need to consider improvement in other aspects of care delivery.

INTRODUCTION

Excitement has grown about the potential of learning health care systems (LHSs), whereby routinely collected clinical care data are shared and secondarily leveraged to promote quality improvement and research.^1-4 Many ethical considerations have arisen as such systems have developed,⁵ and never before has it been more apparent that earning and fostering patients' trust is essential for the integrity and success of these systems.⁶

Despite efforts to respect patients through transparency and accountability in LHSs,^7-9 concerns remain that LHSs might interfere with the doctor-patient relationship^10,11 or lead to exploitation and harm if for-profit companies were to have unchecked access to patients' information.^12-14 Substantial evidence suggests that trust in health care organizations and physicians is critical in this regard, especially since patients are generally wary of sharing their data with third parties perceived to be motivated by profit.^15,16

Our own work examining patients' attitudes and opinions regarding data sharing within the context of CancerLinQ,¹⁷ a LHS developed by ASCO, has underscored the central importance of fostering trust in health care providers and organizations, in addition to alleviating patients' apprehension regarding access to and use of their data by for-profit entities. In two previous reports,^18,19 we described patients' perspectives on consent and data use in this context, drawn from a rich data set obtained through large-scale democratic deliberations.²⁰ In this final analysis of that data set, we delve more deeply into patients' reports of their encounters with physicians, insurance companies, and others during the course of their cancer care to explore how these personal experiences might have shaped their trust and ultimately their preferences for data sharing, hoping ultimately to illuminate best practices to foster the patient trust that is so essential for LHSs' success.

MATERIALS AND METHODS

The University of Michigan Institutional Review Board determined this research to be exempt. As described more fully elsewhere,^18,19,21 the research team hosted four democratic deliberations around the country from June 2017 to May 2018 with 217 patients with cancer notified of the sessions by their oncology providers. To generate informed and considered opinions,²⁰ experts educated the patients about LHS, with CancerLinQ presented as an example. Patients then engaged in facilitator-led small group discussions that were recorded and transcribed. The first round of discussions centered on patients' preferences regarding disclosure and consent,¹⁸ and the second focused on recommendations for governance of users and uses of data.¹⁹ This article further explores patients' unsolicited descriptions of their own encounters with physicians, insurance companies, and others that emerged during the discussions, specifically examining how patients' personal experiences might have informed their level of trust and opinions regarding who should have access to LHS data.

Data Analysis

A coding scheme was established by two analysts (R.D.J. and C.K.), who reviewed and discussed the transcripts. The transcripts were coded independently using Dedoose (version 8.0.35), after which the coders met to resolve any disagreements and refine the codebook. The analysts and the principal investigator (R.J.) organized the finalized codes into themes using interpretive description²² and a thematic analysis approach.²³ Additional details on methods can be found elsewhere.^18,19,21

RESULTS

Patient characteristics have been detailed previously.¹⁸ In brief, 67% were female, 22% Black, and 6% of Hispanic or Latino origin and 18% had a high school education or less. The mean age was 60 years, 23% had fair to poor health, and 30% had metastatic or incurable cancer.

Patients often vividly described their personal lived experiences of the health care system and their cancer treatment. In many instances, such personal experiences were brought up in the discussion when focusing on LHS policies regarding sharing data with insurance companies. Patients were asked if the policy should be to allow or deny the following types of requests from insurance companies: an insurance company interested in making sure that patients with cancer receive the most recommended care and an insurance company interested in determining which cancer treatments are eligible for coverage or reimbursement. The policy options that were discussed during the session have been published elsewhere.¹⁹

We identified four domains of experience that arose in conversation among many patients: (1) the quality of care received, (2) the impact of health care costs, (3) the transparency and communication displayed by a provider or an insurer to the patient, and (4) the extent to which care coordination was either hindered or facilitated because of the interchange between a provider and an insurer. Each of these domains appeared as a spectrum of experiences ranging from negative to positive (Fig 1; exemplary quotes in Table 1). When describing many of these experiences, patients often discussed their trust in health care decision makers (Table 2) and their attitudes and opinions about LHS data sharing (Table 3), illuminating the relationship between these experiences and patients' perceptions of physicians and insurance companies.

FIG 1.

