Abstract
Background:
Respiratory distress protocols (RDPs) are protocolized prescriptions comprised of 3 medications (a benzodiazepine, an opioid, and an anticholinergic) administered simultaneously as an emergency treatment for respiratory distress in palliative care patients in the province of Quebec, Canada. However, data on appropriate use that justifies the combination of all 3 components is scarce and based on individual pharmacodynamic properties along with expert consensus.
Objectives:
Our study aimed to evaluate the conformity and the effectiveness of RDPs prescribed and administered to hospitalized adult patients.
Methods:
This was a prospective and descriptive study conducted in a single center. Prescription and administration conformity were assessed based on predefined appropriateness criteria.
Results:
A total of 467 adult patients were prescribed a RDP, 175 administrations were documented, and 78 patients received at least 1 RDP. Prescription conformity was assessed on 1473 separate occasions over the trial period. Overall prescription conformity was found to be 37% (95% confidence interval [CI]: 33.6-40.4), and administration conformity was 37.7% (95% CI: 26.2-50.7). Low administration conformity was primarily explained by incorrect indications for RDP use. Seemingly important determinants of higher conformity were prescriber’s speciality in palliative care, use of preprinted orders, pharmacist involvement, and hospitalization in the palliative care unit.
Conclusion:
This study highlights important gaps in the use of RDPs in our institution. Health-care provider training appears necessary in order to ensure adequate conformity and allow for further evaluation of RDP effectiveness.
Keywords
Introduction
Dyspnea is encountered frequently in palliative care patients. Prevalence varies between 10% and 70%, 1 -12 most likely due to subjective evaluation criteria and lack of a universal definition. This number may increase to 90% in patients affected by a terminal lung illness. 6,9 In the palliative care setting, respiratory distress can be defined as a respiratory emergency during which the magnitude of dyspnea reaches a point where it subjugates all aspects of the patient’s life (physiological, social, cultural, environmental, emotional, and cognitive). 6,13 Furthermore, dying of suffocation represents the most important fear expressed by patients who are nearing the end of life. 1,13 Comprised of a benzodiazepine, an opioid, and an anticholinergic, respiratory distress protocols (RDPs) are administered subcutaneously or intravenously as an emergency treatment of respiratory distress. All 3 medications are dosed by the physician according to patient’s weight and current medication (including opioids, benzodiazepines, and anticholinergics). They are administered simultaneously in order to quickly alleviate associated symptoms and provide transient sedation. Relevant pharmacodynamic properties include reduction in anxiety and retrograde amnesia (benzodiazepines), reduction in respiratory drive and alteration in dyspnea perception (opioids), and decrease in respiratory secretions and bronchospasms (anticholinergics), which are frequently encountered in terminally ill patients. 6,13 -15 Furthermore, all 3 medications contribute to rapid sedation. Respiratory distress protocols are prescribed on an “as-needed” basis and can be administered following nursing staff or practitioner assessment. They are repeatable every 10 to 30 minutes, depending on symptom relief. Despite a paucity of data evaluating the use, safety, and efficacy of RDPs, they remain a standard of practice in palliative care in the province of Quebec, Canada. To the best of our knowledge, only 2 published retrospective studies have evaluated RDPs in palliative care in the province of Quebec. 16,17 One study suggested that RDPs provided appropriate sedation within 30 minutes following administration in all patients, without having any notable impact on survival. However, 30% to 35% of the patients were unconscious at the time of the administration. 16 A second study demonstrated that RDP prescription conformity varied between 66% and 82%. 17 In our hospital (CIUSSSE-CHUS), a preprinted order form was instituted as a tool to optimize RDP prescriptions, and a training video was developed in an attempt to standardize the administration of RDPs, but an assessment of the use of RDPs has not since been completed. This study aimed to evaluate the conformity of RDP prescriptions and their administration. An exploratory analysis evaluating effectiveness was also conducted.
Methods
Study Design and Objectives
We conducted a prospective, descriptive, cohort study in a single tertiary teaching hospital between March 1 and September 8, 2017. The study was approved by both the scientific and the ethics committees of the Centre hospitalier universitaire de Sherbrooke-CHUS.
