Clinicians Consistently Exceed a Typical Person’s Short-Term Memory During Preoperative Teaching

Abstract

INTRODUCTION: 

Patient education is a critical part of preparation for surgery. Little research on provider-to-patient teaching has been conducted with systematic focus on the quantity of information provided to patients. This is important to assess because short-term memory capacity for information such as preoperative instruction is limited to roughly seven units of content.

METHODS: 

We studied the information-giving practices of anesthesiologists and nurse practitioners during preoperative teaching by examining transcripts from 26 tape recorded preoperative evaluation appointments. We developed a novel coding system to measure: 1) quantity of information, 2) frequency of medical terminology, 3) number of patient questions, and 4) number of memory reinforcements used during the consultation. Results are reported as mean ± sd.

RESULTS: 

Anesthesiologists and nurse practitioners vastly exceeded patients’ short-term memory capacity. Nurse practitioners gave significantly more information to patients than did physicians (112 ± 37 vs 49 ± 25 items per interview, P < 0.01). This higher level of information-giving was not influenced by the question-asking behaviors of the patients. Nurse practitioners and physicians used similar numbers of medical terms (4.0 ± 2.4 vs 3.7 ± 2.8 explained terms per interview), and memory-supporting reinforcements (2.3 ± 3.0 vs 1.4 ± 2.0 reinforcements per interview).

DISCUSSION: 

Given the known limits of short-term memory, clinicians would be well advised to carefully consider their patterns of information-giving and their use of memory-reinforcing strategies for critical information.

In Brief

IMPLICATIONS: Patient education is a fundamental component of outpatient preoperative screening. We studied the volume of information given during patient education and discovered that anesthesiologists and nurse practitioners typically exceed the limits of short-term memory by an order of magnitude, indicating a disconnect between current practice and basic understanding of human memory.

Relationships between health care providers and patients ultimately center on the dialogue created during their interaction. This communication is essential for developing a diagnosis as well as for creating and performing a treatment regimen. As early as 40 yrs ago, researchers established that providers’ inadequate communication skills frequently result in patient misunderstanding.¹ Without effective communication, a patient may not fully comprehend a diagnosis, understand the proposed treatment, or effectively consider options that are presented. Therefore, improving communication is vital for fostering a healthy client-practitioner relationship, as well as for improving health care outcomes.

Effective communication between patients and health care providers typically includes information exchange regarding the patient’s history, symptoms, diagnosis, and prescribed treatment.^2,3 For the clinician, successful communication involves asking open-ended questions, providing informative material and engaging the patient in joint decision-making.⁴ Information-giving in the preanesthetic evaluation is important, as patients may have misconceptions and incomplete knowledge about important, basic concepts in preanesthetic preparation and the role of the anesthesiologist in their care.^5,6

Information-giving may be enhanced through techniques such as repetition, explicit categorization, simplification of the language used, and the use of specific rather than general statements.⁷ However, there seems to be a disconnection between patients’ satisfaction with the quality of communication and our perception of how well this skill is taught. For example, many residency and fellowship programs rated their trainees as well prepared in communication issues,⁸ yet research indicates that medical workers frequently use unexplained medical terminology (e.g., scientific descriptions of physical symptoms)^9,10 when they speak with patients.

Much research has focused on the clinician’s use of empathic and open-ended questions as a predictor of positive clinician–patient communication experiences.^4,10,11 In training, clinicians are taught about the social and interpersonal aspects of interviewing skills. Though these factors are essential for establishing rapport and trust, they represent an incomplete conceptualization of clinician–patient communication. There is little emphasis on basic human cognition issues such as comprehension and memory, despite a rich body of data to guide training and inform our understanding.

Many decades ago, cognitive psychologists introduced the notion that the short-term storage of information is significantly limited. In his landmark article, Miller demonstrated that the capacity of the short-term memory store is 7 ± 2 items.¹² In other words, people can remember roughly seven pieces of information for a short period of time. That information is quickly forgotten unless it is actively moved into long-term memory through techniques such as rehearsal or elaboration.

Given that limited amounts of information can be retained for short periods of time, and that longer retention is predicated on effortful processing, one wonders whether anesthesia clinicians exceed their patients’ short-term memory capacity during preoperative interviews.

