A Qualitative Exploration of Triangulated Shared Decision Making in Rheumatoid Arthritis

Abstract

Objective:

Treat-to-target implementation in rheumatoid arthritis (RA) requires a shared decision making (SDM) process. However, ability to pay is a major determinant of patient choice, but how this factor affects SDM is under explored.

Methods:

Visits at four RA clinics during which patients faced a decision to change their treatment were audiotaped between May 2016 and June 2017. Audiotapes were transcribed verbatim and analyzed using qualitative Framework Analysis.

Results:

156 visits were analyzed. Most RA patients, except those with effective insurance coverage, had deliberations disrupted or sidelined by third-party insurance providers having power to authorize the preferred DMARD choice. This triangulated SDM complicated efficiency in deliberations and timely treatment, and was a barrier to shared engagement about health risks and symptom improvement typically found in patient-provider dyads.

Conclusion:

Rheumatology care providers should aim to incorporate treatment costs and ability to pay into their deliberations so individualized out-of-pocket estimates can be considered during triangulated SDM at point-of-care.

Early detection of rheumatoid arthritis (RA) and early intervention with disease modifying anti-rheumatic drugs (DMARDs) significantly improve short- and long-term therapeutic outcomes (1–3). Moreover, cardiovascular morbidity and the need for total joint replacement have significantly decreased since aggressive care for RA has become the norm (4). Maintaining tight control of disease activity for all RA patients requires frequent monitoring and adjusting medications to minimize inflammation (5,6). However, numerous contemporary studies on treatment adherence highlight patient-related barriers to DMARD escalation. Patient reluctance to add or change medications is influenced by low motivation or personal beliefs, advanced age, poverty, limited health education, and the inability to cope with high drug costs (2,7–10). Some studies conclude that patients’ reluctance to accept the risks of side effects associated with newer treatments impacts providers’ ability to adhere to treat-to-target strategies (11,12). Yet, systematic reviews identify cost barriers as the principle factor underlying RA patients’ reluctance to change treatment – a problem that has worsened as uncertainties related to insurance coverage increase (including deductibles and co-pays) (8,13) as illustrated by one of the subjects in our study:

PT: See, I don’t dread taking my medicine… My main is … the insurance company because you always have to fight with them to get them to give me my Etanercept.

(PT ID 145; Provider ID 5)

Numerous treatment options are now available for patients with RA, and shared decision making (SDM) is considered best clinical practice. SDM is a process during which patients are informed of all available options, and subsequently deliberate with physicians to arrive at treatment choices that are most consistent with their values and priorities (14,15). Yet, despite the apparent influence of cost on patients’ adherence to treatment recommendations, few studies have examined this factor at point-of-care or whether ability to pay influences SDM processes and outcomes.

The drug approval process has been shown in other chronic conditions as influenced by tiered insurance regulations that may negatively impact preferred treatment choice, SDM, and timely decision making (16). O’Connor et al. (17) suggest that physicians must confront difficulties in creating a comprehensive treatment plan, despite using SDM, when patients’ insurance plans limit options and/or offer no coverage for patient’s drug of choice. Although ‘third-party’ insurance payers and Medicaid reimbursement schemes help cover the high costs of medications, their restrictive drug formularies can decrease rather than expand choice options and have an adverse influence on medication management (18).

The literature on third-party influences and medication communication has advanced in pharmacy research, exploring shared patient-pharmacist decision making (19). However, the third-party concept in SDM has yet to capture widespread attention between patients and their physicians.

The objective of this study was to gain deeper understanding of insurance coverage as it relates to treatment escalation in RA by identifying the perspectives of RA patients and their rheumatologists at point-of-contact and while engaged in SDM. The study design supports an assumption that effective deliberation between patients and physicians positively influences decisional determination, i.e., the arrival at a final decision (20). Elwyn and Miron-Shatz (20) explain SDM as gaining sufficient information, appraising new knowledge as sufficient, imagining counter-factual information, and affective (emotional) forecasting, all of which need to be integrated into choosing a treatment option.