The primary objectives were to describe the conformity of prescriptions and administrations of RDPs in our institution. First, overall prescription conformity was defined as compliance to all prescription-related criteria, that is, (1) choice of agent and dosage, (2) route of administration, (3) permitted dosing interval between administrations, and (4) number of repetitions (detailed criteria can be found in the Supplemental Appendix 1). Second, overall administration conformity was determined by 2 criteria, considering both an appropriate indication of respiratory distress and the administration of all 3 prescribed pharmaceutical agents (detailed criteria can be found in the Supplemental Appendix 2). For the purpose of this study, a working definition of respiratory distress was derived from expert opinion and best available evidence in order to assess the appropriateness of indication. Respiratory distress was defined as the combination of at least 1 respiratory symptom (dyspnea, respiratory discomfort, intercostal retractions, tachypnea, and increased effort required to speak) and 1 psychological symptom (anxiety, panic, agitation, feared expression, desperation for air, and impression of sudden suffocation) occurring in a conscious patient. An episode of respiratory distress was defined as a period of time during which 1 or multiple RDPs were administered less than 60 minutes apart.
Secondary objectives include the conformity to each individual criterion for RDP prescriptions and administrations specified earlier. Respiratory distress protocol prescription conformity at the time of administration and description of prescription conformity over time were also assessed. Symptoms documented by nursing staff preceding RDP administrations were collected. Finally, RDP effectiveness and determinants of conformity were evaluated as exploratory analyses.
Patients
Patients were included if they were prescribed an RDP, were older than 18 years of age, and hospitalized at our institution during the trial period. Patients receiving methadone were excluded because equianalgesic opioid conversion is required to assess RDP opioid dosage conformity and is not unanimous among prescribers.
For the purpose of this study, 2 cohorts were defined. Cohort 1 was composed of all study participants, while cohort 2 included only the patients from cohort 1 for who an RDP was administered. Upon administration, patients were censored from cohort 1 and recruited into cohort 2. Other reasons for censoring included hospital discharge, death, discontinuation of the RDP, and end of the study period.
Data Collection and Assessments
During the trial period, investigators identified eligible patients on a daily basis by selecting those with an active RDP prescription using lists generated by a pharmacy software. Once included, medical charts and electronic medical records were reviewed daily throughout hospital stay until patients were censored. Baseline characteristics (age, gender, allergies, weight, height, creatinine, primary diagnosis, and care unit), prescription characteristics (date and time, use of preprinted order form, prescriber’s specialty, agents and dosage, medication taken on a regular basis, and adjustment of the RDP by a pharmacist), and administration characteristics (date and time, symptoms before and after, administered agents, and effectiveness as collected by nursing staff) were collected. To ensure standardization of data collection, the first 10 patients were assessed independently by all investigators. Interindividual variability was analyzed and resolved before the study began. Since data were collected by 4 different investigators over the trial period, 10% of the data were subjected to double entry to control for potential errors.
For patients in cohort 1, prescription conformity was evaluated in 3 predefined situations: (1) new RDP prescriptions, including changes in existing RDPs; (2) changes in the patient’s regular opioids, benzodiazepines, or anticholinergics (outside those included in RDPs); and (3) significant weight fluctuations. For patients in cohort 2, prescription conformity was assessed at the time of RDP administration. For administration conformity, standardized evaluation tools were not available; therefore, conformity criteria were elaborated by investigators and revised by local experts in the field of palliative care prior to the study in an attempt to control for information bias.
Administration conformity and effectiveness were assessed following RDP administration. For administration conformity, nurses were asked to assess consciousness and collect symptoms present before and after RDP administration using a questionnaire specifically designed for this purpose and revised by 7 palliative care specialists prior to use. A multitude of symptoms potentially justifying RDP administration, both respiratory and psychological, were listed, and nurses were asked to identify those leading to RDP use. Using the same list, reassessment of symptoms was also required 15 to 20 minutes after RDP administration. In order to account for recall bias, questionnaires completed more than 48 hours following the administration of an RDP were excluded and considered as missing data. Once completed, investigators collected questionnaires and evaluated conformity. For effectiveness, nurses were also asked to indicate whether or not they perceived RDPs to be effective for symptom relief (yes or no). Effectiveness as determined by investigators was evaluated for conform administrations only and was based on 2 objective criteria: relief of all respiratory or all psychological symptoms 15 to 20 minutes after RDP administration and the absence of RDP readministration in the hour following prior administration (details are available in Supplemental Appendix 2).
