Patient Perspectives on Evidence Supporting Drug Safety and Effectiveness: “What Does It Mean for Me?”

Introduction

Older adults are frequently underrepresented in randomized controlled trials (RCTs) testing investigational drugs, but after regulatory approval often constitute a majority of patients.¹ After drugs’ approval, routine clinical use data, also called real-world evidence or real-world data, emerge from insurance claims, patient registries, and electronic health records.² Such data can help address remaining uncertainties about drug effectiveness and safety for older populations.^3,4 This study aimed to explore the attitudes of patients 65 years and older towards the use of routine clinical use data in making treatment decisions.

Methods

As a case study, we used direct oral anticoagulants (DOACs), a therapeutic class in which early RCTs showed improvements in stroke rates compared to warfarin. By contrast, routine clinical use data in populations older than 75 years indicated higher risk of gastrointestinal bleeding with dabigatran (Pradaxa) versus warfarin.⁵ We conducted four focus groups in March-April 2023 via videoconference with people in the US who were 65 years or older and took a DOAC (Supplementary Methods S1). Focus groups followed a semi-structured guide and began with questions about participants’ decisions to initiate a DOAC, followed by presentations of novel evidence tables we generated comparing dabigatran to warfarin showing incidence of stroke, clot, and major bleeding/death using evidence from either RCTs or routine clinical use (Figure 1). The final table presented routine clinical use data outcomes by age group.

Figure 1.

Example evidence table shared with participants

After each presentation, participants were asked what they thought of the drugs’ benefits and risks and how the evidence would influence their anticoagulant choice. The focus groups were recorded, transcribed, and coded using thematic analysis.^6,7 Codes were then reviewed to construct themes.

Results

Fifteen people participated in four focus groups (3–5 per group). Median participant age was 73 (interquartile range 68.5–76 years) (Supplementary Methods Table S1). We identified five themes, two relevant to how participants perceived routine clinical use evidence (Table 1; Supplementary Methods Table S2). Most participants (n=9, 60%) took apixaban. All but one knew someone taking warfarin and expressed negative views about it.

Table 1.

Themes about routine clinical use data and illustrative quotes

Theme	Quotation (participant number)
Perceptions of evidence from randomized controlled trials vs. routine clinical use data Participants questioned the relevance of RCT evidence to them but also worried about the robustness of routine clinical use data	I like to know efficacy percentages. It’s nice to have an idea of what you’re dealing with, where things can go. [] My [medical] team has been really good about keeping me informed, letting you know if there’s anything new, and how things are working, so I trust my team. But as I said it’s trust and verify. (Participant #13)
	I would be looking at [] the difference between an observational study and a random, double-blind study [] the patients are much more closely selected for a random study. And so you have controls of other factors in the outcome, whereas an observational say, where you just look at insurance claims, and how many people had this, it doesn’t tell you anything about other aspects of their health, or medicines, or illnesses, or anything like that. [] If we had one or the other, I’d rather have the clinical trial than the observational. But when you put two of them together it looks pretty darn good. (#3)
	All of clinical trials, of course, have a time limit to them, and this [RWD] is showing how things work in the field, which is of notable interest to me and something I do ask doctors about it [] the experience in the broader population, it can theoretically be different than the clinical trials and errors and problems can come up. So this is actually quite supportive of the clinical trial and the calculation between Pradaxa and warfarin. (#6)
Evidence for decision-making: what does it mean for me? Participants wanted to know how the data should be interpreted to apply to them as individuals	Statistically there’s not much difference [] 1% is not a big number when you’re looking at [] the patient population this was done on. [] You don’t know if it’s classified by age group or severity of condition, things like that. But other than that it still goes back to [] it’s still your doctor, they’ve got to be like yeah, here’s my experience and here’s what I know about the data. And here’s what I still recommend for you, based on X. (#1)
	The question is, what does that mean for me? And so, that would always be my question with my physician [] given my situation and given my [] specific biology, what is my risk over that specific risk pool. [] Let me put it this way: more information wouldn’t really help me, because I would need that in the context of a conversation with someone who is a physician or an expert. Otherwise, I feel like I’m following the Internet for facts and figures [] that might be legitimate but may not apply to me. Of course, it always comes down to what does it mean to me when we’re taking these drugs. (#14)
	It would just be one bit of data, if I looked at this, this wouldn’t—neither of these would make me choose. I’d really need to weigh this with my doctor and have my doctor’s opinion and really trust my doctor. And I do trust my doctors a lot. (#5)

Quotes have been edited slightly for clarity.

Perspectives on trials vs. routine clinical use evidence

Participants expressed interest in clinical use evidence because they wanted more information on how drugs worked in patients like them. Over two-thirds (n=11) of participants had previously looked for additional evidence and information on DOACs and expressed a desire for trustworthy evidence. Several noted that the large populations from which the routine clinical use data were derived increased their trust in findings.

However, participants recognized the limitations of routine clinical use data. One noted, “I trust the data more because it’s more people, although...it’s not a clinical study...there may be some variation that actually makes it less useful.” Participants raised concerns about confounders, like whether populations were sicker, older, or longer-term DOAC users than themselves. Some specifically mentioned preferring RCTs for their rigor and controlled variables.

For both evidence types, participants raised concerns about common minor side effects (minor bleeding and bruising), two outcomes largely unreported in routine clinical use data—as one said, “How many people file an insurance claim for bruising on warfarin? That doesn’t happen.” Participants noted the evidence did not include information on lifestyle restrictions they cared about such as stopping activities with fall risks or regular testing with warfarin.

“What does it mean for me?”

The groups’ consensus was that they found useful accessible, reliable summaries of evidence on each anticoagulant that included side-by-side comparative information and age-relevant information. Yet, participants were uncertain whether the evidence on outcomes by age-group was relevant to them, and none would make medication decisions based on the evidence tables alone. All agreed that they would need their physicians to determine whether the evidence applies to them and recommend a medication. Participants concluded that although they wanted more information, they would need to talk over any drug-related decisions with their physicians, whom they hoped had access to age-relevant clinical use data.

Discussion

Older patients valued detailed information about the medications they take, specifically data from patients like them. RCT and routine clinical use data do not include some outcomes that mattered to participants’ day-to-day lives, like bruising.⁸

A limitation of the study was possible over-representation of patients with apparent high health literacy (Supplementary Methods S1). Even among this sample, no one would make prescription decisions on his/her own. Therefore, though routine clinical use data has the potential to reduce uncertainty about drug safety and efficacy for older populations, there is a need for clinician-oriented prescribing information to translate this new evidence into practice.