“Is there any way I can get something for my pain?” Patient strategies for requesting analgesics

Abstract

Objective

We examined the direct and indirect means by which patients express a desire for analgesic medication.

Methods

Back pain patients presenting to an academic ED were invited to participate in a study of patient-provider communication. Audio-recorded encounters were transcribed verbatim and transcripts analyzed using a qualitative approach based on conversation analysis.

Results

Requests for analgesics were documented in 15 out of 74 interactions (20%). We identified three basic patterns: direct requests, in which the patient explicitly asked for medication; indirect requests, in which the patient hinted at a desire for medication but did not ask for it outright; and no request, in which the provider discussed a prescription without the patient requesting it.

Conclusion

Most patients did not request analgesics. When they did so, they utilized strategies of mitigation, indirection, and deference that presented themselves as deserving patients while upholding the physician's autonomy.

Practice Implications

Patients come to the clinical encounter with a variety of expectations, of which a desire for an analgesic may be only part of the picture. Rather than focusing on strategies for inuring providers to inappropriate patient requests, it may be useful to devote clinical resources to examining patients' priorities and expectations for treatment.

1. Introduction

Pain is a pressing public health problem, affecting approximately 100 million US adults, at an estimated cost of $560-635 billion per year.¹ Although under-treatment of pain is still a problem in some clinical settings,² policy attention has increasingly focused on the role of opioid prescriptions in the growing epidemic of opioid abuse and accidental death.^3–5 This quandary generates numerous clinical challenges: providers report growing pressures to prescribe opioids despite concerns about misuse and diversion,^6–8 while patients face the stigma of being disbelieved and possibly labeled as “drug-seekers.”^9–12

One critical, understudied area is the role of patient-provider communication in pain treatment decisions. Recent experimental evidence suggests that physicians are more likely to prescribe analgesics when patient-actors request them by name.¹³ While this finding is intriguing, whether this experimental setup serves as an accurate proxy for actual patient behavior is unknown. In this paper, we examine patient requests for analgesics in a real-world clinical context. The data are drawn from audio-recordings of patient-provider communication about back pain in a hospital emergency department (ED). Building on insights from Brown and Levinson's politeness theory,¹⁴ we develop a typology of patient requests for analgesics.

1.1. Patient-provider communication about pain

Pain treatment is beset by numerous interactional challenges: the culture of biomedicine privileges the treatment of objective, verifiable symptoms,^15–17 patients and providers come to the clinical encounter with different goals and expectations,^18,19 and background concerns about deception and addiction can undermine the establishment of rapport.^20,21 At the micro-interactional level, the language of pain medicine relies on morally loaded terms that evoke judgmental meanings, such as “narcotic”²² and “drug-seeking,”²³ while pain itself can be difficult to express verbally.^24–26 Consequently, providers describe communication about pain as time-consuming, uncomfortable, and frustrating.¹⁵

Despite widespread acknowledgment of communication difficulties, we know very little about patient-provider communication about pain. Pain researchers have called for the analysis of communication data,^1,27 yet only a few studies have examined audio-recordings of clinical interactions.^19,28–31 Eggly and Tzelepsis describe communication about pain treatment as a competition for control. Notably, while physicians possess a range of strategies for gaining interactional control, patients have more limited means, focusing primarily on expressing disagreement with physicians' proposals.²⁸ Kenny found that both patients and physicians face obstacles to developing trust and credibility, and observed that patient-provider communication is marred by fundamentally different views about the causes and meanings of pain.¹⁹

Communication about pain treatment is not always tension laden, however. In their study of pain treatment in VA primary care, Matthias and colleagues illustrate that both patients and physicians express fears about opioid use and engage in supportive reassurance.²⁹ Identifying a need to characterize productive aspects of patient-provider communication alongside the negative, they note that many patients appreciate their physician's cautionary stance toward opioids and interpret such actions benevolently.³¹

1.2 Theoretical framework: politeness theory

Brown and Levinson's sociolinguistic theory of politeness¹⁴ offers a useful framework for understanding interpersonal communication in healthcare.³² Politeness theory suggests that people enter interactions with a desire to preserve “face,” a concept drawn from the work of sociologist Erving Goffman referring to the public manifestations of self and identity.³³ Brown and Levinson argue that face is structured by two universal desires: the desire to be liked (positive face) and the desire for autonomy (negative face). Speech acts that risk social disapproval or impede one's interpersonal agenda are termed “face-threatening acts.”