Statistical Analysis
Baseline characteristics were described using means and standard deviations. Prescription conformity, administration conformity, and effectiveness were dichotomous variables (ie, conform/effective or no) and assessed by calculating proportions and associated 95% confidence intervals (CIs). Mean prescription conformity time was determined by dividing the total number of days during which an RDP prescription was deemed conform by the total number of patient follow-up days, multiplied by 100. Backward Generalized Estimating Equation (GEE) multivariate logistic regression was used to identify potential determinants of conformity. Determinants included in the RDP prescription conformity model were age, gender, creatinine clearance (Cockroft-Gault equation), weight, height, primary diagnostic, practitioner specialty, preprinted order utilization, and pharmacist involvement (adjustment of the RDP prescription). Age, nursing staff work shift (day, evening, and night), and the care unit in which the RDP was administered were those included in the administration conformity model. All statistical analyses were performed using SAS software v9.3.
Results
Participant Characteristics
A total of 467 patients were screened between March 1 and September 8, 2017. All patients (467) screened were included in cohort 1 and were followed for a mean duration of 8.3 ± 13.5 days. One patient was censored during the study after being prescribed methadone. Cancer was the primary diagnosis at admission in 49.7% of cases. Investigators completed 1473 RDP prescription conformity evaluations, including 672 new RDP prescriptions, and 801 reevaluations over time. Situations justifying reevaluation, which could occur simultaneously, were opioid changes (n = 642), benzodiazepine changes (n = 183), anticholinergic changes (n = 58), and weight changes (n = 4). Respiratory distress protocols prescribed by palliative care specialists accounted for 50.6% of new prescriptions. Cohort 2 was comprised of 78 patients, 24.4% of which were previously diagnosed with cancer, totaling 175 RDP administrations. Details on patient censoring can be found in Figure 1. Following administration, median survival time was 4 hours, with 88% of patients dying within 24 hours. Only 2 patients survived and were discharged from hospital after RDP administration. Baseline characteristics of individual patients are listed in Table 1.
Patient flowchart.
Baseline Characteristics.
Abbreviations: RDP, respiratory distress protocol; SD, standard deviation.
aCockcroft-Gault Equation = [(140 − age (years)) × weight (kg)]/ (49 × serum creatinine (µmol/mL)) [ × 0.85 if the subject is female]
Respiratory Distress Protocol Prescriptions
Overall prescription conformity was found to be 37% (95% CI: 33.7-40.4) across all 1473 evaluations. Conformity was higher (45.7% [95% CI: 41.9-49.6]) for new RDP prescriptions yet lower at the time of administration (31% [95% CI: 21.9-42.0]). For palliative care specialists, new RDP prescription conformity was higher when compared to other specialties (47.4% [95% CI: 41.7-53] vs 44.2% [95% CI: 39-49.5]). However, prescription conformity at the time of administration was lower (23.6% [95% CI: 8.9-49.4] vs 32.7% [95% CI: 22.3-45]). Conformity for each individual criterion can be found in Table 2. Rates were superior to 90% for most criterion except for benzodiazepine and opioid doses where conformity was 56.6% (95% CI: 52.8-60.4) and 68.7% (95% CI: 65.6-72), respectively. Among benzodiazepines doses that were not conform, dosage was low with respect to hospital protocol in more than 90% of cases. Moreover, rates of benzodiazepine conformity tend to decrease as total daily benzodiazepine dose increases (74.1% if benzodiazepine-naive versus 18.8% if receiving more than 4 mg of equivalent lorazepam per day). Inadequate RDP opioid dosing was equally distributed regarding supra- or subtherapeutic doses with little variability in conformity for patients receiving regular daily treatment (61.7% for opioid-naive patients vs 83.3% for those with a regular dosing schedule). For patients included in cohort 1, mean prescription conformity time was 52.2%.
