INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Prior to World War I, battlefield surgeons routinely performed surgical procedures, including limb amputations, without anesthesia (1). Though these procedures may have been successful from the physician's perspective, the patients had to endure extreme pain. Even after anesthetics had been introduced to surgical practice, many clinicians continued to believe that pain was a necessary part of the healing process (1). Although invasive procedures may be technically feasible without premedication, there is increasing awareness that high-quality medical care includes adequate control of pain related to the procedure.

Flexible bronchoscopy (FOB) is a commonly performed invasive diagnostic procedure that has been considered safe whether or not performed with anesthesia, sedation, and analgesia (2). However, assessment of pain during bronchoscopy has received relatively little attention. FOB has been shown to cause dysphagia (3) and nose pain and throat pain (4) in a substantial proportion of patients. In one study 54% of patients were afraid of pain from the procedure (5), but because studies have shown that FOB can be performed without sedation, some bronchoscopists have argued against the use of premedication (6). Studies and editorials that have argued the use or nonuse of prebronchoscopy sedation and analgesia (11) have focused on whether FOB can be completed safely without sedation and analgesia, but not whether medication-free procedures are optimal from the patient perspective. There are few systematic studies about the extent to which patients experience pain during FOB, and whether there are certain patients at higher risk of pain. The purpose of this study was to assess the extent to which patients undergoing FOB experience pain and to identify patient factors and process of care factors that are associated with pain.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We conducted a prospective cohort study of adults undergoing flexible bronchoscopy at the Johns Hopkins Hospital and Johns Hopkins Bayview Medical Center in Baltimore, Maryland. Our study is part of the Bronchoscopy Quality Improvement project (BRONCHQI), which was designed to improve the quality of bronchoscopy services by systematically assessing the patient experience with FOB. At the outset of the project we interviewed 30 patients with open-ended questions in order to learn about patient concerns regarding bronchoscopy. We found that pain during the procedure was a frequent concern. The goals and methods used for this project have been described previously (4, 15).

Patients eligible for the study included all adults (18 yr of age or older) undergoing FOB from February 1997 to June 1998, performed or supervised by full-time Pulmonary and Critical Care Medicine faculty members. In 83% of FOBs, fellows in-training assisted in the procedure. Reasons for exclusion from the study included intubation and mechanical ventilation, inability to speak English, other communication deficits that precluded answering questions, and death within 48 h following FOB. The Johns Hopkins Joint Commission on Clinical Investigation approved this study.

Data collection involved the use of standardized report forms that were completed by physicians and patients. Physicians reported patient characteristics, indications for the procedure, details of how the procedure was performed, and outcomes of the procedure. Patients completed two questionnaires. The first questionnaire, completed just prior to FOB, assessed baseline symptoms prior to bronchoscopy. A second questionnaire was completed 48 h after the procedure. This questionnaire assessed symptoms experienced during and after the procedure, and assessed satisfaction with several aspects of their care. The questionnaires were either self-administered or administered by research staff in-person or by phone. We employed a protocol of serial phone calls and mailings to promote a high response rate.

Outcomes

The primary outcome of interest was patient reported pain control during the procedure on a five-point scale (poor, fair, good, very good, excellent).

Patients reported other outcomes from FOB. These outcomes included symptoms associated with specific aspects of the procedure, satisfaction with various components of the FOB experience, and changes in symptoms caused by FOB. Patients were asked how much they were bothered by numbing of their nose, taste of the throat spray, and insertion of the bronchoscope on a three-point scale (not at all, a little, a lot). Satisfaction with care included ratings of the doctors (skill, ability to comfort, gentleness, courtesy), nurses (skill of intravenous catheter insertion, ability to comfort, courtesy), and information provided about the procedure (information ahead of time, what to expect afterward, how to obtain results) on a five-point scale (poor, fair, good, very good, excellent). The patients also rated the bronchoscopy environment (helpfulness of escort, helpfulness of receptionist, privacy, safety, comfort for family, and overall assessment of facilities) and the amount of time they waited at various steps (waiting room [outpatients] or hospital room [inpatients], bronchoscopy suite prior to FOB [outpatients, inpatients], bronchoscopy suite post-FOB [outpatients, inpatients]). They reported how much they remembered about the FOB on a five-point scale (remember nothing at all, no details, some details, most details, or every detail). The survey administered to patients immediately prior to bronchoscopy included 13 items that assessed pre-FOB symptoms. Symptoms included nose pain, throat pain, chest pain, and swallowing pain on a four-point scale (none, mild, moderate, and severe). In addition, the frequency of coughing, coughing up blood, coughing up yellow or green phlegm, nose bleed, shortness of breath, difficulty swallowing, wheezing, shaking chills, and fever on a four-point scale (never, once or twice, several times, and all the time) were assessed. Patients answered the same questions 48 h later, reporting their symptoms in the 24-h period immediately following FOB.

