Patient Involvement in Anesthesia Decision-making: A Qualitative Study of Knee Arthroplasty

This study took place at a large urban American academic medical center. We used qualitative methods. Qualitative research addresses causality through directly identifying connections in order to generate in-depth accounts of process, in contrast to quantitative approaches, which infer causality by identifying differences in frequency or intensity between groups along a dimension of interest.¹⁹  Qualitative methods are thus apt for the description and explanation of complex social, cultural, and cognitive processes like those involved in anesthesia decision-making. Further, qualitative methods are not typically structured by a hypothesis, instead seeking to capture any dynamics relevant to the process of interest. They are consequently able to identify unanticipated yet important factors more easily than can quantitative techniques (e.g., surveys) that must delineate factors of interest a priori. The open-ended nature of qualitative methods was suitable for this study given the dearth of previous work that directly examined anesthesia consultations.

Specifically, we used the qualitative approach of focused ethnography.²⁰  Popularized by anthropology and sociology, ethnography is an immersive method using mainly observation and interviewing to describe social and cultural processes in their ordinary, everyday settings. Focused ethnography adapts ethnography to the study of topics that are specialized, thus demanding discrete observation sessions and targeted interviews that focus on specific settings and individuals.

The University of Pennsylvania institutional review board approved this project. Participating patients provided written informed consent, while participating anesthesiologists provided verbal informed consent. Data collection was performed by J.J.C. (a research assistant trained by J.T.C, an experienced qualitative researcher) from February 2018 to July 2018. We used purposive sampling to enroll patients. Eligible patients were those undergoing primary knee arthroplasty, a procedure that typically presents a decision to use spinal or general anesthesia.²¹  We equally stratified patients across participating anesthesiologists. Patients were recruited in the preoperative area on the day of surgery, just before the anesthesia consultation, which occurs directly before surgery in this surgical center. Patients then participated in audio-recorded semi-structured interviews (see Supplemental Digital Content 1 for interview guides, https://links.lww.com/ALN/C611) in which they were asked about previous experiences with surgery and anesthesia, with whom they had spoken about this procedure, other sources of information they had accessed, concerns they had about anesthesia, and expectations about the preoperative consultation. Questions were open ended, giving little priming information to avoid altering the behavior of patients in the consultation. If an interview was not completed before the consultation, any remaining questions were asked directly after the consultation. Consultations were observed and audio recorded. Anesthesiologists participated in audio-recorded semistructured interviews offsite, focusing on their general considerations in deciding on anesthesia type, their approach to the consultation, and their impressions of how patients think about anesthesia. Data were collected until theoretical saturation²² —when additional data neither altered our coding schema nor changed the explanation we were developing to explain trends made apparent by coding (see Qualitative Analysis section below).

Audio recordings of interviews and consultations were transcribed by a professional service. Coding was managed using NVivo 12 qualitative analysis software (QSR International, Australia). First, V.G. (a practicing anesthesiologist and clinical researcher), J.T.C., J.J.C., S.J.H. (a research assistant), and M.M. (a research assistant) all annotated two randomly selected patient interview transcripts and their associated consultation transcripts and one randomly selected anesthesiologist interview transcript. Annotations were discussed to generate themes, which were then formalized into a codebook (a taxonomy for thematic categorization of data).²³  J.J.C., S.J.H., and M.M., supervised by J.T.C., used this codebook to double-code four randomly selected patient interviews and their associated consultations and two randomly selected anesthesiologist interviews (S.J.H. coded all; J.J.C. and M.M. coded two patient interviews/consultations and one anesthesiologist interview each). Coding was compared and all discrepancies rectified through consensus, and the codebook was revised to refine ambiguous categories, eliminate those lacking utility, and create new categories to capture missing themes. A second, identically structured round of double-coding and codebook iteration was performed using a different set of randomly selected transcripts. Having achieved a refined codebook and agreement about how it should be applied, S.J.H. then coded all remaining files. Finally, M.M. and J.J.C. coded an additional randomly selected set—identical in size to previous rounds of double coding—to verify the consistency of the coding performed by S.J.H. All codebook revisions were applied to previously coded transcripts. After this basic coding process, we performed focused coding,²⁴  prioritizing themes most pertinent to our research question, refining these themes, and combining related themes. Using this final set of themes, we developed an overarching explanation using an abductive approach,^25,26  during which we generated potential explanations, assessed their levels of empirical support, and—through this assessment—revised until arriving at explanations that best accounted for our data.

We summarized the characteristics of our participants using descriptive statistics, counting the number of participants in each category and calculating percentages.