Patients and Methods

The study was approved by our institution’s Human Research Protection Program. Data were collected between May 2016 and June 2017 by the first author, a medical anthropologist, at four academically-affiliated rheumatology clinics as part of a broader study on SDM. The parent study involved a 6-month intervention phase in which all RA patients and their clinicians utilized an empirically-based decision tool meant to encourage open deliberation about medication choices (21). This post-test tool is a paper brochure meant to elicit patient perspectives on such characteristics as onset of action, mode of administration, common and very rare side-effects and quality of life, infection risk, time on market, and treatment affordability. All English-speaking RA patients were approached for consent to be audio-recorded after having been roomed by nursing staff for their follow-up rheumatology appointment. The data collection researcher, the first author, was not in the room at the time of the consultation. The study included 10 attendings, one advanced practice registered nurse (APRN) and seven rheumatology fellows. All patients initially seen by fellows were also seen by an attending physician. The care providers were recruited together as a group by the last author after being given information about the study. Participating physicians also consented to have their consultations audio-recorded with RA patients.

The mean (SD) length of time of recorded consultations was 29.9 (11.6) minutes in the pre-test phase and 25.1 (10.7) minutes in the post-test brochure phase. Recordings were transcribed verbatim by the second coauthor, a masters-level nurse practitioner student with several years’ experience assisting in clinical research. Transcripts were initially screened by the first author for presence of DMARD names used during the visits, and for patient biometrics and treatment appraisal by the last author, a rheumatologist with substantial research experience in the field. Biometrics were validated from the medical record by the third coauthor, a rheumatology fellow. A handful of transcripts were excluded due to the presence of a coexisting condition with systemic lupus or dementia. The mean (SD) disease duration was 12.5 (10.2) years, and the majority of patients had moderate disease activity with mean (SD) Clinical Disease Activity Index (CDAI) of 18 (9.6) and mean (SD) Routine Assessment of Patient Index Data (RAPID3) of 12.5 (7.6), as described in Hsiao et al. (21). Among the participants considered for this study, 153 patients refused participation, and were neither consented nor recorded. Their objections were explained as a wish to maintain privacy or not feeling well.

In-depth analysis of the transcripts began with reading and re-reading by all members of the research team. Data-driven phrase-codes for DMARD type, escalation and cost factors were identified by interrater agreement and matrixed onto a spreadsheet using Microsoft Excel. Examples of treatment escalation phrase-codes included “Discussion of DMARD pros and cons,” “Change of DMARD type was deliberated,” “DMARD type,” and “Patient expressed disfavor of DMARD based on cost.” Examples of cost factors included “ability to pay,” insurance coverage,” Medicaid,” “copay,” and “afford.” The matrix was then analyzed using Framework Analysis (22), which is a method that maintains the hermeneutic perspectives of study participants and interprets them within the context of a guideline or policy – in this case, the professional ethos in rheumatology to adjust treatment medications to minimize inflammation and erosive activity using best practice SDM (14). Constant comparison descriptively identified emerging patterns (23) that were interpreted as support of a preferred DMARD option according to key principles of SDM, i.e., the presence of “supported autonomy,” “self-determination,” and “option deliberation,” during deliberation (24).

Results

A total of 156 patient recordings advanced to analysis because the consultation involved statements from the rheumatologist about worsening symptoms and a need to escalate treatment. Participation of rheumatology providers was arbitrary according to their schedule of RA patients, influencing the relatively low number of patients seen by fellows with a senior attending physician versus the senior physicians and APRN alone. Table 1 provides the overview of participants. The number of female to male participants was nearly 6-fold, and the total age range was 26 to 85 years, with 55% of between 45-64 years. The majority of participants were White (65.4%) followed by African American/Black (16.7%) and Latino (10.3%). Nearly 19% of total participants identified themselves as Hispanic. Just as many participants were employed full-time as on disability (19.2%, respectively), and 23% were retired. About 30% had private insurance. Those with Medicare and/or Medicaid alone comprised the majority of participants, almost 38% and 21%, respectively. Nearly 8% of participants were underinsured or had no known source of insurance.