Four domains of patient experience: (1) the quality of care received, (2) the impact of health care costs, (3) the transparency and communication displayed by a provider or an insurer to the patient, and (4) the extent to which care coordination was either hindered or facilitated because of the interchange between a provider and an insurer. Each of these domains appear as a spectrum of experiences ranging from positive to negative.

TABLE 1.

Patient Experiences

TABLE 2.

Relationship Between Experiences and Trust

TABLE 3.

Relationship Between Experiences, Trust, and Data Sharing Preferences

Domain of Experience #1: Quality of Care Received

Some patients perceived that they had received the best, preferred, or recommended treatment. A few described negative experiences with what patients considered to be medical errors or providers who seemed to the patients to be ill-informed about up-to-date treatments or innovations. Others recalled negative experiences with insurance companies denying coverage for physician-recommended treatment, resulting in the patient's perception that they failed to receive important treatments or that cheaper, less effective treatments were substituted. A number of these patients further described anxiety, pain, and suffering that they felt as a result. Some even recounted stories about other people that they knew who had died or suffered long-term disability because of what they perceived as a lack of adequate care.

Domain of Experience #2: Impact of Health Care Costs

Although a few patients commented that their treatments were covered or that their payments were expected and reasonable, others recounted being unable to pay for what they perceived to be the best or preferred treatment if the claim was denied by the insurance company. Some patients expressed frustration with having to deal with unexpected payments or having to go without treatment if they could not afford the costs.

Domain of Experience #3: Transparency and Communication Displayed by a Provider or an Insurer

Positive experiences with transparency and communication of providers or insurers seemed especially important in mitigating any unfavorable effects arising from negative experiences in other domains. For example, some patients recalled instances when the reasoning behind the denial of a treatment was explained adequately by an insurance company or provider. Others described being able to ask questions and discuss alternative treatment options with their providers to the point that they felt relatively satisfied with their care.

Domain of Experience #4: Extent to Which Care Coordination Was Hindered or Facilitated

Some patients recounted positive experiences of care coordination between providers and insurers, with their provider being cognizant of how treatments would be covered by an insurer and even interfacing with the insurer as a patient advocate. Others recalled positive experiences with the insurance company being willing to work with the provider to facilitate patient care. Conversely, some described negative experiences with providers who were unaware or inconsiderate of insurance coverage requirements and with insurance companies that were unresponsive to providers.

Trust in Health Care Decision Makers

In general, patients explicitly referenced their personal experiences related to one or more domains when detailing their levels of trust in certain health care decision makers (Table 2). Patients' trust appeared related to the medical expertise demonstrated by a provider or other decision makers and the degree to which the decision maker's motive was reflective of the best interests of the patient. Patients expressed distrust of those who were not perceived to be medical experts and those perceived to be driven by personal gain or profit. Insurance companies were generally perceived to be less trustworthy than health care providers as they were thought to be driven by financial considerations and not by the best interests of the patient. Many patients discussed their experiences with being denied coverage for their treatments, which appeared to prime them to view all insurance company decisions with skepticism. Moreover, it seems that a bad experience with an insurance company decision might even reflect poorly on the individual health care provider. For instance, one patient expressed that she no longer trusted her doctor's office because she left dissatisfied with their coordination with her insurance coverage and their transparency and communication with her, which the patient described as an “ethical issue.”

Attitudes and Opinions About LHS Data Sharing

Patients largely seemed comfortable sharing data with insurers and drug companies and between hospitals and/or physicians to improve care, both in general and for themselves. Several comments demonstrated the general sentiment that it was important to trust in the entity receiving their shared data (Table 3). Again, insurance companies were generally regarded with suspicion because of the perception that such entities deny treatment for the sake of finances. Because of this, when asked about their opinions about releasing LHS data to such entities, some patients expressed that they would tend to err on the side of caution, either by voting to enact policies that outright restrict sharing or by suggesting heightened governance and oversight of such entities, including reporting requirements and fees and/or penalties.¹⁹ For example, even when an insurance company eventually changed their initial decision about denying coverage after the provider sent in more information, one patient still perceived the insurance company's initial unwillingness as suspicious and, subsequently, was unsure about whether he would be comfortable allowing LHS data to be shared with such entities.