We analyzed anesthesia consultations of 36 primary knee arthroplasty patients, and also interviewed these patients. Of this sample, 25 (69%) ultimately underwent spinal anesthesia and 11 (32%) underwent general anesthesia. Consultations were carried out by eight different anesthesiologists—four consultations observed per participating anesthesiologist—all of whom were also interviewed. Four consultations were carried out mainly by resident anesthesiologists, with attending anesthesiologists signing off on the decisions made. (See table 1 for sample characteristics.) Hereafter, we describe the perceptions of patients and anesthesiologists and their interactions in the consultation.

When discussing the goal of the preoperative consultation in interviews (see table 2), some anesthesiologists stressed that the purpose of the consultation was primarily educational, describing their role as informing the patient about the anesthetic plan that was most medically appropriate in order to instill comfort (table 2, row 2.1). Others identified the interaction as a decision-making situation in which their role was to present options to the patient, educate them on these options, and elicit a choice between them—the ultimate intent being to arrive at a plan that accords with what the patient wants (table 2, row 2.2). Among anesthesiologists who construed the preoperative consultation as an opportunity for facilitating patient choice, there was variation both within and across accounts about how to do so. At times, these anesthesiologists stressed that so long as a patient was not contraindicated for a particular anesthesia type, they approached the consultation with genuine equipoise and thus tried to present the anesthetic options in neutral terms, allowing the patient to choose (table 2, row 2.3). Other accounts exhibited an uneasy vacillation between neutral presentation and a desire to shift the patient toward a particular anesthetic option preferred by the anesthesiologist without bluntly overruling the patient’s inclination (table 2, row 2.4).

When describing what they expected from the preoperative interaction (see table 3), patients’ accounts varied in ways similar to those apparent in the anesthesiologists’ interviews. Some patients expected not only to be informed about the details of the anesthetic procedure, but also to be actively involved in the process of deciding on anesthesia type (table 3, row 3.1). Others did not expect or desire to use this information to contribute to the selection of an anesthetic approach, as they did not feel qualified, comfortable, or interested in doing so (table 3, row 3.2); their desire for information stemmed instead from not wanting to be unpleasantly surprised by the anesthesia experience. Patients who did not perceive the preoperative consultation as a decision-making situation often talked of their commitment to undergoing surgery after a lengthy period of pain unmitigated by other interventions (table 3, row 3.3). Anesthesia for these patients was an inevitable component of a procedure that they had long been anticipating.

As seen in table 4, patients displayed patterned predispositions about anesthesia stemming from a range of previous experiences. Patients who had previously undergone surgery commonly came into this surgery with an initial preference about anesthesia, which could be derived from an uneventful or negative experience with a particular approach (table 4, row 4.1). Another common source of predispositions was previous discourse with family and friends who had undergone surgical procedures or had relations who had done so (table 4, row 4.2). In the latter scenario, patients were part of sometimes long chains of interactions by which ideas about anesthesia were disseminated. Experiences had to be noteworthy in order to be disseminated in this way, and so these experiences were predominantly negative and contributed to patient anxiety. Patient predispositions were frequently brought up in strong terms early in preoperative consultations (table 4, row 4.3). Negative notions about anesthesia were more likely to be brought up and occupied more discursive space than did neutral or positive stances. Because of the frequent appearance of these predispositions in consultations, all participating anesthesiologists described them in detail during interviews (table 4, row 4.4).

In six consultations, the anesthesiologist told the patient that they were contraindicated for spinal anesthesia and would have to undergo general anesthesia. The remaining 30 consultations exhibited several distinct patterns by which the physician and patient arrived at an anesthetic plan (see table 5). In eight consultations, the anesthesiologist did not present the discussion as an opportunity for decision-making; rather, they told the patient that they would be undergoing a certain type of anesthetic, followed by a check for understanding or elicitation of questions, and the patient did not subsequently bring up any issues that derailed the anesthesiologist’s plan (table 5, row 5.1). Conversely, two consultations saw the anesthesiologist present the consultation as a choice, maintain an explicit stance of equipoise throughout, and defer to the patient’s inclination to undergo a particular anesthesia type (table 5, row 5.2). The remaining preponderance of consultations occupied a middle ground between the scenarios just described. In seven consultations, the anesthesiologist described the operation as appropriately done with either anesthesia type but in doing so expressed a preference for one, and the patient quickly agreed to the anesthesiologist’s preferred method (table 5, row 5.3). Twelve consultations similarly featured the physician presenting anesthetic options and expressing preferences; however, in these cases, the patient did not quickly acquiesce to the anesthesiologist’s approach (table 5, row 5.4). In response, the anesthesiologist continued to justify their preference through describing its advantages and/or assuaging the patient’s discomfort with it until the patient consented to its use. Finally, in one1 consultation, the anesthesiologist presented options and conveyed a preference, the patient was disinclined to pursue that preference, and the anesthesiologist’s brief subsequent attempt to persuade was unsuccessful (table 5, row 5.5).

