Screening and managing cannabis use: comparing GP’s and nurses’ knowledge, beliefs, and behavior

The study sample was comprised of delegates attending a Healthed seminar in one of five Australian cities (Sydney, Melbourne, Adelaide, Brisbane, and Perth) in 2011. Each Healthed seminar featured approximately sixteen lectures on women and children’s health and represented a rare opportunity to meet with several hundred GPs and nurses in one location. GPs and nurses received professional education points for attending the lectures. Of the 1,925 delegates who attended the Healthed seminars, 664 (35%) completed the study survey; 503 surveys were completed by GPs (76%) and 162 were completed by nurses (24%). This distribution was representative of the Healthed population, in which 71% were GPs and 26% were nurses. The study sample was predominately female (n = 527, 79%), which again reflects the Healthed population, as only 41% of Australian GPs are female [26]. The sample had a mean age of 48 years (SD = 9.88, range 21-81) and most practiced in a metropolitan area (n = 463, 71%). GPs were significantly more likely to be male (27% vs. <1%; χ²(1) = 51.23, p < 0.001), and younger (GPs M _age = 47.61 years, SD = 10.25; nurse’s M _age = 49.93 years, SD = 8.04; t(343.77) = -2.95, p = 0.003), and reported spending significantly more hours in the clinic per week than did nurses (GP M _hours = 30.45, SD = 12.65; nurse’s M _hours = 26.56, SD = 12.04; t(646) = 3.35, p = 0.001.

The survey was devised primarily for use in this study, and thus, had not been psychometrically validated. The survey had 31 items; of which 25 were relevant to the current study. Five items addressed basic demographic information (career, gender, age, physical location of practice, and weekly clinical hours), and two items covered exposure to cannabis issues (relevant training and personal experience with a regular cannabis user). Five items assessed knowledge regarding cannabis use (e.g., Cannabis users are more likely to have a mental health problem than those who do not use cannabis); four of these items were rated on a Likert scale from 0 (disagree completely) to 4 (agree completely). The fifth knowledge item was answered with a percentage (What percentage of people who try cannabis will someday develop cannabis dependence?).

The remaining 13 items used in this study were based on previous surveys assessing health practitioners’ beliefs, attitudes, and behavior regarding substance use and its treatment. Three items required GPs and nurses to assess their own knowledge about cannabis use and their skills related to screening and managing cannabis use [16‐18, 27]. These three items were rated on a Likert scale from 0 (very poor) to 4 (very strong). In addition, three items assessed beliefs regarding role legitimacy—factors related to the appropriateness of a GP or nurse to intervene with someone’s cannabis use [21, 25, 27, 28] and were rated on a Likert scale from 0 (disagree completely) to 4 (agree completely). One item assessed attitudes regarding cannabis use policy (Cannabis use should be illegal / decriminalized / available for medical use) [22, 29]. Four open-ended items assessed behavior related to screening, intervention, and referral [22, 28, 30]. Participants responded to these items with numbers (e.g., I have screened ____ patients for cannabis use in the last month). The final two items assessed barriers and facilitators to screening and treatment provision [18, 20, 25, 27‐29]. These items were answered by checking all applicable barriers or facilitators. Options referred to attitudes, motivation, confidence, and support.

Differences between GPs and nurses on categorical variables were analyzed using 2 Χ 2 χ²- tests (e.g., Hypothesis 1), while two-sample t-tests were used for continuous variables (e.g., Hypothesis 2, Hypothesis 3). When Levene’s test for equality of variances indicated that the variances between GPs and nurses responses were significantly different, a two-sample t-test was performed that did not assume equal variances. Logistic regressions using a generalized linear function examined the prediction of service provision (Hypothesis 4; 0 = none, 1 = at least once; reference category = 0). Before conducting the logistic regressions, zero-order correlations were conducted to determine which variables were appropriate for model inclusion. Only items statistically significant at the univariate level were included at the multivariate level. The significance level for all statistical tests was set at p < 0.01 to reduce Type I errors associated with multiple testing.

Table 1 contains summary data for cannabis-related training and exposure to cannabis users. GPs were significantly more likely to not have any personal experience with cannabis use compared to nurses (p < 0.001), while nurses were more likely to have a relative (p < 0.001) or know a friend (p < 0.001) who regularly uses cannabis. Further, GPs were significantly more likely to have received a small amount of training in cannabis-related issues (p = 0.006), while nurses were more likely to have not received any cannabis-related training (p < 0.001).

As set out in Table 2, GPs were more likely than nurses to have agreed that effective psychological treatments exist for reducing cannabis use (p = 0.001) and that cannabis users are more likely to have a mental health problem than non-cannabis users (p < 0.001). Further, 34% (n = 221) of health practitioners somewhat or completely agreed that effective pharmacological treatments exist for cannabis use.