DISCUSSION

In this qualitative study, we delved deeply into patients' personal lived experiences of the health care system, particularly during their cancer treatment, to try to better understand how these encounters might have shaped their trust and preferences for health data sharing. This study highlights the direct personal experiences that illuminate how patient trust can be either cultivated or squelched and areas with potential for improvement. We identified four domains of experience that appeared to shape patients' trust in certain health care decision makers and, possibly, their attitudes and opinions about LHS data sharing. These merit further consideration as policy makers refine regulations governing insurers and allocate resources to support providers.

This study suggests the importance of dedicated attention to patient experiences and lends support to previous observations regarding the need for physicians and health care systems to focus on transparency and communication, care coordination, reasonable health care costs, and quality of care.^24-26 This underscores the importance of ongoing support for targeting these key areas when proposing advancements in medical school education, such as a curriculum that teaches adherence to health care quality and safety standards along with communication and relationship-building skills to promote patient advocacy and shared decision making.²⁷

Other compelling work further confirms the significance of the patient-physician relationship in building patient trust in the overall health care system experience. Upon conducting a structured exercise to identify interventions, a group of health care leaders and patient advocates recommended approaches to build trust via a coordinated mobilization effort among physicians, the health care system, and insurance companies. These include standards and measurements such as formal physician training in communication and relationship skills, accountability systems that can gather patients' feedback regarding their care coordination, and quality improvement systems driven by payer data that can increase transparency regarding patient experiences and outcomes.²⁴ Our findings are in line with their proposals and serve to further emphasize the central role that the provider may play in solidifying patient trust by facilitating such approaches, not just during the initial encounter but also throughout the entire duration of the patient's care.

This qualitative study of patients' experience of cancer care also demonstrates how trust in certain health care decision makers can, potentially, shape attitudes and opinions about LHS data sharing. It highlights that patients often look toward their physicians as trusted experts when it comes to such matters and that trust can be affected by how patients experience the quality, cost, communication, and coordination of their care. Indeed, as Platt et al²⁸ posit, health care providers act as “representatives of the health system who manage relationships with the various institutions in the health system as well as with the public,” and thus, patients' trust in them as “health information brokers” is integral to their trust in the healthcare system as a whole, particularly when it comes to issues of data sharing. This reinforces the importance of ensuring that oncology practices are reimbursed adequately to include sufficient staff necessary to support providers in navigating the complex process of care coordination with health insurers, while simultaneously developing policies that encourage greater transparency, responsiveness, and streamlining of interactions between patients and insurers in the uniquely complex American system.

These findings also underscore and illuminate the basis for concrete policy recommendations that we have previously articulated specifically for LHS data sharing, including the need for informed consent¹⁸ and specific provisions for governance of users and uses of data.¹⁹ Because trust appears to influence patients' desire for transparency and control over their data, providers should distribute LHS informational materials that are easy to understand and give their patients the choice to opt out of participation. Policy makers and administrators should consider restricted access, user fees, and penalties when designing LHS systems to foster patients' trust and prevent users perceived by patients to be financially motivated, including insurers, from misusing data. To further cultivate transparency and patient trust, LHS users should be required to publish their findings and LHS governing boards should include patient representation.

Most importantly, the current work builds upon our previous findings by providing a rich, nuanced understanding of the ways in which patients' personal health care encounters shape their trust and the role that providers can play in building trust among their patients. Providers and practices should seek to augment protocols, training, and resources to maximize patients' quality of care, reduce health care costs, improve transparency and communication, and facilitate care coordination with insurance companies. Billing departments should explain costs and copays in ways that are transparent and understandable. Services should be offered in-house, if possible, to give patients more cost-effective options. Providers unable to afford having their own patient navigators might consider working with local institutions to provide that resource where possible, as evidence suggests that patient navigators can play important roles as advocates, facilitators of communication and patient education, and providers of practical support, such as helping to secure services related to finances, medication, or equipment.^29,30 Also valuable are innovations such as the oncology medical home model, in which care delivery is coordinated, efficient, accessible, and evidence-based with a focus on continuous quality improvement.³¹ Such approaches have the potential to elevate levels of trust in patients by highlighting sincere efforts to ensure that they receive the best care, unhindered by financial considerations or profit-driven motives. Efforts by providers should be complemented by regulations that promote timely review and greater responsiveness of insurers to providers.