In consultations where the anesthesiologist presented multiple procedures, recommended one, and a patient expressed a negative inclination about it (see table 5, rows 5.4 and 5.5), a great deal of remaining time was devoted to the anesthesiologists’ attempts to allay these concerns. These efforts relied heavily on “grading,” as anesthesiologists characterized the intensity of some aspect of the anesthesia procedure through comparisons to other procedures—medical or nonmedical—that also had this aspect (see table 6).^27–29  Likely due, in part, to the relative lack of patient familiarity with regional anesthesia and in part to the preference of many participating anesthesiologists for its use in knee arthroplasty, most of this grading was done to address patient hesitancy about spinal anesthesia. Grading, for example, took the form of comparing the degree of sedation used under spinal anesthesia to that used during more familiar procedures like colonoscopy (table 6, row 6.1). Grading also compared the size of the needle used to deliver spinal anesthetic and the pain it would cause to more mundane experiences like receiving a tattoo (table 6, row 6.2). The risk of complications resulting from spinal anesthesia was downplayed by characterizing the risk as low and the complications as minor relative to other procedures (table 6, row 6.3). The frequency of the use of spinal anesthesia for this type of surgery was graded as high relative to alternative approaches, emphasizing the mundanity of spinal anesthesia.

In this study of anesthesia consultations, we found variation across the accounts of anesthesiologists and patients as to whether the consultation was an opportunity for a collaborative decision or an activity whose purpose was to provide information to patients before moving forward with an often long-awaited surgery. Patients sometimes had strong predispositions about anesthesia and frequently brought them up to anesthesiologists. Consultations displayed a number of decision-making patterns, from the anesthesiologist not framing the visit as a discussion of options to the anesthesiologist adhering steadfastly to a stance of equipoise; however, most consultations fell between these two poles, with the anesthesiologist presenting multiple procedures, recommending one, and persuading patients who expressed any disinclination. Anesthesiologists made patients feel more comfortable with their proposed approach by comparing its elements to more familiar procedures or experiences.

Faced with a patient who hesitates about or refuses a recommendation, a clinician can either reformulate the recommendation or justify it; the tendency to do one or the other varies by clinical setting.³⁰  In the preoperative consultations we examined, anesthesiologists usually sought to persuade patients to accept recommendations. With the exception of consultations in which they did not tell patients there were options, anesthesiologists sought to direct patients toward certain treatment plans without coming off as paternalistic, a dynamic observed in other contemporary studies of medical consultation.^31,32  For decades, Western medical training has taught practitioners to avoid overt paternalism by managing consultations in a way that simultaneously satisfies patients attuned to patient advocacy and medical consumerism, achieves medically sound plans, and meets the demands of time efficiency.^33,34 

Reliance on self-report data or on assessment of clinical communication with a tool that measures the content of medical discourse might lead to the conclusion that the consultations we studied show patient collaboration in decision-making. After all, in most cases, the patient was notified that there were multiple possibilities for anesthesia and was given opportunity to voice opinions about them. However, care must be taken not to characterize the patient’s role in determining treatment without fully considering the circumstances.³⁵  Only one patient caused the anesthesiologist to reformulate the treatment plan, despite many patients expressing initial discomfort about this plan. To simply describe the function of these preoperative consultations as involving the patient in a choice would thus be a mischaracterization.^36,37  The findings of this study combined with prior work on medical consultations suggest explanations for why patient influence on the anesthetic plan was limited. For one, physicians do not present a neutral ledger of risks/benefits or pros/cons of various options to patients, who then form preferences.³⁸  Rather, the facts of the situation are continually under construction during the consultation. Physicians, through their power to indicate degree of medical necessity and determine whether a patient’s experiential input is relevant to a given situation, have asymmetrical authority in these encounters.³⁹  To expect a patient’s experiential authority to overrule the anesthesiologist’s medical authority—except perhaps in situations where a patient is adamant and the anesthesiologist’s time to persuade is short—may be unrealistic. Second, the preoperative consultation occurs as part of ongoing medical care.⁴⁰  In orthopedics, surgery is typically cast as a solution in contrast to alternative treatments that are cast as palliative.⁴¹  Our study and others⁴²  have shown that patients, already committed to operations that they believed were their best hope for the long-term alleviation of pain and enhancement of function, often did not approach the preoperative consultation as a choice, but rather saw anesthesia as one inevitable technical component of much-anticipated surgery. Although many patients who participated in this study expressed strong preferences about certain anesthetic options, such inclinations usually faded quickly during the consultation.