People manage face threats through five interactional strategies that are transmitted through implicit sociocultural learning. (See Table 1.) The selection of strategy is influenced by the degree of social distance between conversational actors and relevant power asymmetries. For example, in a Taiwanese pediatric setting, parents adopted more off-record strategies and commonly demonstrated support for physicians' statements before raising a concern of their own.³⁴

Table 1. Brown and Levinson's politeness strategies^*.

Politeness Strategy	Example
Refraining:
Avoiding particular conversational topics
Off-Record:	P= 54-year-old white female
Producing ambiguous statements to avoid taking responsibility for one's position	D:	Before we begin, is there anything I can do to make you more comfortable?
	P:	I don't know. I'm in a lot of pain.
	D:	I can help with that.
	P:	((Laughs))
	D:	Ok as soon as I leave I will order some pain medication for you.
	P:	Okay. Okay.
Negative Politeness:	P=33-year-old white female
Minimizing the imposition placed on one's interlocutor through indirectness, hedging, deference, apologies, or self-effacement	D:	Alright, well the latest note from the pain clinic which you saw just about a week ago, um, they do not mention that the pain contract had been dropped. Okay, so I'm happy to offer you pain medications while you're here in the ED, um.
	P:	Okay, well can you get a hold of Dr.___ (name)?
	D:	I can try to.
	P:	Because that's who the pain contract is under and that that was revived, revised, whatever- revived, whatever that word is
	D:	Mm hm.
	P:	I'm no longer on any pain medication narcotics with him and that was the only reason that pain contract um was brought up. And I stopped with my narcotics with him.
	D:	Mm hm.
	P:	So. Whatever you want. Whatever you want to do, doc.
	D:	Okay.
	P:	I am in pain.
	D:	Okay.
	P:	And if you could give me, and if you could give me something for my pain, then yes, I would appreciate that, but, um, if you also can't, maybe get in, uh, contact with Dr.____ (name), um, I would appreciate it.
Positive Politeness:	P=25-year-old white male
Demonstrating affection, affiliation, or cooperation	NP:	I can give you more medication if you want.
	P:	Um, that's your call. Um, I mean I'm hurtin like crazy.
	NP:	I'm not hurting at all. So you have to let me know.
		[…]
	NP:	I mean we can repeat the medication again. Sometime you just need that.
	P:	We can go with a shot then. One of them shots in my back they're always talking about?
	NP:	No, no. It would just be a pain shot. It's just for pain.
	P:	Um. Alright, let's try it.
Bald-on-Record	P=55-year-old black male
Use of a direct, unambiguous message, conveying less concern for politeness	NP:	You have any questions about anything? All right.
	P:	I need- you reckon you can give me somethin for
	NP:	Yeah, we'll give you somethin for [pain
	P:	[Pain
	NP:	Yeah. Yep, we can do that.

^*Adapted from Brown P, Levinson SC. Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press; 1987:1-135.

Politeness theory is particularly pertinent to understanding patient requests for analgesics for two specific reasons. First, politeness strategies are especially salient in highly sensitive interactional arenas due to the close relationship between stigma and face. Goffman famously defined stigma as a threatened social identity on the basis of one's departure from established social norms.³⁵ Face-threatening acts are thus both precipitated by, and constitutive of, stigma. Consequently, politeness strategies abound in conversations about stigmatizing issues in medicine.³⁶

Second, because the act of requesting and its logical response creates vulnerabilities for both speakers and hearers, requests are prototypical face-threatening acts.^37,38 Patients have expressed reluctance to request specific analgesics due to concerns that their requests will be dismissed or their motives distrusted.³¹ As we will demonstrate, patients mobilize politeness strategies to assert their agendas in such a way that they will not be perceived as threatening or usurping the provider's authority.