Of note, responses analyzed here were generated during a process of democratic deliberation to elicit patients' considered opinions on LHS policy issues.²⁰ This is a particularly compelling strength of this study since patients were not prompted to cite their own personal encounters with providers and others in health care as part of the policy discussions that comprised the main study,^18,19 but rather they brought up their own experiences organically as reasoned justifications for their policy opinions. It is telling that these deliberations roused patients to communicate specific examples of their personal health care experiences to establish a shared context when discussing their views surrounding LHS policies.

The limitations of this study have been discussed in detail elsewhere.^18,19,21 In brief, the conclusions of this study are hypothesis-generating given the limitations of qualitative research and should be confirmed in larger quantitative studies. The results of this study might not be generalizable to a broader patient population. Since the deliberation sessions required all-day, in-person attendance and travel to the study location, patients had to be healthy and/or financially stable enough to engage in such efforts. The complex nature of the study topic itself might have inordinately attracted patients with a better understanding, ability, interest, or willingness to discuss opinions regarding data sharing implications because of their educational background or previous experiences with medical care and/or research. Nevertheless, the deliberative sessions featured a diverse sample of patients who all participated in an educational component developed by a diverse group of experts and their discussions were facilitated by trained discussion moderators to help ensure equitable subject participation and common background content on which to draw upon.

In addition, although there were exceptions, a considerable number of patients who recounted negative experiences were Black and/or categorized as having a lower education level. A large body of evidence emphasizes the importance of reducing health disparities by addressing social determinants of health, such as low income, lack of employment, poor education, or racial discrimination.^32,33 Moreover, trust in health care institutions and health care providers may reasonably vary by race and ethnicity given egregious historical trespasses³⁴ and current discrimination against certain groups.³⁵ Future research is needed to further evaluate whether particular domains of experience are more relevant or prevalent for certain demographic groups and the impact on their trust in health care systems.

Finally, although this study was conducted before the outbreak of the COVID-19 pandemic, it has implications for current policy and practice. First, the COVID-19 pandemic has highlighted the importance of learning in real time from every patient's clinical care experience.³⁶ Second, disparities in treatment outcomes and vaccine hesitancy show that trustworthiness of and trust in health care institutions and health care providers is essential for public health.³⁷ Finally, in these urgent and dire circumstances, it has become clear that building upon and expanding the learning health system framework is key to ensuring that the accumulated data are effectively contributing to high-quality care for all. In response to the current crisis, Romanelli et al³⁸ call for the development of a learning health network that prioritizes “the patient's voice” and “the broad inclusion of diverse health care and community settings to address potential inequities.”

In summary, this study suggests a need for additional resources, infrastructure, regulations, and practice innovations to improve patients' experiences with and trust in the health care system in the United States. We cannot expect patients to trust in a system that they have directly experienced to be untrustworthy. Thus, those who seek to build LHSs for oncology care must focus not only on practices directly related to data sharing but also on improvement in other aspects of care delivery. For patients to have trust in LHS data sharing, it is important that they have trust in the health care system as a whole.

AUTHOR CONTRIBUTIONS

Rochelle D. Jones, Chris Krenz, Kent A. Griffith, Rebecca Spence, Angela R. Bradbury, Raymond De Vries, Sarah T. Hawley, Robin Zon, Sage Bolte, Navid Sadeghi, Richard L. Schilsky, Reshma Jagsi

Conception and design: Rebecca Spence, Angela R. Bradbury, Raymond De Vries, Sarah T. Hawley, Richard L. Schilsky, Reshma Jagsi

Financial support: Reshma Jagsi

Administrative support: Rebecca Spence

Provision of study materials or patients: Sage Bolte

Collection and assembly of data: Rochelle D. Jones, Chris Krenz, Rebecca Spence, Sage Bolte, Navid Sadeghi

Data analysis and interpretation: Rochelle D. Jones, Chris Krenz, Kent A. Griffith, Rebecca Spence, Angela R. Bradbury, Raymond De Vries, Sarah T. Hawley, Robin Zon, Sage Bolte, Richard L. Schilsky, Reshma Jagsi

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Patient Experiences, Trust, and Preferences for Health Data Sharing

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).