Given these conditions, the preoperative consultations we observed were less oriented to the involvement of patients in decision-making and more to addressing affective and relational concerns. Patients do not necessarily desire biomedical information in consultations so that they can contribute to decisions. The information provided allows them to feel involved in their care⁴³  and—perhaps most important for anesthesia—calm and assured. Indeed, prior research suggests that patient satisfaction with anesthesia consultation is based on a summative judgment of the encounter more so than its specific ability to involve them in decisions.⁹  It has long been recognized that the preoperative visit plays an important role in addressing surgical patients’ anxiety.⁴⁴  Vulnerable and in an alien environment,⁴⁵  surgical patients can be deeply uncomfortable. Anesthesiologists have relatively little time to address this discomfort. The grading observed in this study, done through calibrating the intensity of various aspects of the anesthesia procedure to more familiar situations, is an efficient means of decreasing patients’ sense of alienation.

The findings of this study have implications for the pursuit of shared decision-making in anesthesia consultations. Implementation of shared decision-making would most clearly benefit the minority of consultations in this study where patients were not told that there were multiple anesthetic approaches. Even if anesthesiologists are more comfortable performing particular approaches, patients should be made aware when there is more than one clinically viable option, should the differences between the options possibly be of material concern to them. In most consultations we examined, however, anesthesiologists indicated that there were options and patients were able to express their predispositions. Nevertheless, patients typically deferred to, or were persuaded to accept, the anesthesiologists’ initial recommendations. It appears unlikely that the provision of more detailed medical information would have enhanced patient involvement in these decisions, given that the patients we studied mainly valued this information for its ability to prepare their expectations and decrease preoperative anxiety. The provision of greater opportunity for patients to introduce their preferences seems similarly unlikely to enhance their role in determining the anesthetic approach, since patients freely articulated comfort and discomfort with proposed procedures. As such, the potential benefit of behavioral interventions designed to enhance patient involvement in decisions should be carefully considered alongside their potential effect on the other functions accomplished by preoperative consultation. Such interventions may seem innocuous, but they are not costless.⁴⁶  Their use may result in less attention being devoted to important humanistic aspects⁴⁷  of preoperative communication observed in this study, such as lessening anxiety.

This study has several notable limitations. It was performed at a single academic medical center where preoperative arrangements may differ in important ways from other settings. Anesthesia consultations for orthopedic surgery at this surgical center occur directly before surgery. Though this arrangement is shared by many other centers, it differs from sites where surgical patients have separate preoperative visits at earlier time points. The practitioners at this center may also differ in important ways from those elsewhere. The anesthesiologists practicing at this site were mainly young and tended to favor regional anesthesia for primary knee arthroplasty. Relatedly, surgeons at this center were open to operating on patients under regional anesthesia and, in some cases, discussed this possibility with them before the anesthesia consultation. We did not have sufficient data on these conversations—patients were able to recall little about them in interviews—to accurately characterize their contribution to anesthesia decision-making, which is an interesting direction for future research. Finally, the study focused on a single surgical procedure. While this focus allowed us to achieve theoretical saturation, the medical inclinations of anesthesiologists, predispositions of patients, and interactional characteristics of the anesthesia consultation undoubtedly vary by surgery to some extent.

In conclusion, this qualitative study stresses that even a short preoperative consultation is a complex, multifaceted interaction that serves several functions. The involvement of patients in determining the anesthetic approach may not be the most important of these functions in many cases. Furthermore, this involvement is limited by a number of forces that extend beyond the consultation, including the authority of the physician to determine the relevance of the patient’s experiential input, the positioning of the anesthesia consultation after a patient has already committed to surgery, and the feelings of anxiety and alienation that often come with this commitment. This suggests that interventions to increase patient involvement in anesthesia decision-making would be most impactful if aimed beyond the preoperative conversation itself, instead altering how this interaction fits into the overall surgical trajectory. However, given our findings, broad consideration should be given not just to the feasibility of increasing patient involvement in anesthesia decisions, but also to the implications of such efforts given the other important functions that the consultation accomplishes.

The authors would like to thank Mark D. Neuman, M.D., Horatio C. Wood Associate Professor of Anesthesiology at University of Pennsylvania Perelman School of Medicine (Philadelphia Pennsylvania), for guidance on study design.

Support was provided solely from institutional and/or departmental sources.

The authors declare no competing interests.