2. Data and methods

This paper reports findings from an audio-recorded study of patient-provider communication in an academic ED that sought to characterize communication about pain and analgesics. EDs are important sites for pain treatment because they are open 24 hours per day and provide a critical safety net for access to care.^39,40 Due to the exploratory nature of this study, we wanted to focus on a single type of pain. We chose to focus on back pain because it is one of the most common pain complaint of ED patients⁴⁰ and a leading cause of functional disability.⁴¹ The study received approval from the Institutional Review Board at UNC – Chapel Hill.

2.1 Participants

Prescribing providers (attending physicians, medical residents, and nurse practitioners (NPs)) were recruited via a combination of methods, including a presentation at a staff meeting, an informational email explaining the study's goals, and an in-person review of study protocol. The study was described to providers as an investigation of patient-provider communication in the ED. Providers received a $5 Starbucks gift card each time an encounter was recorded. At the end of the study, providers received a debriefing form that explained that the study sought to characterize communication about analgesics.

Potential participants were identified using the electronic medical record and screened prior to enrollment for the following eligibility criteria: 1) age 18 or older; 2) English speaking and literate; 3) indicate back pain as a primary complaint. Patients were excluded from the study if they were: 1) unconscious or disoriented; 2) immobilized using a backboard; 3) febrile; or 4) receiving dialysis. Only patients of providers who had agreed to be in the study were approached. ED clinical staff approached patients who met initial criteria and asked if they would be willing to hear about a research study while they waited to see a provider. The Principal Investigator (PI) or a research assistant (RA) then approached patients to explain study procedures and obtain informed consent, including a HIPAA authorization. Patients were invited to participate in an audio-recorded study of patient-provider communication but were not informed that communication about analgesics was a specific focus. Patients received a $25 gift card for participating.

2.2 Data collection

The PI and five RAs collected data over an eight-month period (September 2012 – April 2013). Five-hour research shifts were scheduled to synchronize with participating providers' shifts, with the goal of balancing daytime, nighttime, weekday, and weekend shifts. Hand-held digital audio-recorders were used to record all communication between patients and their prescribing provider. The PI or RA was present in the exam room during the patient-provider encounter to take notes, which provide an important context for recordings in the ED due to frequent interruptions.⁴² Information about visit characteristics and prescriptions was abstracted from the electronic medical record.

2.3 Data analysis

Audio-recordings were transcribed verbatim and de-identified by a trained transcriptionist. Two coders reviewed the transcripts to identify any instance during the visit in which a patient or provider mentioned pain medication. After an initial trial period, in which both coders and the PI (MB) reviewed each transcript and any discrepancies were discussed and resolved, the remaining transcripts were divided in half and reviewed by one coder. Coding was performed using Dedoose software.⁴³

Discussions of analgesics were then reviewed by MB and RW to further characterize patient request behavior. Categories were refined iteratively to accommodate and resolve discrepancies. We identified three basic patterns: direct requests, in which the patient explicitly asked the provider for medication; indirect requests, in which the patient hinted at a desire for medication but did not ask for it outright; and no request, in which the provider discussed a prescription without the patient requesting it. Although much of the clinical policy literature focuses on opioid analgesics,^3,44 we did not differentiate between different types of analgesics in our analysis because patients rarely made such distinctions themselves; thus, it is not clear that patients differentiate between requests for different kinds of analgesics to the extent that providers do.

We subsequently analyzed this corpus of request interactions to identify patterns and themes in patient request behavior (or avoidance thereof) using a fine-grained, inductive approach based on conversation analysis (CA).⁴⁵ CA is a micro-analytic approach to understanding social interaction that recognizes conversation as a form of social action. An important tenet of CA is that one participant's conversational utterances have logical consequences for conversational partners. Attention to the sequential organization of conversational turns is thus an important facet of the CA approach. In CA, transcripts are read line-by-line to examine how talk is structured by the prior conversational turn, with a particular focus on the social action accomplished by specific interactional framing.