This concern is valid because anesthesia clinicians today discuss increasingly complicated technical medical information with patients. Levinson and Chaumeton¹³ suggest that, “the complexity of these diagnostic and therapeutic interventions may act as a barrier to effective communication and produce a more distant relationship between the physician and patient.” Communication with patients often entails lengthy periods of information-giving by the clinician, leaving little time for patients to process and comprehend the information or to request clarification. Simply reflecting on our own experiences as patients demonstrates the face validity of this assertion. Ley¹⁴ found that, “many patients do not understand what they are told, and many do not remember what was said.” He additionally found that in the few instances when information is reinforced by means of written communication, this information is often, “not noticed, not legible, not read, not understood, not believed, and not remembered.”¹⁵ Thus, we should be concerned that preoperative patients may find it difficult to remember the information they have been given— potentially leading to confusion, concern, misinterpretation of facts, and failure to recall information and instructions.

Research pertaining to provider-to-patient teaching has not been conducted with a systematic focus on the quantity of information provided to patients. Therefore, we sought to quantify the amount of information given by anesthesia clinicians during preanesthetic interviews. To better characterize the information flow, we also sought to quantify the number of memory reinforcement techniques used by providers, the number of medical terms explained by providers and the number of questions asked by patients during interviews.

Because of a unique organizational feature of our preanesthesia clinic, wherein anesthesiologists and nurse practitioners both conduct preanesthetic interviews in the same clinic, we were also able to compare the information-giving practices of these two different groups of clinicians. Given the differences in their educational and training backgrounds, we wondered whether nurse practitioners and physicians would differ in the quantity of information they presented to patients, or in the use of memory reinforcements during the interview. This question has some basis, as Kinnersley et al.¹⁶ found that patients seeking same-day medical consultations had significantly longer interactions with nurse practitioners than with physicians. Therefore, nurse practitioners in our setting might use more memory reinforcement, explain medical terms more often, and receive more patient questions than the physicians.

METHODS

This study was approved by the Human Research Committee of the Partners Healthcare System, (Massachusetts General Hospital, Boston, MA). Patient participants were recruited from the Pre-admissions Testing Area (PATA) of the hospital. Those who agreed to participate gave written informed consent. At the PATA, patients are seen on an unscheduled basis for preoperative consultations. Patients were approached in the PATA waiting room and asked if they would be willing to participate in a study of the natural variations in health care provider and patient communication. Exclusionary criteria for potential participants were: individuals younger than 18 years, individuals not capable of providing their own consent for participation, and individuals who planned to have a friend or family member with them during the consultation. Twelve patients, 7 women and 5 men, received a physician consultation; an additional 14 patients, 7 women and 7 men, received a nurse practitioner consultation.

Eight anesthesiologists and five nurse practitioners volunteered as clinician participants. The clinicians’ experience in the medical field ranged from residents in their second year of postgraduate training to multiple years of experience.

Design and Procedure

We audiotaped the consultations between participating patients and clinicians. Recording devices were positioned discretely and recording began when the practitioner entered the room. The clinicians were aware of the recording, but were given no guidelines for their provider–patient interactions and were unaware of the study objectives and hypotheses.

The consultations (ranging from 30 to 90 min in length) all contained a medical history interview, a brief examination of the patient’s airway, a summary of anesthesia for the operation, and a discussion/signing of the anesthesia consent form. For nurse practitioners, the supervising physician along with the nurse practitioner witnessed the signing of the consent form. Immediately after the consultation, confirmation of permission to use the audiotape was obtained from the patient. All patients agreed to allow their recording to be used in the study.

We attempted to contact patients by telephone the evening after the preanesthetic interview. Contacted patients were asked to recall the information they had been given about fasting before surgery, directions for medication prior to surgery, and risks of anesthesia.

Coding System

A coding system was developed to assess clinicians’ communication patterns. Transcripts were coded using four nonexclusive categories: 1) quantity of information given to the patient by the clinician, 2) numbers of explained and unexplained medical terms used by the clinician, 3) number of questions asked by the patient, and 4) number of memory-enhancing reinforcements used during the consultation.

The first category, quantity of information, was defined as the sum of the number of instructions, the number of descriptions of procedures or complications, and the number of interpretations of the patient’s symptoms. Because our focus was on the educational, information-giving, portion of the consultation, this code did not include information-gathering done by the clinician during the clinical medical history segment of the consultation.

The second coding category, the frequency of explained and unexplained medical terms, was a subset of the information category. Names of medications, procedures, and possible complications that were elaborated in common terms were classified as explained medical terms. Medical terms were categorized as “unexplained” if they were accompanied by patient queries about the statement, indicating that the patient did not understand it.

The third category, patient questions, included all patient requests for information, including clarifications about past information discussed in the consultation, requests for new information, and other queries relevant to the medical encounter.

Finally, in the fourth coding category, the number of memory reinforcements of information was extracted from the transcripts. These included rehearsal of information, elaborations of previously given information, and written back-up of verbally communicated information. Reinforcements were sub-categorized as clinician-driven or patient-driven.

