The landscape of acute and chronic pain management has been evolving rapidly because of the increased importance placed by patients and government agencies on adequate analgesia, as well as mounting concerns about the efficacy and safety of commonly used pain medications such as opioids and nonsteroidal anti-inflammatory drugs.1 This, coupled with expanding social and legal acceptance of cannabinoids in the United States for medical indications as well as recreational use, has created a favorable environment for the integration of these compounds into evidence-based medical practice.2 Although there is growing evidence on the role of the endocannabinoid (EC) system in nociceptive processing, clinical outcomes have been mixed.3
Further research is needed to legitimize the use of cannabinoids as analgesics in the scheme of evidence-based medicine. Strikingly, patient views regarding the use of cannabinoid compounds for acute and chronic pain have not been meaningfully reported in the literature. As groundwork for the much needed clinical outcomes research, we believe it is imperative to assess patients' belief structures regarding the medical use of cannabinoids, as it is likely that patients have been exposed to this issue because of the ubiquitous presence of these compounds in the fabric of our culture.
This study sought to prospectively elucidate via an anonymous survey of patients presenting for elective surgery at a large metropolitan hospital both their perceptions of potential effectiveness and acceptance, if prescribed, of cannabinoid compounds for analgesic indications.
The study protocol was approved by the institutional review board of Icahn School of Medicine at Mount Sinai. An anonymous questionnaire (survey) was designed to collect data on patient demographics, presence of pain, pain severity, use of pain medication, history of illicit-drug use, tobacco use, and cannabis use. Using a 4-point Likert scale, all patients were asked to describe their beliefs about the potential effectiveness of marijuana for acute and chronic pain and their willingness to use cannabis for pain if prescribed by a physician. A Likert scale was used because it is the standard instrument for assessing attitudes about a particular topic.
Self-identified cannabis users were further asked to describe the route of administration, the amount and frequency of use, whether the cannabis was prescribed, its efficacy, and adverse effects. The survey form is displayed in Appendix 1 (Supplemental Digital Content 1, http://links.lww.com/AAP/A223).
The questionnaire was administered to patients in the preoperative registration area prior to surgery at Mount Sinai Hospital, New York, New York, from May 2016 until July 2016. All patients in the preoperative registration area were scheduled to undergo elective surgery. Types of surgeries included those typically performed at a major academic center main operating room complex including general, neurologic, orthopedic, and gynecologic, among others. The survey was not administered during each day of this 2-month interval; the survey days were subject to availability of the 2 medical students tasked with administering the questionnaire. The 2 medical students trained for the project advised subjects of the anonymous nature of the study, and institutional review board–approved information sheets were made available prior to survey administration. The medical students were asked to cease survey administration once 500 surveys were completed. Survey acceptance rate was 100% because there were no patients who refused to participate in the survey after being asked to do so. Patients were given ample time to complete the surveys in private.
Study data were managed with a secure REDCap (Research Electronic Data Capture) electronic data capture tool housed on a secure institutional server.
Descriptive information for the sample's study measures included both frequency distributions and, where appropriate, means and SDs. Normality of distributions for continuous variables was assessed with skewness and kurtosis measures. Once sample descriptive information was generated, a logistic regression model was used to assess the demographic and medical characteristics of marijuana users compared with nonusers. In the initial univariate model building, we adopted a somewhat liberal P value (P < 0.15) for potential predictor variables and included only those variables in the final multivariate model that met that initial criterion. Once included in the final model, only those predictors whose P < 0.05 were retained.
Because there were study questions about sample attitudes toward marijuana and, for marijuana users, questions about the effectiveness of marijuana in dealing with pain and adverse effects associated with its use, we examined the interrelationships among these questions using exploratory principal component analysis (PCA). In the PCA analyses, the principal components were obliquely rotated, which allowed them to be correlated with one another. If the data supported a 3-component solution, demographic, medical, and pain variables as predictors of each component cluster were assessed using a general linear model approach (analysis of variance). All analyses were conducted using version 9.4 of the Statistical Analysis System (SAS).