3. Results

3.1 Overview and description of sample

Of the 40 providers approached, 32 (80%) agreed to participate in the study. The most common reason cited for declining was that the provider did not want to be recorded. Two providers were not assigned an eligible patient; the final sample thus included 30 providers (12 attending physicians, 7 medical residents, and 11 NPs).

Eighty of the 104 patients approached agreed to participate in the study. The most common reasons patients cited for declining were being in too much pain and not wanting to be recorded. Six patients dropped out or were excluded before the study procedures could be completed, for a final sample of 74. The patient sample was 50% female, 60% white, and 35% black. A quarter of all patients did not have the equivalent of a high school degree, 49% were currently unemployed, and 53% had no medical insurance.

Nearly everyone in the study left the ED with an analgesic prescription, regardless of whether they requested it. Rates of medication receipt during ED stay were somewhat lower because hospital policy stipulated that patients who had driven themselves to the ED could not receive opioids within six hours of discharge. While this policy did not preclude patients who had driven themselves from receiving non-opioid analgesics in the ED, in many cases, providers opted to offer an opioid analgesic prescription to go rather than offering a non-opioid treatment in the ED. Of the six patients who were not given a prescription, only one case entailed the rejection of a patient's request. This request was refused because the patient had an opioid treatment agreement, which stipulate requirements for opioid treatment and often limit opioid prescriptions to being filled by a single primary care provider, on file in the electronic medical record. Requests for analgesics were documented in 15 of the 74 interactions (20.3%). Below, we describe our typology of patient request behavior, which includes direct requests, indirect requests, and no requests.

3.2 Direct requests

Eight patients expressed a direct request for an analgesic. Two patients had misplaced a prescription received during a recent ED visit. Only two patients specifically requested drugs by name. One requested ibuprofen, an over-the-counter (OTC) medication; the other was a patient with chronic pain who had recently relocated from another state and requested refills of several opioid medications. In the latter case, the patient indicated medication as the primary reason for her visit and said that she was running out and had not yet identified a pain manager. The NP gave her a prescription for a one-time, two-week supply of Soma, Dilaudid, Klonopin, and Percocet and a referral to a primary care provider. The NP also told the patient, “I won't be doing it again.”

The following example constitutes a prototypical direct request.

Example 1. P=44-year-old white male
1	D:	Ok, um, do you get have you ever had your prostate checked?
2	P:	Yeah.
3	D:	Ok and that's always been [normal?
4	P:	[Last year.
5	P:	Yeah. (.) Oh, can you give me anything? Anything to help with this pain?
6	D:	Yeah, I can give you- we can get you medicine for pain.

In relation to the rest of our corpus, this bald-on-record request is unusual in several respects. First, it identifies the NP (“you”) as the agent of the action implied by the request (“give”), creating an unavoidable entailment for the provider (line 5). Second, the question structure is optimized for a “yes” response.⁴⁶ Third, the patient does not employ any strategies to mitigate the imposition on the NP.

Most patients in our sample adopted one or more contrasting strategies when making a direct request for an analgesic. In the following extract, for example, the patient avoids naming the NP as the agent of the requested action:

Example 2. P=26 year-old white male
1	P:	That's where I went first. It's right by where I live.
2	NP:	Okay.
3	P:	Is there a way I could have some ibuprofen or somethin?
4	NP:	Yeah. Okay. I'm gonna take a look at you.

Although the patient is likely aware that the NP would need to authorize the receipt of medication, the indirect framing of his request (line 3) does not position her as such. A similar strategy can be found below, where the patient maintains agency over the desired action:

Example 3. P=39-year-old black male
1	NP:	Are you breathing okay? Chest pain? Cough? Shortness of breath?
2		Palpitations? Heart irregular?
3	P:	Other than my back I'm in perfect health, I think.
4	NP:	Okay.
5	P:	Is there any way I can get something for the pain?
6	NP:	Yeah, once I examine you, I'll write for um some pain medicine.

