Development of a Framework and Tool to Facilitate Cost-of-Care Conversations With Patients During Prenatal CareFREE

To draft a cost-of-care conversation tool, we used a human-centered design (HCD) approach. The field of HCD includes methods for examining the real-world context and behaviors of individuals, deeply engaging stakeholders, and rapidly prototyping solutions optimized for specific end-user needs (12). The methods of HCD have been recognized in the health care literature as relevant to program implementation, patient services, and promotion of chronic disease management (13–15). For this project, we adapted an HCD development process first used to design a patient-facing health education tool for use with a similar population—low-income, minority, and Chicago-based—in a Patient-Centered Outcomes Research Institute (PCORI)–funded clinical trial (16). Our adapted approach engaged a 5-step process (Figure 1).

We engaged 4 stakeholder groups at a university medical center obstetrics-gynecology (ob-gyn) clinic in Chicago's predominantly African American south side: 1) pregnant and recently postpartum patients presenting for pregnancy-related care; 2) ob-gyns, nurses, medical assistants, and technicians who provided direct services to patients in that clinic; 3) frontline clinic staff who interact with the patient population; and 4) executive administrators responsible for workflows in the clinic or other services that address patient cost issues (for example, billing and call center). Original eligibility criteria and recruitment focused on women diagnosed with a high-risk pregnancy; designated “near poor” or “nearly poor” (approximately 138% to 400% of the federal poverty level [FPL]); Medicaid-covered or privately insured through the Affordable Care Act; and seeking prenatal care at the participating site. However, early patient and clinician interviews counseled that cost burdens were a significant factor in low-risk pregnancy care among low-income women, as well as in high-risk pregnancy care for women outside the 400% FPL range. As a result, we expanded our income range to include a small sample of women with household incomes above the 400% FPL and low-income women with low-risk pregnancies for comparison.

Semistructured interview guides were prepared with stakeholder input. The patient guide was informed by community health workers who work in underserved black and Hispanic communities in Chicago, to align questions with the social and cultural context of patients. The clinician guide was informed by an expert ob-gyn and a third-year ob-gyn resident to align questions with clinic practices and culture. Executive and administrative staff guides were informed by participatory research in which researchers called to schedule appointments and documented the patient experience from call center to insurance review.

American College of Obstetricians and Gynecologists guidelines, pregnancy books, and pregnancy apps were reviewed to design a “provotype” of a cost-of-care conversation tool for use in initial interviews. Provotypes, or provocative prototypes, are early-stage designs that are not focused on accuracy but instead are crafted to provoke discussion of current or future practice (17). Review of provotypes in interviews can draw out acceptability, appropriateness, feasibility, and adoption of a proposed change (18), and suggest productive design directions.

Patients were recruited in clinic by front desk staff using fliers and posters. Patients who called the study phone number were screened and scheduled by a research team member. Interviews were conducted at a location of the patient's choice for 60 to 120 minutes and were compensated at $75. Patients were asked about prenatal care experiences in the home and clinic, pregnancy costs and cost-of-care conversations with clinicians and family, and experiences with insurance. In addition, during patient interviews, patients were asked to diagram their care experience over time and connect it to accrued expenses; were shown photographs and asked to select which best depicted their feelings about being pregnant; and were asked to sketch “stick figure” drawings to illustrate all of the people important to their pregnancy or providing cost-related support (Supplement 1).

Employees of the medical center were recruited using direct outreach by the research team and snowball sampling. All interviews lasted 30 to 60 minutes, occurred in the clinic or administrative offices, and were compensated $25. Questions targeted the frequency of, nature of, institutional support for, and receptivity to cost-of-care conversations. Clinicians were asked to map typical appointments and treatments for low- and high-risk pregnancies as a context for cost discussions. Administrators were asked about financial and social services, patient population, and barriers to providing care or services to their population (Supplement 1).

All participants were shown at least one prototype of a cost conversation support tool (Supplement 2), with at least one half reviewing 2 prototypes as new design concepts emerged. With permission, interviews were audio-recorded and transcribed; drawing activities were photographed. All participants were assigned an anonymous study identification number.

Two patient focus groups included participants from earlier interviews. In 2-hour sessions held at a local YMCA, patients reviewed the cost-of-care support tool and discussed benefits of its adoption.

A physician focus group was recruited from a safety-net hospital on Chicago's West Side through an open e-mail to clinical staff; they met for 30 minutes to review the final prototype.

Additional feedback was gathered during departmental grand rounds at the primary research site, where clinicians voluntarily engaged in review of the final prototype, using sticky notes to annotate individual copies.

The project was approved by the institutional review boards (IRBs) at all participating institutions: IRB 2016-134 at the Illinois Institute of Technology, IRB 16-175 at the University of Chicago, and IRB 17-24 at Mount Sinai Hospital.

All work was funded by Robert Wood Johnson Foundation, which had no role in study design, data collection, analysis, interpretation, or writing the report.

When prompted, clinicians and patients typically talked about different costs. Clinicians identified tests, procedures, and drugs as drivers of patient cost burden, and voiced frustration that such costs are unknowable when formulating care plans and sometimes never known unless patients return with a bill. Lack of precise cost information, owing to variations in insurance types and coverage, was a primary reason for skepticism about cost-of-care conversations. Patients, on the other hand, were more likely to highlight indirect costs of appointments (such as lost income, transportation, and childcare). Patients characterized these costs as burdensome and stressful, particularly when unexpected.