Details of PCA Analysis
The questionnaire included a total of 13 items asking about attitudes toward marijuana (6 questions), marijuana effectiveness (3 items), and marijuana adverse effects (4 items). It was expected that the interrelationships between and among these questions would reveal 3 correlated principal components. These components would include the 6 attitudinal questions (component I) as a group, a cluster of the 3 effectiveness items (component II), and a third component comprising the 4 adverse effects items (component III). However, although all patients responded to the attitude questions, only the marijuana users were asked about the effectiveness and adverse effects because these questions would not be relevant to nonusers. To test the expectation that there were 3 principal components, we first conducted a PCA of the 13 items for the marijuana users only. Next, a PCA of the 6 attitudinal items for the nonuser group was conducted for comparison purposes.
Five hundred one surveys were distributed, with all surveys collected. A summary of the patient demographics can be seen in Table 1. A minority of patients did not respond to all survey items; however, the items to which they did respond were included in the analysis.
There were an approximately equal number of men and women, although there were more white participants (52%), followed by Latino and African Americans. Approximately 62% of the participants were older than 45 years, and most had completed either high school or college. There was a considerable income spread, with more than 35% reporting annual incomes greater than $100,000 and 12% reporting incomes less than $25,000 per year.
With regard to the pain variables, approximately 46% of the patients reported having pain that exceeded everyday kinds of pain, and most reported that the pain in the last 24 hours was moderate. Twenty-four percent were taking prescription pain medications; however, only 10% reported using opioids for pain relief. A plurality (43%) reported that there was less than a 25% reduction in their current pain, and less than 6% reported their medications were 75% to 100% effective. A summary of the pain-related variables can be seen in Table 2.
Slightly more than a quarter of the sample (27.4%) reported being marijuana users, whereas a smaller percentage smoked tobacco (approximately 10%). The percentage of those reporting the use of different types of illicit drugs (eg, amphetamines, cocaine, heroin, phencyclidine, synthetic cannabinoids [K2/Spice], methamphetamines, and/or 3,4-methylenedioxymethamphetamine [Ecstasy]) within the last year was exceptionally small (≤1%) except for amphetamines (3%) and cocaine (2.2%). A summary of tobacco and drug use is shown in Table 1.
To determine the appropriateness of the PCA for both samples, the Kaiser-Meyer-Olkin (KMO) index was used.4 Values between 0.80 and 0.99 indicate that PCA is appropriate. For the marijuana user sample, the KMO index was 0.85. Appendix 2 (Supplemental Digital Content 2, http://links.lww.com/AAP/A224) summarizes the resulting obliquely rotated PCA. It can be seen that the 3-component structure appears to characterize the data adequately. The first component represents attitudes toward marijuana, the second component describes marijuana effectiveness, and the third component focuses on adverse effects. As expected, there was a positive correlation (r = 0.50) between marijuana attitudes (component I) and reports of its effectiveness in dealing with pain, sleep, and mood (component II). Increases in reported effectiveness were associated with increases in positive attitudes toward marijuana. There was a weak negative relationship (r = − 0.25) between effectiveness and adverse effects. Not surprisingly, increases in reports of negative adverse effects were associated with decreases in reported effectiveness. Finally, there was essentially no relationship between marijuana attitudes and reports of its adverse effects (r = −0.09).
In addition to the PCAs, the reliabilities of each of the scales representing each of the principal components were assessed using Cronbach α, an index that ranges between 0 and 1.0 (higher values indicate greater reliability). The reliability values of the marijuana attitudes scale were 0.91, 0.92 for the effectiveness scale, and 0.76 for the adverse effects scale. All α measures would indicate that the scales were sufficiently reliable.
As noted previously, a PCA analysis of marijuana attitudes for patients who are not marijuana users was also conducted. The KMO index for this sample was 0.89. As expected, all the marijuana attitude items clustered together as a single component (Appendix 3, Supplemental Digital Content 3, http://links.lww.com/AAP/A225). As was the case for marijuana users, increased agreement with any 1 attitudinal item was associated with increased agreement with the remaining items. The scale's reliability was 0.94.
Marijuana Attitudes and Perception of Effectiveness in All Patients
Eighty-one percent of patients expressed the belief that marijuana could be at least somewhat effective for pain after surgery, and 81.5% reported that they would use it if prescribed by a physician for pain after surgery or acute injury. Eighty-three percent of patients expressed the belief that marijuana could be at least somewhat effective for chronic pain, and 82% indicated that they would use it if prescribed by a physician for chronic pain. Eighty-nine percent of the cohort affirmed that marijuana should be legalized for medical use, whereas only 60% believed that it should be legalized for recreational use. Data regarding marijuana attitudes of the total cohort are depicted in Figure 1.