It follows from the descriptions that the second, third and fourth categories are overlapping. For example, a clinician might make a statement containing a medical term. The patient may then ask for a clarification of the term. The clinician might then explain the term using lay language and reinforce the explanation by repetition. We would score this three-utterance exchange as two items in category 1 (medical term plus new lay terms to explain medical term), 1 item in category 2 (unexplained medical term), 1 item in category 3 (patient question) and 1 item in category 4 (clinician-driven reinforcement—repetition).

All of the transcripts were independently coded by two investigators. Each rater identified codable content and recorded the frequency of the categories occurring in each preanesthetic consultation transcript.

Statistical Analyses

All statistical comparisons were planned in advance of the study. Comparisons were performed using Statistica, version 6.1 StatSoft, Inc., Tulsa, OK. Pearson Product Moment Correlations were conducted to analyze inter-rater agreement. Information quantity, the number of unexplained medical terms, patient questions and the number of reinforcements from the interviews are presented as descriptive statistics. Results are reported as mean ± sd. Physicians and nurse practitioners were compared on these axes using provider type as the independent variable. We also analyzed the quantity of information provided during the interview with patient gender as the independent variable, and again with provider gender as the independent variable. All such comparisons were made using independent samples t-tests. Throughout the study, a P value of <0.05 was considered significant.

RESULTS

For the category of quantity of information given, there was a high level of agreement between raters, r = 0.86, P < 0.05. The raters’ agreement for patients’ questions was also strongly, positively correlated, r = 0.97, P < 0.05. The categories of explained medical terms and practitioner reinforcements maintained adequate reliabilities, r = 0.73, P < 0.05, and r = 0.71, P < 0.05, respectively. Although the correlations were statistically significant, the coding for unexplained medical terms and patient-driven reinforcements categories were less well correlated, r = 0.63, P < 0.05. We concluded that the coding scheme was suitably reliable to perform further analyses.

A sample transcript fragment, demonstrating the coding of information quantity, medical terminology and patient questions is provided in the Appendix. The quantity of new information in this transcript fragment is representative of the interviews. Systematic analysis of all of the transcripts revealed that clinicians, whether physician or nurse practitioner, grossly exceed the 7 ± 2 pieces of new information that cognitive science has long since established as the limit of the average person’s short-term memory capacity¹² (Fig. 1).

Figure 1.: Communication behaviors.

The number of unique information items presented by nurse practitioners was significantly higher than the number given by physicians (112 ± 37 vs 49 ± 25, P < 0.01). Based on this initial result, one might predict that there would be more patient questions during the nurse practitioner consultations, if one assumed that the rate of questions generated per unit of information given is roughly constant. Although the number of patient questions during the nurse practitioner consultations was higher than in the physician group, these differences were not statistically significant, (9.00 ± 16.18 vs 6.58 ± 7.84, P = 0.64). Nurses’ information-giving was not driven by patient questions.

No other comparisons between provider groups were significantly different. The average number of explained medical terms presented by nurse practitioners was similar to the physicians’ average for explained medical terms (4.0 ± 2.4 vs 3.7 ± 2.8, P = 0.74). Nurse practitioners’ use of unexplained medical terms was also similar to the physicians’ (0.5 ± 0.8 vs 0.4 ± 0.9, P = 0.80). Although the average number of memory reinforcements used by nurse practitioners was slightly higher than physicians’, the results were not significantly different, (2.3 ± 3.0 vs 1.4 ± 2.0, P = 0.40). Furthermore, the patient-driven reinforcements did not vary between the nurse practitioners and the physicians (0.4 ± 0.5 vs 0.7 ± 1.2, P = 0.40).^*

Finally, an independent samples t-test yielded nonsignificant results for the quantity of information given to male patients versus female patients (89 ± 50 vs 77 ± 41, P = 0.51). The quantity of information given by male clinicians did not significantly vary from that given by females (54 ± 26 vs 88 ± 46, P = 0.17).

We succeeded in contacting 12 patients for telephone follow-up of their preanesthetic interview. Ten patients were able to recall something about the instructions for preoperative fasting, while two could not recall that information. Only 7 of the 12 could recall anything about medication instructions prior to surgery. All 12 patients were able to recall a conversation about complications of anesthesia, although the potential complications recalled varied widely among patients.

DISCUSSION

At baseline, an average individual can recall approximately seven “chunks” of new information. ^†12 Against this backdrop, we observed an extreme tendency toward information overload by health care providers, coupled with a failure to use memory-enhancing techniques. Our study is novel in that it specifically examines the quantity of information transmitted in a preanesthetic interview. Others have studied the impact of strategies to improve patient recall after preanesthetic teaching, and patients’ ability to recall information has generally been disappointing.^5,17 These prior studies did not quantify the amount of information provided, but if they are representative of our own experience then information overload may explain their results.