On average, GPs believed that 21% (SD = 21.58) of people who try cannabis will someday develop cannabis dependence, which was significantly less than nurses believed (M = 25%, SD = 21.58), t(131.51) = -1.14, p = 0.001. Of the 516 participants who responded to this item, only 24% (n = 126) responded with an estimate close to 9% (6-12%) [6, 7]. Twenty-one percent of participants did not answer the question.

Nurses reported significantly less role legitimacy about providing cannabis-related services. Compared to nurses, GPs more strongly agreed that conducting a 10-minute brief assessment of someone’s cannabis use can lead to reductions in cannabis use (p < 0.001) and that people in their position are effective in treating patients with cannabis use problems (p < 0.001). Almost all GPs and nurses agreed that people in their position should receive education regarding cannabis (See Table 2).

Few GPs and nurses felt confident in their knowledge and skills about cannabis use and its treatment. As shown in Table 3, GPs rated their knowledge about cannabis (p < 0.001), and their skills in screening (p < 0.001) and managing cannabis use (p < 0.001), higher than nurses rated their own knowledge and skills.

A significantly greater percentage of GPs than nurses reported that cannabis use should be illegal, χ²(1) = 14.59, p < 0.001, and a significantly greater percentage of nurses than GPs reported that cannabis should be available for medicinal purposes, χ²(1) = 30.11, p < 0.001, as shown in Figure 1.

In the previous month, GPs reported screening an average of 3.84 (SD = 7.82, n = 498) patients for cannabis use and treating an average of 1.71 (SD = 4.17, n = 500) patients for cannabis use; although, approximately half (46%, n = 230) had not engaged in any screening or intervention provision (51%, n = 254). Nurses reported screening an average of 2.22 (SD = 10.00, n = 158) patients for cannabis use and treating an average of 1.33 (SD = 8.40, n = 159) patients for cannabis use in the last month. Most nurses, however, had not screened (85%, n = 134) or provided an intervention to any patients for cannabis use (87%, n = 139). In addition, GPs and nurses had referred less than one patient in the last month to an alcohol and other drug (AOD) facility (GPs: M = 0.48, SD = 1.42, n = 500, 374 (75%) had not engaged in referral; Nurses: M = 0.84, SD = 4.62, n = 160, 139 (87%) had not engaged in referral) or to a mental health service (GPs: M = 0.76, SD = 1.78, n = 500, 318 (64%) had not engaged in referral; Nurses: M = 0.44, SD = 2.39, n = 160, 143 (89%) had not engaged in referral). GPs were statistically significantly more likely to have engaged in screening (χ²(1) = 72.45, p < 0.001), treatment (χ²(1) = 64.85, p < 0.001), and referral (AOD: χ²(1) = 10.21, p = 0.001; Mental Health: χ²(1) = 38.24, p < 0.001) at least once (versus not at all) compared to nurses.

In order to identify if any variables had a unique predictive variance regarding increased provision of screening or intervention for cannabis use, only variables correlated with screening, intervention, and referral to AOD or mental health services at a significance level less than 0.01 were entered into the logistic regression models (Table 4). Interaction terms between health profession and other predictor variables were examined; however, as none were significant they were not included in the reported models. When the significant univariate correlates were included in the regressions, profession remained a significant predictor of screening, intervention and referral to mental health services. GPs were 4.73 times more likely to have engaged in screening provision than nurses, were 4.22 times more likely to have provided an intervention, and were 3.47 times more likely to have referred someone to a mental health service, after controlling for other predictors.

Confidence in the provision of screening was important, with every unit increase in perceived screening skills associated with a 3.28 times greater chance of having engaged in screening provision in the previous month and an 8.69 times greater chance of having referred a patient to an AOD service. Neither the degree of training, knowledge, nor perceived skills in cannabis intervention, however, offered a significant unique prediction of screening, intervention, or referral.

Health professionals who had a friend who regularly used cannabis were 1.80 to 1.83 times more likely to have referred someone to an AOD or mental health service than those without such a friend. In regards to attitudes, participants who believed cannabis should not be available for medicinal purposes were 2.06 times more likely to screen for cannabis use than those who believed it should be available. While that attitude was a significant univariate correlate of intervention and referral to a mental health service, it did not provide a unique prediction of those actions.

As shown in Table 5, GPs more often than nurses reported not having enough time for screening and intervention provision (p < 0.001), not being personally interested in screening and treatment provision (p = 0.001), not believing that cannabis use is the most important clinical issue (p < 0.001), not wanting to attract more cannabis users to their clinic (p < 0.001), and believing that the effort required to motivate a cannabis user to quit is not justified by the likelihood of a positive outcome (p < 0.002). Nurses more often reported that cannabis users should only be treated by specialists in the field (p < 0.001). In respect to facilitators, GPs more so than nurses reported that having more referral options would facilitate their provision of screening and intervention for cannabis use (p < 0.001).