Table 1:

Demographic information describing RA patient and Rheumatologists

RA patients	N (%)
Total	156 (100)
Sex
Female	133 (85.3)
Male	23 (14.7)
Age
26-44	18 (11.5)
45-64	86 (55.1)
65-85	52 (33.3)
Race
African American/Black	26 (16.7)
Asian	1 (0.60)
Latino	16 (10.3)
American Indian/First Nation	3 (1.90)
White	102 (65.4)
Unknown	8 (5.10)
Ethnicity
Hispanic	29 (18.6)
Non-Hispanic	119 (76.3)
Unknown	8 (5.10)
Employment
Full-time	30 (19.2)
Part-time	11 (7.10)
Student	3 (1.90)
Retired	36 (23.1)
Unemployed	41 (26.3)
On disability	30 (19.2)
Unknown	5 (3.20)
Insurance type
Private	47 (30.1)
Medicare	59 (37.8)
Medicaid	32 (20.5)
Military (Tricare/VA)	6 (3.80)
Indian Health Service	3 (1.90)
Inactive coverage	4 (2.60)
Unknown	5 (3.20)
Rheumatology care providers	N
Total	18
Senior attending	10
Fellows	7
APRN	1

Insurance Coverage and SDM

Certainty about insurance coverage for potential new DMARDs significantly influenced the SDM process (Figure 1). Patient-physician dyads expressing confidence about the patient’s insurance coverage experienced Advantaged SDM and deliberations did not appear impacted by issues of affordability. Most patients with drug coverage plans, however, were somewhat less certain about future third-party approval and engaged in Contingent SDM with their rheumatologists. Yet, although deliberations did involve the topic of cost, they remained constructive as both patients and physicians expressed confidence their preferred choice would be approved. The patients having authorization restrictions from third-party payers expressed worry that their insurance coverage would limit the treatment options available to them, experiencing At-Risk SDM, where the quality of SDM was at risk because choices were restricted. Disrupted SDM occurred for patients having inadequate or no insurance coverage. Physicians were unable to offer options based on medical necessity and deferred discussing appropriate treatment until a payment option was found. There was little deliberation during these visits, despite the need to escalate care.

Figure 1:

Certainty of insurance coverage negatively influences the effectiveness of autonomous SDM, illustrated by the small blue arrows. As cost coverage becomes less certain, the patient-physician dyad becomes less autonomous and increasingly dependent upon third-party decision making for a preferred DMARD.

Advantaged SDM

About a quarter of patients were actively engaged in employment or education despite their worsening RA symptoms. Most of these engaged in in-depth conversations with their physicians about DMARDs, which included detailed information about risks and side-effects and the patients’ preferences. Because the majority had secure private insurance coverage, the deliberations involved little to no discussion about DMARD affordability. Also missing was uncertainty about whether a prescription might be denied in favor of the pros and cons of available treatment options. One dyad engaged a 12-minute conversation about possible options that allowed the patient to autonomously express curiosity and ask meaningful questions, with no expressed concerns about costs. The positive deliberation appeared strengthened by the manageability of the DMARD cost:

Doctor (DR): Okay, we will stop the certolizumab…. I will give you some information on tofacitinib…. Tofacitinib is the name of the [stat-free] drug…. Your [employer] health plan is generally pretty good about this stuff. So, it is one 5 milligram tablet twice a day.

PT: Those are actually expensive?

DR: Oh yeah.