Candidate Predictor Variables

Patients provided information on their health status, education, number of prior bronchoscopies, height, and weight.

Physicians recorded the underlying medical conditions (including chronic obstructive pulmonary disease, asthma, interstitial lung disease, and sarcoidosis), the indications for the procedures, and specific sampling techniques used: bronchoalveolar lavage (BAL), transbronchial biopsy (TBBx), endobronchial biopsy, proximal airway brushing, peripheral airway brushing, and transbronchial needle aspiration (TBNA). Physicians also reported the total dose of sedatives, analgesics, and topical anesthetic administered in each case. Administration of medication was determined by the bronchoscopist, and usually included a combination of lidocaine, midazolam, and fentanyl. The duration of the procedure was recorded (time with the bronchoscope in the patient, in minutes).

Analyses

We described baseline patient characteristics using the proportion of categorical or ordinal values. To understand what contributes to pain during FOB, we developed two separate models. The first model examined patient factors and care factors and their relationship to pain. The second model, designed to understand what other patient experiences may have contributed to pain, examined the association of other patient-reported outcomes with pain. In this second model, for example, we were interested in determining whether discomfort from insertion of the bronchoscope was associated with overall pain.

Models were developed in a two-stage process. In the first stage, factors that may influence the degree to which pain is experienced and reported by patients were selected based on our clinical understanding and findings from the medical literature (16). These factors were examined by bivariate analysis; factors that were associated with the outcome (p < 0.20) were retained for inclusion in multivariate models. Regardless of significance, age, sex, race, and medication doses used were entered into the multivariate models because of their clinical importance.

Multivariate analyses were performed using multiple logistic regression. Overall control of pain during bronchoscopy was the dependent variable (excellent pain control versus not excellent). We began with a model that included all of the variables from bivariate analysis that were significant and then removed all variables that were not significant at p  0.05. This process was repeated until only significant predictors of pain control remained in the model. All analyses were performed with SAS statistical software (17).

In some cases, several questions for patient-reported outcomes were conceptually related. For example, there were three questions about the quality of information and five questions about the quality of the physicians. These groups of questions were analyzed using Spearman's correlation coefficient (r_s). When responses to these questions were highly correlated (r_s  0.60) the questions were combined into a single scale in order to avoid collinearity in the multivariate models. This process resulted in development of scales of physician quality (16 point), nurse quality (12 point), information quality (12 point), waiting time (12 point), and the bronchoscopy environment (24 point). For the bivariate analysis, the scales were standardized by dividing the results for each scale into even quartiles.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population and Characteristics

During the study period of February 1, 1997 to June 30, 1998, 840 eligible patients were identified. A total of 584 patients (69.5%) completed the pre-FOB survey and 519 patients (61.8%) completed the post-FOB survey. We confined our analysis to the 481 patients (57.3%) who completed both questionnaires. Patients who completed both questionnaires were similar to those who did not with respect to age and sex. Nonwhites were less likely than whites (51% versus 64%, p < 0.03) to complete both forms. The study population included 50% men, 63% whites and 37% nonwhites (Table 1). Over half of the group reported their health status as poor or fair and nearly one-third required supplemental oxygen prior to the procedure. Patients reported overall control of pain as excellent (36%), very good (26%), good (27%), fair (8%), and poor (2%).

Sampling methods used to obtain clinical specimens and medication usage are reported in Table 2. Bronchoalveolar lavage (BAL) was the most commonly performed sampling method (73% of cases) and transbronchial biopsy was performed in 35%. There were wide variations in the amount of medications used. In cases in which fentanyl was used (95%), the mean dose given was 105 µg (range 12.5-250 µg). The mean dose of midazolam, in cases in which midazolam was used (93%), was 3.7 mg (range 0.5-10.0 mg).

Patient Characteristics Associated with Better Pain Control Ratings

Patient sex and race were not significantly associated with pain. Age and undergoing FOB for the first time were associated with greater pain, though these associations were not statistically significant. However, self-reported health status, education, and asthma were significantly associated with pain control (Table 1). Patients with excellent health status were more than twice as likely to report excellent pain control than those with poor or fair health status (65% versus 31%, p < 0.05). College-educated patients had a higher frequency of excellent control of pain than those who had not graduated from high school (45% versus 29%, p < 0.05). Patients with asthma were less likely to report excellent control of pain than those without asthma (24% versus 37%, p < 0.05).

Process of Care Factors Associated with Better Pain Control Ratings

Compared with other sampling methods, patients who had BAL performed were significantly more likely to report excellent pain control whereas those who had brushings of the proximal airways were significantly less likely to have excellent pain control (Table 2). Patients who had an FOB performed as a research procedure reported less pain than patients who had a clinical indication for it (47.1% excellent versus 34.2%, p < 0.003) (Table 2). Higher doses of fentanyl and midazolam were associated with better pain control (p < 0.03), though the association for midazolam was not statistically significant. There was no significant association between procedure duration and control of pain.