Predictors of Marijuana Attitudes in All Patients
Because the PCAs for marijuana attitudes for both marijuana users and nonusers yielded the same results, the information from both groups was used in the identification of possible demographic, medical, and pain relief predictors of attitudes. An analysis of variance was used for this purpose. The resulting analysis suggested that whether the patient was a marijuana user, the patient's race, and whether they experienced pain other than everyday kinds of pain were significant predictors. The overall model was statistically significant (F4,445 = 39.00, P < 0.01, R2 = 0.26). Patients reporting having pain other than everyday pain had significantly (F1,445 = 3.87, P = 0.05) more positive attitudes toward marijuana (mean, 2.98) compared with those not reporting any pain beyond everyday pain (mean, 2.83). Compared with patients of other races/ethnicities, whites reported significantly (F1,445 = 6.02, P = 0.01) more positive attitudes toward marijuana (mean, 3.01) compared with nonwhites (mean, 2.80). Not completely surprisingly, marijuana users were significantly (F1,445 = 135.84, P < 0.01) more positive (mean, 3.41) compared with nonusers (mean, 2.40). The significant findings for whites and marijuana users should be interpreted cautiously because there was a significant (F1,445 = 6.17, P = 0.01) interaction between white race and marijuana use. White marijuana users were significantly more positive (mean, 3.41) compared with white nonusers (mean, 2.62). The same pattern held for nonwhite respondents (meannonwhite users = 3.41 vs meannonwhite nonusers = 2.19). Both white and nonwhite marijuana users were almost equally positive toward marijuana (meanwhite users = 3.41 vs meannonwhite users = 3.41). However, when comparing nonusers with one another, race/ethnicity did make a difference. White nonusers were significantly more positive to marijuana (mean, 2.62) than nonwhites who were not users (mean, 2.19). Significant predictors of positive marijuana attitudes in the total cohort are summarized in Table 3.
Cannabis Use Detail
Of the marijuana users, the majority (77%) said it was for recreational purposes, and most said they smoked it rather than vaped or ate it. Most reported they took between 1 to 2 and several puffs when smoking. Of the 53% who reported rare use of marijuana, 83% said they last used it either 1 month to a year ago or more than 1 year ago. Details regarding marijuana users, effectiveness, and adverse effects are shown in Table 4.
Characteristics of Marijuana Users
A logistic regression model was used to determine the demographic, pain, and pain treatment relief variables that were associated with marijuana use. The overall model that was developed was highly significant (Wald χ24 = 30.4, P < 0.01, R2 = 0.11). Higher levels of average pain within the last 24 hours were associated with an increase in the odds of marijuana use (Wald χ21 = 4.90, P = 0.03; odds ratio [OR], 1.10; 95% confidence interval [CI], 1.01–1.19). Increased age was associated with a decrease in the odds of reported marijuana use (Wald χ21 = 15.36, P < 0.01; OR, 0.66; 95% CI, 0.54–0.82), whereas reported tobacco use was significantly associated with an increase in the odds of marijuana use (Wald χ21 = 8.32, P < 0.01; OR, 3.22; 95% CI, 1.45–7.11). Finally, decreased levels of relief from the pain treatment or medications provided within the last 24 hours were associated with a decrease in the odds of marijuana use (Wald χ21 = 3.27, P = 0.07; OR, 0.76; 95% CI, 0.57–1.02). A summary of patient factors predictive of marijuana use, effectiveness, and adverse effects is shown in Table 3.
Marijuana Effectiveness in Marijuana Users
Approximately half of marijuana users believed that it is very effective in decreasing pain and improving both sleep and mood. There was only a single variable that was predictive of marijuana effectiveness for marijuana users—average pain in the last 24 hours. As pain in the last 24 hours increased, there was a significant (F1,112 = 15.99, P < 0.01, R2 = 0.13) increase in reported marijuana effectiveness. A summary of marijuana effectiveness is shown in Table 4, with factors predictive of effectiveness found in Table 3.