PT: Is it more expensive than injections? (laughs) … I do not know if it is, like, since it is a pill, I guess I would expect it is not more expensive. (laughs)

DR: They are all expensive…. (laughs)

PT: Mhm. (laughs)

DR: Um, tofacitinib. Let’s see… 60, for one month: $2,713.

PT: Oh my God. (laughs)

DR: And 32 cents. That is why insurance exists.

(Patient ID 67; Provider ID 10)

Patients articulated insights into symptoms and treatments, with physicians responding in-kind with open, matter-of-fact information. Importantly, discussions could include options recommended in practice guidelines to each patient. In a few instances mandated drug lists defined by the patients’ insurance carrier restricted choice. And yet, there appeared to be no negative influence of cost on SDM because the insurance coverage included both physicians’ and patients’ preferred choices.

Contingent SDM

Other patients with good, but less certain insurance coverage, had to devote variable amounts of time to discussing the issue of cost:

PT: And then, I guess, after [we learn it is approved] that it would still be the cost because if I cannot afford it I cannot take it. So, that would be a factor….

DR: Right.

PT: And then?

DR: Right, so… Right. We have to make sure whatever we end up going with, the first thing will be get an authorization.

(Patient ID 150; Provider ID 16 and 8)

Yet, attitudes remained positive. For example:

DR: So … we are going to start you on the Etanercept and the pharmacy is going to help us with this.… And they will do our authorization stuff for us. And so, that is that. Etanercept and click [the button], you are good… With those two insurance cards, you should be okay.

PT: Okay.

(Patient ID 125; Provider ID 10)

Deliberations were supported by presumed adequate insurance coverage. This confidence also afforded time to engage the patient in deeper discussions about options, for example:

DR: Uh, so Infliximab and Abatacept. Uh, they were before we switched to Rituxan, right?

PT: Yes.

DR: …So was that an insurance issue that you got the infusion rather than the injectable? It was better covered?

PT: Uh, I- I guess so. I mean, I did not know I had a choice with the Infliximab.

DR: Yeah. Well there are now five different medications that work like Infliximab.

PT: Mhm.

DR: Adalimumab, there is also Certolizumab and Golimumab in addition to the Infliximab. So, they work like Infliximab but each is a little different… But they all do basically the same thing…. So … you’re on Medicare now?

PT: Yes.

DR: Okay. So, we have to do is deal with the injectables because that’s the insurance wrinkle here. The injectables go on your Part D plan.

PT: Okay.

DR: You have the Medicare … drug plan?

PT: Um, I have [private insurance] with the drug plan, yeah.

DR: …Okay. So… we will have to find out through the prior authorization process whether it is covered or whether there are alternatives…. Everything else looks pretty good.

(Patient ID 100; Provider ID 10)

Knowing that pre-authorization was required was viewed as less straightforward, and took more time, compared to patients in the Advantaged group. Yet, the deliberations still addressed deeper question/answer sessions about DMARD options, and about risks and side-effects. For example:

DR: And then as far as adding another medicine to the methotrexate and sulfasalazine… Usually the first one we go to would be something like an Etanercept…

PT: Yeah.

DR: Um, it is a type of medicine that suppresses part of your immune system that causes inflammation.

PT: Mhm…. But is it dangerous with all those [side-effects]?

DR: The biggest risk is infection.

PT: Right…. All that stuff makes me nervous, but I just don’t know why I’m not getting better.

DR: Yeah…. The disease is just- is very active right now. Um, and it is really common. A lot of people need this kind of combination, need something like Etanercept.

PT: So, you recommend it.

DR: I would give it a try…

PT: … I just want something safe. I’m afraid of all those side effects…

DR: … I would probably recommend, you know, trying the Etanercept first…. Tofacitinib [costs] a little bit more. What we usually do is we send it to the outpatient pharmac. And, then they get the authorization for you…. We could try the Tofacitinib first… and see if your insurance will cover that.