Prescription Conformity.
Abbreviation: CI, confidence interval.
aLorazepam or midazolam.
bPatients not receiving benzodiazepines on a regular basis and weighing under 70 kg: lorazepam 2 mg or midazolam 5 mg; patients receiving once daily dosing benzodiazepines or weighing over 70 kg: lorazepam 4 mg or midazolam 10 mg; patients receiving twice or more daily dosing benzodiazepines or with known benzodiazepine tolerance: lorazepam 6 mg or midazolam 15 mg.
cExcluding any “if-needed” prescriptions (PRNs).
dMorphine or hydromorphone.
ePatients receiving 0-3 mg subcutaneous morphine or 0-1 mg subcutaneous hydromorphone every 4 hours on a regular basis: 5 mg subcutaneous morphine or 1 mg subcutaneous hydromorphone; patients receiving > 3 mg subcutaneous morphine or > 1 mg subcutaneous hydromorphone every 4 hours on a regular basis: 1.5× multiple of subcutaneous dose received every 4 hours.
fScopolamine.
gPatients not receiving regularly daily doses of an anticholinergic (scopolamine or glycopyrrolate): 0.4-0.6 mg scopolamine; patients receiving regularly daily doses of an anticholinergic (scopolamine or glycopyrrolate): 0.6-0.8 mg scopolamine.
hSubcutaneous or intravenous.
iBetween 10-30 minutes.
j1-2 (maximum of 3 administrations total).
Respiratory Distress Protocol Administrations
Of the 175 RDP administrations for patients in cohort 2, eighty-six were excluded from the analysis due to incomplete data collection. The 89 remaining administrations were determined to be conform in 37.7% of cases (95% CI: 26.3-50.7). Results were driven primarily by incorrect indications for RDP use. Indication of use was deemed inappropriate due to unconsciousness in 55 (93.2%) of cases, lack of a psychological symptom in 38 (64.4%), and lack of a respiratory symptom in 14 (23.7%). All 3 pharmaceutical components were administered 88.2% of the time (95% CI: 79.3-93.6). Important variability in administration conformity was apparent between different care units. Overall conformity was numerically higher for patients admitted to the palliative care unit (62.9% [95% CI: 35.6-83.9]) when compared to the intensive care unit (13.9% [95% CI: 4.6-35.1]) and all other units combined (42.3% [95% CI: 25.9-60.7]) as presented in Table 3. Symptoms leading to RDP administration, as collected by nursing staff, can be found in Figure 2.
Administration Conformity.
Abbreviation: CI, confidence interval.
aPresence of at least one psychological and one respiratory symptom in a conscious patient.
bAdministration of all 3 prescribed pharmaceutical agents (benzodiazepine, opioid, and anticholinergic).
Signs and symptoms leading to respiratory distress protocols (RDPs) administration.
Effectiveness
Effectiveness evaluation by nursing staff was recorded for 69 (39%) administrations. Of the available data, RDPs were found to be effective in 66.7% of cases (95% CI: 55.2-76.5). Among conform RDP administrations (n = 89), 62 (70%) could not be evaluated for the following reasons: patient unconsciousness preceding RDP administration (61.3%), missing data (25.8%), and death before evaluation following RDP administration (12.9%). Available data (n = 27) indicated that 7.4% (95% CI: 1.9-26.7) of cases achieved complete symptom relief. Results can be found in Table 4.
RDP Effectiveness Following Administration.
Abbreviations: CI, confidence interval; RDP, respiratory distress protocol.
aDocumented by nursing staff following RDP administration (yes or no) then collected and analyzed by investigators.
bRelief of all psychological or all respiratory symptoms present before administration AND absence of readministration of a RDP in the 60 minutes following prior administration.