General practitioners and nurses represent the overwhelming majority of primary care providers in Australia. They have enormous potential to reach a broad range of people for early intervention, where the focus is on prevention and health promotion. Although Australian GPs are being encouraged via government reimbursement schemes to provide brief advice in many prevention areas, including drug and alcohol use [13], efforts to increase service provision for cannabis users may need to be directed at both nurses and GPs, in order to utilize the opportunity for collaboration between GPs and nurses in the primary care setting. There are several reasons why nurses should be involved in this process. First, nurses were consistently less likely to have engaged in service provision than GPs. Second, GPs in this study were substantially more likely to identify barriers to service provision that are difficult to resolve, such as not having enough time and cannabis use not being the most important issue. Third, in 2010 the Australian Government announced a $523 million investment in Australia’s nurses in order reduce GP’s clinical burden [32]. Such an investment increases the need to improve nurses’ skills to ensure that they are fully able to undertake effective preventive health activities. Although the aim of the initiative is to allow GPs to focus on diagnosis, while nurses focus on tasks such as care co-ordination, health assessments, and health education [32], the most suitable model of practice for nurses has not been established. Thus, when developing training models, researchers will need to keep in mind that nurses may be operating from a substitution model (only doing those tasks delegated to them) or from a collaborative model (working autonomously within a practice). In addition, researchers will need to find an effective method for resolving the low legitimacy that nurses feel for managing cannabis use.

Hence, part of a training strategy may require education and empowerment of nurses in terms of their clinical potential to assist in managing cannabis use problems. In a substitution model, it may be advantageous to educate both GPs and nurses in how best to utilize their respective clinical strengths. This might involve nurses performing the screening and assessment of patients’ readiness for change and then referring motivated patients on to GPs for a brief intervention. GPs may develop clinical plans that involve utilizing nurses in their role as health educators. This would allow GPs to remain in the referral process, but minimize their clinical burden, while appropriately managing patients concerns. In those circumstances where nurses and nurse practitioners are working in a relatively autonomous context, nurses could be trained in both screening and intervention, referring to the GP only when the extremity of clinical features might warrant it.

Regardless of whom efforts are focused on, combining motivational interviewing with screening and intervention training for cannabis use may lead to greater provision of services to individuals who use cannabis. Research has found that motivating GPs to provide an intervention does not need to involve long professional development sessions. For example, a 20-minute motivational discussion with GPs has substantially increased their likelihood of discussing methods for managing dependent cannabis users [30]. Such motivational training may circumvent the negative effect that health practitioners’ pre-existing low motivation has on service provision. Providing GPs with training and support has been found to only increase provision of alcohol screening and brief advice for those GPs who were already motivated to work with patients with alcohol problems prior to receiving training [28]. Almost all GPs and nurses in the current study reported it was part of their role to receive education about cannabis use, and over three-fourths reported that receiving more training would increased their service provision. As such, providing training in screening (that includes information about screening resources, practice guidelines, and referral options) and in brief motivational interventions (that includes information about treatment resources, and discussion of personal experience and medicinal cannabis use) during professional development seminars may increase service provision for cannabis users.

When evaluating the effects of training on GPs’ and nurses’ service provision, researchers should carefully monitor increases in knowledge and attitudes toward service provision. Previous research found that training in a brief intervention led to increases in the understanding of a brief intervention for nurses, but not GPs [33]. Furthermore, while nurses’ knowledge about alcohol increased as a result of training, their positive attitudes towards discussing alcohol with patients decreased. Thus, training needs to be carefully developed and administered to avoid potential negative consequences.

The strength of the study is its large sample size. Limitations include the sole recruitment of health practitioners from Healthed seminars, reliance on self-reports of service provision, using a survey without established psychometric properties, and the low response rate (35%). Notably, the obtained survey response rate was similar to other surveys conducted with Australian GPs [34, 35], and thus, may not be indicative of non-respondents being less interested in cannabis use than responders. An additional limitation of this study is the high amount of missing data regarding the percentage of patients who will develop dependence (21% missing data). Research suggests that the high nonresponse on this item may have been due to the cognitively demanding nature of the question [36]. In order to reduce the non-response rate for this item, future researchers may want to adapt the open-ended item to include response options. Lastly, respondents were not provided with operational definitions for screening, brief interventions, or referrals; therefore, these items may have been interpreted differently. In light of these limitations, results may only generalize to GPs and nurses interested in women’s and children’s health who also have an interest in contributing to research. Future research with professionals interested in men’s health and that uses operational definitions and a validated survey may obtain different results.