Even considering memory-enhancing factors, such as personal relevance and scripted sequencing (e.g., making a “story” of how an anesthetic might be conducted), how can an individual possibly be expected to encode 50 to >100 medical descriptions and instructions, or to filter and recall the most relevant ones? Although our study was not designed to assess patients’ recall, our limited follow-up information demonstrates that they cannot remember all that they are told. Additionally, we speculate that patients may have trouble distinguishing information that is critical to recall (such as instructions concerning eating and drinking the night before surgery) from nonessential information. Both speculations lead to testable hypotheses that should be topics of future research.

Traditionally, increased information-giving has been favorably viewed in studies of the clinician–patient relationship. In the preoperative setting, patient ratings of general satisfaction with a preoperative clinic correlated with their ratings of factors related to information giving, although the authors of this study did not study information quantity.¹⁸ Perhaps the goal of the clinician giving a great deal of information is not to reliably inform, but to provide detailed information in the belief that it may enhance rapport and trust, and thereby increase patient satisfaction or reduce anxiety.

Anesthesiologists first documented the value of the preanesthetic visit for the relief of patients’ anxiety through the work of Egbert et al.¹⁹ in 1963. Prior to this work, Sheffer and Greifenstein²⁰ had documented, in 1960, that patients wanted to know more about what would happen to them during anesthesia, and that they wanted reassurance from their anesthetists. Egbert et al. also discovered that patients wanted more information about their anesthetics. In the second part of the 1963 study, they attempted to compare a purely fact-giving style of preanesthetic interview with a style in which the provider allowed the patient to draw them into a closer rapport by answering patient questions and providing emotional support. The study was abandoned because patients on the “facts only” arm invariably drew the clinicians into the sympathetic/reassuring style of information-giving.¹⁹ Thus, relief of anxiety and information-giving seem to be linked.

Additional evidence that the preanesthetic interview reduces patient anxiety continues to accumulate.^5,18 However, some research suggests that increased information-giving alone does not reduce patient anxiety.^21–23 Therefore, the rapport with the anesthesiologist must also play a role.²² Regardless of the final outcome with respect to anxiety, the possibility of information overload interfering with recall of information had not been considered until our study. Reduction of anxiety is a worthwhile goal, but recall of important directions is also important.

We were fortunate in our setting to be able to compare the information load presented by two different groups of practitioners. Compared with physicians, nurse practitioners give approximately twice as much information to patients during their consultations. This result confirmed previous research indicating that nurse practitioners, on average, provide more information to patients than do doctors.¹⁶ It has previously been proposed that patients may perhaps regard nurses as more approachable, or regard a nurse’s time as more available to them than a physician’s.¹⁶ However, our patients were not asking more questions during appointments with nurse practitioners than in appointments with physicians. Since the number of patient questions was not significantly higher for nurse practitioners, the amount of information given by nurses cannot be attributed to higher patient demands for information.

Application of basic behavioral science elements to the field of clinician-to-patient communication suggests that reducing information overload and application of memory-enhancing techniques should be joint goals when conducting the preanesthetic consultation. Patients’ encoding of information can be improved through explicit organization of information by the clinician. One technique that has been useful in enhancing patient’s recall was informing the patient of how the information was organized—e.g., “now, I am going to tell you what is wrong with you” or “here is what you must do to help yourself get better.” Communicating in this way gives the patient an explicit structure within which to organize new information.²⁴

Reinforcement is another technique that can improve memory capacity. Structured verbal educational content, reinforced by concise written or video-based material, could bring the necessary information load within the limits of a patient’s ability to remember. This need not occur during the preanesthetic interview itself. ^‡ For example, printed instructions can allow for future reference to forgotten directives or information.²⁵ Video-based instruction can also improve memory. Done and Lee⁵ assessed performance on an 18-item, true/false test about anesthesia procedures, risks and misconceptions in two groups of patients. One group viewed a standardized video that provided teaching on most of the questions presented in the test, while the other group had no intervention. Both groups then had a standard preanesthetic interview with an anesthesiologist who was unaware of group assignment. The group who viewed the video and then had a preanesthetic interview performed significantly better than the group who had only the preanesthetic interview. Unfortunately, only about 10% of the best performing group answered all of the questions correctly. This is not surprising given the limits of short-term memory, and supports the notion that short term memory is just as limited in the preanesthetic setting as in Miller’s experiments. The authors conclude that the video is a useful adjunct to the standard preanesthetic interview because it improves patient knowledge about anesthesia. We suggest that the combination of the video and the standard interview provided many opportunities for reinforcement through repetition of material (same material in video and then the interview), while the interview-only patients did not have the benefit of such reinforcement.