(Patient ID 121; Provider ID 15)

At-Risk SDM

Uncertain insurance coverage threatened high quality SDM. However, some patients were not aware of gaps in coverage or high deductibles, or how these factors could limit their choices, which contributed to uncertainty about whether a preferred DMARD would be approved:

DR: Tofacitinib is very expensive.

PT: Oh, it is? … Will my insurance cover it?

DR: It should. But, you know, it will throw you right into a donut hole.

PT: What you mean?

(Patient ID 114; Doctor ID 12)

Because cost overshadowed these deliberations, physicians reduced time from presenting appropriate DMARDs or discussing benefits and risks to explain the Medicaid coverage gap. The following is from a 6-minute conversation that left the physician providing literature instead of engaging SDM:

DR: I will give you some literature, maybe on that hydroxychloroquine…. And let’s see what apothecary says about the etanercept copay. I will try to research to see if there is anything else, any other assistance programs to see if there is anything else…. Let’s continue etanercept for now….

PT: No, I am just, I am a little nervous about what that is going to cost me, if I am in the donut hole now.

DR: Yeah. So, when you are usually in the donut hole, do you usually come out of it eventually after a few months? Like once you paid for your medications?

PT: Once I get up to my limit…

DR: $6800 you said it was?

PT: Yeah … up to $6800…

DR: Yeah, yeah. That just stinks. I am sorry you have to go through that. I wish things were different.

(Patient ID 99; Provider ID 10)

Uncertain insurance coverage restricted autonomy and choice, as illustrated in the following passage:

DR: Right. And then, did you also lose insurance for… insurance coverage for the etanercept?

PT: No, cause I, no. I mean, but I think it is after 65 or 66, it is not covered anymore. There is no ins…

DR: Oh, because of the…

PT: Medicaid.

DR: Okay.

PT: You know, so I would have to pay full price for that, and I cannot afford it.

DR: Right, right.

PT: (laughs) $1000 a month, if not more.

DR: If we could get you the… so you did well on the etanercept.

PT: Yup….

DR: Okay…. Um, would you be, if insurance issue was not an issue, would you be willing to go back on that?

PT: Oh yeah. I did not want to go off of it.

(Patient ID 10; Provider ID 15)

For participants with uncertain, limited insurance coverage the problem of cost dominated available consultation time, deprioritizing effective deliberation about risks, drug side-effects, or the patient’s concerns about treatment escalation. The need to defer to third-party authorization initiated the consultations rather than being a tentative end-product of deliberation.

Disrupted SDM

SDM and deliberative choice-making were essentially absent for patients with little or no drug coverage. This disruption influenced delivery of care by eliminating choice and deliberation about competing harms and benefits. In these cases, the ultimate choice was fully dependent upon the third-party payer. For example:

DR: So, I want you to please, please hear me, and think about going to a TNF inhibitor.

PT: I’m really trying….

DR: Etanercept or Adalimumab.

PT: Listen, when my pharmacist called me I was considering that I might take it, but he said it was gonna cost me $1700 a month…. He asked, Do you really want me to fill this prescription?

COMPANION: Doesn’t insurance cover any of that?

DR: Of course, it does.

PT: My pharmacist actually called me…. I- I- I-I was shocked…. I thought, I don’t think so.

DR: [We will] do an etanercept pre-authorization…. We specifically look to see if those side effects happen, and then we decrease the dose of methotrexate or stop it. We’ve had to do that once.

PT: No, I decreased it myself.

DR: You decreased it yourself….

PT: (laughs)

DR: Actually, if you think about it, [etanercept] is a really much better option for you…. I can’t push your methotrexate [because of the liver problem].

PT: Okay…. I’m interested, in what that insurance coverage is…. Cause that was a lot of money.