Marijuana Adverse Effects in Marijuana Users
Most marijuana users did not report any adverse effects. We queried patients regarding 4 common adverse effects that include anxiety, dizziness, dry mouth, and paranoia. When reported, adverse effects were generally classified as mild, with dry mouth reported as the most frequent untoward effect. As was the case for marijuana effectiveness, there was only a single variable that predicted marijuana adverse effects—patient income. Higher patient income levels were significantly associated (F1,111 = 4.80, P = 0.03, R2 = 0.04) with an increase in reported marijuana adverse effects. A summary of marijuana adverse effects is shown in Table 4, with factors predictive of adverse effects shown in Table 3.
Despite a general willingness by health care providers to consider the use of cannabinoids in clinical care, significant knowledge gaps and lack of comfort, both clinically and medicolegally, among practitioners with recommending cannabinoid compounds likely hinder both research and clinical practice.5–7 In addition, the perception of a large cohort of patients regarding the potential efficacy of cannabinoids and willingness to take such compounds if prescribed for the treatment of postoperative and chronic pain have not been previously described. Smaller surveys have been done on the prevalence of cannabis use and analgesic effectiveness in patient populations with underlying pain conditions, but these are less representative of the broader demographic that our study aimed to describe.8–10 Overall, we have found that more than 80% of patients presenting for elective surgery believe that marijuana may be at least somewhat useful for the treatment of pain and are willing to take cannabinoid compounds for analgesia, if prescribed by a physician.
Various forms of cannabis have been used for thousands of years to treat a variety of conditions, and modern research efforts have recently identified the endogenous system involved in mediating its actions. The EC system is composed of 2 major receptors, CB1 and CB2, with the former found at presynaptic sites in the peripheral and central nervous systems and the latter mostly on immune cells. The signaling via these receptor subtypes has been implicated in normal nociceptive processing, acute pain, and chronic pain states in both clinical and preclinical models.3 Ongoing research of this system has yielded the concept of the “endocannabinoidome,” a complex system of receptors, ligands, EC-like mediators, and specific metabolic enzymes, with each presenting potential areas for pharmacological targets.11
A recent systematic review and meta-analysis found that there was moderate quality evidence to support the use of cannabinoids for the treatment of chronic pain and spasticity, with other recent systematic reviews supporting this assertion.12 In addition to pain, exogenous cannabis compounds and formulations, such as ajulemic acid, cannabidiol, cannabis, dronabinol, levonantradol, nabilone, nabiximols, cannibadiol, ECP002A, tetrahydrocannabinol, and tetrahydrocannabinol/cannabidiol combinations, have been studied for the treatment of multiple conditions such as psychosis, anxiety, depression, nausea and vomiting, spasticity, and insomnia.12
Marijuana Attitudes and Perception of Effectiveness in All Patients
Although overall the majority of patients in this study believed that marijuana could be effective for treating pain and would use it if prescribed by a physician, a more detailed look at the data in Table 5 reveals some ambiguity in patients' views, as approximately 1 in 4 patients only “somewhat agreed” with these premises. In addition, approximately 1 in 5 patients did not believe that marijuana could be an effective analgesic and may not be comfortable with taking cannabinoid compounds for pain relief, even if prescribed by a physician. This, coupled with the fact that the majority (72.3%) of the patients in this cohort have never used marijuana, signals the need for careful patient selection, consent, and education in both trials and clinical practice because a significant number of patients may have reservations regarding the utility of cannabinoids in medical practice.
The incidence of marijuana use in our study is similar to that reported in previous surveys.8 Age, sex, level of education, and income did not seem to effect patient beliefs regarding potential effectiveness or acceptance of marijuana. This signifies that cannabinoids may be accepted as analgesics by patients across a broad range of demographic and socioeconomic backgrounds.