Other Outcomes Associated with Better Pain Control Ratings

Several patient-reported outcomes were associated with pain (Table 3). When patients were bothered by insertion of the bronchoscope, they were unlikely to report excellent control of pain (10% versus 47%, p < 0.001). Many patients were bothered by numbing of the nose (46%) and this was also associated with overall pain (31% versus 41%, p < 0.02). A majority of patients (69%) were bothered by the taste of lidocaine spray, but not surprisingly, this complaint was not significantly associated with pain. Favorable ratings of control of cough and ability to breath during FOB were associated with less pain.

Pain control was better in patients with less memory of the procedure, though this did not reach significance (remember nothing, 52% excellent versus remember every detail, 34% excellent, p < 0.08). Positive significant relationships were seen between higher ratings of physician quality, nursing quality, waiting time, the bronchoscopy environment, quality of information, and excellent control of pain (Figure 1). Of patients who rated nursing quality highest, 67% reported excellent pain control, whereas only 10% of patients reported excellent pain control if they rated nursing quality on the lower third of the scale (data not shown).

View larger version (25K): [in this window] [in a new window]   Figure 1.   Relationship of pain control to indicators of quality of care. The association between scales of physician quality, waiting time, bronchoscopy environment, and information about FOB and pain control. First quartile is 25% of patients who gave the lowest ratings on each scale. p < 0.01 for the association between quartile of each scale and patients reporting excellent pain control by 2.

Most symptoms that were new or worsened in the 24 h after FOB were not related to pain during the procedure (data not shown). New or worsened chest pain and swallowing pain after bronchoscopy, however, were significantly associated with worse overall pain control during FOB (p < 0.05).

Multivariate Analyses

In the model that examined patient and process of care factors, there were three factors that were significantly associated with pain: lower health status, less education, and a diagnosis of asthma (Table 4). There was a lower odds of excellent pain control in patients who considered themselves to have poor health status, compared with those in excellent health (OR 0.16: 95% CI, 0.06-0.40). The odds of excellent pain control were lower in patients with a high school education, compared with patients with college education (OR 0.58: 95% CI, 0.35- 0.95). Patients with asthma had significantly lower odds of excellent pain control compared with those without asthma (OR 0.35: 95% CI, 0.14-0.92). These results show that health status, education, and asthma are independent predictors of pain control after adjusting for other factors including medication use, age, sex, and race.

Based on this model, a 40-yr-old white male who is in poor health and has an eighth-grade education had a 20% chance of reporting excellent pain control, whereas a 40-yr-old white man with excellent health and a college degree had a 75% chance of reporting excellent pain control.

In the model that examined the association of other patient-reported outcomes with overall pain (Table 5), the risk of pain was related to discomfort of scope insertion, less amnesia, and lower rating of information. For example, a patient who had no memory of the procedure and was not bothered by scope insertion had an 88% chance of reporting excellent control of pain, whereas a patient who had some memory of the procedure and was bothered by scope insertion had a 61% chance of reporting excellent pain control. Patients who reported receiving better information about the procedure were more likely to rate pain control excellent.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This is the first prospective study of patients undergoing bronchoscopy to report patient and process of care factors associated with pain during the procedure. This study demonstrated that among patients undergoing FOB at a center in which premedication use is routine, only 36% reported pain control as excellent and 10% reported pain control as poor or fair. Patients at risk for greater pain included those with worse health status, less education, and a history of asthma. This study is unique in that we identified specific aspects of the procedure that were associated with pain, including insertion of the instrument and having more memory about the procedure. Patients who reported more pain had lower ratings of information received prior to the procedure.

There is a great deal of variation in the technique bronchoscopists use for sedation and anesthesia (18). In a survey of bronchoscopists in North America, intravenous sedation was used routinely by 50.7%, sometimes by 23.2%, and rarely by 23.8% (18). In our study intravenous sedation was almost always given, but the doses varied widely. Such wide variations in medication usage may be related to the lack of information available on the experience of patients undergoing FOB.

Our study demonstrated in bivariate analysis that higher doses of fentanyl were associated with better control of pain, though fentanyl was not significant in multivariate analysis. Midazolam dose was positively, but not significantly, associated with better pain control in bivariate analysis. However, amnesia for FOB was significantly associated with excellent pain control. These findings suggest that the absolute doses of medication given are not as important as achieving the desired effect. Because individual patients respond to and metabolize drugs differently, bronchoscopists should not adhere to firm dosing protocols, but should attempt to titrate medications to each individual. This may be accomplished best with short-acting intravenous medications.