(Patient ID 94; Provider ID 1 and 8)

Whereas physicians typically entered the consultation room asking about how a patient was doing and their levels of pain and potential treatment options, the conversation was soon hijacked by restrictions imposed by cost. Some physicians took a relatively large portion of consultation time to encourage and educate patients on how to sort out their financial issues so they could access treatment. For example:

DR: And they would all be available to you based on your insurance. We are going to check the insurance thing first.

PT: What insurance? I had to drop my [insurance]. But I think I am getting a prescription plan from Medicaid. But, I had to drop the secondary insurer because I need to see this [other doctor regularly for another condition].

DR: Oh okay…. Well then maybe we should talk about that. Do you not have a drug plan?

PT: I don’t….

DR: … if you don’t have a Part D plan that covers Tofacitinib, then you may not be able to get it cause obviously you won’t be able to afford it out of pocket.

PT: Oh, yeah? How much is it out of pocket? Too much?

DR: …I would say, um, something on the order of $3-4000 a month. I said that with a straight face, but it’s not funny.

PT: No, it isn’t.

(Patient ID 96; Provider ID 10)

Emphasis on affordability shortened visits and appeared to negatively influence deliberations. Those experiencing disrupted SDM also appeared to have lower expectations for being able to identify a suitable means of escalating care.

Discussion

SDM for treatment escalation in patients with RA can be negatively influenced by uncertain, or inadequate, insurance coverage. As the need to deliberate about insurance factors increases, the more intrusive triangulated decision making becomes. In a triangulated scenario, a third party, e.g., a health insurance payer, significantly contributes to the final treatment decision and has authority to prevent access to preferred choice. Triangulated SDM created a barrier that limited the ability of the rheumatologists to fully implement SDM.

Triangulated SDM identifies new areas for policy debate about the influence of third-party decision making on patient adherence, and might also influence the RA patients’ perspectives about who takes final responsibility for the treatment choice. The latter is shown by Nota and colleagues to be the physician (25). Our findings suggest the third-party decision-maker is conceptually situated above both patient and rheumatologist during the clinical encounter, acting as a top-down decisive voice. This influence can significantly and adversely impose on the physicians’ best efforts to involve patients in the treat-to-target process.

There is also apparent danger in the possibility that both patients and clinicians will feel ignored, which is antithesis to patient-centered care, and can erode trust (24, 26). In addition, triangulated decision making is routinely required for some biologic DMARDs but may erode the foundations of health equity for patients with uncertain or limited insurance coverage because decision interference occurs before real deliberation about treatment choices takes place.

The necessity of adequate, stable, and predictable insurance coverage defines effective decisional deliberation and, by proxy, best clinical practice. Appeals for greater cost transparency as it supports effective SDM at the point-of-care should be made. One-size-fits-all managed care drug formularies may lack necessary sensitivity to capture individual RA patient needs, which these findings suggest can seriously dismantle SDM. Our findings support patient uncertainty about affordability as a problem that can substantially work against effective SDM during treatment deliberations, and can upend the overall quality of the clinical encounter in RA.

One limitation of this study is that the post-test analysis included only data from RA patients confronting escalation of treatment, and the opportunity to evaluate the brochure for non-biologics was missed. Yet, the randomization procedure of asking all patients to participate may still be considered a strength mainly because patients were not preselected with unknown insight into their current disease status. In addition, the final treatment choice was not assessed, which is a potential limitation since better follow-up as to whether the medication was approved by the insurance payor could have provided an accurate estimate of out-of-pocket costs during the SDM process (27). This qualitative exploration nevertheless offers a baseline for future studies on triangulated SDM in other settings to demonstrate coherence to and validate its findings (23).

Significance and Innovations.

1. For patients with limited to no insurance coverage, the focus of deliberation during shared decision making prioritizes insurance factors over discussions of risk and treatment options, and triangulates the decision-making process.

2. Triangulated shared decision-making limits the opportunity to embrace assured self-determination and relational autonomy during deliberations, and imposes uncertainty about who takes final responsibility for choice, despite best efforts to involve patients in the treat-to-target process.