Patients with a history of significant pain and those who reported marijuana use had more positive views toward marijuana. This is likely due to the fact that patients with ongoing pain have failed other analgesic therapies and are more open to exploring alternative options with a desire for these options to be effective. It is also encouraging to see that patients with a history of marijuana use have had an overall positive experience. This is likely due to the overall effectiveness and low rates of adverse effects reported by marijuana users in this study. Acceptance was more pronounced in white patients without a history of marijuana use (vs nonwhite patients without a history of marijuana use). Although this difference is statistically significant, it is unclear if the presence of racial disparity, which has been documented in the perceived effectiveness of other treatments, in acceptance of marijuana effectiveness for pain, should be a factor to consider when presenting cannabinoids as an analgesic option to patients without history of marijuana use.13–15
Eighty-nine percent of the cohort responded affirmatively to the idea of having marijuana being legalized for medical use, only 60% thought it should be legalized for recreational use. These numbers are strikingly similar to a recent national Quinnipiac University Poll and may support the generalizability of our overall findings to larger patient populations.16 The findings signal that a significant proportion of the population believes that marijuana is a medication and therefore should be taken only under physician supervision. Overall, the prevalence of illicit-drug use was lower than previously reported in perioperative patient populations. This may have been related to our cohort being older because increasing age is known to be correlated with a decrease in illicit-drug use.17
Predictors of Marijuana Use
Younger patients who use tobacco with higher levels of pain in the last 24 hours were more likely to be marijuana users in our cohort. Preexisting pain has been previously found to be significantly associated with marijuana use.18 As younger age and preexisting pain are well-known risk factors for difficult-to-control pain, these patients were likely searching for alternative methods of analgesia after finding typical methods of pain control to be inadequate.19 The inadequacy of existing analgesic strategies for many patients is supported by our findings, which revealed that 43% of the overall cohort reported less than a 25% reduction in their pain with medications, and less than 6% reported their medications were 75% to 100% effective. Inadequacy of analgesia with use of typically available modalities is an unfortunately common occurrence in clinical pain practice and a strong impetus for the introduction of novel analgesic agents into the clinical arena.20–22
Cannabis Use Detail
Of the marijuana users, the majority (77%) said it was for recreational purposes mostly by smoking. In other surveys, patients were more likely to identify their marijuana use as medicinal (vs recreational); however, our cohort was less restrictive than the previously studied pain clinic population perhaps because of the anonymity of our study and the changing landscape today regarding the use of marijuana in society.8 Among patients who admitted to marijuana use, most found it to be effective in decreasing pain and improving both sleep and mood. These findings are consistent with a recently published study of patients using cannabis for therapeutic purposes.23 Adverse effects of marijuana use were generally mild and corresponded to those described in similar reports.8 The only variable significantly associated with an increase in reported adverse effects was higher income. The relationship between patient income and clinical outcomes likewise has been supported by other studies.24–26 Overall, it seems that marijuana smokers in this cohort report a generally favorable effect on their symptoms and experienced mostly no or mild adverse effects.
There were several limitations of our study. First, sample selection may have been biased because patients were not selected at random, but only patients approached in the preoperative registration area by 2 medical students during certain days within a specific 2-month interval were surveyed. Because of the large sample size, however, and the stability of the surgical case volume and mix, it is likely that the sample is representative of our center's overall ambulatory surgical population. Our cohort consisted entirely of preoperative patients and therefore may not be representative of the general population. This was not a major concern of the present study because we purposefully sought to categorize this specific population as groundwork for planned future clinical trials of cannabinoids during the perioperative period. However, the overall demographics of the group, as well as the close correspondence of the sample's opinions to existing data on legalization, suggest that our findings in fact may be relevant to broader patient populations. The survey queried personal information regarding illicit-drug use, which may have resulted in underreporting of this dimension. Great care was taken to explain to the patients that the surveys were completely anonymous, as well as to demonstrate anonymity by the method of survey distribution and collection processes. Ideally, a postoperative survey would have been administered to track patient views after experiencing postoperative pain. Unfortunately, the completely deidentified nature of this survey made correlation of the preoperative and postoperative cohorts prohibitive. Overall, the exceptionally high participation rate signals patients' confidence regarding the anonymity of the survey. It should be noted that for both the linear and logistic regression models the R2's are quite modest. This result suggests that there are unmeasured explanatory variables that could have increased the predictive value of the models considered in this study and that their identification is a task for future research.
In conclusion, the results of this novel survey reveal that most patients believe that marijuana may be potentially effective for the management of pain and exhibit a willingness to have cannabinoid compounds incorporated into their care, thus underlining the need for further research to responsibly allow such integration. It is our hope that the acceptance by patients of cannabinoids for the management of pain will empower the medical community to engage in the considerable effort required to properly define the appropriate roles, efficacy, and safety of these compounds in legitimate medical practice.