Determinants of Conformity
Increased age, lower body weight, prescriber specialized in palliative care, prescription adjustment by a pharmacist, use of preprinted orders, and evaluation for new RDP prescriptions were found to be statistically significant determinants for higher prescription conformity. Changes in daily benzodiazepine doses were shown to negatively affect prescription conformity. For patients included in cohort 2, those hospitalized in the palliative care unit were found to have a higher likelihood of conform RDP administration. Detailed results are available in Table 5.
Backward Generalized Estimating Equation (GEE) Logistic Regression.
Abbreviation: CI, confidence interval.
Discussion
Overall, this study demonstrates that RDP prescription and administration conformity in our institution are remarkably poor (< 40%), highlighting a lack of standardized practice in the care of respiratory distress. Although limited by missing data, effectiveness was also found to be low, with notable variation between objective evaluation by investigators and nurse perception.
Results for overall RDP prescription conformity were driven predominantly by suboptimal dosing for benzodiazepines and opioids. Conformity was higher for new RDP prescriptions when compared to overall results and those at the time of RDP administration. This may indicate a more in-depth evaluation of patient characteristics initially, followed by a neglect to reevaluate. Repeated evaluations allowed to account for changes in patient-specific factors during hospital stay, which is relevant considering prescription conformity may vary over time, and patients are subject to develop respiratory distress at any moment during hospital stay. Mean conformity time of RDPs for all patients included in this study was only 52.2%, suggesting that patients with respiratory distress would receive an optimal RDP in approximately half of cases, which could in turn lead to suboptimal symptom relief and potentially unnecessary readministration.
Altogether, prescription conformity seems inferior in comparison to previous studies, but results are difficult to compare due to important differences in methodology. 16,17 Benzodiazepine dosage had the lowest conformity of all individual RDP prescription–related criteria, primarily due to suboptimal doses. Although difficult to assess, this may be a result of suggested benzodiazepine doses being too high. Interestingly, utilization of 15 mg of midazolam in RDPs was rare (3.7% of orders), despite being recommended for patients with frequent benzodiazepine dosing (more than twice daily) or known benzodiazepine tolerance. 6,13 However, studies indicate that tolerance is a result of numerous factors, including exposure time and dose and varies depending on the molecule and the pharmacologic property of interest, making it difficult to evaluate. 18,19 Furthermore, clinicians may not necessarily rely on frequency of administration of benzodiazepines when prescribing midazolam in RDPs, as frequent dosing does not necessarily correlate with higher total daily doses. In fact, mean daily doses of benzodiazepines (in equivalent lorazepam) were 7.56 mg in patients having a prescribed midazolam dose of 15 mg in their RDPs compared to 2.78 mg for patients with midazolam 10 mg. In order to account for this gray area, an adjusted subset analysis was performed to consider 10 mg as an appropriate dose for all patients in who 15 mg should have been prescribed, showing an increase in relative benzodiazepine conformity by 12.5%.
Opioid dosage had the second lowest conformity rate for prescription-related criteria. However, it is interesting to note that conformity increased to 83.3% in patients receiving opioids on a regular basis. This may be because practitioners pay greater attention to adjusting the opioid in RDPs to account for opioid tolerance in patients receiving regular doses.
As opposed to what may have been expected, new RDP prescription conformity was only slightly higher for palliative care prescribers when compared to other specialties. Expertise in palliative care most certainly plays a role in optimizing patient care, however, may also favor deviation from preprinted order forms.
A total of 17% of patients in this study received at least 1 RDP. Administration conformity was low, driven by an inappropriate indication for use, primarily due to unconsciousness (>90%). These results were consistent with a previously published study evaluating RDPs. 16 However, in general practice, evaluation of consciousness is not performed using a validated scale during respiratory distress, which may create interevaluator variability. Administration conformity in the palliative care unit was superior to other care units, showcasing the positive impact that expertise can have on patient-centered care. Conformity was lower in the intensive care unit, as RDPs were often used repeatedly following extubation of terminal and already unconscious patients, which did not correspond to the study definition of respiratory distress.