In addition, frequency and duration of appointments multiply indirect costs. Guideline-based prenatal care is visit-intensive, requiring 13 to 27 visits for low- and high-risk pregnancies, respectively. Patients and staff reported that a visit can last 4 hours. Cumulatively, even when care is “free” because of Medicaid coverage, indirect costs can exceed $2500 during high-risk prenatal care. To estimate these costs, we used multiple scenarios for transportation (bus, car, taxi), 1 small meal, and potential lost earnings for each visit (calculated at the 2016 Illinois minimum wage). We excluded child care and calculated parking at the lowest discounted rate, although patients expressed difficulty getting these rates. Estimates for all scenarios ranged from $2500 to $2800. Although a patient may not necessarily pay $2500, this is the financial hurdle she must negotiate to attend appointments. Clinicians and clinic staff, when presented with this cost table, expressed shock but concluded the estimates were both accurate and conservative. They also requested that patients not be shown the cost table for fear it might affect patient commitment to pregnancy care.

Multiple staff stated that ob-gyns should understand that patients say they miss appointments because they cannot afford to attend, not because they don't want to attend. Staff also reported that patient strategies to mitigate indirect costs—circling for free parking, leaving to feed the parking meter, or bringing children to appointments—caused clinic operations issues. Both patients and staff reported that the frequency and duration of appointments caused stress with participating family members and partners. Patients expressed frustration that they didn't understand the burden of visits in advance. Low-income patients expressed a greater need than higher-income patients to forecast indirect costs so they could budget and make tradeoffs in other spending to attend visits. One ob-gyn observed that his patients with resources (such as family and job flexibility) but poor insurance demonstrated more ability to mitigate the burden of high-risk care than patients who had quality insurance but inflexible jobs or no family.

All clinicians and staff reported that cost-of-care conversations are already occurring. Nurses, social workers, and front desk staff were more likely to report addressing indirect costs with patients.

Obstetrician-gynecologists voiced reluctance to initiate cost conversations, citing lack of time in the appointment; lack of training, resources, and information to offer patients; and discomfort bringing up problems without having solutions. They feared that adverse events from accommodating cost-related requests (such as skipping a test) could cost them their license. Some ob-gyns raised concerns that talking about costs would cause patients to avoid care or end their pregnancy.

Patients and nurses stated that the absence of cost discussions made nonadherence more likely, not less. Low-income African American women, while reporting disproportionate financial stress from visit-intensive care plans, voiced the most concerns about cost conversations. These concerns included embarrassment at revealing need to doctors, avoiding stigma, and not wanting to prompt lesser care. Some worried that if they brought up costs, they might be seen as not belonging there. These same patients expressed pride in receiving care at a respected institution and “getting the best care for their baby.” Low-income African American patients expressed interest in services and programs that were clearly available to everyone (such as free breast pumps), but less receptivity to supports that might mark them as different.

Stakeholder requirements are presented in the Table, using the Multilevel Requirements Framework from design (18). Thirteen iterations of a cost-of-care conversation tool were generated before concluding in a design that stakeholders deemed acceptable, appropriate, and feasible. Each iteration was reviewed by at least 4 participants; one half of the participants were shown 2 concepts when a second design emerged. Both concepts were refined through subsequent interviews, and the most successful elements from each were merged into a single, refined concept (Figure 2).

The final prototype uses a checkbox format that fits both low- and high-risk pregnancy care (Figure 2, section 1). Writing and design choices are women-centric, not baby-centric, to show sensitivity for women with at-risk pregnancies. Text is written at a 6.4 reading grade level (Flesch-Kincaid). Descriptions of tests and procedures include estimated times and preparation tips to help patients prepare for appointments (Figure 2, section 2). A timeline organizes and sequences all appointments (Figure 2, section 3). A planning worksheet orients patients to indirect costs but is placed on the back page, away from the treatment section, so patients can introduce cost conversations when they are ready (Figure 2, section 4). Insurance and treatment questions help patients initiate cost conversations with clinicians and payers (Figure 2, section 5). Support phone numbers for services are placed on the front cover, not the back, to normalize that all patients may need assistance (Figure 2, section 6).

Our findings informed the following framework for designing cost-of-care conversations:

1. Create a shared understanding of the care plan that includes the frequency and duration of appointments.

2. Build awareness of the indirect costs of the care plan among both patients and clinicians. A shared model of burden permits discussion and, where possible, negotiation.

3. Build the belief that talking about money is allowed and a part of quality care.

4. Provide a means to initiate and standardize cost conversations to ensure they are applied routinely and equally, and so avoid the appearance of judgment to patients. A good solution will fit within the limited time of a typical visit and be initiated by a trusted provider.

Twenty-seven of 69 comments collected at grand rounds offered praise for content and approach: for example, “These explanations are simple and clear,” “I love this, and if a patient had a $$ constraint, I would pick which appts to cancel/skip,” and “At least we are acknowledging costs and can maybe help.” The remaining 42 comments contained suggestions for additional content, corrections, requests for more visuals, and use of colloquial language. Safety-net hospital physicians noted a strong fit with their patient populations, for whom costs are a chronic issue. They observed that the tool would be easy to integrate into standard care, where cost discussions are already taking place. One physician proposed adding 15 minutes to the first prenatal appointment to go over the tool and map out what to expect throughout the pregnancy.

Patient focus group participants praised the tool for “prompting things I wouldn't know to ask” and “making me a partner in my care plan”; providing a treatment road map; helping coordinate care with other providers; as proof of visit frequency to their partners; and as a means of initiating a budget conversations with a partner.