Studies in other patient populations have identified factors that are related to patient-reported pain (16, 19). In a study of severely ill, hospitalized patients, several factors were found that influence patient reports of pain. Patients who were older, had more comorbid conditions, greater anxiety and depression, and lower income reported less satisfaction with pain control (16, 19). Thomas and coworkers (20) found in postoperative pain control that presence of anxiety, being female, and younger age were associated with more severe pain. Desbiens and coworkers (22) in a study of pain in elderly patients found, in the oldest hospitalized patients, that depressed mood, admission diagnosis, and level of activity prior to admission were predictive of reported pain. In our study of a single invasive procedure, we also found evidence that overall health status was significantly related to pain. There are several possible explanations for this finding. There may be differences in drug metabolism in patients with worse health. Sicker patients may have other comorbidities that contribute to pain, or there may be differences in the way less healthy patients are treated that were not measured in our study. In contrast to the other studies, we did not find significant associations with patient age, sex, and race. One possible explanation for this inconsistency is patient response bias. The response rate was lower in nonwhites than in whites. If the group that experienced more pain did not complete satisfaction questionnaires we may not have found an association that actually existed.

Lower education level may be associated with increased pain due to suboptimal communication with physicians and nurses. If clinicians use technical terms or jargon, explanations may be less well understood by people with less formal education. One study of health literacy demonstrated that 35% of patients admitted to an inner city hospital had inadequate health literacy (24). Poor health literacy leads to impaired communication with clinicians and is associated with adverse health outcomes including decreased life expectancy, more frequent hospitalization and inappropriate preparation for tests and procedures (25). A new finding specific to bronchoscopy is that patients with asthma were less likely to have optimal pain control than patients without asthma. There is evidence that patients with severe asthma have increased proliferation of substance-P-containing afferent nerves in their airways (29). It is plausible that differences in airway innervation could make patients with asthma more susceptible to irritation from bronchoscope trauma. Thus, patients with asthma may need more premedication than other patients to achieve optimal comfort during FOB. Our findings demonstrate that manipulation of the upper airways and proximal lower airways is an important cause of pain, as bronchoscope insertion, numbing of the nose, and brushing of proximal airways were all significantly associated with pain. The finding that insertion of the bronchoscope can be uncomfortable has been reported in previous studies (8, 30). In our practice, numbing of the nose often occurs before sedation is administered. Our findings suggest that patients might tolerate FOB better if intravenous medications are administered prior to anesthesia of the nares or if improvements were made in the technique of bronchoscope insertion. Because most of our patients had the bronchoscope inserted via the nose, we cannot comment on whether the oral route is less painful. Nevertheless, these findings suggest that avoiding trauma to the upper airway and using adequate local anesthesia may improve the patient's pain control.

We were interested to find that patient reports of better information about FOB were related to less pain. In a study of how precolonoscopy information influences pain and anxiety, Morgan and coworkers (21) reported decreased anxiety, shortened recovery, and improved behavioral indices of pain with a precolonoscopy information protocol. A clinical trial of postoperative patients demonstrated that improved patient instruction led to decreased ratings of pain and less use of analgesics (31). Patients who received better or more complete information may have had less anxiety and felt more prepared about what to expect during FOB. However, our present study does not allow us to determine whether better information reduced pain or whether, perhaps, patients who experienced less pain were more likely to judge the pre-FOB information favorably. We are currently planning an intervention trial to assess the impact of improved information on pain.

This study has several unique strengths. It is the first large, detailed, prospective study of the association of pain with patient and process of care factors in FOB. We evaluated specific aspects of the procedure and measured the use of medications. These results will have implications for prebronchoscopy preparation, sedation protocols, and anticipation of patient needs.

There are several limitations to this study. Because the patients interviewed were all from one academic medical center in the mid-Atlantic region, their experiences may not be characteristic of patients elsewhere. In particular, we performed nearly all procedures on patients who were white or African-American. These results may not be readily applic-able to patients of other races and ethnicities. All of the procedures were performed by physicians at one institution. Their techniques and experiences may differ from bronchoscopists elsewhere. Additionally, reducing pain does not guarantee maximal comfort. To do so would require minimizing other components of discomfort including anxiety, fear, and unpleasant but not painful experiences such as the bad taste of lidocaine throat spray.

Pain during FOB is a frequently reported symptom among patients in our study population, especially those with less education, lower health status, and asthma. Although there is room to improve care for all patients, those at increased risk may benefit from more attention from bronchoscopists prior to and while undergoing this procedure. This attention may include providing more complete, understandable information about the procedure, and through more liberal use of anesthetic, analgesic, and amnestic agents. Further investigation on improving bronchoscope insertion techniques is also needed. Although it may not be feasible for all patients to undergo FOB painlessly, we believe there are significant opportunities to improve control of pain in many patients.