Whereas in Example 2, the request for medication is optimized for a “yes” response (“Is there a way”), in Example 3, the question is polarized negatively (line 5). Insofar medical questions that use “any” are designed for a “no” response,⁴⁶ the patient's question demonstrates deference toward the NP's authority. As a negative politeness strategy (see Table 1), this question is particularly significant for the way in which it sublimates the patient's autonomous goals (i.e. to receive medication) to a desire for social acceptance (i.e. maintaining positive face).

Another way of minimizing the imposition of a direct request entails downplaying the hypothetical burden of a requested action. Below, the patient diminishes the significance of his request by referring to medication obliquely:

Example 4. P=46-year-old white male
1	D:	So, anything else been going on other than what we've talked about?
2	P:	No sir.
3		(.)
4	P:	Would it be possible to maybe get a little something you know to help with my
5		back pain a little?
6	D:	Yeah I mean that's what we're going to do.

By describing the object of his request with a diminutive form instead of naming it directly (line 4), the patient depicts his request as unproblematic and non-threatening. Furthermore, by qualifying the intended therapeutic action with “a little” (line 5), the patient minimizes the extent of the intended therapeutic action, thereby reducing the burden of request.

3.3 Indirect requests

Eight patients in our sample requested a medication through indirect means.¹ The example below is typical of this pattern insofar as the patient avoids going “on-record” as asking for medication.

Example 5. 52-year-old white female
1	NP:	And what would you like for me to do for you today?
2	P:	Get me out of pain.
3	NP:	Okay, so the pain medicine?
4	P:	Yeah.

Here, the patient expresses a desire to be “out of pain” without specifying the agent of pain relief which could be construed as a face-threatening act.

In formulating indirect requests for analgesics, four patients referred to a specific drug by name. Patients displayed several strategies for framing indirect requests, including establishing the relative efficacy of a medication consumed in the past, inquiring about a treatment that a different physician had mentioned previously, or expressing a desire in a hypothetical manner. These strategizes were frequently characterized by an ambivalent orientation toward analgesics and framing analgesics as a last resort.

Example 6. P=45-year-old white female
1	NP:	Have you been taking ibuprofen or anything for your back pain?
2	P:	Um, I have not. Dr.___ (name) had gave me some Valium and Percocet and I
3		honestly only took them when I needed them.
4	NP:	Mm hm.
5	P:	And now the Valium did really help. The Percocet's (.) Oooh. Yeah.
6	NP:	Okay.
7	P:	Mm, um, I mean, it helped and I kept them for the longest time. Like, I just didn't
8		want to take them, like, just only as I needed them.

Here, the patient attests to the efficacy of the Valium and Percocet prescribed by another physician. While she does not request these drugs directly, she provides implicit support for the NP to repeat this already proven treatment. Furthermore, the patient exhibits judiciousness with respect to medication, chiefly, by asserting that she has withheld medication unless absolutely necessary. The NP responded by prescribing Percocet, Valium, and Motrin, in the ED and to take home.

The patient below adopts a similar strategy to the patient in Example 6.

Example 7. P=37-year-old white male
1	P:	I was in the middle of moving when this happened.
2	NP:	Okay.
3	P:	Started looking for a new doctor down here some.
4	NP:	Gotcha. I'll talk to you about that in a second. And that's what you do the
5		gabapentin for and the Robaxin and the Cymbalta, is that correct?
6	P:	Yeah.
7	NP:	Okay.
8	P:	And I was prescribed fifty milligrams of Oxycodone but I was trying to get off of
9		it but I hadn't had it for a couple of months so I was doing good without it. And I
10		really wished I had some of that. ((Laughs))

Here, as in the previous example, the patient affirms the efficacy of a prior treatment (line 8) and casts himself as a deserving patient by referencing successful attempts “to get off of it” (lines 8-9). Moreover, rather than ask for Oxycodone directly, he expresses his request more obliquely, as a “wish” (line 10) that he had it. With this hypothetical framing, he minimizes the imposition on the NP to respond to his desire, a negative politeness strategy. The NP responded by prescribing Oxycodone in the ED and to take home.