The reinforcing power of adjunct teaching tools apparently depends on contextual factors that cannot be easily discerned from the studies in which they were used. For example, Snyder-Ramos et al.²⁶ confirmed that either showing patients a video or having them review a brochure about anesthesia before the preanesthesia interview improved performance on a post-interview knowledge test, with the video giving the best performance. On the other hand, a similar study by Zvara et al.¹⁷ demonstrated that patients who had seen a video prior to the preanesthetic interview performed only slightly better than patients who did not receive this reinforcement. The use of written material or video as a preanesthetic teaching tool has been reviewed and subjected to meta-anlysis.²⁷ These authors demonstrated that the quality of trials was generally fair, with adequate subject allocation concealment and blinding during outcomes assessment in only a few studies. The authors concluded, nevertheless, that media presentations do provide useful reinforcement of preoperative teaching.

Our finding that clinicians, both nurse practitioners and physicians, vastly exceed patients’ ability to learn new information further justifies the effort and expense of reinforcements during teaching. Moreover, it suggests that it is time for a paradigm shift in preoperative education, since there is little hope that patients can recall all of the facts presented in a typical preoperative interview. We suggest that preanesthetic teaching should always describe the obvious structure for new information, to include “Now I am going to tell you how things will happen on the day of your surgery,” with explicit rehearsal of that information. Finally, each patient should receive written instructions specifically tailored to their condition and the planned procedure, introduced at the end of the interview with the statement “Now let’s review together this comprehensive written list of the most important instructions for you to follow to get ready for your operation.”

APPENDIX: EXAMPLE OF CODED TRANSCRIPT

The following transcript is an example of a physician–patient exchange. Each separately coded piece of information, as explained in the Methods section, is numbered with a subscript. The sample was also coded for medical terms. These were classified as unexplained when the patient indicated (through statements and questions) lack of understanding of the term. Examples of both {unexplained} and {{explained}} medical terms are set off by brackets as illustrated.

Patient

What if you don’t use the epidural then? How do they handle the pain?

Physician

They give you what is called the {{PCA}}₍₁₎ a little pump that you hold on to₍₂₎ and push a button₍₃₎ and you give yourself an injection₍₄₎ of medicine in the IV₍₅₎. {{A PCA stands for a Patient Control Anesthesia}}₍₆₎ and those are pretty good₍₇₎. I want you understand that some of the most common things with an epidural₍₈₎ is a 1–3% chance of a headache₍₉₎. It comes on when you get up and walk around₍₁₀₎. I don’t think that you are going to be up and around too much₍₁₁₎. Injuries, {permanent injuries to the nerves}₍₁₂₎ are extremely rare₍₁₃₎ – less than 11,000 but it does happen₍₁₄₎.

Patient

What does that mean?

Physician

{Permanent injury to a nerve?} It means that the …, a back nerve would be damaged₍₁₅₎. Once again, I am trying to get you understand here. The general anesthesia has the most common complications₍₁₆₎ meaning nausea and vomiting_(17)- maybe a 20–25%_(18)- it could last a couple of hours, maybe longer₍₁₉₎ and they can give you something to cut down on that₍₂₀₎ so that you might not get it but they can’t totally -

Patient

Okay, you lost me, I understand the thing in the back where they do that – as I have pain they would give it to me and that would automatically make sure that I am out of pain-

Physician

Correct, if you don’t do that, they would give you the PCA-the pump.

Patient

What’s that?

Physician

That goes into the IV, you will have it in you.

Patient

Oh, I see okay and then if I feel pain, I just push the button-

There were 20 new pieces information introduced in this interview fragment. There were two medical terms introduced: “PCA” and “permanent injuries to the nerves”(we assumed that “IV” and “general anesthesia” were commonly understood by patients). “PCA” is an example of a medical term that was introduced and immediately explained. “Permanent injuries to the nerves” is an example of an unexplained medical term, in this case by prompting a question from the patient. This question prompted an explanation from the physician, although the patient’s third and fourth utterances indicate that they do not fully understand their choices for pain management. The transcript contains two patient questions. The final two lines contain examples of clinician-driven and patient-driven reinforcements: “That goes in the IV” is a clinician reinforcement of “medicine in the IV” by repetition, while “… I feel pain, I just push the button-” is a patient-driven reinforcement “push a button … give yourself … medicine” by rehearsal.