In this study, few observations can be derived from effectiveness analysis. Nursing staff generally perceived RDPs to be effective following administration (66.7%), but complete symptom relief appears to be rare. Differences in the definition of respiratory distress between nursing staff and investigators may explain this gap. Nursing experience with end-of-life care may also play a role due to discrepancies in symptom recognition. In 63% of respiratory distress episodes, only 1 RDP was administered despite the fact that only 7.4% of administrations were determined to be effective by investigator criteria which are defined by complete relief of all respiratory or all psychological symptoms. This suggests that the effectiveness observed by nurse staff might be driven by the decrease in the intensity of symptoms rather than complete relief of symptoms. However, it cannot be confirmed because the intensity of symptoms before and after RPD administration was not collected. In addition, the relief of all symptoms may not be representative of the clinical relief of a respiratory distress episode. Although comparative data are lacking, results from a similar study demonstrate higher rates of effectiveness, possibly due to differences in methodology, evaluation parameters, and patient population. 16
Specialization in palliative care, pharmacist involvement, and admission to palliative care unit all had a positive influence on conformity. Appropriate training with regard to recognizing and managing respiratory distress as well as understanding the importance of optimizing the pharmacologic aspect of treatment also plays a key role. Use of preprinted orders, accounting for 54.6% of RDP prescriptions, also seems to play a predominant role in prescription conformity by standardizing prescriber practice. However, this result is influenced by the fact that prescription conformity criteria were derived from the preprinted order form.
To the best of our knowledge, this is the first prospective study evaluating conformity of prescription and administration of RDPs in a clinical practice setting. Multiple factors capable of influencing conformity were considered in order to provide as much information as possible to a growing body of literature. Prospective design was meant to reduce missing outcome data. However, close to 50% of nursing staff questionnaires were not completed following RDP administration. Results may be limited by information bias owing to the lack of a universally approved definition of respiratory distress and the absence of validated evaluation tools. Subjective evaluation by nursing staff before and after RDP administration also introduces significant variability for the administration conformity results. However, data collection and evaluation of conformity were completed with tools that were developed and approved by local experts in the field of palliative care. Study objectives were also known by several prescribers and certain members of nursing staff, potentially introducing desirability bias. Patient characteristics were diverse and representative of most palliative care populations, most likely due to lenient exclusion criteria that contribute to improving external validity. However, RDPs seem to be a concept applied uniquely in the province of Quebec, which limits generalizability.
Conclusion
Rates of conformity found in this study highlight important gaps in the evaluation of respiratory distress and the use of RDPs in our institution. Updated recommendations in the local protocol may be beneficial in order to adapt to day-to-day clinical practice. Appropriate, in-depth training for all health-care professionals implicated in respiratory distress management is also an essential step in ensuring adequate conformity and proper evaluation of RDP efficacy, which would hopefully lead to more convincing data supporting the use of RDPs on a larger scale.
Supplemental Material
Supplementary_material_RELIEVE_revised_20181219_xyz13441d6886433 - Conformity in Prescription and Administration of Respiratory Distress Protocols in a Tertiary Care Hospital in the Province of Quebec: RELIEVE Study
Supplementary_material_RELIEVE_revised_20181219_xyz13441d6886433 for Conformity in Prescription and Administration of Respiratory Distress Protocols in a Tertiary Care Hospital in the Province of Quebec: RELIEVE Study by Camille Dufort-Rouleau, Benjamin Martin, Vincent Barré, Véronique Bédard, Émilie Dufort Rouleau, Marie-France Beauchesne, Julie Quenneville and Mathieu Berteau in Journal of Palliative Care
Footnotes
Authors’ Note
All authors contributed equally to the publication process.
Acknowledgments
The authors would like to thank the nursing staff and the administrators of the participating units and Brigitte Perreault’s foundation.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was made possible through the support of Brigitte Perreault’s foundation who help improve the healthcare quality at Centre Hospitalier Universitaire de Sherbrooke and from the Université de Montréal for the help in statistical services. MFB has received research grants and professional fees from AstraZeneca, Boehringer Ingelheim and Novartis.
Supplemental Material
Supplemental material for this article is available online
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