In the next case, which involved a patient who had been in the same ED three weeks ago for the same problem, rather than uphold the efficacy of a prior prescription medication, the patient maintains the inefficacy of OTC treatments.

Example 8. P=33-year-old white male
1	D:	How do we want to manage this?
2	P:	I mean I'm hurtin really bad.
3	D:	Okay.
4	P:	It's hard to move. It's hard to do anything.
5	D:	Right.
6	P:	I mean it's like- it's mainly like right here.
7	D:	Okay.
8	P:	Uh, I don't know. I mean like I said Ibuprofen or Tylenol you know any of the
9		cheap pain relievers, ain't nothin helping nothin.
10	D:	Okay. Well, I'm willing to give you another prescription here. Cuz we've got you
11		know obviously an issue with your pain. I'm gonna make a note that we've talked,
12		had this conversation and this is the last time you're going to get a prescription
13		here.

Here, the physician displays positive politeness by inviting the patient's input into the treatment decision (line 1). As in Example 5, the patient avoids going “on-record” as wanting medication by offering an ambivalent answer (line 8). At the same time, by affirming the severity of his symptoms (lines 2 and 4) and asserting the inadequacy of Ibuprofen and Tylenol (line 8), he makes an implicit bid for a stronger medication. Particularly noteworthy here is the patient's probable awareness, in light of his recent ED visit, that this could be a delicate request. This presupposition is later borne out by the physician's determination that this will be “the last time” that the patient receives a prescription in this ED (line 12-13).

3.4 No request

In most cases in our study, providers prescribed analgesics or offered to do so without receiving a request. Patient responses to such offers are revealing of their orientations toward analgesics and deference to physician authority. In the positive politeness example presented in Table 1, the patient attempts to cede decisional control to the provider, even though he was “hurtin like crazy” and was explicitly invited to weigh in on the decision. In several cases, patients do not directly accept the offer of analgesics, as the following example demonstrates.

Example 9. 53-year-old black male
1	NP:	Do you have any, um, do you need some pain medicine?
2	P:	Well it's hurtin.
3	NP:	Okay.
4	P:	It's kind of hurtin.
5	NP:	Hurting like you want a Tylenol or Motrin or hurtin like you want something a
6		little bit more?
7	P:	Uh, I might have to have something a little more.
8	NP:	Okay. I'm okay with that. That's why I asked.

By replacing “any” with “need” (line 1), the NP reframes his initial offer of medication to a question design that favors a “yes” response. At first, the patient is equivocal, downgrading “it's hurtin” (line 2) to “it's kind of hurtin” (line 4). The NP responds by giving the patient dichotomous options and presenting the prescription medication option in a non-threatening way by mitigating it with the diminutive “little bit more” (line 6). Finally, the patient agrees to the stronger of the two medication options, to which the NP responds supportively.

While the patient above ultimately agreed with the NP's treatment recommendations, several patients demonstrated resistance to the provider's prescription decision in ways that seemed less motivated by politeness than by concerns about side effects or dependence. One patient said, “I'm sorry this is going to sound real unorthodox, but I would really much rather smoke a joint than take them crazy pills.” Another patient, when offered Percocet, responded, “Hate that shit.” These examples serve as important reminders that not all patients with pain seek medical attention because they desire medication. Other common patient requests included requests for MRI imaging and requests for a sick note for work.

4. Discussion and Conclusion

4.1 Discussion

This qualitative study of patient-provider communication about analgesics in an academic ED adds to a small but growing body of literature that provides a different perspective on mounting concerns about the opioid epidemic and its effects on the clinical encounter. Lewis et al found that many VA patients take opioid medications less than their prescribed dose, to minimize intake, reduce the risk of addiction, or decrease financial burden.¹⁰ The authors of that study theorize that patients may accept opioid prescriptions against their better judgment so as not to offend their doctors. Along with this work, our study calls into question research and media portrayals that characterize pain patients as “drug-seekers” who generally desire opioid analgesics.

Despite the popular conception of pain patients as “drug-seekers,” only 20% of the back pain patients in this study requested analgesics. Furthermore, those who did so displayed sophisticated strategies of indirection, deference, and mitigation, suggesting that they were attuned to the delicacy of this communicative action. The reluctance to request analgesics implies that patients perceive asking for analgesics to be a delicate and potentially stigmatizing act, revealing a sensitivity to the operation of authority and asymmetry in medicine.^47,48

A striking finding of this study is that patients who requested an analgesic by name were more likely to do so indirectly. One possible explanation is that patients felt more comfortable requesting specific medications indirectly as this impinged less on provider autonomy. This finding is particularly important in light of the experimental study by McKinlay et al,¹³ in which the independent variable was the “active request” of a specific analgesic. Although the experimental prompts in their study (reference to a spouse's medication and reference to a pharmaceutical advertisement) more closely mirror what we refer to as indirect requests, we did not find evidence of their request strategies. Our participants were more likely to appeal to the efficacy of drug they had previously taken or to assert that OTC treatments were insufficient.

This article represents one of only a few studies that examine patient-provider communication about pain, and the first to focus specifically on requests for analgesics. Although nearly all patients in our study received an analgesic prescription, prescription rates vary with institutional culture.⁴⁹ In one multi-center study of ED pain treatment, based on patient recall rather than on recorded interactions, 42% of patients who did not receive an analgesic felt they needed one, yet only 31% of this group had actually requested it.⁵⁰ Further attention to patient requests may enhance understandings of analgesic treatment disparities.

This study has several limitations. First, because participants were recruited from one academic ED, our findings may not be generalizable to other clinical settings. Patients in our study were typically meeting their providers for the first time, but patients may be more or less motivated to employ face-saving politeness strategies with providers with whom they have a long-term relationship; plausible arguments can be made in both directions. Although politeness theory has been mobilized to analyze a variety of clinical interactions,^37,38,51,52 further work is necessary to elucidate the application of our findings to pain encounters outside of the ED. Second, patients who declined to be in the study may have been more likely to request analgesics directly than those who agreed to be recorded. Nevertheless, because only 23% of patients we approached declined, we can be reasonably sure that direct requests were not dominant in this population. Third, patients and providers may have altered their behavior due to the presence of an observer in the room. Fourth, we did not ask patients directly about their goals or expectations for the visit, which would have provided useful contextual information about their requests (or lack thereof). Fifth, although men in our study made more requests than women, our sample was not large enough to quantitatively investigate gender differences in patient request styles. Despite these limitations, this study demonstrates that examining audio-recordings of clinical interactions can yield valuable insights about how patients and providers talk about analgesics in a real-world clinical context.

4.2 Conclusions

This study contributes to work on the interactional challenges of pain medicine by examining patient requests for analgesics in an academic ED. Contrary to popular and clinical depictions of pain patients as “drug-seekers,” most patients did not request either opioid or non-opioid analgesics. When they did so, they utilized strategies of mitigation, indirection, and deference that presented themselves as deserving patients while upholding the physician's autonomy.

4.3 Practice implications

Attending to patient-provider communication in a naturalistic context can inform clinical guidelines and policies and illuminate fresh perspectives on challenging issues. Patients come to the clinical encounter with a variety of hopes and expectations, of which a desire for an analgesic may be only part of the picture. Rather than focusing on strategies for inuring providers to inappropriate patient requests, it may be useful to devote clinical resources to examining patients' priorities and expectations for treatment.

Patient-provider communication about pain is challenging yet understudied.

We examine how patients ask for pain medications in a real-world clinical context.

We audio-recorded clinical conversations in an academic emergency department (ED).

Contrary to ideas about “drug-seekers,” most patients did not request analgesics.

When they did request pain medications, they did so indirectly and deferentially.

Appendix: Transcription Conventions

NP:	Nurse practitioner speech
D:	Physician speech
P:	Patient speech
F:	Friend or family member speech
[	Denotes overlapping speech
[…]	Transcript excerpted
( )	Uncertain content
(.)	Pause
(name)	Person or place removed for anonymity
(( ))	Nonverbal action